You are on page 1of 309

Advances in Drilling Technology

E-proceedings of the First International Conference on Drilling Technology


(ICDT - 2010) and National Workshop on Manpower Development in
Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010

First International Conference on


Drilling Technology (ICDT - 2010)
and National Workshop on
Manpower Development in
Petroleum Engineering (NWMDPE
- 2010), November 18-21, 2010

Editors
R. Sundaravadivelu, SP. Subramanian and R. Sharma

Indian Institute of Technology Madras, Chennai (TN) - 600 036, India


18 – 21, November, 2010
Advances in Drilling Technology - E-proceedings of the 2010 International Conference on Drilling
Technology

Indian Institute of Technology Madras, Chennai (TN) - 600 036, India, 18 – 21, November 2010
R. Sundaravadivelu, SP. Subramanian and R. Sharma, eds.

Preface

We are honored to bring you this e-proceedings on first International Conference on “Drilling Technology:
(ICDT - 2010)” and “National Workshop on Manpower Development in Petroleum Engineering (NWMDPE -
2010)” held at the Indian Institute of Technology Madras, Chennai, India, November, 18-21 2010.

The primary focus of ICDT-2010 is on new ideas and the latest technologies related to the important topics that
are relevant for the industry, practicing engineers and academicians and will bring latest research results and
state-of-art technologies to the focus of the participants.

The main objective of this conference is to share and disseminate the information and design base being built by
the researchers in the area of drilling technology and make available to desired knowledge to the participants to
understand the advanced and emerging drilling technologies.

The international conference is followed by a national workshop on “manpower development in petroleum


engineering (NWMDPE-2010)”. Oil and gas – non renewable energy sources are fast depleting worldwide.
Search of these have entered into deeper waters which calls for new concepts and structures. The manpower
availability is not adequate in the field of upstream discipline of ‘Petroleum Engineering’. In order to provide
adequate manpower and create research facilities the Department of Ocean Engineering has introduced a
postgraduate course on Petroleum engineering from the academic year 2009. The primary focus of NWMDPE-
2010 is on identifying key areas that are hampering the proper and just development of manpower in petroleum
engineering and help Indian and global educational institutes and industries to develop their strategies
accordingly.

We would like to thank the keynote speakers for their excellent contributions in research and leadership in their
respective fields. We express our gratitude to the workshop organizers, who have been a very important part of
the conference and workshop, and, of course, to all participants and technical contributors of ICDT-2010 and
NWMDPE-2010.

R. Sundaravadivelu, SP. Subramanian and R. Sharma


Department of Ocean Engineering,
Indian Institute of Technology Madras,
Chennai, Tamil Nadu, India.

1
ISBN: 978-93-80689-03-6

Acknowledgements
We are indebted to many people who have helped with the conference:

Organizing Committee

Chief Patron
Prof. M. S. Ananth, The Director, IIT Madras, India

Patron
Prof. Job Kurian, The Dean (IC and SR), IIT Madras, India
Prof. S. K. Bhattacharyya, IIT Madras, India

Technical Committee

Prof. Nicholas Marinos Patrikalakis, Massachusetts Institute of Technology, USA


Prof. Brian Francis Towler, University of Wyoming, USA
Prof. Ramadan Ahmed, University of Oklahoma, USA
Prof. Anuj Gupta, Texas A&M University at Qatar, Qatar
Prof. Norbert Willenbacher, Universität Karlsruhe, Germany.
Prof. Pål Skalle, Norwegian University of Science and Technology, Norway
Prof. Tae-wan Kim, SNU, Republic of Korea
Prof. Ridha Gharbi, Kuwait University, Kuwait
Dr. Sudjati Rachmat, Institut Teknologi Bandung (ITB), Indonesia

Advisory Committee

Prof. Tarkeswar Kumar, ISMU, India


Prof. P. K. Banik, School of Petroleum Technology, PDPU, India
Prof. Ethirajan Govinda Rajan, PRC(P)L, India
Dr. Harish Mukundan, FloaTEC, LLC, USA.
Mr. Apurba Saha, ONGC, India
Mr. J. S. Sekhon, ONGC, India
Ms. S. Rajashree, GE Infrastructure (Oil & Gas), India
Dr. P. Gopalakrishnan, Consultant Reservoir Engineer, Mangalath Pariyarath, India
Dr. D. R. Babu, Iran Marine Industrial Company (SADRA), Iran

ICDT-2010 Keynote Speakers

Dr. Tarun Chakraborti, ONGC Academy, ONGCL, India


Prof. Y. S. Rajan, ISRO, India
Shri N. M. Borah, Oil India Limited, India
Dr. Anuj Gupta, Texas A&M University at Qatar, Qatar
Prof. T. Kumar, ISMU, Dhanbad, India
Shri Apurba Saha, ONGCL, India
Prof. E. G. Rajan, PRCPL, India
Dr. Sudjati Rachmat, Bandung Institute of Technology, Indonesia

2
Prof. Tae-wan Kim, SNU, Republic of Korea
Dr. Pål Skalle, NTNU, Norway
Dr. Peter Bergodes, Shell Hazira Private Limited, India
Prof. S. K. Bhattacharyya, IIT Madras, Chennai, India
Dr. G. A. Ramadass, NIOT, India
Dr. S. Rajashree, GE Oil & Gas, Bangalore, India

ICDT-2010 Workshops Organizers


Prof. G. Ali Mansoori, University of Illinois at Chicago, USA
Dr. Ridha Garbi, University of Kuwait, Kuwait
Dr. J. S. Sangwai, IIT Madras, Chennai, India
Dr. G. S. Kumar, IIT Madras, Chennai, India

NWMDPE-2010 Keynote Speakers


Shri Sudhir Vasudeva, The Director (Offshore), ONGCL, India
Shri B. K. Namdeo, The Executive Director (International Trade and Supply), HPCL, India
Shri D. Rajkumar, The Managing Director, BPRL, India
Shri Anand Kumar, The Director, Petrotech, India
Dr. P. Vijayan, Vice Chancellor, IMU, India

Trademarks and copyrights


All rights are reserved with IIT Madras, Chennai, India

3
4
Contents

Page number
Technical Sessions 1(A): Computer applications in drilling

1. Title: Characterization of Khoraghat field of South Assam shelf using well log data
Authors: D. Srikant, P. Shanmugam and SP. Subramanian 9-35

2. Title: Selecting attribute to standing out in the competitive objects


Authors: Murlidher Mourya and K.V Raghavender 37-41

3. Title: Image Segmentation Using both Color and Texture


Authors: Sreevani Raichur and K. Ramanaidu 43-48

Technical Sessions 1(B): Completion fluid/packer fluids and flow assurance

4. Title: Kinetic Study of wax deposition in the flow lines due to Indian crude oil
Authors: Vikas Mahto, Dharmendra Verma and Harveer Singh 49-54

5. Title: Effect of high pressure, high temperature and crude oil composition on the stability of crude oil
emulsions in the presence of low level of Asphaltene
Authors: Jitendra Sangwai 55-65

Technical Sessions 1(C): Design and analysis of marine risers

6. Title: Event analysis for offshore riser failure


Authors: S. K. Bhattacharyya, Srinivasan Chandrasekaran and M. V. S. Ram Prasad 67-79

7. Title: Risk assessment of offshore pipelines


Authors: S. K. Bhattacharyya, Srinivasan Chandrasekaran and M. V. S. Ram Prasad 81-91

8. Title: HAZOP study for crude oil pipe line


Authors: Srinivasan Chandrasekaran, J.Ramesh Babu and Arshad Ayub 93-97

9. Title: MPD application ranging from deep water to gas hydrate


Authors: M. Thangarasu 99-104

Technical Session 2 (A): Deepwater drilling platforms

10. Title: Design of new-generation drill ship with ice operating capabilities
Authors: Bo-ram Kim, Tae-wan Kim and Seung-chul Kang 105-116

11. Title: Deepwater drilling platforms


Authors: Reetanshu Kumar, Shiva Dayal, Rahul Ranakoti and Astha Gupta 117-124

12. Title: Numerical studies on deepwater dry tree semisubmersible


Authors: R. Sundaravadivelu, N. Srinivasan and V. P. K. Kaundinya 125-130

5
13. Title: A new pipe handling approach for deepwater slim drillship
Authors: Alex Kah Keong Tan, Asbjorn Mortensen, Sudhan Vincent and Abhijit D. Dimble 131-140

14. Title*: Conceptual design of an ultra-low motion new-age semi-submersible platform


Authors: Manuel Carmona Martínez and R. Sharma *-*

15. Title*: A NURBS based element for marine riser analysis


Authors: M. Goel, R. Sharma and S. K.Bhattacharyya *-*

Technical Session 2 (B): Drilling fluid design and analysis

16. Title: Treated fly ash: A drilling fluid


Authors: Chandan Guria, Akhilendra Kumar Pathak and Tarkeswar Kumar . 141-152

17. Title: Study the effect of polymers on rheological properties and filtration properties
of water based drilling fluids 153-160
Authors: Vikas Mahto, A.K. Pathak and V.P. Sharma

18. Title: Conceptual development of a computer model for rheological analysis of oil based drilling fluids at
HT/HP
Authors: S. Nainar, N. Willenbacher and R. Sharma 161-176

Workshop (A): Reservoir modeling and simulation

19. Title: Numerical modeling of fractured hydrocarbon reservoirs 177-182


Authors: G. Suresh Kumar

20. Title: Role of reservoir fluid sampling and phase behavior studies in reservoir management
Authors: Anil Kumar and Chaman Lal 183-186

21. Title: An assessment of non linear AVA inversion for reservoir parameter estimation using two global
optimization methods
Authors: Rajesh R. Nair, Yudhvir Singh and Suresh Ch. Kandpal 187-196

22. Title: Gas Hydrate stability and saturation estimation on Cascadian margin using ODP legs 889 & 891
inferred from petrophysics modeling
Authors: Yudhvir Singh, Ramakrushna Das and Rajesh R. Nair 197-205

Technical Session 2 (C): Others

23. Title: An investigation into enhancement of surface finish of pulse electrochemical machined surface using
rotating electrode
Authors: D. S. Bilgi, P. V. Jadhav and A. S. Harel 207-214

24. Title: Marginal gas field development using directional / multilateral drilling
Authors: S.R. Joshi 215-220

6
25. Title: Elastomeric and metallic stator progressive cavity pump: A comparative study
Authors: Ramasamy Karthikeshwaran and Abdus Samad 221-226

26. Title: Current state-of-the art well stimulation technology starting from types of near-wellbore damages to
nano ferro fluid thermal well stimulation
Authors: Sudjati Rachmat and Astra.A. Pramana 227-232

27. Title: Deep water drilling platform - zenith for efficient operations.
Authors: Priyadharshini Saptharishi 233-240

Technical Session 3(A): Others – 2


28. Title: Downhole pump selection methods
Authors: Abdus Samad 241-248

29. Title: Transfer of experience for improved oil well drilling


Authors: Pål Skalle, Agnar Aamodt and Odd Erik 249-257

30. Title: Computer aided design of air/gas drilling system for tight gas reservoirs
Authors: S. M. Mushoor and R. Sharma 259-264

31. Title: Design of integrated casing drilling system for deepwater applications
Authors: D. Vamshi Krishna and R. Sharma 265-286

32. Title: Drilling to exploit groundwater in “Vedanattam” Model village in Malattar Basin, Tamilnadu
Authors: R. Ramasamy, R. Sundaravadivelu, SP. Subramanian and P. Shanmugam 287-292

33. Title: Drilling problems in Lakadong sandstone – practical experience at the captive limestone mine of M/s
ACL, Meghalaya
Authors: P. Pal Roy and C. Sawmliana 293-298

Workshop (C): Advanced techniques for EOGR (Enhanced oil and gas recovery)

34. Title: Cementing practices in in-situ combustion wells to enhance heavy-oil recovery
Authors: Pradip Chakravarty, K. M.Suyan and S. Dutta 299-306

35. Title: Experimental investigations on water alternate gas injection methodologies for enhanced oil recovery
for an Indian Brownfield application
Authors: Jigar Bhatia, J. P. Srivastava, Jitendra Sangwai and Abhay Sharma 307-316

36. Title*: Performance of miscible gas injection in a middle-eastern reservoir.


Authors: Ridha Garbi *-*

__________________________________________________________________________________________
*
Only presentation at ICDT-2010.

7
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Characterization of Khoraghat field of South Assam Shelf using well log data
 
D. Srikant1, P. Shanmugam2, and SP. Subramanian3
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
E-mails: 1doddapaneni_srikanth@yahoo.co.in, 2pshanmugam@iitm.ac.in, 3sps@iitm.ac.in

ABSTRACT

The South Assam Shelf is a part of the petroleum province of the North East India. The area Khoraghat in SAS was selected as oil and gas producing
since 1989 and has a greater challenge as far as spatio – temporal distribution of reservoirs and its characterization is concerned. Thus the present
study has been performed with a view to elucidate the different litho units present in the study area by means of well log motifs with an emphasis on
the hydrocarbon prospect by formation evaluation. The reservoir parameters are estimated using the multi – mineral software called “Elan Plus”
(Elemental Analysis) of Schlumberger. The basic logs used as an input for this software are SP, GR, BS, CALIPER, DLL – MSFL, LDL – CNL and
DT (Acoustic logs). Elan Plus uses INVERSE MODEL to determine various reservoir properties such as water saturation (Sw), effective porosity (Φe),
lithology and hydrocarbon saturation (Sh). The Assam and Assam – Arakan basin is a polyhistoric basin as three tectonic styles viz. extentional,
compressional and resultant foredeep are present in this basin. The concerned area comprises mostly sediments of extentional regimes superimposed
by foredeep sediments adjacent to the compressive tectonics related to the east. All the well completion reports, lab data and all sets of log data have
been considered for making maximum geological information of 29 drilled wells. Khoraghat has interesting pay types. For formation evaluation, the
petrophysical properties of the various reservoir sands within Bokabil have been accessed for 16 producing wells. Of these 15 wells are oil producing
and one is gas producer. Number of sands increases from eastern side to western side within Bokabil. The thickness of sand increase basinward, i.e.,
towards east. Lower Bokabil sands are the major reservoir, though sands in Middle and Upper Bokabil are producing at some places. The effective
porosities of different sand range from 25 – 35 percent with occasional deviations. Volume of clay or shale ranges from 20 – 40 percent with
exceptions. The water saturation is 40 – 60 percent with occasionally higher values. Hydrocarbon saturation is accordingly calculated. As far as
petroleum system is concerned, the Kopili shales act as major source rocks. Middle Bokabil shale itself act as a cap rock. Lower Bokabil sands are the
major reservoir, though sands in middle and upper Bokabil are producing at some places. Type of entrapment is stratistructural. Major fault act as
probable migratory paths in the study area. The linear NE – SW trending horst in the southeastern part appears to have the promising locale of oil
entrapment. Using Hingle plot Rw is calculated by plotting Rt verses rhob on a semilog paper. The calculated Rw is used as an input for tne
determination of water saturation (Sw) by Archie’s equation. Pickett plot generated for the estimation of petrophysical constants a,m and n which are
used as an input for Elan Plus. Histograms are also generated for more realistic estimation of GRsh, SPsh, Rsh, RHOBsh, NPHIsh and DTsh which are
used as a control parameter for log processing. The petrophysical parameters of the hydrocarbon bearing sands of Khoraghat field are tabulated for
0.5m interval and a continuous color plot of the volumes of the rock and fluids are also presented.

INTRODUCTION

Logs are used for a variety of purposes depending on the nature of the data gathered. Correlation from well to well is the oldest and probably the most
common use of logs. It allows the subsurface geologist to map formation depths and thicknesses and then to identify conditions that could trap
hydrocarbons. Correlation is usually based on the shapes of the recorded curves versus depth. Correlation in complex geologic areas may be difficult
or impossible, and in any event requires corroboration from actual rock samples for the initial correlations in an area. After the curve shape patterns
are recognized, they can often be used in subsequent wells without relying too heavily on rock sample data. Identification of the lithology of the rock
sequence is another important use of logs. A log shows many variations from top to bottom. Each variation has significance, but it can be related to the
rocks being logged only by comparing the log with actual rock samples or a core from the well. After acquiring experience in an area it is possible for
a log analyst to make an educated guess as to lithology by looking at the log. Modern analytical methods permit more accurate lithology identification,
but this requires charts or mathematical solutions in addition to the curve shapes.
One of the important uses of logs today is the determination of rock porosity. This measurement is significant because it tells how much storage space
a rock has for fluids. No log actually measures porosity directly, but many analytical methods are available to help estimate this important property.
Another of the routine uses of logs is the determination of the water, oil, or gas saturation in the rock pores. When the porosity, oil or gas saturation,
the thickness and extent of the reservoir are known, then it is possible to tell how much hydrocarbon is in place in the reservoir. Again, no log actually
measures the fluid saturation directly, so analysis of indirect measurements is required. The logs most often run for the above purposes are resistivity,
sonic travel time, density, neutron, gamma ray and spontaneous potential logs. One of the older, but very useful, surveys is the caliper log. In open
hole logging, it is used to determine hole volume and aids the engineer in designing a cementing program. It also indicates mud cake build-up and hole
wash-out. Both of these indications are of interest to the log analyst when he considers the other logs. In cased hole work the caliper is often used to
find casing damage and separated casing.

A more recent development in logging is fracture-finding. It is important because fractures will often produce large quantities of fluid even though the
rock the fractures are in would not otherwise produce commercially. Many open-hole logs have some artifacts caused by fractures, but the formation
micro-scanner and borehole televiewer are the most useful. The temperature log is commonly used to indicate the top of cement behind a newly
cemented string of casing. The setting cement liberates heat and warms up the well bore, which is thus recognizable on a temperature log of the well.
Another use for the temperature log is the location of points of fluid entry in a well bore or of fluid flow behind casing. As the fluid enters the well it
expands and cools creating abnormally low temperature in the well at the point of entry. Acoustic noise logs also find flow entry and flow behind pipe
by the noise caused by the flowing fluid. The most significant change in the use of logs, in recent years, is production monitoring. The thermal decay
time log (often called a pulsed neutron log) allows for the interpretation of porosity and fluid saturation behind casing. The fluid saturation will change
over time as a reservoir is depleted by production, and the changes may be monitored by logging at regular intervals, say once a year. If the production
pattern is not as predicted, remedial action may be possible. The log is also used to provide porosity and fluid saturation data in wells which are not, or
could not, be logged in the more conventional open-hole manner. A large suite of logging instruments is available to evaluate fluid type, fluid flow,
and mechanical conditions in producing or injecting wells, in addition to those already mentioned. These are generically called production logs and are
usually run in cased holes, but some are also effective in open hole or "bare-foot" completions.

1.2 What Does Quantitative Analysis Really Mean?

Quantitative analysis is a matter of data reduction and summary. Analysis and interpretation should not be confused or construed to be one and the
same. Some people use the words log evaluation to mean either or both analysis and interpretation. We prefer to think that analysis and evaluation
mean data reduction and that interpretation involves trying to understand these results in light of the assumptions made, and other known facts not
used in the analysis.

Objectives of Quantitative Analysis:


1. Observation of log response in a known environment (calibration).
2. Assumptions concerning the present environment of the logging instrument (parameter selection).
3. Observation of log readings in the present environment (reading the log).
4. Calculation of results based on log readings, the assumptions and the log calibration (analysis).
5. Understand what the results mean (interpretation).

Although some theoretical research has been done on log response, much work is based on observation and empirical relationships derived from
correlation of measured rock properties versus well log readings of the same rocks. These empirical equations are the heart of quantitative analysis.
Unfortunately the formulae are often derived from very limited data sets. For example, one of the original papers on log response was by G.E. Archie.
In it, he described 48 core samples and a corresponding number of porosity and resistivity measurements.

Other analysts have run many more such samples through the lab, and have come up with different relationships, depending on where the samples
originated. Archie's equation, and the interpretation parameters he suggested, is still in use nearly 60 years later, when it is known that the data is only
an average result of a very limited local sample, not even a world-wide average.
With the advent of modern, inexpensive, multi-function, programmable calculators, pocket computers, and desktop micro-computers, chart book
methods are quickly disappearing. Charts are occasionally referred to when working in complex lithology, where the pattern or position of points on
the chart or graph may be helpful. Even this can be quantified by appropriate equations.

Different types of logging tools


Invaded zone
Most wells are drilled with drilling mud having a hydrostatic pressure opposite layers that is greater than the pressure of these same layers. Part of
the fluid from the mud filters into the layer, and the solid parts are deposited to form a mud cake on the well wall.
The situation is schematically diagrammed in the figure. The squares show the parameters read by the resistivity sonde, Rt or Ro and Rxo as well
as Rsh which is the resistivity of the shale. The circles show the parameters normally obtained or checked by measurement on a sample:
Rw = resistivity of water in the uninvaded pores.
Rmf = resistivity of the mud filtrate.
Rmc = resistivity of the mud cake.
Rm = resistivity of the mud.

The triangle contains parameters that have been calculated, with the most important being Sw, Sxo and Ø.
di = diameter of the flushed zone, i.e., the zone through which several pore volumes of filtrate have calculated, leaving behind a residual
hydrocarbon saturation.

  10
dj = diameter of the virgin or uninvaded zone, i.e., the one not attained by the filtrate.
dh = diameter of the borehole.
h = thickness of the layer.

Figure 1.1: The Data Reduction/Analysis/Interpretation Model.

SPONTANEOUS POLARIZATION (SP):


1. The spontaneous potential (SP) log records the naturally occurring electrical potential (voltage) produced by the interaction of formation
connate water, conductive drilling fluid, and shale.
2. The SP curve reflects a difference in the electrical potential between a movable electrode in the borehole and a fixed reference electrode at
the surface.
3. Though the SP is used primarily as a lithology indicator and as a correlation tool, it has other uses as well: permeability indicator, shale
volume indicator, porosity indicator, and measurement of Rw (hence formation water salinity).

GAMMA RAY LOG:


1. Gamma Rays are high-energy electromagnetic waves which are emitted by atomic nuclei as a form of radiation.
2. Gamma ray log is measurement of natural radioactivity in formation verses depth and can be used for Identifying lithologies: Correlating
zones (b/w wells), Evaluating shale content within a formation, Determining boundaries, Defining shale bed when the SP is featureless. It
measures the radiation emitted by the formation that contain radioactive materials such as U, Th, and K.
3. Shale-free sandstones and carbonates have low concentrations of these minerals and give low gamma ray readings.
4. However, clean sandstone (i.e., with low shale content) might also produce a high gamma response if the sandstone contains K-feldspars,
micas, glauconite and uranium-rich waters.
5. If a zone has a high potassium content coupled with a high gamma ray response, the zone might not be shale. Instead, it could be a
feldspathic, glauconitetic or micaceous sandstone.

  11
Figure 1.2: Schematic diagram of invasion.

Figure 1.3: Origin of SP potential.

GAMMA RAY LOG:


6. Gamma Rays are high-energy electromagnetic waves which are emitted by atomic nuclei as a form of radiation
7. Gamma ray log is measurement of natural radioactivity in formation verses depth and can be used for Identifying lithologies: Correlating
zones (b/w wells), Evaluating shale content within a formation, Determining boundaries, Defining shale bed when the SP is featureless.
8. It measures the radiation emitted by the formation that contain radioactive materials such as U, Th, and K.
9. Shale-free sandstones and carbonates have low concentrations of these minerals and give low gamma ray readings.
10. However, clean sandstone (i.e., with low shale content) might also produce a high gamma response if the sandstone contains K-feldspars,
micas, glauconite and uranium-rich waters.

  12
11. If a zone has a high potassium content coupled with a high gamma ray response, the zone might not be shale. Instead, it could be a
feldspathic, glauconitetic or micaceous sandstone.

POROSITY LOGS

SONIC LOG:
1. The sonic tool consists of one or more ultrasonic transmitters and two or more receivers.
2. Designed to measure the time it takes for a pulsed compressional sound wave to travel one foot (interval transit time, ∆t or DT) along the
axis of the borehole.
3. The acoustic pulse from a transmitters detected at two or more receivers. The time of the first detection of the transmitted pulse at each
receiver is processed to produce ∆t.
4. The interval transit time (µsec/ft) is the reciprocal of the velocity of a compressional sound wave and depends on the elastic properties of the
formation which are related to lithology and porosity.
5. In general, waves travel faster through denser formations. Therefore, an increasing travel time for a given type of material indicates
increased porosity.

Interpretation goals: Porosity, Lithologyidentification (with Density and/or Neutron), Synthetic seismograms (with Density), Formation mechanical
properties, (with Density), Detection of abnormal formation, pressure, Permeability identification (from waveform), and Cement bond quality.

DENSITY LOG:
1. Gamma rays emitted from a chemical source (Ce137, Co60) interact with electrons of the elements in the formation.
2. Two detectors count the number of returning gamma rays which are related to formation electron density
3. For most earth materials, electron density is related to formation density through a constant
4. Returning gamma rays are measured at two different energy levels
a. High energy gamma rays (Compton scattering) determine bulk density and therefore porosity
b. Low energy gamma rays (due to photoelectric effect) are used to determine formation lithology
c. Low energy gamma rays are related to the lithology and show little dependence on porosity and fluid type Symbol for density:
ρ(rho)

Interpretation goals: Porosity, Lithologyidentification (from PEF and/ or with Sonic and/ or Neutron, Gas indication (with Neutron), Synthetic
seismograms (with Sonic), Formation mechanical properties (with Sonic), Clay content (shaliness) (with Neutron), and Determine Hydrocarbon
density.

NEUTRON LOG:

1. Neutron logs are porosity logs that measure the hydrogen content in a formation. In clean, shale free formations, where the porosity is filled
with water or oil, the neutron log measures liquid filled porosity (φN, PHIN, NPHI).
2. Neutrons are emitted from a chemical source (americium – beryllium mixture). At collision with nuclei in the formation, the neutrons loses
energy. With enough collisions, the neutron is absorbed and a gamma ray is emitted.
3. Because the hydrogen atom is almost equal in mass to the neutron, maximum energy loss occurs when the neutron collides with a hydrogen
atom.
4. Materials with large hydrogen content like water or hydrocarbons become very important for slowing down neutrons. Since hydrogen in a
porous formation is concentrated in the fluid-filled pores, energy loss can be related to the formation’s porosity.

Gas detection: Density porosity is too high, Neutron porosity is too low, Neutron porosity < Density porosity, Cross-over

Note – Be aware, cross-overs may also be caused by lithological differences as an affect of the scaling

RESISTIVITY LOGS

TYPES OF RESISTIVITY TOOLS:

Two general types:


• Electrode (Laterolog): forces a current through the rock and measures resistivity
• Induction (Induction Logs): Uses a fluctuating electro-magnetic field to induce electrical currents in the rock; it measures conductivity which is
converted to restivity.

Basic Differences:

1. Electrode logs need a conductor in the well bore so they don’t work with fresh water and oil based mud, or air.
2. Induction tools do not need a conductor in the bore hole and canbe used with fresh water and oil based mud and air. They are designed to
read intermediate and deep resistivities and are labeled ILm and ILd respectively.
3. Induction tools do not work well in formations with low conductivity (high resistivity) because only weak currents will be induced in the
rock and are hard to measure.

  13
4. The flushed zone resistivityis normally measured with an electrode log.

INTRODUCTION TO KHORAGHAT FIELD

The khoraghat oil field is the major hydrocarbon producing field in Dhanshri valley of South – Assam Shelf, Assam and Assam Arakan basin. The
South Assam Shelf is a part of the petroleum province of North – East India. The sands of lower bokabil are the main producers. In addition sands
within sylhet, barail and middle bokabil are also producers of hydrocarbons. The study area is located in the southern part of the Golaghat district of
Assam. The data input in the area are 29 wells. The geological and geophysical leads have been taken from khoraghat for structural mapping and
hydrocarbon entrapment model in the area. The hydrocarbons encountered in the lower bokabil sands have proved beyond doubt that hydrocarbon
migration has taken place to the westerly major fault bounding khoraghat field. In addition the porous sands within sylhet i.e. hydrocarbon bearing is
also present in most of the study area. Recent microbial studies in the north of khoraghat area have indicated areas of higher hydrocarbon
concentrations corresponding to the surface expression of some deep seated faults which are the conduits for hydrocarbon migration.

Objective of Study

The area of khoraghat in South Assam Shelf was selected as oil and gas is being produced since 1980 and as a great challenge as far as spatio –
temporal distribution of reservoir sands and their study is concerned. The present study has been performed with a view to elucidate the different litho
units present in the study area by detailed well log correlation deciphering the distribution of the reservoir sands by well log motives with an emphasis
on the hydrocarbon prospect by formation evaluation.

Approach

All the well completion reports, formation evaluation reports and well logs have been used for estimation of reservoir parameters namely total porosity
(Ø), effective porosity (Øe), water saturation (Sw), hydrocarbon saturation (So), formation water resistivity (Rw), pay thickness (h) have been derived
by established techniques of well log interpretation. During formation evaluation an iso - porosity map of the reservoir sand encountered in several
producing wells is prepared.

Regional Tectonic Setup: The Assam and Assam Arakan basin represents a classic example of poly history basin with a superimposition of both
extensional and compressional regimes. As a result passive margin sequences are overlain either by foredeep sequences (resultant of compressive
tectonics) or by deep basinal sequences as a result of thrusting, particularly in the Schuppen Belt area.

The Dhansiri valley is situated in the southern part of Assam shelf within the Assam and Assam Arakan basin. It is a classical asymmetric foreland
basin bounded by the Mikir hills in the west and the Naga hills in the east and the south east, the east – west trending Jorhat fault separating it from the
Upper Assam part of the shelf and with the Dauki – Haflong fault as its southern limit, separating it from the folded belt part of the Assam and Assam
Arakan basin. The easterly prograded Paleogene sequences (Sylhet, Kopili and barail) were deposited during marginal sag cycle of divergent margin
while the westerly/southwesterly prograded Neogene sequences represent the foredeep packages of the convergent phase.

Singh et al, 1996 have explained the evolution history of the basin as follows: This basin was initiated during Late Jurassic/ Early Cretaceous in
divergent margin setup, the basin went through interior fracture cycle and margin sag cycle until the beginning of the collision of the Indian plate with
the Eurasian plate during the early Eocene. During the interior fracture cycle, rift related horsts and grabens were superimposed which parallel the
ENE – WSW Satpura trend. During the marginal sag cycle, basin subsided passively to great depths aided by numerous down to basin faults, which
are sub parallel to the Dharwar and Aravalli trends. Collision of the Indian plate with Eurasian plate, initiated in the north western part of the Indian
plate, advanced progressively towards the east transforming the marginal sag of Assam – Arakan basin into interior sag basin in early Miocene. The
continental collision advanced progressively from north to south. The progressive continental collision and anti – clockwise rotation of the Indian plate
may end up the tectonic elements of the Surma depression in a situation similar to the one prevailing rift faults, down to basin faults and their
associated antithetic and synthetic faults. The Gondwana sediments outcropping in ENE – WSW trending grabens bordering the Shillong platwau and
encountered in few wells represent the pr-rift stratigraphic units. The plateau encountered in few wells represent the pre – rift stratigraphic record of
the early rift phase. The easterly prograded Paleogene sequences (Sylhet, Kopili and Barail formations) were deposited during the marginal sag cycle
of divergent margin development. The westerly/southerly prograded Neogene sequences (Surma and younger units) represent the sedimentary
packages developed as a foredeep basin during convergent time. Dasgupta et al, 2000 in their recent publication have explained the different tectonic
phases, sedimentation and stratigraphy of Assam in the following manner:

Table2.1: different tectonic phases, sedimentation and stratigraphy of Assam


Phase Age Event
V Early Eocene Collision with Myanmar, Tibet & Mishmi, Major uplift of Shillong –
Mikir hills Plateau,
IV Lower Jurassic / Early Cretaceous Marine Sedimentation in sothern & eastern Assam. Garo, Sylhet, Mikir
traps & basic – ultrabasic intrusive
III Late Triassic/ Early Jurassic Marine sedimentation in northern fringe of India.
II Permo – Caboniferous Adjoining areas rifted, coal bearing Gondwanas deposited. Local
volcanism.
I Proterozoic Part of Gondwana Super – Continent

  14
Kanungo et al, 2002 opined that the tectonic evolution of NE India could be explained with an Oblique Collision and Tectonic Wedging Model. The
present Assam foredeep adjoining to subduction margin represents the updip extension of an ancient passive margin. The Naga – Schuppen belt was
formed during Miocene with deposition of molassic sediments of tipam formation in upper Assam. Surma basin, the southern extension of the
foredeep, remained as a Remnant Ocean Basin, which was subsequently closed by the southwardly moving orogenic front.

Figure 2.1: Location map of study area.

Figure 2.2: Satellite imagery map of NE India showing the study area.

  15
Tectonic Evolution of North – Eastern India: A Synthesis

The complete tectonic scenario of the region is intricately linked with the movement of the India plate and its relationship with the tectono –
chronology of the Himalayan Origen. Based on the broad geodynamic background of Indian Plate, the tectonic history of the north eastern India has
been discussed in three main phases of the ‘Wilson Cycle’ of ocean opening and closing (Plate: ). Broadly, the spatico – temporal evolution of the
north eastern India is explained with an ‘Oblique Collision and Tectonic Wedging Model’ (Naik, 1994), (Plate). The tectonic history seems to vary
along the strike and progressively propagates from north – east to south – west. The Ophiolites and the associated older rock suggest the development
of the basin during Cretaceous time. Further, the Ophiolite line marks the paleo subduction belt. Prior to collision and overriding, the north – eastern
India was almost certainly a passive margin and fronted by Mesozoic Oceanic lithosphere that was continuous with today’s eastern coast passive
margin. The present day upper Assam foredeep adjoining to the A – subduction margin represents the up – dip extension of an ancient passive margin.
The Mesozoic sediments in the southeast are presently involved in fold and thrust belt. Subsequent to the thrust loading, flexural sag i.e. the foreland
basin was initiated latest by late Eocene. The flexural process continued with migration of the foreland basin towards NW.

STRATIGRAPHY AND DEPOSITION

The Dhansiri Valley is bounded by the Naga hills on the east and the Mikir hills on the west. Its northern limit is marked by the jorhat fault and the
southern limit by the Dauki fault (Dasgupta et. al. 2002). The basin was evolved from an east – facing passive – margin setting in Paleogene to present
day foreland basin that superposed during Neogene. The Dhansiri valley contains a tertiary fill sequence of about 3.5 km thickness overlying the
Precambrian Gneissic basement. Sporadic occurrence of Gondwana sediments in NE – WS trending linear belts are present in between basement and
the tertiary fill. The khoraghat – Nambar area is located in the central part of the Dhansiri Valley. The generalized litho – stratigraphic succession
(Ref. ONGC), encountered in wells drilled in Dhansiri Valley is given in Plate:3.

Generalised Stratigraphy and Depositional Setup:

Basement: In Assam and Assam – Arakan basin, the Pre – Cambrian crystalline rocks from the floor for the sedimentary succession and consists of
granite, granite gneiss and metamorphic. It is referred to as Metamorphic Basement Complex (MBC) and group all the crystalline rocks floor the
sedimentary sequence. These rocks exposed in the Shillong Plateau and Mikir hills and also encountered in subsurface in number of wells in
Meghalaya, Nagaland and Assam.

Gondwana / Rift Sediments: The overlying Gondwana sediments are present in the basin as remnant patches confined to paleograbenal parts. These
sediments comprises dominantly sand / shale alterations, rift fill continental sediments with occasional marine intercalations on the lower unit. Based
on their composition, log signatures and palynofossil assemblage, the Gondwanas here are also differentiated into two units viz. Upper and Lower.
Structural closures formed wherever Gondwanas are present in substantial thickness make them good exploration targets.

Tura Formation: The Tura formation (Late Paleocene to Early Eocene) is commercial producer in the Borhola – Champang field. The thickness of
this formation varies from less than 5m to more than 120m (maximum thickness of 123m encountered in a particular well in Borholla). This formation
unconformably overlies the metamorphic basement.

Figure 2.3: Tectonic map of NE India.

  16
This formation consists mainly of sandstone with a few shale intercalations. The presence of kaolinitic sandstones in association with coal and
carbonaceous shale is indicative of fluvio – deltaic depositional setting at its initial phase, being gradually taken over by a brackish to marginal marine
condition in the final phase (RGL, Sivasagar, 1999). This is supported by the palynofloral assemblages and the textural changes of sandstones
(indicated by the frequent presence of calcareous cement and glauconitic patches).

Sylhet Formation: The sylhet sediments in the study area consists of silica – clastic and carbonate sediments and vary in thickness from 44m to more
than 100m. They range from calcareous sandstone to sandy/ silty limestone, calcareous shale/ marl along with pure limestone. A near uniform
thickness of this formation is seen throughout Dhansiri valley in North – South Electro – log Profile along certain wells that passes through western
fringe of Dhansiri Valley. Another North – South Profile along certain wells passing through the eastern part indicates increasing thickness towards
south. East – West electrolog profiles show thickness variation is in the order of 50m to 100m indicating the basinal slope towards east and south –
west. The sediments as well as the faunal assemblages indicate deposition in a shallow open to restricted ramp setting (K. Dutta et. al. 1996). Presence
of echinoids and brachiopods in abundance in the faunal assemblage indicates deposition under normal marine condition (Wilson and Jordan, 1982),
whereas presence of marine calcareous algae represents water depths within the photic zone. However, common blue – green are more tolerant to
salinity variation and generally occur in water of less than 5m of depth (Wilson and Jordan, 1982).

Kopili Formation: This formation overlies the carbonate dominated Sylhet formation with sharp / conformable throughout the area. It is overlain
unconformably by dominantly argillaceous Bokabil formation to the west of the area and by arenaceous barail group of sediments in the rest of the
area.

Figure 2.4: Tectonic map of study area.

  17
Figure 2.5: Tectonic model for hydrocarbon generation in NE India.

Thickness of the kopili formation varies from 200m to 350m that increases towards south – east with maximum thickness of 650m encountered in a
particular well. In khoraghat – Namber area and in some places along the weatern margin, thickness of kopili is lesser as compared to adjoining areas.
This may be due to prolonged exposure and subsequent weathering during Oligocene time. Though kopili is mostly argillaceous, small sand bodies are
present towards lower part of kopili. At places towards khoraghat area limestone bands are also present in the lower part of this formation. The kopili
formation comprises of two members, viz. Lower Charali Member and Upper Amguri Member (Despande et. al. 1993). The Charali Member
comprises dominantly of shale with thin streaks of limestone / marl and thin sandstone layers. The Amguri Member is a shale dominated section
sandstone intercalations. The depositional setting was shallow marine during Charali sedimentation that gradually changed to marginal marine to
brackish water during Amguri sedimentation (RGL, Sivasagar, 1999). The total absence of Barail Group and the absence of Upper part of kopili
formation in certain wells of Naojan, Barpathar, Kasomarigaon and Jamuguri indicate an erosional unconformity in the western part of the area.

Barail Group: The Barail Group is totally absent in the western part of the area. The thickness of the group attains more than 400m in the eastern part
of the area close to the Naga Thrust around Mekrang and Borholla areas. The group comprises of three formations viz. Disangmukh, Demulgaon and

  18
Rudrasagar in Upper Assam (Deshpande et. al. 1993). The sandstone – shale alteration of Disangmukh formation and Arenaceous dominant
Demulgaon formation except in the area very close to the Naga thrust as seen in Khoraghat area. Barail formation overlies the dominantly argillaceous
kopili formation with a sharp comformable contact. Lower part of this formation is represented by inert – bedded sequence of sandstone and shale,
occurring by inter bedded sequence of sandstone and shale, occurring in sub – equal proportion and designated as Disangmukh formation.
Occasionally thin coal streaks are present. Compositionally the sandstones are very fine grained calcareous quartzose arenite stippled with pyritic
inclusions at places. Enormous thickness variation is observed within this formation. The formation becomes dominantly arenaceous towards south.
The Disangmukh formation is interpreted to have been deposited in distal delta front (RGL, Sivasagar, 1999) environment. The upper unit Demulgaon
formation is a dominantly arenaceous section with thin, subordinate shale. Compositionally the sandstones are calcareous quartzose arenite with
calcite and ferruginous cement. A proximal delta front environment of deposition has been envisaged for this formation. On the logs, Barail is marked
by blocky, serrated nature of the log. At places top of Barail is marked by presence of coal.

Bokabil Formation: Bokabil formation unconformably overlies Barail group sediments and is overlain conformably by Tipam group. On the basis of
variation in the litho – association, textural analysis, microfacies analysis, micro – fauna, micro – floral assemblages and electrolog characters, the
Bokabil formation of Dhansiri valley area that has been divided into three distinct units viz. Upper arenaceous, lower sand intercalations in shale and
middle argillaceous litho units. The lower Bokabil unit is partially eroded by tidal inlets leaving lower Bokabil relict features on which the argillaceous
middle bokabil is deposited providing lateral as well as vertical seal.

Lower Bokabil: The lower Bokabil arenaceous litho unit consists of five (5) distinct petrofacies (Mahanta et. al. 2000). They include fine to very fine
grained, well sorted calcareous lithic arenite, calcareous quartzose arenite, micro laminated sandstone, siltstone and shale, calcareous quartzose wackes
and calcareous glauconitic sandstone. The lower part of this unit is characterized by interlaminated sandstone, siltstone and claystone with parallel and
wavy laminations, lenticular bedding, small scale ripple cross bedding, bioturbation, serrated cylindrical and bell shaped log characters are indicative
of tidal flat deposit. In the middle part of the lower unit, fine grained, well sorted sandstone plane and wavy laminations, vertical increase in grain size,
high grain – matrix ratio, presence of glauconite, blocky electro log motif, tapered spike (basinward) and simple spike (landward) log character
suggest a barrier bar deposit. Elongated sand body parallel to depositional strike and log probability plot further support this view – point (Mahanta et.
al. 2000). The sand body has been cut by fine grained, well sorted sandstone body showing bell shaped log motif, which is interpreted by as tidal inlet
facies. Interlaminated claystone, siltstone and thin sandstone bands dominate the upper part of the lower unit. The presence of lenticular bedding,
plane laminations, small scale ripple cross bedding, presence of glauconite, bioturbations indicate an intertidal mudflat deposit. Two depositional lows
– one in the north and the other in the south have been mapped in the area.

Middle Bokabil: The lower part of the middle argillaceous unit is interpreted to be inner shelf deposit as evidenced from its typical litho – association,
dominantly greenish grey claystone with minor siltstone, parallel and lenticular laminations, small scale graded bedding and colour layering in clays.
Typical log character of almost flat GR without any lithological variation and presence of planktonic forminifera indicate that sediment has been
deposited in 30 – 50m bathymetry (Mahanta et. al. 2000).
The upper part of the unit is composed of siltstone that show gradual coarsening upward sequence as evidenced from funnel shaped log motif, and if
interpreted to be distal delta front deposits (Mahanta et. al. 2000).

Upper Bokabil (Khoraghat Sandstone Member): The lower part of upper arenaceous unit of bokabil formation comprises of sandstone with minor
siltstone and claystone. The presence of sedimentary structures like parallel laminations, ripple cross lamination sand stacked, serrated, blocky and
funnel shaped electrolog motif suggests a delta front deposition comprising of delta front slope sands and mouth bar sands (Mahanta et. al. 2000).
Towards the upper part of this unit, it becomes argillaceous and the sandstone bodies associated with clay pellets, planar cross bedding, streaks of
carbonaceous matter and coal , display bell shaped log motif. The litho and electrolog characteristics along with sand geometry suggest delta plain
deposit comprising distributary channel sands and inter – distributary channel sands and inter – distributary marsh deposits (Mahanta et. al. 2000).

Hydrocarbon prospects of Bokabil Sands: Bokabil formation in Khoraghat area is interpreted to be product of transgressive barrier bar – tidal inlet
sedimentation. The barrier bars are envisaged to be formed along NE – SW ridges. The absence of flood tidal delta deposits on landward side, relative
stability of tidal inlet are suggestive of transitional to tide dominated nature tidal inlet. In such inlets the sand bodies are localized near mouth of inlet
and may extend seaward. These barrier and inlet sands are covered by fine clastics of intertidal mud flat constituting cap rock for the same. The fluid
distribution and behaviour of reservoir pressure point to hydrodynamic continuity of these reservoirs. The sedimentation model predicts the orientation
of this reservoir in NW – SE direction. The tidal delta deposits constituting upper paysand also form excellent stratigraphic trap due to lateral and
vertical sealing by mud flat deposits.

Tipam Group: In upper Assam, the Tipam Group is litho – stratigraphically composed of four formations. From bottom to top they are divided into
Geleki sandstone, Lakwa sandstone, Girujan caly and Nazira sandstone units. However, except in the Mariani – Borholla Block, where the tipam
group can be differentiated into the above formations, it is not possible to sub – divide it in other parts of Dhansiri Valley (Despande et. al. 1993). This
group of sediments was deposited in a fluvial setup. The sands are mostly braided bars and meandering channel point bar deposits. The Girujan clay is
interpreted to have been deposited in a flood plain/ lacustrine setup. Thickness of arenaceous Tipam unit is increasing towards north whereas thickness
of overlying Namsang formation is increasing towards south which may be attributed to basin inversion.

  19
Figure 2.6: Generalized Litho – Stratigraphy of Dhansiri Valley.

Figure 2.7: Petroleum system.

  20
STRATIGRAPHY AND SEDIMENTATION

Sediments ranging from tural to Namsang formation of Paleocene to recent age over lie on Achaean basement in Dhansiri valley, comprising both of
passive margin and foredeep setup. In some places, presence of Gondwana sediments have been established in wells East Lakibhari, Barpathar,
Jamuguri and Furkating. Again, in Naga – Schuppan, Passive margin sequences / foredeep sequences comprising Tipam, Namsang and Dhekiajuli
formations are of fluviatile origin. All the passive margin sequences viz. Sylhet, Kopili, Barail and Bokabil and Barail sediments are present in most of
the wells except in Borholla area where , Bokabil formation of marine origin is absent and equivalent sediments of fluviatile origin are present.
Bokabil sediments, where hydrocarbons have been encountered, are overlain comformably by Tipam sediments and underlain unconformably by un –
differentiated barails. Based on the presence of prominent coal layers, this unit is considered to have been deposited in shallow to marginal marine
environment (fossil evidences indicate inner neritic depositional setting). Within Dhansiri valley the tipam group is differentiated, except in the
Borholla – Mariani Block where the Tipam Group can be differentiated into four formations viz. Geleki sandstone, Tipam sandstone, Girujan clay and
Nazira sandstone as in Upper Assam, by Deshpande et al. 1993. The groups of sediments were deposited in a fluvial setup mostly as briaded bars and
meandering channel point bar deposits. The unit is characterized by massive sandstones with minor intervening shales and shows a distinctive blocky
nature on electro – log motif. The Girujan clay was deposited in a flood plain/ lacustrine environment. The Namsang formation consists of
unconsolidated fluvial sands and clay lignite, whereas the post Namsang sediments are made up of sand and clays. The Tipam and post Tipam
sediments thicken towards east and southeast within Dhansiri valley. The generalized stratigraphy of Dhansiri Valley is given below and stratigraphy
chart and environment give in Plate:

STATUS OF EXPLORATION

Systematic study for hydrocarbon in Dhansiri valley was initiated with acquisition of gravity data in the year 1961 – 62 including aero – magnetic
survey. Subsequently, conventional seismic data acquisition was carried out in 1965 – 66, followed by CDP seismic coverage in 1975 – 76. The new
era of seismic data acquisition commenced with 3D seismic survey during 1987 – 88 field season. Till date, the different campaigns of seismic
investigation carried so far include conventional 2D, CDP survey and 3D seismic coverage. Satellite imagery is also available for analysis. Major play
in Dhansiri valley lies in the basement, sylhet, kopili, bokabil and some gas presence in shallow Tipam sands and Girujan lenses. The major
discoveries of fields are tabulated below:

Table 2.2: The major discoveries of fields.


S.NO Field Name Year of Discovery Potential pay zones
1 Borholla-Champang 1970 Basement, Tura, Sylhet, Girujan
2 Uriamghat 1988 Bokabil
3 Khoraghat 1989 Bokabil,Barail,Sylhet
4 Mekrang 1996 Sylhet
5 Nambar 1999 Bokabil
6 East Lakhibari 2002 Sylhet,Kopili,Namsang
7 Kalyanpur 2006 Middle Bokabil
8 Babejia 2008 Middle Bokabil

PRESENT STUDY

The South Assam Shelf is a part of the petroleum province of North East India. The area Khoraghat in SAS was selected as oil and gas producing
since 1970 and has a greater challenge as far as spatio – temporal distribution of reservoir sands and their prospect evaluation is concerned. The Assam
and Assam – Arakan basin is a polyhistoric basin as three tectonic styles viz. Extensional, compressional and resultant foredeep are present in this
basin. The concerned area comprises mostly sediments of extensional regimes superimposed by foredeep sediments adjacent to the compressive
tectonics related sediments to the east. Petroleum System: There are predominantly two petroleum systems existing in the Assam shelf. The oldest one
is the Paleocene-Eocene petroleum system. Source rock of this system is organic rich carbonaceous shales of Sylhet formation and matured shales of
Kopili from structurally down grabenal areas. The reservoir rocks are the clastic units of Tura and Sylhet. The interbedded shale units act as cap rocks
for entrapment of the oil. As regards geochemical analysis, the organic matter of this petroleum system is predominantly Type II, with subordinate
Type I materials. The second one is the Oligocene-Miocene petroleum system. In this particular system source and reservoir rocks are often
intercalated. Source facies occur within the coal-shale units of Barail group and matured shales of Kopili. The reservoir facies are principally the
sandstones of Barails, lenticular silty sands within Bokabils and fluvial sands within Tipams. The thick shales of Bokabils (mostly in South Assam
Shelf) and flood plain deposits of Girujan clay act as cap rocks. Burial and loading related to later compressional phase were responsible for the
cooking of hydrocarbons.

Future Exploration in Khoraghat area: The main reservoirs in Khoraghat area are found in Bokabil, Barail and Sylhet formations and the entrapment is
predominantly strati-structural in nature. Sizable Barail anomaly in paleogeomorphic high capped by Bokabil shales can be of good exploratory target.
Apart from that, tracing of existing hydrocarbon pool against west hading longitudinal fault towards southern and northern directions warrants
exploratory interests.

INTRODUCTION TO SOFTWARE (ELAN PLUS) - GEOFRAME

GeoQuest Software is an integrated business unit of Schlumberger Information Solutions (SIS). GeoFrame is the GeoQuest solution to the increasing
difficulty of discovering and developing viable oil & gas reserves. GeoFrame helps geoscientists and engineers with prospect interpretation and well

  21
planning by providing the efficiency and productivity of a single project database through the entire prospect life cycle. The GeoFrame system is both
a project database and framework for application development. GeoQuest software applications encompass all major areas of exploration and
development including geology, petrophysics, reservoir engineering, visualization, and project management. The GeoFrame open architecture is
designed to support the entire GeoQuest suite of integrated exploration and production applications, as well as client applications. With the GeoFrame
Developers Toolkit, application developers can incorporate the GeoFrame interface and database into their own applications. GeoFrame is also
designed to work with Finder Enterprise, GeoQuests master data storage solution. GeoQuest is fully committed to being an Open software provider.
GeoQuest has consistently worked towards helping to define and using data standards in the Oil & Gas industry. Therefore, many of the GeoFrame
applications have been designed to allow open access to the data contained in the GeoFrame projects (for instance, inter-operability between
GeoFrame and OpenWorks) utilizing Open Spirit software. The ORACLE database, which underlies all GeoFrame applications, is driven by a
standard user interface that provides graphical access to all data, full query capabilities, and a suite of user-friendly data management tools. The single
project database ensures that changes made in one application are immediately reflected in another. GeoFrame also provides Inter-Task
Communication (ITC) to allow two or more applications running in the same session to communicate with each other. This tight integration between
applications and the GeoFrame project database streamlines every interpretation workflow process.

PURPOSE OF ELAN PLUS APPLICATION: The ELANPlus computer program is designed for quantitative formation evaluation of cased and open-
hole log level by level. Evaluation is done by optimizing simultaneous equations described by one or more interpretation models. Single-well
ELANPlus can be run anytime after preliminary data editing (such as patching, depth matching, environmental correction) is complete. Most users
think the purpose of the ELANPlus application is solving the so called inverse problem, in which log measurements, or tools, and response parameters
are used together in response equations to compute volumetric results for formation components. In reality, that aspect of the program is only one side
of a three-way relationship among tools, response parameters, and formation component volumes. The relationship is often presented in a triangular
diagram, as in figure below.

 t  v
t = Rv


Figure3.1: Petrophysical model used by the ELANPlus application.

In the above figure, the t represents the tool vector—all logging instrument data and synthetic curves. The v is the volume vector, the volumes of
formation components. R is the response matrix, containing the parameter values for what each tool would read, given 100% of each formation
component. Given thedata represented by any two corners of the triangle, the ELANPlus program can determine the third. In the inverse problem, t
and R are used to compute v. As stated before, solution of the inverse problem is often considered the main job of the ELANPlus program.

The forward problem, also known as log reconstruction, uses R and v to compute t. A log reconstruction problem is computed for each inverse
problem, or Solve process. The reconstructed logs are compared against input data to determine the quality of volumetric results from the inverse
problem. Using t and v to compute R is called the calibration problem. Here the question is, “What response parameter value(s) should I use to obtain
the best fit between the observed logging instrument readings and some believed formation component volumes (often core results)?” Note that the
inverse problem solves for formation component volumes only. Other traditional log interpretation program results (such as water saturation, matrix
grain density, and so on) are provided by the Function process. That approach allows the program user to control the definitions of the additional
output types rather than having the definitions hard-coded in the program.

CROSSPLOTS

DENSITY – NEUTRON CROSSPLOT

The response of porosity logs can be expressed in ideal case by the simultaneous equations of the following form:
ΦN = f1 (ΦN1) + f2 (ΦN2) + f3 (ΦN3) + Φ (1)
ρB = f1 (ρ1) + f2 (ρ2) +f3 (ρ3) + Φ (2)
∆T = f1 ∆T1 + f2 ∆T2 + f3 ∆T3 + Φ (3)
Apply shale corrections as per the formula given below for PHIN and PHID prior to using the crossplot for determination of lithology and effective
porosity.
9 Calculate corrected PHID porosity by:

  22
ΦD,corr. = ΦD – ΦDSH VSH (4)
9 Calculate corrected PHIN porosity by:
ΦN,corr = ΦN – ΦNSH VSH (5)
The most widely used formula for calculating effective porosity from density and neutron is given by:
Φe = (2ΦNC + 7ΦDC) / 9 (6)
Porosity and rock matrix are variables in these equations. These equations can be solved for lithological mixtures and porosity. Graphical method
known as litho – porosity crossplot is used for obtaining the lithological mixture and porosity. Apparent porosity from density log and CNL log are
plotted in figure to obtain the lithologic mixture and porosity.

Figure 4.1: Bulk (Log) Density versus Neutron Porosity.

HINGLE CROSSPLOT

With the Hingle crossplot 1/ (Rt)1/m is plotted versus porosity. The straight line drawn through the water zone. Points gives Rw and the matrix
characteristics, i.e., ρma for density porosity, ∆tma for sonic porosity, ΦNma for neutron porosity.
Rt = F * Rw / Sw2 = (a Rw / Φm) (1 / Sw2) (7)
Φ = (a Rw)1/m * (1 / Rt Sw2)1/m (8)
For example with the density:
ρb = Φ ρmf + ( 1 – Φ) ρma
= ρma – (ρma – ρmf ) Φ
ρb = ρma – (ρma – ρmf ) (a Rw)1/m * (1 / Rt Sw2)1/m (9)
If Sw = 1, ρb or Φ is a linear function of (1 / Rt)1/m and for Rt = ∞, we have ρb = ρma. Equations similar to (2) can be written with ∆t and ΦN.

Properties of Hingle Plot:


9 It provides information on matrix parameters ( ρma, ΦNma, ∆Tma) and Ro provided the lithology is constant and Rw is also constant.
9 In clean formations, the interpretation results are reliable to a good approximation.
9 Use of neutron or density logs as porosity logs gives erroneous results in gas bearing formation.
9 Straight line corresponding to Sw = 1 is established by using the point Φ = 0 and Rt = ∞ as pivotal points and extending the line through the
most extreme north – westerly plotted points.
9 The slope of the line Sw = 1 provide value of Rw and that of the line Sxo =1 value of Rmf.
9 Straight line, for different value of Swcan be drawn by assuming a value of Φ and value of Rt corresponding to the desired value of Sw
obtained by substitution of assumed value of Φ and Swin the equation (2).

  23
9 The equation relating F and Φ (Archie or Humble etc) must be appropriately selected depending upon prevailing geological conditions to
make Hingle plot.
9 Assumption of values of n = 2, m = 2 and a = 1 restricts the application of this plot to certain specific geological conditions.

MANUAL INTERPRETATION

Depth ΦDSH ΦNSH


Interval RHOB PHID PHIN VSH VSH ΦDC ΦNC Crossplot Φ Lithology
3110 - 50 % Q,
3114 2.375 16.7 15 0.018 0.0485 0.149 0.102 0.129 50%C
3028 - 45% D,
3035 2.375 16.7 25.5 0.021 0.0582 0.146 0.1968 0.18 54%C
2676 - 98% C,
2683 2.225 21.8 25.5 1.76 4.85 20 20.65 20 2% D
2672 - 98% C,
2676 2.225 21.8 26.4 2.3 6.47 19.5 19.93 20 2% D
2665 - 16.66C,
2672 2.25 24.2 22.5 2 5.5 22.2 17 20.1 83.3 Q
2558 -
2565 2.35 18.2 27 4.8 13.25 13.4 13.75 13.5 100% C
2523 -
2528 2.35 18.2 27 5 13.9 13.2 13.1 13.1 100% C
2519 - 68.2% C,
2522 2.425 13.6 27 5.5 15.18 8.1 11.82 11 31.8 D
2515 - 68.2% C,
2518 2.41 14.5 27 5.3 14.53 9.2 12.47 11.8 31.8 D

QUICK LOOK INTERPRETATION

Manual interpretation has been done by recording the log values namely:
9 Gamma Ray (GR)
9 Resistivity (Rt)
9 RHOB
9 PHIN
9 Delta – T (∆T)

Against depths by using the following formulae and the results are tabulated below:
Sw =
Depth Interval GR Rt RHOB PHID PHIN DT Rwa √Rw/Rwa

3110 - 3114 45 60 2.375 16.7 15 67.5 0.45 71


3028 - 3035 49.5 9 2.375 16.7 25.5 90 0.612 61
2676 - 2683 45 5 2.225 21.8 25.5 87.5 0.2975 88
2672 - 2676 52.5 4.5 2.225 21.8 26.4 85 0.23 100
2665 - 2672 48 5 2.25 24.2 22.5 85 0.255 94
2558 - 2565 84 24 2.35 18.2 27 80 0.864 52
2523 - 2528 87 20 2.35 18.2 27 85 1.02 47
2519 - 2522 93 19 2.425 13.6 27 82.5 0.827 53
2515 - 2518 90 18 2.41 14.5 27 82.5 0.783 54

Determination of Rwa (apparent resistivity of the formation water) method

The apparent resistivity of the formation water (Rwa) can be calculated so as to find the minimum Rwa in aquifer beds. Indeed:
Rw =( Rt / F) * Sw2
and Rwa = ( Rt / F) ( assuming Sw = 1)
In eastern region ‘F’ is calculated from sonic porosity transform (this transform has been generated over the years by using the sonic travel time and
the core porosity). When several aquifer intervals have the same value of (Rwa)min, there is a fair amount of certainty that this value corresponds to Rw.
Calculate water saturation which is given by: Sw = √Rw/ Rwa.

  24
Table 4.1: Sonic porosity calculated using ERBC Transformation.

RW DETERMINATION FROM CROSSPLOTS

Hingle Crossplot

Rw was calculated in Khoraghat field of South Assam Shelf and following observations were noted:
i. Rw in Kopili formation: 0.26 ohm – m
ii. Rw in Sylhet formation: 0.27 ohm – m

Well # B:

Sylhet formation - 2867 – 2964

Depth Interval GR Rt RHOB NPHI DT

2945 - 2954 63 30 2.6 7.5 60


2938 - 2945 48 35 2.6 7.5 55
2928 - 2935 48 20 2.65 9 60
2923 - 2928 45 20 2.55 7.5 60
2905 - 2923 45 30 2.6 6 55
2896 - 2905 45 15 2.45 15 65
2883 - 2891 48 15 2.35 37.5 60

To estimate Rw, use the following equation, substituting Φ and Rt from a point at which Sw = 100% we have,
Rw = ( Rt * Φm) / a
For Sylhet formation we take the values of a = 1 and m = 2, Rw = 7 * (0.185)2 = 0.24 ohm – m.

  25
Well # A

Kopili formation - 2845 – 3105

Depth Interval GR Rt RHOB NPHI DT

3035 - 3028 47 8 2.375 22.5 90


2938 - 2932 17 9 2.25 19.5 85
2925 - 2920.5 30 7 2.31 21 82.5
2991 - 2987 60 7 2.275 27 85

To estimate Rw, use the following equation, substituting Φ and Rt from a point at which Sw = 100% we have,
Rw = ( Rt * Φm) / a
For Kopili formation we take the values of a = 0.81 and m = 2
Rw = ( 7 * (0.22)2) / 0.81 = 0.41 ohm – m.

  26
Indonesian Equation

This equation is used when high quantity of shale is present in the reservoir rock. It revolves around the shale parameters such as shale volume Vsh,
shale conductivity (Csh), shale density (ρsh) and porosity (Φ) which enabled to quantify the effect of shale on log measured values while evaluating the
shaly sands. Vsh is defined as the volume of the wet shale per unit volume of the reservoir rock. The procedure essentially begins with the estimation
of Vsh by using different logs individually or in combination as clay indicators. A good number of Vsh models have evolved. The notable features of
these models are that the mathematical expression used for each Vsh model contains two pertains:
- One pertains to clean sand matrix
- And the other is known as clay term

While sand term obeys Archie’s equation, the shale term is defined by Vsh and Rsh values. The most widely used equation is known as Indonesian
equation suggested by Poupon et. al. (1971).
(1/Rt) = [( Vclay (1- Vclay) / √ Rclay) + (Φm/2 / √a * Rw)] * Swn/2
Where,
m = n = 2 and the value of ‘a’ varies from 0.62 to 1.
Using this equation the reservoir parameters for Well # A in the Sylhet formation are calculated.

OBSERVATIONS MADE FROM PROCESSED OUTPUT

WELL # A

Bokabil Section: In the Bokabil section sanstones developed are having an average resistivity of the order of 10 – 20 ohm – m, and gamma ray of 80 –
90 API. The sands developed in the intervals 2523 – 2527.5m and 2557.5 – 2566m are interpreted to be hydrocarbon bearing, with average effective
porosity of 24% and average water saturation of 45% for the first interval and average porosity of 20 - 24 % and average water saturation of 40 %
respectively. The processed results are presented in the form of Elan Plus output.

Sylhet Section: Sylhet section is mainly characterised by well developed, massive, tight limestone in the interval from 3201 – 3126m, not interesting
from hydrocarbon point of view. The sandy limestone / limestone developed in the interval 3105 – 3112m (top of Sylhet) are also tight and not
interesting. The interval from 3112 – 3126m is separately discussed below.

Interval: 3112 – 3126m: This layer is mainly sandstone having an average true resistivity of 20 – 40 ohm –m, is cleaner at the top 5m i.e., from 3112 –
3117m compared to the remaining interval, which is an alternation of shale and sandy limestone. The top 5m has an average effective porosity of 20 -
22% and average water saturation of 30 - 35% is interpreted to be hydrocarbon bearing. Neutron – Density (ND) are not reliable against this interval

  27
due to bad borehole (caving). Sonic log was not affected even though the hole is caved. Hence water saturation (Sw) calculated from sonic log gave
good results when compared to water saturation (Sw) calculated from Neutron – Density (ND) logs. Therefore Elan processing in the Sylhet could not
be carried out. The water saturation (Sw) was therefore calculated by using Indonesian equation developed in Microsoft Excel which is shown below.

Figure 5.1: Indonesian solution to Sylhet section.

Stratigraphic Sequence of Well # A

S.NO STRATIGRAPHIC UNIT NAME INTERVAL (M) LITHOLOGY

1 POST TIPAM Surface to 1123 CLAY/CLAYSTONE/SANDSTONE

2 TIPAM 1123 – 2002 SANDSTONE/SHALE

3 BOKABIL 2002 – 2660 CLAY/CLAYSTONE/SILTSTONE/S


ANDSTONE

4 BARAIL 2660 – 2845 SHALE/SANDSTONE

5 KOPILI 2845 – 3105 SHALE/SANDSTONE

6 SYLHET 3105 – 3201 LIMESTONE/SANDSTONE

7 BASAL SANDSTONE 3201 – 3227.5 SANDSTONE/SHALE

8 BASEMENT 3227.5 - 3283 GRANITE

  28
WELL #K

Bokabil Section: In the Bokabil section sanstones developed are having an average resistivity of the order of 15 – 20 ohm – m, and gamma ray of 90 –
100 API. The sands developed in the intervals 2464 – 2476m is interpreted to be hydrocarbon bearing, with average effective porosity of 22 – 25%
and average water saturation of 40 – 50%. The gas – oil contact (GOC) is at a depth of 2472m (indicated by red color) and water – oil contact (WOC)
is at a depth of 2476m. And hence the oil is in the interval of 2472 – 2476m. The processed results are presented in the form of Elan Plus output.

Stratigraphic Sequence of Well # K

S.NO STRATIGRAPHIC UNIT NAME INTERVAL (M) LITHOLOGY

1 POST TIPAM Surface – 1325 CLAY/CLAYSTONE/SANDSTO


NE

2 TIPAM 1325 – 2184 SANDSTONE/SHALE

3 BOKABIL 2184 – 2553 CLAY/CLAYSTONE/SILTSTON


E/SANDSTONE

4 BARAIL 2553 – 2622 SHALE/SANDSTONE

5 KOPILI Not drilled ------

6 SYLHET Not drilled ------

7 BASAL SANDSTONE Not drilled -----

8 BASEMENT Not drilled -----

  29
WELL # C

Bokabil Section

In the Bokabil section sanstones developed are having an average resistivity of the order of 15 – 18 ohm – m, and gamma ray of 105 –120 API. The
sands developed in the intervals 2602 – 2611m is interpreted to be hydrocarbon bearing, with average effective porosity of 20 – 25% and average
water saturation of 58 – 64%. The processed results are presented in the form of Elan Plus output. The entire sand is hydrocarbon bearing. This is no
OWC in this particular zone. There is a oil – shale contact at a depth of 2611m. The processed results are presented in the form of Elan Plus output.

Stratigraphic Sequence of Well # C

S.NO STRATIGRAPHIC UNIT NAME INTERVAL (M) LITHOLOGY


1 POST TIPAM Surface – 1313 CLAY/CLAYSTONE/SANDSTONE

2 TIPAM 1313 – 2181 SANDSTONE/SHALE

3 BOKABIL 2181 – 2674 CLAY/CLAYSTONE/SILT


STONE/SANDSTONE

4 BARAIL 2674 – 2955 SHALE/SANDSTONE

5 KOPILI 2955 – 3035 SHALE/SANDSTONE

6 SYLHET Not drilled ------

7 BASAL SANDSTONE Not drilled -----

8 BASEMENT Not drilled -----

  30
Well # L

Barail section: The well is drilled upto 2653m in the barail formation. The drilled section of the barail formation is predominantly sand some siltstone
and claystone. Of the several sands developed in barail five sand intervals 2478.5 – 2496m, 2499 – 2501.5m, 2508 – 2511m, 2523 – 2534m and 2536
– 2539m are hydrocarbons bearing. The topmost sand in the interval 2478.5 – 2496m is both oil and gas bearing with GOC at 2488.5m. The gas
bearing interval 2478.5 – 2488.5m is having resistivity 50 – 190 ohm –m, porosity 17 – 22% and water saturation 16 – 31%. The bottom interval
2488.5 – 2496m is oil bearing with average resistivity of 35 – 45 ohm – m, porosity 18 – 24% and water saturation 17 – 45%. The formation from
2496 – 2499m is very shaly with resistivity 9 – 11 ohm – m with volume of shale 80 – 90%. The sand interval from 2499 – 2501.5m is oil bearing
with resistivity 10 ohm – m, porosity 11 – 12% and water saturation 58 – 65%. The formation from 2501.5 – 2508m is shaly with resistivity 6 – 7 ohm
– m and volume of shale from 80 – 90%. The sand interval from 2508 – 2511m is interpreted to be oil bearing with resistivity 40 – 48 ohm – m,
porosity 20 – 27% and water saturation 25 – 43%. The interval 2523 – 2534m is an ensemble of thin sand shale alterations with better porosity
development towards top. This interval is having a resistivity of 11 – 20 ohm – m, porosity 18 – 20% and water saturation 40 – 51%. The interval is
2536 – 2539m is clean and homogeneous and interpreted to be oil bearing with resistivity 25 – 37 ohm – m, porosity 23 – 25% and water saturation 31
– 41%. The sand interval 2543 – 2547m is having comparatively less resistivity of 5 – 7 ohm – m, porosity 20 – 23% and water saturation 70 – 80%.
This sand is interpreted as water bearing. The processed results are presented in the form of Elan Plus output.

Stratigraphic Sequence of Well # L

S.NO STRATIGRAPHIC UNIT NAME INTERVAL (M) LITHOLOGY


1 POST TIPAM Surface to 1204 CLAY/CLAYSTONE/SAND STONE

2 TIPAM 1204 - 1961 SANDSTONE/SHALE

3 BOKABIL 1961 – 2478.5 CLAY/CLAYSTONE/SILTSTONE/SANDSTONE

4 BARAIL 2478.5 – 2653 SHALE/SANDSTONE

5 KOPILI Not drilled ------

6 SYLHET Not drilled ------

7 BASAL SANDSTONE Not drilled -----

8 BASEMENT Not drilled -----

  31
Well # J

Bokabil formation: The bokabil formation is encountered in the interval 1585 – 2324m. Lithologically it comprises of siltstone, claystone and
sandstone. In the lower part of this section, the sandstone layers are having resistivity of the range 7 – 10 ohm – m except in the interval 2023 – 2029m
where it is of the order of 12 ohm – m. The interval 2023 – 2029m is interpreted to be hydrocarbon bearing with effective porosity of 18 – 20% and
water saturation of 35 – 45%. In the upper part of this section, the sandstone layers in general are having resistivity 8 – 14 ohm – m with high sonic
time and are not interesting from hydrocarbon point of view. The processed results are presented in the form of Elan Plus output.

Stratigraphic Sequence of Well # J

S.NO STRATIGRAPHIC UNIT NAME INTERVAL (M) LITHOLOGY


1 POST TIPAM Surface to 941 CLAY/CLAYSTONE/SAND STONE

2 TIPAM 941 – 1455 SANDSTONE/SHALE

3 BOKABIL 1455 – 2502 CLAY/CLAYSTONE/SILTSTONE/SANDSTONE

4 BARAIL 2502 – 2605 SHALE/SANDSTONE

5 KOPILI 2605 – 2820 SANDSTONE/SHALE

6 SYLHET 2820 – 2893 LIMESTONE/SANDSTONE/ SHALE

7 BASAL SANDSTONE 2893 – 2914 SANDSTONE

8 BASEMENT 2914 – 2971 GRANITE

  32
Well # H

Bokabil formation: Bokabil formation is developed in the interval 1410 – 2148m. It consists of clay and sandstone layers and some of the sandstone
layers are hydrocarbon bearing. The sand pack developed in the interval 1993 – 2001m is having good porosity of 20 – 25% and characterised by
resistivity of 10 – 12 ohm – m. The sand is evaluated as hydrocarbon bearing with shaliness 25 – 50% and oil saturation from 30 - 50%. Oil – shale is
observed at 2001m.

Stratigraphic Sequence of Well # H:

S.NO STRATIGRAPHIC UNIT NAME INTERVAL (M) LITHOLOGY

1 POST TIPAM Surface to 730 CLAY/CLAYSTONE/SAND STONE

2 TIPAM 730 - 1410 SANDSTONE/SHALE

3 BOKABIL 1410 - 2148 CLAY/CLAYSTONE/SILTSTONE/SANDSTONE

4 BARAIL 2148 - 2177 SHALE/SANDSTONE

5 KOPILI 2177 - 2319 SANDSTONE/SHALE

6 SYLHET 2319 - 2399 LIMESTONE/SANDSTONE/ SHALE

7 BASAL SANDSTONE Not drilled --------------

  33
8 BASEMENT Not drilled --------------

DISCUSSION AND CONCLUSION

¾ For identifying hydrocarbon zones it is necessary to find out the resistivity of the formation water (Rw). Crossplots have been generated (Hingle)
taking porosity on abscissa and true formation resistivity on ordinate for eight wells. The average Rw value has been taken as an input for Elan
processing.
¾ Elan processing has been extensively used for determination of reservoir parameters of the wells of Khoraghat field. The output from this
processing provided a lithology distribution, oil, gas and water. Thicknesses of gas sands, oil sands and water sands have been determined. Gas –
oil contacts (GOC) and Oil – Water contacts (OWC) have also been identified. The output is provided in coloured plot.
¾ The major oil producers of Khoraghat field are Barail sands and Lower Bokabil sands. The porosities are in the range of 20 – 25% and oil
saturation of 60 – 70%. They are good producers of hydrocarbons.
¾ Hydrocarbons are also encountered in few wells in Sylhet, Kopili and Middle Bokabil. These reservoirs have also got good petrophysical
properties namely porosity (15 – 20%), hydrocarbon saturation (50 – 60%). They are moderate producers of hydrocarbons.
¾ In most of the wells Middle Bokabil shale and Kopili shale has acted like a cap rock for hydrocarbon accumulation. Middle Bokabil shale is
acting as a cap rock for hydrocarbon accumulation in Lower Bokabil sands and Barail Sands.
¾ Kopili shale is acting as a cap rock for hydrocarbon accumulation in Sylhet and sands within Kopili.
¾ Mostly faults are acting as a seal for oil entrapment. There is no clear anticlinial feature in Khoraghat field.

ACKNOWLEDGEMENT

We are indebted to the ONGC-Assam for providing data for this work.

REFERENCES

[1] Encyclopedia of Well Logging – by Robert DESBRANDES.

  34
[2] PRATICAL FORMATION EVULATION by ROBERT C. RANSOM.
[3] GEOPHYSICAL WELL LOGGING – by J.P.VAISH.
[4] Archie, G.E.: The Electrical Resistivity Log as an Aid in Determining Some Reservoir Characteristics,Trans., AMIE(1942), 146; 54 – 62.
[5] Wyllie, M.R.J.: A Quantitative Analysis of the Electrochemical Component of the S.P. Curve, Trans., AIME, (1949), pp. 17 – 26.
[6] Creighton, H.J.: Principles and Application of Electrochemistry, Vol. 1, Principles, 4th ed., Wiley, New York, (1943).
[7] Doll, H.G.: The SP Log in Shaly Sands, Trans., AIME (1950), 189; 205 – 14.
[8] Hilchie, Dougles W.: Applied Openhole Log Interpretation, Dougles W. Hilchie, Inc., Golden, CO (1978).
[9] Schlumberger Well Services; Log Interpretation, Volume 1 – Principles (1972).
[10] Wahl, J.S., and Johnstone, C.W.: The Dual Spacing Formation Density Log J.Pet. Technol., AIME, (Dec., 1964), pp. 1411 – 1416.
[11] Wyllie, M.R.J., Gregory, A.R., and Gardner, G.H.F.: An Experimental Investigation of Factors Affecting Elastic Wave Velocities in Porous
Media, Geophysics (July, 1958), vol.23.
[12] Hingle, A.T.: Use of Logs in Exploration Problems, 29th International Annual Meeting SEG, Log Angeles, CA (Nov., 1959).
[13] Waxman, M.H. and Smits, L.J.M.: Electrical Conductivities in Oil – Bearing Shaly Sands, SPE J. (June, 1968), pp. 107 – 122.
[14] Ransom R.C.: Methods Based on Density and Neutron Well Logging Responses to Distinguish Characteristics of Shaly Sandstone Reservoir
Rock. The Log Analyst, SPLWA (May – June, 1977), pp. 47 – 62.
[15] Ransom R.C.: A Contribution Towards a Better Understanding of the Modified Archie Formation – Resistivity – Factor Relationship, The Log
Analyst, SPWLA (March – April 1948), pp. 7 – 12.
[16] Burke, J.A., Campbell, R.L. Jr., and Schmidt, A.W.: The Litho – Porosity Crossplot, The Log Analyst, SPWLA (Nov – Dec., 1969), pp. 25 – 43.
[17] Poupon, A., Clavier, C., Dumanoir, J., Gaymard, R., and Misk, A.: Log Analysis of Sand – Shale Sequences – A Systematic Approach, J. Pet.
Technol. (July, 1970).

  35
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Selecting attribute to standing out in the competitive objects


Murlidher Mourya and K.V Raghavender
Department of Computer Science and Engineering, JNTUH University.
TKR College of Engg & Technology
murli_cool9@yahoo.com, raghukotl@gmail.com

ABSTRACT
Mining of frequent item sets is one of the most fundamental problems in data mining applications. My proposed algorithm which guide the seller to
select the best attributes of a new product to be inserted in the database so that it stands out in the existing competitive products, due to budget
constraints there is a limit, say m, on the number of attribute that can be selected for the entry into the database. Although the problems are NP-
complete. The Approximation algorithm are based on greedy heuristics. My proposed algorithm performs effectively and generates the frequent item
sets faster.

KEYWORDS: Data mining, Mining frequent itemsets, Marketing.

INTRODUCTION:

In recent years there has been development of ranking functions and efficient top-k retrivel algorithms which help the users in mining Frequent items
set which plays a major role in many data mining applications .examples include: users wishing to search databases and catalogs of products such as
homes, cars, cameras, or articles such as news and job ads. Users browsing these databases typically execute search queries via public front-end
interfaces to these databases. Typical queries may specify sets of keywords in case of text databases or the desired values of certain attributes in case
of structured relational databases. The query-answering system answers `such queries by either returning all data objects that satisfy the quer
conditions, or may rank and return the top-k data objects, or return the results that are on the query’s skyline. If ranking is employed, the ranking may
either be simplistic—e.g., objects are ranked by an attribute such as Price; or more sophisticated—e.g., objects may be ranked by the degree of
“relevance” to the query.

Attributes selection: There are two types of users of these databases. Buyers of products who search such database trying to locate objects of interest
,while the latter type of user are sellers of products who insert new objects into these databases in the hope that they will be easily discovered by the
buyers i.e it must stands out in the existing competitive products. To understand it a little better consider the following scenario :If a real estate seller
wants to give an add on the news paper about sale of flats, He has to choose the best features of the flats, that are the most of the customers are
interested. If he has given an add with some features (or attributes), and if no customer is interested on those features, then the add may not add value
to his advertisement.If he has a system, that can suggest top_k attributes (or features) of the product, then he can give a very good add, and that add
will be referred by more number of customers. General problem also arises in domains beyond e-commerce applications. For example, in the design of
a new product, a manufacturer may be interested in selecting the 10 best features from a large wish-list of possible features—e.g., a homebuilder can
find out that adding a swimming pool really increases visibility of a new home in a certain neighborhood.
The problem here is selecting the proper and the best attributes of the flats, to give a good advertisement that is more number of customers are
interested.

To define our problem more formally, we need to develop a few abstractions. Let D be the database of products already being advertised in the
marketplace (i.e.,the “competition”). let Q be the set of search queries that have been executed against this database in the recent past—thus Q is the
“workload” or “query log.” The query log is our primary model of what past potential buyers have been interested in.For a new product that needs to
be inserted into this database, we assume that the seller has a complete “ideal”description of the product.But due to budget constraints, there is a limit,
say m, on the number of attributes/keywords that can be selected for entry into the database. Our problem can now be defined as follows.
Problem. Given a database D, a query log Q, a new tuple t, and aninteger m, determine the best (i.e., top-m) attributes of t to retainsuch that if the
shortened version of t is inserted into thedatabase, the number of queries of Q that retrieve t is maximized.

PRELIMINARIES

First we provide some useful definitions

Boolean database. Let D = {t1 . . . tN} be a collection of Boolean tuples over the attribute set A = {a1 . . . aM}, where each tuple t is a bit-vector
where a 0 implies the absence of a feature and a 1 implies the presence of a feature. A tuple t may also be considered as a subset of A, where an
attribute belongs to t if its value in the bit-vector is 1.

Database D

Car AC Four Turbo Power Auto Power


ID Door Doors Trans Brakes

T1 0 1 0 1 0 0
T2 0 1 1 0 0 0
T3 1 0 0 1 1 1
T4 1 1 0 1 0 1
T5 1 1 0 0 0 0
T6 0 1 0 1 0 0
T7 0 0 1 1 0 0

Query Log Q
Query AC Four Turbo Power Auto Power
ID Door Doors Trans Brakes

Q1 1 1 0 0 0 0
Q2 1 0 0 1 0 0
Q3 0 1 0 1 0 0
Q4 0 0 0 1 0 1
Q5 0 0 1 0 1 0

New AC Four Turbo Power Auto Power


Door Doors Trans Brakes
Car

T 1 1 0 1 1 1

New tuple t to be inserted

Tuple domination. Let t1 and t2 be two tuples such that for all attributes for which tuple t1 has value 1, tuple t2 also has value 1. In this case, we say
that t2 dominates t1.
Tuple compression. Let t be a tuple and let t’ be a subset of t with m attributes. Thus, t’ represents a compressed representation of t. Equivalently, in
the bit-vector representation of t, we retain only m 1s and convert the rest to 0s. Query log. Let Q = {q1 . . . qS} be collection of queries where each
query q defines a subset of attributes. The following running example will be used throughout the paper to illustrate various concepts

Conjunctive Boolean Retrieval: We view each query as a conjunctive query. A tuple t satisfies a query q if q is a subset of t. For example, a query such
as {a1, a3} is equivalent to “return all tuples such that a1= 1 and a3 = 1”. Alternatively, if we view q as a special type of “tuple”, then t dominates q.
The set of returned tuples R(q) is the set of all tuples that satisfy q.
In this problem variant as well as in most of the variants defined later, our task is to compress a new tuple t by retaining the best set of m attributes
(i.e., top-m attributes) such that some criterion is optimized.

38
PROBLEM DEFINATION
Conjunctive Boolean - Query Log (CB-QL): Given a query log Q with Conjunctive Boolean Retrieval semantics, a new tuple t, and an integer m,
compute a compressed tuple t′ having m attributes such that the number of queries that retrieve t′ is maximized.
Intuitively, for buyers interested in browsing products of interest, we wish to ensure that the compressed version of
the new product is visible to as many buyers as possible.

EXAMPLE
To illustrate this problem variant, consider the example in which shows a new tuple t that needs to be inserted into the database. Suppose we are
required to retain m = 3 attributes. It is not hard to see that if we retain the attributes AC, Four Door, and Power Doors (i.e., t′ = [1, 1, 0, 1, 0, 0]), we
can satisfy a maximum of three queries (q1, q2, and q3). No other selection of three attributes of the new tuple will satisfy more queries.

This problem also has a per-attribute version where m is not specified; in this case we may wish to determine t′ such that the number of satisfied
queries divided by |t′| is maximized. Intuitively, if the number of attributes retained is a measure of the cost of advertising the new product, this
problem seeks to maximize the number of potential buyers per advertising dollar. Notice that in CB-QL, it is the query log Q that needs to be analyzed
in solving the problem; the actual database D (i.e., the “competing products”) is irrelevant.

NP – COMPLETENESS RESULTS

Theorem 1: The decision version of CB-QL problem is


NP-hard. Proof: An instance of decision version of CB-QL is similar to an instance of CB-QL, except that it has in addition a target parameter X, and
is satisfied if there is a compressed tuple t′ having m attributes such that the number of queries that retrieve t′ is at least X. Clearly, the decision version
of CB-QL is in NP. To prove that it is NP-complete, we reduce from the Clique problem. Given a graph G = (V, E) and an integer r, the task in the
Clique problem is to check if there is a clique of size r in G.
We transform this to an instance of decision version of CB-QL as follows. The attribute set A will be V, and the query log will contain one row for
each edge. If e = (u, v) is an edge, then the query log Q contains the conjunctive query {u, v}, i.e., the query retrieving all tuples with u = 1 and v = 1.
The new tuple t has all the attributes in V set to 1. Let m = r and X = m(m- 1)/2. Now t has a compressed representation with m attributes that satisfies
X number of queries if and only if the graph has a clique of size r.

ALGORITHMS FOR CONJUNCTIVE BOOLEAN WITH QUERY LOG

Optimal Brute Force Algorithm Clearly, since CB-QL is NP-hard, it is unlikely that any optimal algorithm will run in polynomial time in the worst
case. The problem can be obviously solved by a simple brute force algorithm (henceforth called BruteForce-CB-QL), which simply considers all
combinations of m-attributes of the new tuple t and determines the combination that will satisfy the maximum number of queries in the query log Q.
However, we are interested in developing optimal algorithms that work much better for typical problem instances.

Optimal Algorithm Based on Integer Linear


Programming
We next show how CB-QL can be described in an ILP framework. Let the new tuple be the Boolean vector t ={a1(t),…..,am(t)}, S be the total number
of queries in the query log, and let x1; . . . ; xM be integer variables such that if ai(t)= 1, then xi €{0,1},else xi =0. Consider the tasks m
Maximize ∑ ∏ xj subject to ∑ xj ≤m
I=1 aj aqi=1 j=1
It is easy to see that the maximum gives exactly the solution to CB-QL. The objective function is not linear, however, and thus, we next show how this
can be achieved. We introduce additional 0-1 integer variables y1,…, yS, i.e., one variable for each query in Q. The formulation is
s
Maximize ∑ yi, subject to
i=1
m
∑ xj ≤m and yi ≤xj
J=1
for each j and i such that ajqi = 1:

yi ≤ xj actually means yi =xj =1 if yi =1, that is, if qi is satisfied. Thus, the variable yi corresponding to a query can be 1 only if all the variables xj
corresponding to the attributes in the query are 1. This implies that the maximum remains the same. We refer to the above algorithm as ILPCB-QL.
The integer linear formulation is particularlyattractive as unlike more general IP solvers, ILP solvers are usually more efficient in practice.

Selecting the threshold value:

There are two alternate approaches to setting the threshold. One approach is essentially a heuristic, where we set the threshold to a reasonable fixed
value dictated by the practicalities of the application. threshold enforces that attributes should be selected such
that the compressed tuple is satisfied by a certain minimum number of queries. For example, a threshold of 1 percent means that we are not interested
in results that satisfy less than 1 percent of the queries in the query log.

Two-phase random walk to compute maximal frequent item sets

A maximal frequent item set is a frequent item set such that none of its supersets are frequent. The set of maximal frequent item sets are much smaller
than the set of all frequent item sets. For example, if we have a dense table with M attributes, then it is quite likely that most of the maximal frequent

39
item sets will exist very high up in the Boolean lattice over the attributes, very close to the highest Possible level M. Fig. 2 shows a conceptual
diagram of a Boolean lattice over a dense Boolean table _Q. The shaded region depicts the frequent item sets and the maximal frequent item sets are
located at the highest positions of the border between the frequent and infrequent item sets.
There exist several algorithms for computing maximal frequent item sets, e.g.two-phase random walk based algorithm in , which
starts from a random singleton item set I at the bottom of the lattice, and at each iteration, adds a random item to I (from among all items A-I such that
I remains frequent), until no further additions are possible. At this point, a maximal frequent item set I has been discovered. If the number of maximal
frequent item sets is relatively small, this is a practical algorithm: repeating this random walk a reasonable number of times will, with high probability,
discover all maximal frequent item sets. However, since this
algorithm is based on traversing the lattice from bottom to top, it implies that the random walk will have to traverse a lot of levels before it reaches a
maximal frequent item set of a dense table.
Two-phase random walk can be divided into two phases: 1) Down Phase:
starting from the top of the lattice (I = {a1a2 . . . aM}), walk down the lattice by removing random items from I until I becomes frequent, and 2) Up
Phase: starting from I, walk up the lattice by adding random items to I (from among all items A _ I such that I remains frequent), until no further
additions are possible. At this point, a maximal frequent item set I has been discovered.
Figure shows an example of the two phases of the random walk. What is important to note is that this process is much more efficient than a bottom-up
traversal, as our walks are always confined to the top region of the lattice and we never have to traverse too many levels. Complementing the query log
eventually results in a dense data set. In a dense data set, maximal frequent item sets are usually at the top of the lattice. That is, they are close tolevel
M _ m, where m is comparatively much smaller than
M. If a bottom-up approach is used to find maximal frequent item sets, it will have to traverse a long portion of the lattice (i.e., too many levels) and
will be inefficient. Whereas, in top-down approach, the first phase tries to find the first frequent item set along the path from the top which is usually
close to a maximal frequent item set, the walks are confined to the top region of the lattice and we never have to traverse too many levels..

Fig : Two-phase random walk.

Mining tree algorithm

The main idea of the my algorithm is to maintain the frequent itemsets tree of the database.the algorithm involves two phase.in phase -1 it scan the
database and and derives the set of frequent items is stored in the order of descending support count. First we create a root of the tree with a labeled
with “null”.the items in each transaction in processed and a branch is created.In phase-2 we not use the database anymore but we uses the mining tree
which contains all the information about frequent itemsets.

Algorithm TwoPhase-Random-Walk(~Q, r)
execute Down Phase random walk
execute Up Phase random walk
return itemset reached after Up Phase

Algorithm Mining Tree: a compact form of DB/sub-DB


Key steps: FI(N,N-tree)
if N-tree is a single path
then a possible fi is found
else { extend N with x∈FX(N)
construct Nx-tree
FI(N∪{x},Nx-tree)}

Mining tree vs. Two phase Random walk:

100 Data set T25I20D10K


90
80

70
Run time(sec.)

60

50
40

30

20
10

0
0 0.5 1 1.5 2 2.5 3
Support threshold(%)

29

40
Mining tree vs.two phase random walk:
Scalability with Support Threshold

Data set T25I20D100K


140

120

100

Runtime (sec.)
80

60

40

20

0
0 0.5 1 1.5 2
Support threshold (%)
30

RELATED WORK
A large corpus of work has tackled the problem of ranking the results of a query. In the documents world, the most popular techniques are tf-idf based
ranking functions, like BM25, as well as link-structure-based techniques like Page Rank if such links are present (e.g., the Web). In the database
world, automatic ranking techniques for the results of structured queries have been proposed. Also there has been recent work on ordering the
displayed attributes of query results.

Both of these tuple and attribute ranking techniques are inapplicable to our problem. The former inputs a database and a query, and outputs a list of
database tuples according to a ranking function, and the latter inputs the list of database results and selects a set of attributes that “explain” these
results. In contrast, our problem inputs a database, a query log, and a new tuple, and computes a set of attributes that will rank the tuple high for as
many queries in the query log as possible.
Although the problem of choosing attribute is related to the area of feature selection,our work differs from the work on feature selection because
our goal is very specific—to enable a new tuple to be highly visible to the database users.and not to reduce the cost of building a mining model such as
classification or clustering.

CONCLUSIONS

While the problems considered in this paper are novel and important to the area of ad hoc data exploration and retrieval, we observe that our specific
problem definition does have limitations. After all, a query log is only an approximate surrogate of real user preferences, and moreover, in some
applications neither the database, nor the query log may be available for analysis; thus, we have to make assumptions about the nature of the
competition as well as about the user preferences. Finally, in all these problems, our focus is on deciding what subset of attributes to retain of a
product

REFERENCES

D. Burdick, M. Calimlim, and J. Gehrke, (2001)“MAFIA: A Maximal Frequent Item Set Algorithm for Transactional Databases,” Proc. Int’l Conf.
Data Eng. (ICDE), 2001.
M.R. Garey and D.S. Johnson (1979), “Computers and Intractability: A Guide to the Theory of NP-Completeness”.
W.H. Freeman, K. Gouda and M.J. Zaki, (2001) “Efficiently Mining Maximal Frequent Itemsets,” Proc. Int’l Conf. Data Mining (ICDM),
J. Han, J. Pei, and Y. Yin, (2000) “Mining Frequent Patterns without Candidate Generation,” Proc. SIGMOD Conf., pp. 1-12,.
J. Han, J. Wang, Y. Lu, and P. Tzvetkov, (2002) “Mining Top-k Frequent Closed Patterns without Minimum Support,” Proc. Int’l Conf. Data Mining
(ICDM),.
M.D. Morse, J.M. Patel, and H.V. Jagadish, (2007) “Efficient Skyline Computation over Low-Cardinality Domains,” Proc. Int’l Conf.Very Large
Data Bases (VLDB),.
M. Miah, G. Das, V. Hristidis, and H. Mannila, (2008) “Standing Out in a Crowd: Selecting Attributes for Maximum Visibility,” Proc. Int’l Conf.
Data Eng. (ICDE), pp. 356-365, 2008.

41
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Image Segmentation Using both Color and Texture


Sreevani Raichur1 and K.Rama naidu2
Electronics and Communication engg., JNTU
Anantapur,Andhra Pradesh, INDIA
E-mail:sreevani.raichur@gmail.com

ABSTRACT

This paper is about the development of image segmentation frame work (referred to as C- Tex) that is based on the adaptive inclusion of Color and
texture in the process of data partition. An important contribution of this work is the extraction of Color features that evaluates the input image in
multi space representation. To achieve this have used anRGB image, converted into an HSV image and its characteristics, where the key component
was the inclusion of K-means clustering in the computation of dominant colors, and the Estimation of optimal number of clusters in the image. The
texture features are also computed using texture decomposition Scheme based on K- means clustering algorithm. The Major contribution of this work
resides in the integration of color and texture features. This can also be performed by using K-means clustering algorithm . The experimental results
indicate that the proposed technique is accurate in capturing the color and texture characteristics when applied to images.

KEY WORDS: K means clustering, color and texture segmentation.

INTRODUCTION

IMAGE segmentation is one of the most investigated subjects in the field of computer vision since it plays a crucial role in the development of high-
level image analysis tasks such as object recognition and scene understanding. A review of the literature on image segmentation indicates that a
significant amount of research has been dedicated to the development of algorithms where the colour and texture features where analysed alone [1]-
[4]. The early colourtexture segmentation algorithms were designed in conjunction with particular applications [5], [6] and they were generally
restricted to the segmentation of images that are composed of
scenes defined by regions with uniform characteristics. Segmentation of natural images is by far a more difficult task, since natural images exhibit
significant inhomogeneities in colour and texture [7], [8]. Thus, the complex characteristics associated with natural images forced researchers to
approach their segmentation using features that locally sample both the colour and texture attributes [9], [10]. The use of colour and texture
information collectively has strong links with the human perception, but the main challenge is the combination of these fundamental image attributes
in a coherent colourtexture image descriptor [11]-[16]. In fact, if we take into consideration that the textures that are present in natural images are often
characterised by a high degree of complexity, randomness and irregularity, the simple inclusion of colour and texture is not sufficient and a more
appropriate segmentation model would encompass three attributes such as colour, texture and composite elements that are defined by a large mixture
of colours. Although this model is intuitive, there is not a generally accepted methodology to include the texture and colour information in the
segmentation process. In this regard, Mirmehdi and Petrou [17] approached the segmentation of colour images from a perceptual point of view. In
their paper, they calculate a multi-scale perceptual image tower that is generated by mimicking a human observer when looking at the input image
from different distances. The first stage of their algorithm deals with the extraction of the core colour clusters and the segmentation task is defined as a
probabilistic process that reassigns the non-core pixels hierarchically starting from the coarsest image in the tower to the image with the highest
resolution. The main limitation of this algorithm is the fact that the colour and texture features are not explicitly used and this causes problems in the
analysis of their contribution in the overall segmentation process. Deng and Manjunath [18] proposed a different colour-texture segmentation
technique (also known as JSEG) that consists of two independent computational stages: colour quantization and spatial segmentation. During the first
stage the colour information from the input image is quantized into a representative number of classes without enforcing the spatial relationship
between pixels. The aim of this process is to map the image into a structure where each pixel is assigned a class label. The next stage of the algorithm
enforces the spatial composition of the class labels using a segmentation criterion (J value) that samples the local homogeneity. The main merit of this
paper is the use of colour and texture information in succession and the authors argue that this approach is beneficial since it is difficult to analyse the
colour similarity and spatial relationship between the neighbouring pixels at the same time. Other researchers adopted different strategies regarding the
inclusion of texture and colour in the segmentation process. Tan and Kittler [19] developed an image segmentation algorithm that is composed of two
channels, one for texture representation and one for colour description. The texture information is extracted by applying a local linear transform, while
the colour is sampled by the moments calculated from the colour histogram. This approach is extremely appealing, since the contribution of colour and
texture can be easily quantified in the segmentation process and it has been adopted by a large number of researchers [20]- [22]. Another related
implementation has been proposed by Carson et al [23]. They developed a Blobworld technique that has been applied to the segmentation of natural
images in perceptual regions. The central part of this algorithm is represented by the inclusion of the polarity, contrast and anisotropy features in a
multi-scale texture model that is evaluated in parallel with the colour information that is sampled by the CIE Lab colour components. The main
advantage of the Blobworld algorithm consists in its ability to segment the image in compact regions and it has been included in the development of a
content-based image retrieval system.

In this paper we propose a flexible and generic framework (CTex) for segmentation of natural images based on colour and texture. The developed
approach extracts the colour features using a multi-space adaptive clustering algorithm, while the texture features are calculated using a multi-channel
texture decomposition scheme. Our segmentation algorithm is unsupervised and its main advantage resides in the fact that the colour and texture are
included in an adaptive fashion with respect to the image content.

OVERVIEW
The main components of the proposed segmentation framework are depicted in Fig. 1. The colour and texture features are extracted independently, on
two different channels. The colour segmentation algorithm is the most sophisticated part of the CTex framework and involves a statistical analysis of
the input image in the RGB and HSV colour representations. The original image (RGB colour
space) is first pre-filtered with a neibhour hood low pass filter that is applied to eliminate the image noise, artefacts and weak textures and improve the
local colour coherence. The second step is to extract the color features by applying the derived HSVimage to to k-means clustering framework.

ORIGINAL RGB TO HSV Hue K-means Color features


RGB IMAGE [h,s,v]
clustering

Hue

K-means Final segmented image


PRODUCT clustering

K-means
Texture features
FILTERING
clustering

Fig. 1. Outline of the proposed CTex colour-texture image segmentation framework.

Next, the filtered RGB image is clustered using a K-Means algorithm where the cluster centres are initialised with the dominant colours to get the
texture features. The number of texture segmented images, color segmented images are given by the k- means clustering i.e. k. and the product of
hue and smoothened texture image is input to k-means clustering framework, which returns the final color texture segmented image.

RGB IMAGE

An RGB color image is an MXNX3 array of color pixels, where each color pixel is a triplet corresponding to Red, Green, Blue components of an RGB
image at a specific spatial location. An RGB image can be viewed as a stack of three grayscale images that, when fed into the Red, green and blue
inputs of a color monitor, produce a color image on the screen. by convention three images forming an RGB color image are referred to as the red,
green and blue component images. The data class of the component determines the range of the values. If an RGB image is of class of Double the
range of values is [0,1].Similarly, the range of values [0,255] or [0,65535]for RGB images of uint8 or uint16 respectively. the number bits used to
represent the pixel values of component images determines the bit depth of the RGB image .for ex:- if each component image is an 8-bit the
corresponding RGB image is said to be 24bits deep. generally the number of bits in all components is same. In this case the number of possible colors
in an image is (2b)3where b is the number of bits in each component of image.The RGB image is usually shown graphically as an RGB color cube.
The vertices of the cube are the primary(red,green,blue),and secondary(cyan,megenta and yellow).

44
Fig.2: RGB image model
HSV IMAGE
HSV(hue,saturation,value) is one of several color systems used by people to select colors from a color wheel .this color system is considerably closer than
the RGB system to the way in which humans experience and describe color sensations .
The HSVcolor space is formulated by looking at the RGB color cube along its gray axis which results in the hexagonally shaped palette as shown AS we
move along the vertical axis the size of the hexagonal plane that is perpendicular to the axis changes ,yielding the volume depicted in the figure. Hue is
expressed as an angle around the color hexagon typically using the red axis as the 0o axis. and hue corresponds to the color(orange, blue….). value is
measured along the axis of the cone .the V=0 end of the axis is black. the V=1end of the axis is white. which lies in the center of the hexagon.thus,this axis
represents all shades of gray. saturation is measured as the distance from the V-axis it corresponds to the degree of color(orangeness,blueness…..).
The MATLAB function for converting from RGB to HSV is rgb2hsv whose syntax is
[h,s,v]=rgb2hsv(rgbimage);

Fig.3:HSV image model

FILTERING

All operations on image matrix are done at neighborhood level. neighborhood processing consists of the following steps
(1) defining a center point I(X,Y)
(2) performing an operation that involves only the pixels in a predefined neighborhood about the center point
(3) letting the result of that operation be the response of the process at that point. Hence output pixel value is decided by spatial mask around the
reference pixel; and
(4) Repeating the process for every point in the image the process of moving the center point creates new neighborhoods one for each pixel in the
image this process is called “filtering” .

K- means clustering algorithm

45
Clustering is used to organize data for efficient retrieval. One of the problems in clustering is the identification of clusters in given data. A popular
technique for clustering is based on K-means such that the data is partitioned into K clusters such that similar elements are assigned to same cluster
while elements with different properties are assigned to different clusters. Clustering is used to perform efficient search of elements in a data set.
Clustering is particularly effective in multi-dimensional data that may be otherwise difficult to organize in an effective manner. One of the earliest
clustering techniques in the literature is the K-means clustering method (Anderberg 1973; Romesburg 1984). In this technique, clustering is based on
the identification of K elements in the data set that can be used to create an initial representation of clusters. These K-elements form the cluster means
. The remaining elements in the data set are then assigned to one of these clusters. The most common form of the algorithm uses an iterative
refinement heuristic known as Lloyd's algorithm starts by partitioning the input points into k initial means. It constructs a new partition by associating
each point with the closest mean. Then the means are recalculated for the new clusters, and algorithm repeated by alternate application of these two
steps until convergence, which is obtained when the points no longer switch clusters (or alternatively means are no longer changed).
The algorithm for k-ens clustering is as follows,
• Place K points into the space represented by the objects that are being clustered. These points represent initial group means.
• Assign each object to the group that has the closest mean.
• When two elements are belongs to same group then finding the average of those two elements and upgrade the mean with this average value.

Repeat Steps 2 and 3 until the means no longer change. This produces a separation of the objects into groups from which the metric to be minimized
can be calculated.

RESULTS

The experimental results indicate that the proposed technique is accurate in capturing the color and texture characteristics when applied to images.

a(1) b(1) c(1) d(1) e(1)

a(2) b(2) c(2) d(2) e(2)


Fig. 4. a(1),a(2) – original images, b(1), c(1), d(1), e(1) ,b(2), c(2), d(2), e(2) – segmented images

CONCLUSION

In this paper we presented a new segmentation algorithm where the colour and texture features are adaptively evaluated by a clustering strategy that
enforces the spatial constraints during the assignment of the data into image regions with uniform texture and colour characteristics. The inclusion of
the colour and texture features in a composite descriptor proved to be effective in the identification of the image regions with homogenous
characteristics. The performance of the developed colour-texture segmentation algorithm has been quantitatively and qualitatively evaluated on a large
number of synthetic and natural images and the experimental results indicate that our algorithm is able to produce accurate segmentation results even
when applied to images characterised by low resolution and low contrast.

REFERENCES

M. Tuceryan and A. K. Jain, “Texture analysis”, Handbook of Pattern Recognition and Computer Vision, 2nd Edition, C.H. Chen, L.F. Pau andP.S.P
Wang (eds.), World Scientific Publishing, 1998, pp.207-248.
H. D. Cheng, X. H. Jiang, Y. Sun, and J. L. Wang, “Colour image segmentation: advances and prospects”, Pattern Recognition, 34(12), pp. 2259-
2281, December 2001.
A. Materka and M. Strzelecki, “Texture analysis methods – A review”,Report COST B11, Institute of Electronics, Technical University of.Lodz,
1998.
T. Randen and J. H. Husoy, “Filtering for texture classification: A comparative study”, IEEE Trans. on Pattern Analysis and Machine
Intelligence, vol. 21, no. 4, pp. 291-310, 1999.
K. Y. Song, M. Petrou, and J. Kittler, “Defect detection in random colour textures”, Image and Vision Computing, vol. 14, pp. 667-683,

46
1996.
L. Shafarenko, M. Petrou, and J. Kittler, “Automatic watershed segmentation of randomly textured colour images”, IEEE Trans. Pattern Analysis and
Machine Intelligence, vol. 6, no. 11, pp.1530-1544, 1997.
D. Panjwani and G. Healey, “Markov random field models for unsupervised segmentation of textured colour images”, IEEE Trans. on
Pattern Analysis and Machine Intelligence, vol. 17, no. 10, pp. 939–954, 1995.
C. H. Yao and S. Y. Chen, “Retrieval of translated, rotated and scaled colour textures”, Pattern Recognition, vol. 36, no. 4, pp. 913-929, 2003.
M. A. Hoang, J. M. Geusebroek, and A. W. M. Smeulders, “Colour texture measurement and segmentation”, Signal Processing, vol. 85, no. 2, pp.
265-275, 2005.
J. Freixenet, X. Munoz, J. Marti, and X. Llado, “Colour texture segmentation by region-boundary cooperation”, in Proc. 8th European
Conference on Computer Vision, Prague, Czech Republic, 2004, pp.259-261.
Z. Kato and T. C. Pong, “A Markov random field image segmentation model for colour textured images”, Image and Vision Computing, vol. 24, no.
10, pp. 1103-1114, 2006.
H. Permuter, J. Francos, and I. Jermyn, “A study of Gaussian mixture models of colour and texture features for image classification and
segmentation”, Pattern Recognition, vol. 39, no. 4, pp. 695-706, 2006.
M. Vanrell, R. Baldrich, A. Salvatella, R. Benavente, and F. Tous, “Induction operators for a computational colour–texture representation”, Computer
Vision and Image Understanding, vol. 94, no. 1-3, pp. 92-114, 2004.
Y. G. Wang, J. Yang, and Y. C. Chang, “Colour-texture image segmentation by integrating directional operators into JSEG method”,
Pattern Recognition Letters, vol. 27, no. 16, pp. 1983-1990, 2006.
L. Shi and B. Funt, “Quaternion colour texture segmentation”, ComputerVision and Image Understanding, vol. 107, pp 88-96, 2007.
J. Chen, T. N. Pappas, A. Mojsilovic, and B. E. Rogowitz, “Adaptive perceptual colour-texture image segmentation”, IEEE Trans. on
ImageProcessing, vol. 14, no. 10, pp. 1524- 1536, October 2005.
M. Mirmehdi and M. Petrou, “Segmentation of colour textures”, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 22, no. 2, pp. 142-
159, Febr. 2000.
Y. Deng and B. S. Manjunath, “Unsupervised segmentation of colourtexture regions in images and video”, IEEE Trans. on Pattern Analysis and
Machine Intelligence, vol. 23, no. 8, pp. 800-810, August 2001.
T. S. C. Tan and J. Kittler, “Colour texture analysis using colour histogram”, IEE Proc. Vision, Image, and Signal Processing, vol. 141
P. Nammalwar, O. Ghita, and P. F. Whelan, “Experimentation on the use of chromaticity features, Local Binary Pattern and Discrete Cosine
Transform in colour texture analysis”, Proc. of the 13'th Scandinavian Conference on Image Analysis, Goteborg, Sweden, 2003, pp.186-192.
S. Liapis and G. Tziritas, “Colour and texture image retrieval using chromaticity histograms and wavelet frames”, IEEE Trans. on
Multimedia, vol. 6, no. 5, pp. 676-686, March 2004.
T. Mäenpää and M. Pietikäinen,”Classification with colour and texture: jointly or separately? “, Pattern Recognition Letters, vol. 37, no. 8, pp. 1629-
1640, 2004.
C. Carson, S. Belongie, H. Greenspan, and J. Malik, “Blobworld: Image segmentation using Expectation-Maximization and its application to image
querying”, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 24, no. 8, pp. 1026-1038, August 2002.
D. E. Ilea and P. F. Whelan, “Adaptive pre-filtering techniques for colour image analysis”, in Proc. of the International Machine Vision & Image
Processing Conference, September 2007, Maynooth, Ireland, IEEE Computer Society Press, pp. 150-157.
G. Gilboa, N. Sochen, and Y. Y. Zeevi, "Forward-and-backward diffusion processes for adaptive image enhancement and denoising",
IEEE Trans. on Image Processing, vol. 11, no. 7, pp. 689-703, July 2002.
P. Perona and J. Malik, “Scale-space and edge detection using anisotropic diffusion”, IEEE Trans. on Pattern Analysis and Machine
Intelligence, Washington, vol. 12, no. 7, pp. 629-639, July 1990.
B. Smolka and K. N. Plataniotis, “On the coupled forward and backward anisotropic diffusion scheme for colour image enhancement”, Lecture Notes
in Computer Science, Springer Verlag, vol. 2383, London, 2002, pp. 70-80.
J. M. Pena, J. A. Lozano, and P. Larranaga, “An empirical comparison of four initialization methods for the K-Means algorithm”, PatternRecognition
Letters, vol. 20, 1999, pp.1027-1040.
T. Kohonen, Self-Organising Maps, Springer Series in Information Sciences, 3rd edition, vol. 30, Berlin Heidelberg, New York: Springer Verlang,
2001.
G. Dong and M. Xie, “Colour clustering and learning for image segmentation based on neural networks”, IEEE Trans. on Neural
Networks, vol. 16, no. 4, pp. 925-936, July 2005.
S. H. Ong, N. C. Yeo, K. H. Lee, Y. V. Venkatesh, and D. M. Cao, “Segmentation of colour images using a two-stage self-organising
network”, Image and Vision Computing, vol. 20, pp. 279-289, 2002.
D. E. Ilea and P. F. Whelan (2006), “Colour image segmentation using a self-initializing EM algorithm”, International Conference on
Visualization, Imaging and Image Processing (VIIP), Spain, August 2006.
A. C. Bovik, M. Clark, and W. S. Geisler, “Multichannel texture analysis using localized spatial filters”, IEEE Trans. on Pattern Analysis andMachine
Intelligence, vol. 12, no. 1, pp. 55-73, 1990.
A. K. Jain and F. Farrokhnia, “Unsupervised texture segmentation using Gabor filtering”, Pattern Recognition, vol. 33, pp. 1167-1186, 1991.
T. Randen and J. H. Husoy, “Texture segmentation using filters with optimized energy separation”, IEEE Trans. on Image Processing, vol. 8, no. 4,
pp. 571-582, 1999.
J. G. Daugman,”Complete discrete 2D Gabor transforms by neural networks for image analysis and compression”, IEEE Trans. on
Acoustics, Speech and Signal Processing, vol. 36, no. 7, pp. 1169-1179, 1988.
D. E. Ilea and P. F. Whelan, “Automatic Segmentation of Skin Cancer Images using Adaptive Colour Clustering”, China-Ireland
InternationalConference on Information and Communications Technologies, Hangzhou, China, 2006, pp.348-351.
D. E. Ilea and P. F. Whelan, “Colour Image Segmentation Using a Spatial K-Means Clustering Algorithm”, Proc. of the Irish Machine
Vision & Image Processing Conference, Dublin City University, 2006, pp. 146-153.

47
Y. Rubner, J. Puzicha, C. Tomasi, and J. M. Buhmann “Empirical evaluation of dissimilarity measures for colour and texture”, ComputerVision and
Image Understanding, vol. 84, no. 1, pp. 25-43, 2001.
Vision Texture (VisTex) Database, Massachusetts Institute of Technology, MediaLab. Available online at:
http://vismod.media.mit.edu/vismod/imagery/VisionTexture/vistex.html
D. Martin, C. Fowlkes, D. Tal, and J. Malik, "A Database of Human Segmented Natural Images and its Application Segmentation Algorithms and
Measuring Ecological Statistics", ICCV, 2001, pp. 416-425.
A. Olmos, and F. A. A. Kingdom, McGill Calibrated Colour Image Database, http://tabby.vision.mcgill.ca, 2004.
Outex natural images database: http://www.outex.oulu.fi.
R. Unnikrishnan and M. Hebert, ”Measures of similarity”, Proc. of IEEEWorkshop on Computer Vision Applications, 2005.

48
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Kinetic Study of wax deposition in the flow lines due to Indian Crude Oil
Vikas Mahto, Dharmendra Verma and Harveer Singh
Department of Petroleum Engineering
Indian School of Mines Dhanbad-826004
Email:vmahto@yahoo.com

ABSTRACT

The wax deposition problem during production, transportation and handling of crude oil causes loss of billions of dollars every year to the petroleum
industry. Hence, the study of wax deposition in the tubing and pipelines has become most important issue in petroleum sector particularly in the design
and development of the deepwater fields. In this work, attempts were made to determine the factors responsible for the deposition of waxes in the
flow lines due to Indian crude oil. The study of wax deposition, dispersion and inhibition of neat crude oil and pour point depressant treated crude oil
in the flow lines was carried out using self fabricated set up in our labororatory. In the kinetic study of oil with respect to time, the weights of
deposited wax are observed after thirty minutes and dispersion and inhibition characteristics and effective dispersion time was observed. The kinetic
study of oil with respect to differential temperature is done in two ways. Firstly, the ambient/surface temperature was kept constant and varying the
pumping/reservoir temperature and secondly, pumping/reservoir temperature was kept constant and the varying the ambient/surface temperature. From
these studies ideal differential temperature for this crude oil to flow is determined. The present investigations reveal that the selected pour point
depressant may be very helpful for controlling the wax deposition problem of the crude oil.

KEYWORDS: Waxy Crudes, Wax Deposition, Wax Inhibition, Wax Dispersion, Wax Appearance Temperature, Pour Point

INTRODUCTION
The crystallization of waxes in crude oil produces non-Newtonian flow Characteristics and it increases viscosity of crude oil, yield stress for restarting
the flow and deposition of wax crystallites on surfaces and causes severe problems during production, storage and transportation. The crystallization
behavior of crude oils is mainly dependent on the wax composition and distribution and rheological or flow characteristics of the crude oil are largely
dependent on the temperature. The behavior of waxy crude oil is shown in Figure 1. When the temperature of the crude is higher than the wax
appearance temperature (the temperature when first wax crystal is formed), the waxy crude oil is Newtonian fluid but it posses non-Newtonian
characteristics at the temperature below the wax appearance temperature due to wax crystallization and gel formation of their heterogeneous matrix.
Further lowering the temperature stage is reached when crude oil ceases to flow and is called pour point.(Boukadi and Amri, 2005; Ying-ru and Jin-
Jun, 2007;Fan and Guo quen, 2007; Jin-jun and Xin, 2008; Deshmukh and Bharambe, 2008; Mahto, 2010).

Fig 1: Rheological behavior of waxy crude oil with temperature


In order to decrease the viscosity of crude oil and reduce the congealing temperature and yield strength, chemical additives and pour point depressants
should be used (Jin-jun et al, 2007). Their structure is mostly polymeric and improves the flow properties of crude oils and decrease the surface
deposition of crystals (Mishra et al 1995; Mahto, 2010).The pour point depressants are very sensitive to the crude oil and its performance is closely
governed by composition of crude oil (amount and structure of resin, asphaltene and wax in the crude oil) and structure of pour point depressants
(During the cooling of a pour point depressant-treated crude oil, molecules of pour point depressants and wax separate from the crude oil in the same
time, pour point depressants modify the wax crystal structure and prevent the interlocking of wax crystal, thus reduce the temperature at which a gel
structure is built up and improve the rheological behavior of crude oil (Figure. 2).

Fig 2. Mechanism of pour point depression in waxy crude oil


The crude oil pour point depressants are polymers made from discrete monomers. The polymers have three variable characteristics that may affect
their performance; the polymer backbone, the length of the pendant chain and the polymer molecular weight. The backbone and pendant chain length
can be changed by using different monomers. The polymer molecular weight can be changed by adjusting reaction conditions, amount of initiator used
etc. It is believed that the backbone only provides a structure from which the important pendant chains are suspended. To investigate the effect of the
polymer backbone, pour point depressants with the same pendant chain length, but different backbones were tested. Every effort was made to keep the
molecular weight of the polymers the same. The interaction between the wax control additives and the paraffin in the crude oil is crucial and the
additives work best when they are matched to the paraffin distribution in the crude. As the average carbon number of the pendant chain on the pour
point depressant increase, the pour point of the additized crude drops until it reaches a minimum and then increases again. The molecular weight of the
wax control polymer may affect the interaction of the polymer with the paraffin; a very short, low molecular weight polymer may not have the
molecular volume to disrupt the paraffin crystals as it co-crystallizes within the paraffin matrix. A very long, high molecular weight polymer may be
so large that it interacts with itself instead of the crude oil paraffins or the polymer’s solubility in the crude oil may be limited and actually initiate
paraffin crystallization and thus raise the pour point of the crude oil (Manka and Ziegler, 2001). Undiluted pour point depressants are waxy materials
that are often solids at ambient temperature. To pump these products in the field, they usually need to be drastically diluted with solvent. Therefore,
solvents comprise a very large portion of these finished formulations to make a transportable product. The concentration of the polymer in the solvent
will affect the extent of interaction of the polymer with itself. At high concentration the polymer may interact with other polymer molecules and
become entangled. This entanglement may impact the accessibility of the polymer to the paraffin in the crude and, thus, may impact the performance
of the cold flow modifier package. At low concentration, the polymer is fully solvated and should not interact with other polymer molecules and
should be very accessible to the paraffin in the crude oil (Qian et al, 1996)

The factors, which contribute the extent of wax deposition in flowing system, are flow rate, temperature differential and cooling rate, and surface
properties (Khan, 2008). In Laminar flow, wax deposition increases with flow rate. This can be explained by the availability of more particles for
deposition at the surface. As the flow rate increases to turbulent regimes, wax deposition decreases because of shear dispersion. Thus, wax deposition
gradually decreases with increases in turbulence and flow rate. Shear dispersion is predominant in turbulent flow in all stages. The cooling rate and
temperature differential between the bulk of solution and cold surface is another factor for wax deposition. Wax deposition increases with an increase
in temperature difference. Wax deposition would occur only when the surface temperature is below both the solution temperature and the solution
cloud point. Initially, the rate of wax deposition is higher but slows down as more wax is deposition on the pipe surface. The thickness of the wax
layer increases, and this layer acts as insulation and reduces the effective temperature differential. This lowers the availability of wax crystals for
further deposition. The size and number of crystals formed are also important for wax deposition. At a higher rate of cooling, the wax precipitates out
in smaller crystals and a large number of crystals are formed because of the large number of crystallization sites available. At a lower rate of cooling,
the crystallization process is lore uniform. Thus, more uniformly packed crystals are formed that possess a relatively small surface area and free
energy. Temperature differential also effects the composition of deposited wax. If it is high, cooling is rapid and both lower and higher melting waxes
crystallize simultaneously, forming a weak porous structure (owing to malcrystallization) with cavites full of oil. The wax crystals adhere to the pipe
surface during deposition. So wax deposition can also be a function of surface properties. The presence of certain adsorbed film on a metal surface
would reduce the adherence of paraffin to the surface. As the paraffin wax is deposited on a surface, it is held in place by adsorbtion forces. These
adsorbtion forces are dependent on the free surface energy of both the paraffin and surface (Mishra et al, 1995).

Physicochemical properties provide important clues for findings solution to the problems of flow behaviour and wax precipitation and Wax deposition
studies provides important information about wax deposition potential of an oil and effect of chemical additives on the oil. Hence keeping these in

50
mind the physicochemical properties and kinetic study of wax deposition of Indian waxy crude oil were carried out to solve the flow assurance
problems in the Indian oil fields.

EXPERIMENTAL PROCEDURE

A crude oil sample was collected from Ankleshwar oilfield of Oil and Natural Gas Limited, India and polymeric additive used in the present work was
procured from M/s Krishna Antioxidants, Mumbai, India.

In order to understand the nature and extent of problem in test oils, physicochemical and wax deposition tendencies of test oils were studied. The
physicochemical properties of the waxy crude oil were determined using IP/ASTM methods and other methods reported in the literatures (Kim et al,
1996; Fuhr and Holloway, 1999; Rao, 2001; Khan et al 2008). To simulate the wax depositions of the crude oil to the temperature of oil fields under
dynamic condition kinetic study were carried out. This test was useful to predict the amount of deposition at a particular temperature during
production. In this test, the temperature of 1000 ml crude oil was maintained at elevated temperature through a thermostatically controlled water bath
representing pumping station of oil field (the inlet flow in to pipe line) and the copper tube was kept in another thermostatically controlled water bath
at much lower temperature representing actual field conditions of pipeline near surface. The oil was circulated through U tube by means of a peristaltic
pump. The flow was maintained in the laminar range with uninterrupted circulation. After certain period, the circulation was stopped and the weight of
U tube measured. The difference in weight of U tube before and after the circulation gives the amount of wax deposition. Now the oil treated with
required concentration of chemical dispersant was circulated through U tube having paraffin deposits in the previous step. The difference in weight of
deposition gives the amount flushed out by dispersant i.e. the decrease in the amount of deposition was due to dissolution/dispersion effect of chemical
after treatment.

RESULTS &DISCUSSIONS

The API gravity, wax appearance temperature, pour point and characterization factor are listed in Table 1. The crude oil sample under present study is
light crude oil having API gravity of 19.12. The Kuop value of the sample is 10.3 which as per nomographs of Universal Oil Products Company
indicate that the crude oil is of intermediate base characteristics. The wax appearance temperature is +55oC and the pour point of the crude is + 36oC
with a wax content of 17.1 wt %. The sample is predominantly rich in saturate fractions and ashphaltene is almost negligible.

Table 1. Physicochemical characterizations of Indian Crude Oil

Sl. Parameters Unit Observed Values


No
1 Density at 15oC gm/cc 0.9389
2 API Gravity Deg 19.12
3 Characterization ------ 10.36
Factor(Kuop)
o
4 WAT C 55
o
5 Pour Point C 36
6 Wax Content % wt 17.2
7 Saturates % wt 54.9
8 Aromatics % wt 36.25
9 Asphaltene % wt 0.95
10 Resin % wt 7.88

The Table 2 indicates that in case of neat oil, after circulation for 60 minutes, the further deposited amount of wax debris washed out with flowing oil
stream due to its loose bounding. The process of deposition and removal of wax continued till the oil circulation.

Table 2: Kinetic study of wax deposition with neat crude

Time(min.) U-Tube Weight(gm) Weight of wax deposited (gm)


0 272.4 -------
30 293.4 21.0
60 298.2 25.8
90 295.5 23.1
120 294.9 22.5
150 299.9 27.5
180 293.9 21.5
210 292.8 20.4
240 303.9 31.5
270 298.9 26.5
300 296.7 22.5

51
The treated oil was circulated through cleaned U-tube and the amount of wax deposition against the time of circulation of crude oil has been shown in
Table 3. It indicates that the amount of wax deposition with treated oil was considerably lower than the treated oil

Table 3. Kinetic study of wax inhibition with 1000 ppm pour point depressant treated crude

Time(min.) U-Tube Weight(gm) Weight of wax deposited (gm)


0 272.4 -----
30 278.41 6.01
60 298.9 6.5
90 280.3 7.9
120 279.9 7.5
150 280.9 8.5
180 281.5 9.1
210 282.9 10.5
240 282.8 10.4

The treated crude oil was circulated through U-tube containing the deposited wax in the above stage and the amount of wax remained in the U tube
against the time of oil circulation has been shown in Table 4. The figure indicated that certain amount of wax has been removed/dispersed effectively
within 30 to 60 minutes and further dispersion process was observed ineffective due to little deposition and dispersion process might be taking places
simultaneously.
Table 4: Wax dispersion studies in 1000 ppm pour point depressant treated crude

Time(min.) U-Tube Weight(gm) Weight of wax deposited (gm)


0 294.9 22.5
30 285.4 13.0
60 286.5 14.1
90 284.9 12.5
120 285.5 13.1
150 285.9 13.5
180 285.6 13.2
210 285.7 13.3
240 285.9 13.5

The effect of initial pumped oil temperature into the pipeline against the circulation time was shown in Table 5-6. It was observed that 60 t0 55 was
the ideal temperature of oil to pump into the pipeline where there was almost constant deposition of wax whereas on decreasing the pumping
temperature the deposition increased sharply at 40oC. Both untreated and treated crude shows the same tendency of deposition. It is because all the
hydrocarbons are soluble above cloud point of crude oil but as the cloud point temperature is attained the higher molecular weight hydrocarbons start
coming out of the solution in the form of precipitates which deposited in the walls of pipeline. As the temperature decreases below cloud point more
hydrocarbon molecules start coming out of the solution and deposited on the walls of the pipeline.

Table 5: Effect of oil pumping temperature/reservoir temperature on wax deposition with neat crude

Temperature(oC) U-Tube Weight(gm) Weight of wax deposited (gm)

60-15 294.9 22.5

55-15 294.9 22.5

50-15 300.8 28.4

45-15 311.5 39.1

40-15 320.7 48.3

52
Table 6: Effect of oil pumping /reservoir temperature on wax deposition with1000 ppm pour point depressant treated crude

Temperature(oC) U-Tube Weight(gm) Weight of wax deposited (gm)


60-15 294.9 22.5
55-15 295.9 23.5
50-15 299.5 27.1
45-15 302.6 30.2
40-15 307.5 35.1

The effect of ambient temperature of pipeline on amount of wax deposition against circulation time was shown in Table 7-8.

Table 7: Effect of ambient/surface temperature on wax deposition with neat crude

Temperature(oC) U-Tube Weight(gm) Weight of wax deposited (gm)


55-45 276.9 4.5
55-40 278.4 6.0
55-35 286.5 14.1
55-30 287.9 15.5
55-25 298.5 26.1
55-20 304.5 32.1
55-15 305.41 33.01

Table 8: Effect of ambient/surface temperature on wax deposition with1000 ppm pour point depressant treated crude

Temperature(oC) U-Tube Weight(gm) Weight of wax deposited (gm)


55-45 272.4 3.5
55-40 276.1 3.7
55-35 276.3 3.9
55-30 276.5 4.1
55-25 277.9 5.5
55-20 282.9 10.5
55-15 288.9 16.5

It was observed that on decreasing the ambient/surface temperature the amount of wax deposition increases on further decrease the temperature to
pour point temperature of oil or below that temperature the deposition increased sharply. It is probably due to formation of three dimensional network
of wax crystal lattice on the walls of pipeline which directly in contact of surroundings.

CONCLUSIONS

The following conclusions can be drawn from the present investigations:

1. The sample under present study is of intermediate base characteristics. It is predominantly rich in saturate fraction and the ashphaltene
content is almost negligible.
2. In the kinetic study, it was observed that on decreasing the ambient/surface temperature the amount of wax deposition increases on further
decreasing the temperature to pour point temperature of oil or below the deposition sharply.
3. The sufficient amount of deposition occurred within 60 minutes, after than there was continuous process of deposition and removal of
deposits in U-tube.
4. Considerable amount of wax dispersion is effective initially within 30-60 minutes. After that it is ineffective.
5. Inhibition study shows that the amount of wax deposition with treated oil was considerably lower than the untreated oil.
6. On varying the pumping/reservoir temperature, we observe that 60 to 55oC was ideal temperature of oil to pump into the pipe line.

ACKNOWLEDGEMENT

Authors are thankful to UGC, New Delhi for providing financial support to carry out this work.

REFERENCES

Boukadi, A and Amri H (2005). “Flow improvement of waxy crude oil by chemical additive utilization”, Pak. J. Bio. Sci., 8(8), pp 1131-1136
Deshmukh, S and Bharambe, DP (2008). “Synthesis of polymeric pour point depressants for Nada crude oils and its impact on oil rheology”, Fuel
Process. Technol, 89, pp 227-233

53
El Gamal, IM (1998). “Combined effect of shear and flow improvers: the optimum solution for handling waxy crude oil below pour point”, Colloidal.
and Surfaces. A: Physicochem. and Eng. Asp, 135, pp 283-291.
Fan, Z and Guo-Quen, C (2007). “Viscoelastic properties of waxy crude oil”, J. Cent. South Univ. Technol., S1, pp 0445-04
Fuhr, BJ and Holloway, LR (1999). “Analytical considerations related to asphaltenes and waxes in the some crudes”, Energy and Fuels, 13, pp 336-
339
Jin-jun, Z and Xin, L (2008). “Some advances in crude oil rheology and its application”, J. Cent. South Univ. Technol., 15(1), pp288-292
Jin-Jun, Z, Fan, Z, Qi-yu, H and Da-fan, Y (2007). “Experimental simulation of effect of shear on rheological properties o beneficiated waxy crude
oils”, J. Cent. South Univ. Technol., S1, pp 0108-04
Khan, AR., Mahto, V, Fazal, SA and Laik, S (2008). “Studies of wax deposition onset in case of Indian crude oil”, Petrol Sci. and Technol., 26,
pp1706-1715
Kim, H, Chung, K and Kim, H (1996). “Measurement of the asphaltene and resin content of crude oils”, J. Ind. Eng. Chem., 2(1), May, pp 72-78
Mahto, V (2010). “Study of the rheological properties of Indian waxy crude oil”, Proc. of National Seminar on Recent Advances in Chemical Engg,
G.I.E.T. Gunupur Orissa, India, 30th -31th January, pp 41-44
Manka, JS and Ziegler, KL (2001). “Factors affecting performance of wax control additives”, World Oil, June 2001
Mishra, S, Baruah, S and Singh, K (1995). “Paraffin problems in crude oil production and transportation: A Review” SPE Prod. Fac., February, pp.
50-55.
Qian, JW, Qi, GR, Xu, YL and Yang, SL (1996). “Solvent effect on the action of ethylene-vinyl acetate copolymer pour point depressants in waxy
solutions”, J. App. Pol. Sci., 60, pp 1575-1578
Rao, BKB (2001). “Modern Petroleum Refining Processes”, 4th ed., Oxford & IBH Publishing Co. Pvt. Ltd, New Delhi.
Ying-ru, Z and Jin-jun, Z (2007). “Pipelining characteristics of Daqing waxy crude”, J. Cent. South Univ. Technol., S1, pp 0442-03

54
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Effect of High Pressure, High Temperature and Crude Oil Composition on the Stability of Crude Oil
Emulsions in the Presence of Low Level of Asphaltene
Jitendra S. Sangwai
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
E-mail: jitendrasangwai@iitm.ac.in

ABSTRACT

Crude oil, when in contact with water under high shear may form stable emulsions at high-pressure and high-temperature (HPHT) conditions. These
emulsions pose formidable challenge for flow assurance downhole and at surface facilities due to their highly viscous nature. It is crucial to break
these emulsions either downhole or at surface facilities in order to increase the oil production. Rheological properties of these emulsions are indeed
significant to understand their stability at surface and HPHT conditions and are also required to design and develop efficient lift operations, fluid
handling systems and robust reservoir models. There are numerous studies related to rheology of crude oil emulsions, many of them primarily focus at
surface conditions which are inadequate to understand their behavior at in-situ HPHT reservoir conditions. Additionally, it is well known that some of
the components of the crude oil, such as, asphaltene, resins, etc., act as surfactant to form stable emulsions with water. Very recently, it is observed
that other components of the crude oil, such as, waxes, may act as surfactant to form stable emulsions with water at downhole conditions at low
concentration of asphaltenes. This work will review and present some important observations about the fundamental aspects of crude oil emulsions
and their stability from an open literature with some field observations. In addition to this a study on the waxy crude oil emulsion obtained from an
Indian brownfield is carried out and discussed with respect to the effect of HPHT conditions and the waxy nature of the crude oil on their stability. It is
observed from the investigation on the SARA composition analysis of this crude oil that the crude oil shows the presence of more waxy components
in the oil with little asphaltene. Effects of various parameters such as shear rate, water cut, temperature and pressure on the viscosity of crude oil
emulsion were studied individually. The studies on the rheology of emulsion at surface conditions shows a phase inversion point in the water-cut range
between 75 and 80 %. The zero-shear viscosity of the emulsion mixture (measured at near zero shear value) at surface and at HPHT conditions were
observed to be between 300-350 cP and 140-160 cP, respectively. It is observed that the magnitudes of viscosities of the emulsions are higher at
surface condition than at HPHT condition. It is expected that the increase in the system temperature helps decrease the viscosity of emulsion, however,
the presence of pressurized environment (pressure) weakens the effect of temperature in reducing the viscosity of emulsion, thus keeping emulsion
stable at high pressure environment even at high temperature conditions. To the best of our knowledge similar studies and information are not
published in an open literature and need more investigations on these lines. This may help develop better correlations for viscosity of the crude oil
emulsions at in-situ conditions.

KEY WORDS: Asphaltene, Crude oil Emulsion, High Pressure High Temperature (HPHT), Resins, Viscosity, Wax.

INTRODUCTION

An emulsion is a stable or unstable mixture of two or more immiscible liquids, mostly, water and oil, wherein one liquid is dispersed (dispersed phase)
in another liquid (continuous phase). Emulsions are of great research interest to the scientific community since several decades due to their widespread
presence in everyday life. They may be found in important areas such as food, cosmetics, pulp and paper, pharmaceutical and agricultural industries.
Emulsions are generally classified based on their structure. If the oil is the dispersed phase, the emulsion is termed oil-in-water (o/w) emulsion; and, if
the water is the dispersed phase, it is termed water-in-oil (w/o) emulsion. Other types such as, multiple emulsions of the type o/w/o, may also be
found. In the emulsified state, the interfacial area between the dispersed droplets and the bulk phase represents an increase in the systems free energy,
and thus, make them thermodynamically unstable. They, therefore, seek to minimize the surface area by separating into the different phases. For an
emulsion to separate, the droplets must merge with each other. However, in most of the practical application of emulsions, like in paints, cosmetics,
etc., the droplet coalesce are avoided by an addition of surfactant or polymeric molecules. These molecules get adsorbed on the droplets creating
mechanically strong elastic film that can act as a barrier against coalescence. Such aggregates are also referred as micelles. These molecules also help
reduce surface tension between the dispersed phase and continues phase, making emulsion stable. Emulsions are also been observed in an oil-field
industry, both in upstream and downstream production facilities. In an upstream oilfield industry, emulsion (mainly of polymeric fluids) does play an
important and positive role in most of the oilfield operations, viz., drilling fluids, enhanced oil recovery, flow assurance, etc. However, when the crude
oil containing surfactants (asphaltene, resins, wax, clays, etc.) comes in contact with the formation water forms crude oil emulsions, which shows
higher viscosities than pure oil and water. Such crude oil emulsions create problems for the flow of fluid downhole, well tubing and at surface
facilities. In downstream facilities, crude oil emulsion are observed, mainly, in separators and flow lines due to the presence of unsolvated asphaltene,
waxes, solids and in presence of high shear due to flow phenomena. The emulsified water also causes corrosion in the flow lines, thus adding an extra
cost for cleaning and transportation of crude oil. Water-in-crude oil emulsions can also happen in oceanic spills (Mehta, 2005). These emulsions are
very stable and the oil phase is difficult to recover, leading to great environmental damage. It is therefore required to break these emulsions with
minimal additional cost. In order to devise optimum treatment for water-in-oil emulsions, it is vital to understand their stability in complex reservoir
conditions and at surface facilities.

Crude oil emulsions can be of water-in-crude oil or crude oil-in-water type emulsions, former are quite important in several situations in upstream
industry. A typical water-in-crude oil type emulsion is shown in Figure 1 (a). The crude oil having more of higher hydrocarbon and polar compounds
such as ashphaltene, resins tends to form stable emulsions. Asphaltene molecules are believed to have complex crosslinked network, which along with
resins molecules forms micelles. These micelles trap water inside them resulting in water-in-crude oil type emulsion as shown in Figure 1 (b) (Kokal,
2005). The emulsion stability arises from a physical barrier of micelles that hinders the film to break when insufficient energies are involved in
collisions between water and oil droplets. I discuss below the possible sources of downhole emulsion followed by literature review on crude oil
emulsion studies. A detailed study on the stability of crude oil emulsion for waxy crude oil obtained from a brownfield in India will be presented in
subsequent section.

(a)

(b)

Fig 1 (a): An emulsified crude oil obtained well-ABC from Indian brownfield (water-in-oil type of emulsion) (b): Water droplet in crude oil emulsion
surrounded by asphaltene and resin aggregate (Kokal, 2005)

Sources of Downhole Emulsion

Several types of recovery methods are employed in a life time of a well so as to move hydrocarbon from the reservoir to near well-bore region. These
are mainly categorized into primary, secondary and tertiary (enhanced) oil recovery methods. Primary recovery method is the initial production phase
of the oil well, in which the hydrocarbons from the reservoir are flown near well-bore due to inbuilt reservoir pressure. When the production continues
for long time, the pressure of the reservoir decreases and the oil is no longer in natural flow conditions. Thereafter, secondary recovery methods is
instituted which consist of inducing an artificial drive into the formation to replace the natural drive. The most common secondary recovery method is
water flooding, which involves injecting water through water injection wells in the formation. These wells are drilled in the periphery of the
production well. This helps built the reservoir pressure and maintain the pressure drive. This method is shown schematically in Figure 2. Tertiary oil
recovery methods are employed when secondary oil recovery methods may no longer remain fruitful. Tertiary Recovery involves the use of enhanced
hydrocarbon recovery methods to enable the oil and gas to mobilize and recover from the reservoir, such as Alkaline- Surfactant-Polymer (ASP)
techniques, the injection of gas (such as CO2, N2, hydrocarbon gas), chemical (polymer) flooding, microbial treatment, steam injection into the well-
bore to modify its fluid properties and thereby enhance the movement of hydrocarbons through the formation.

56
When the field is in secondary recovery phase using injection water, the water undergoes changes in its chemical properties when it moves downhole

Production
Injection Well Injection
Well Well

Gas

Hydraulic Fractures
Water Flood Oil Water Flood

Water

Fig 2: Secondary oil recovery method employed in oilfield

in the reservoir as well as in the well tubing. The change in chemical properties of the water may be harmful in creating stable emulsion downhole. If
the crude oil contains sufficient amount of emulsifying agents like asphaltene, resins, waxes, etc., then it may form emulsion with the water in the
reservoir and/or near well-bore region under the influence of shear caused during the fluid flow and lift operations, such as, gas lift, electrical
submersible pumps (ESPs), etc. Additionally, crude oil emulsions may form during fracturing operation if suitable demulsifiers are not used during the
job. In EOR techniques, the use of surfactants and alkali may form downhole emulsion reducing the deliverability of the well (Li et al., 2004). Indeed,
the possibility of emulsion formation downhole is also a function of high pressure and high temperature (HPHT) conditions in the reservoir and hence
need to be investigated at in-situ conditions.

LITERATURE REVIEW

The emulsion stability is generally expressed from the viscosity of emulsion. Various factors are expected to affect emulsion stability. These are,
mainly, shear rate, pH of formation water, temperature, pressure, water cut (WC) and the composition of the crude oil particularly that of asphaltene,
resin. Other factors that usually favor emulsion stability are low interfacial tension, high viscosity of the bulk phase and relatively small volumes of
dispersed phase (Kokal, 2005). The knowledge of crude oil composition, manly, the concentration of asphaltene and resins plays an important role of
understanding whether the crude oil will form stable emulsions or not. The asphaltene and resins are naturally occurring surfactants which forms a
rigid film around the dispersed phase thus making emulsion stable. Fine solids are also responsible of forming stable emulsions. Here wettability of
solids particle plays an important role. Temperature affects the stability of emulsion at greater extend by changing physical properties of crude oil-
water system. Increase in temperature results in increase in the entropy of the systems, decrease in density and viscosity of the oil, thus helping
droplet coalescence. pH of water also have an influence on emulsion stability by affecting rigidity of the film. Effect of pressure on the emulsion
stability is not well studied and understood, though it is expected that pressure has an influence on the stability of emulsion. Droplet size of the
dispersed phase also affect the emulsion stability, in fact, this is the basis of categorizing emulsions as tight, medium and loose type emulsions. Tight
emulsions posses very tiny droplets of dispersed phase, on contrary, loose emulsion have higher size of droplets which helps them break easily. The
concentration of surfactant and shear rate plays an important role in having different sizes of droplet of dispersed phase.

Kokal and Al-Juraid (1999) studied the emulsion stability on Berri field, Suadi Arabia and observed that the temperature (including asphaltene) is
responsible factor to form stable emulsion which gets aggravated in winter seasons. They also observed that mixing asphaltene oil with other oil will
destabilize asphaltene particles and help them precipitate on crude oil water interface forming strong emulsion. The use of proper demulsifier is
required to optimize field performance on yearly basis. Auflem et al. (2001) carried out study on crude oil emulsion at separator conditions and
observed that if the separator pressure is decreased below bubble point, due to degassing, the surfactant molecules are removed from the emulsion
interfaces and help reduce emulsion stability, thus, correlating stability phenomena to the bubble point pressure. Janseen (1998) carried out core flood
experiments on crude oil emulsion and observed that the droplet size of produced emulsion from the core sample are function of pressure gradient and
interfacial tension. They showed that the near well bore condition may enhance the formation of small droplet characterized emulsion, thus pose
challenge handling multiphase fluid flow. Evdokimov et al. (2004) studied the effect of fabrication temperature on the stability of crude oil emulsion.
They observed that at about critical fabrication temperature the memory of emulsion is strong enough to retain their emulsion structure due to
collection of asphaltene molecules with the network of wax crystals formed at oil-water interfaces at low fabrication temperature.

Mclean and Kilpatric (1997) have studied the effect of aromacity, resin to asphaltene ratio (R/A) and the concentration of functional group on the
stability of crude oil emulsion, both by numerical and experimental investigation. They observed that the asphaltene molecules which are near to the
precipitation condition are more surface active than those solvated well in the medium. The R/A ratio play an important role in stabilizing emulsions.
One way to control R/A is to blend crude oil from different fields. Spiecker et al. (2003) studied the effect of petroleum resin and R/A ratio on
asphaltene aggregation and water-in-oil emulsion formation. They observed that at high R/A ratio, smaller aggregates of asphaltene and weak
emulsions are formed. This presumably corresponded to asphaltene monomer strongly interacting with resin molecules. Schorling et al. (1999) studied
the effect of R/A ratio on the stability of water in oil emulsion. They observed that the emulsion stability does not depend on the origin of resin, i.e.,
from the oil it is obtained, in the presence of asphaltene. They also observed that high R/A ratio decrease the emulsion stability. Li et al. (2004)
observed that ASP flooding technique used for enhanced oil recovery can results in the formation of stable crude oil water emulsion in the reservoir

57
and at surface facilities even in the absence of asphaltene. They relate this phenomenon to the formation of sodium salts of carboxylic acids as surface
active agents due to the reaction between sodium hydroxide and paraffin components of the crude oil. Recent studies show that the formation of wax
particles also favor the emulsion stability along with clays and other inorganic materials or napthanates (Ali and Alqam, 2000; Visintin et al., 2008,
Quintero et al., 2008). The crude oil studied by Quintero et al. (2008) presents two kinds of rheological behavior for temperature at or lower than
23°C, which corresponds to the formation point of paraffin crystals. These crystals produce a gel type structure that increases viscosity and
viscoelastic properties. Temperature has a strong influence on the rheological behavior of emulsion mainly due to the presence of paraffinic gel at
lower temperatures. Sjoblom et al. (2003) have presented the review on the current understanding of water in crude oil emulsions and the various
characterizing techniques such as near infrared spectroscopy, high pressure NIR studies, Nuclear Magnetic Resonance, etc.

Rheology is an important tool to define emulsion stability by measuring viscosity of emulsions for different shear rate as the emulsions mainly shows
non-Newtonian behavior. Bulk viscosity and interfacial viscosity (viscosity at oil-water interfaces) are important parameters of emulsion stability and
need to be determined in order to device suitable treatment methods. Weiss and MacClements (2000) studied the influence of Ostwald ripening and
droplet size on the rheology of hydrocarbon oil-in water emulsions stabilized by sodium dodecyl sulfate. Emulsions underwent a liquid-to-solid
transition when the mean droplet radius was decreased below 85 nm because of overlap of the electrical double layers surrounding the droplets. Moran
et al. (2006) studied the viscoplastic properties of crude oil-water interfaces. They developed novel micropipette technique to study the interfacial
rheology. Farah et al. (2005) studied the viscosity of six different waxy crude oil water emulsions for variation with temperature and different
dispersed phase volume fractions and correlated with Newtonian and non-Newtonian behavior of the crude oil-emulsion. Tadros (2004) carried out
extensive review on the application of rheology for the assessment and prediction of long term physical stability of emulsions in general.

Field Observations

Dutta and Ahamed (2003) carried out field study at Baharain field where crude oil emulsion poses challenge for flow assurance downhole. They
suggested a chemical injection to be applied in the gas lift operation which helps to reduce emulsion downhole and improve well performance. Kokal
(2006) carried out extensive study for potential emulsion problem before a cluster of field development plan. The study was aimed for different layers
of the reservoir and for multiple fields. He observed that the selection of pump, mainly, electrical submersible pump environment may tend to produce
emulsion due to high shear environments. Castro (2001) observed that heavy oil reservoirs are also prone to form emulsion, mainly, in the presence of
hot water/steam injection techniques. He observed that the stable emulsion of heavy oil do form in the porous medium of the reservoir thus blocking
the path of production. These emulsion can be treated using water soluble chemical demulsifier before and during the cyclic steam injection operation.
Oliviera and Gonzalvez (2005) carried study on the emulsion rheology for a Brazilian heavy crude oil. They studied the effect of water volume
fraction and shear rate condition on the droplet size and emulsion viscosity. The emulsion viscosity is then used for modeling the flow and pressure
drop calculation. They observed that below the wax appearance temperature, emulsion behave as shear thinning fluids. The viscosity and shear
thinning behavior increases as the droplet size distribution reduced. Aziz et al. (2002) have studied the rheology of crude oil emulsion of Ras-Budran
field near the Gulf of Suez where the crude oil has 15 % of asphaltene content. Due to increase in water production and Sea water breakthrough, tight
emulsions were formed in downhole and at surface facilities. Some of the chemical used to treat scales depositions (mainly calcium carbonate and
sulphate), hydrochloric acid treatment had an adverse effect on emulsion viscosity and asphaltene precipitation near well-bore.

Though several studies have been reported in an open literature on the fundamental aspects on crude oil emulsion rheology, still there is scope to
understand their behavior at in-situ reservoir conditions. Not much information is available on the effect of gas-oil ratio, effect of H2S environments,
and effect of porous media and last but not the least, the effect of pressure on the stability of emulsion at in-situ reservoir conditions. Additionally,
very few correlations have been developed and available for the viscosity of crude oil emulsions which considers the effect of high pressure, possibly
due to the lack of experimental data, which is indeed significant for reservoir modeling of field showing emulsion tendency. One of the important
aspects to be noted here is that there are no unique solutions available to treat these emulsions in fields but need specific studies for each field due to
their varying reservoir fluid compositions and different reservoir environments. Below a study on an Indian brownfield is presented, where crude oil
emulsions were formed near well-bore and at surface facilities. This was significant from the production loss from the well after well treatment and
evidence of emulsion in the separator. The discussion will be followed by suitable treatment methods to treat the well under consideration.

STUDY ON INDIAN WAXY CRUDE OIL EMULSION

A sample crude oil is obtained from the oil well-ABC from an oilfield in India which shows decrease in productivity. The water cut of the concern
well ranges from about 30 to 50 % during the production history of the well. The oil samples are collected from the well head separator of the
respective well and used for the study. The sample obtained does not represent the downhole sample, as the light components (particularly <C7+) from
the oil are expected to be released with the separator gas. For the recent study acquiring downhole sample will not serve the purpose due to the fact
that the well is experiencing pressure drawdown near well-bore region. This might be due to the skin damage that had happened due to emulsion,
scales/sludge, etc. In such case, acquiring downhole sample may not represent true reservoir sample. Hence, a study on the separator sample is carried
out. Figure 1 (a) show the oil sample collected from the candidate well-ABC. A visual inspection of this oil sample shows the presence of a viscous
emulsion layer. The emulsion layer of the oil sample is stable and does not break at surface conditions of temperature and pressure and also by heating
the emulsion sample at 90 oC at atmospheric pressure. It is obvious that the high water cut is causing the emulsion in the oil well. Also the crude oil is
observed to be thick and waxy at surface condition. It is to be noted that the emulsion layer and water layer of the sample were separated from the oil
layer before being used for all studies discussed henceforth.

Experimental Procedure

In order to study the stability of crude oil emulsion it is important to gather data on the composition and the viscosity of crude oil emulsion. The data
on composition will gives us an idea about the presence of surfactants. The data on viscosity can provide clue on the stability of emulsion, phase

58
inversion, and can also provide help in developing robust model for the reservoir management. A procedure to create emulsion of different water cut is
discussed now. A sample crude oil (after removing emulsion layer and associated water) obtained from the flowing well-ABC and injection water is
mixed in a desired ratio in a blender. The mixing of two phases is carried out at about 3000 RPM and for 15 minutes. The mixing temperature and
pressure is 40 oC and atmospheric pressure. An emulsion sample of about 40 ml from the blender is carefully transferred to the annular space of the
cup and bob assembly of the viscometer, so as to avoid emulsion breakout during transfer. Two different type of measurement are carried out. One is
at 40 oC and at atmospheric pressure in Fann-35 viscometer and other is at high temperature of about 150 oC and at 900 psi in Fann®-50 HPHT
viscometer. In case of HPHT studies in Fann-50 viscometer, initial temperature of the cup is maintained at about 40 oC. The values of viscosities are
measured for different shear rates. Indeed, it would be more realistic to form emulsion by mixing the crude oil and water in the pressurized chamber
maintained at reservoir temperature and pressure conditions in order to better simulate the reservoir scenario.

RESULTS AND DISCUSSION

Composition of the Crude Oil of Well-ABC

The crude oil sample collected at the well-head separator from the well is analyzed for Saturates, Aromatics, Resins and Asphaltene (SARA),
elemental composition upto C36+ components (using gas chromatograph), molecular weight and the density of the crude oil using Cryette® measuring
instrument. First, studies on the SARA analysis of crude oil samples are presented, followed by the results on the elemental composition analysis.

SARA analysis

The procedure to determine SARA fraction is briefly highlighted in Figure 3 and the results are presented in Table 1. It is observed from the SARA
compositional analysis that the asphaltene and resin percentage are low, however, saturates are much higher. This is possibly the strong reason of
waxy nature of the crude oil sample.

Elemental Composition Analysis

A Gas Chromatograph (GC) for the heavy oil is used to analyze the oil sample due to their thickness (waxy nature) at the surface condition. Figure 4
shows the graphical representation on the mole % of various components in the crude oil ranging from C1 to C36+.The composition analysis can also
be shown as in breakup of C7-15, C16-20 and C20+ components as given in Table 1. It is observed from the analysis (see Figure 4) that the crude oil
samples have considerable volume of heavier fraction (C15+ components). Therefore, the wax content (C15+ fractions) in the reservoir at present
condition is expected to be more than 30 %. This possibly result in the formation of wax along the well-tubing, and also in HPHT environment as
there is increase in WAT with increase in pressure (Sansot et al., 2005). Additionally, the wax particles can adsorb at the water-oil interface, resulting
in emulsion/gel. These wax particles contribute significantly to emulsion stability even in the presence of low level of asphaltene and resin (Ali and
Alqam, 2000; Visintin et al., 2008). It is also expected that asphaltene, resin and wax in present surface sample might be less as what it would be in the
reservoir, as most of it might have deposited in the reservoir or in the well tubing during flow and fluid sampling. The molecular weight and density is
measured for the oil sample and is given in Table 1. It is to be noted that the current study is utilizing the crude oil sample collected at the well-head
separator for the analysis and that the sample is heated at 60 oC before injecting it to GC, as the sample was very thick (waxy) at surface conditions.

When crude oil (having heavier hydrocarbon fractions such as waxes and asphaltene) in the reservoir comes in contact with formation water, forms
emulsion in the presence of shear (Ali and Alqam, 2000). These emulsions are non-Newtonian in nature even at high pressure and high temperature
conditions in the reservoir. These emulsions pose formidable challenge for flow assurance downhole and at surface facilities due to their highly
viscous nature. It is crucial to break these emulsions either downhole or at surface facilities in order to increase the oil production. Rheological
properties of these emulsions are indeed significant to understand their stability at surface and HPHT conditions and are also required to design and
develop efficient lift operations, fluid handling systems and robust reservoir models. Rheological studies on the stability of sample crude oil emulsion
are carried out at surface and HPHT conditions and presented below.

Crude Oil

Dilute with n-heptane

Maltenes
Adsorbed on Alumina Oxide/Silica by eluting with

n-heptane DCM/Methanol Precipitate


Toulene

Saturates Aromatic Resins Asphaltene

Fig 3: Typical scheme for separating crude oil into Saturate, Aromatic, Resin and Asphaltene (SARA) components

59
Table 1: Analysis of the crude oil sample of the well-ABC

Molecular Density,
Components Saturates Aromatics Resins Asphaltene
C7-15 C16-20 C20+ Inorganic weight gm/cm3 at
, Wt % (S) (A) (R) (A)
(measured) 15.5 oC
SARA 84.68 13.10 5.29 0.21 44.19 18.98 33.19 0.46 261 0.850

10

1
Mole %

Waxy components
0.1

Well ABC

0.01
Component

Fig 4: Composition analysis of the crude oil of well-ABC using Gas Chromatography

Rheology Studies on Crude Oil Emulsion

Firstly, the rheological studies on the crude oil emulsion at surface condition are presented followed by the rheological studies at HPHT conditions. A
study on the demulsification treatment is also presented thereafter.

Viscosity of Emulsion at Surface Condition

Figure 5 (i) and (ii) shows the effect of water cut and shear rate on the viscosity of emulsion. It is observed from Figure 5 (i) that as the water cut
increases, the viscosity of emulsion increases for all shear rates. At low water cut, water remains in the dispersed phase and oil as a continuous phase.
As water cut increases, the water droplets grow in size making resistance to flow. This results in an increase in the viscosity of emulsion. However, it
is observed that the emulsions with low water cut are tighter and more stable than the emulsions with high water cut. This is due to the fact that at the
low water cuts, emulsion have narrow distribution of water droplets, while, at high water cuts, emulsion shows broader distribution of water droplets.
This result in Ostwald ripening, wherein, the large droplets grows in size on the expense of smaller droplets (Sjöblom, 1996). As water cut increase, at
certain water cut of about 75 to 80 %, phase inversion occurs (known as phase inversion point). After this point the water phase becomes continuous
and oil phase become dispersed. The emulsion is less stable onwards and viscosity decreases rapidly with increase in water cut beyond the phase
inversion point. Figure 5 (ii) shows the effect of shear rate on the viscosity of emulsion. The shear rate values are varied from about 2 to 1000 S-1.
Shear rate influence the viscosity of emulsion. At low shear rate, emulsions are highly viscosity and at high shear rate these emulsion shows low
viscosity. This is because emulsions are non-Newtonian shear thinning in nature. In case of emulsion of WC of 20 % and 40 %, data on viscosity for
low shear rate of 50 to 100 S-1 shows increase in viscosity as shear rate increases. This is possibly due to insensitivity of Fann-35 viscometer to
measure such low viscous emulsion at low shear rate (below 100 S-1). The zero shear viscosity (values of viscosity for near zero values of shear rate)
of the emulsion of 80 % WC is observed to be about 300 cP which decreases as shear rate increases. This explains that if the emulsion is not under the
influence of external shear, remains stable donwhole by maintaining high viscosity and does not flow, unless until acted by external shear. The
viscosity decreases to about 50 cP as shear rate increases beyond 600 S-1 and remains there. This is because at high shear rate, the two phases
encounter vigorous mixing, which helps in formation of emulsion, however, such high shear rate also helps breaks films around the water droplets
thereby breaking emulsion. The rates of formation and breakage of emulsion becomes almost equal, making the viscosity of total system constant.
This information on viscosity at varying shear rate can help design suitable lift operations (particularly for ESPs) in the field.

Viscosity of Emulsion at High Pressure and High Temperature

The study on the effect of pressure on the viscosity of emulsion in HPHT Fann-50 viscometer is presented in two ways. One way is to record the
viscosity change with respect to pressure at constant temperature of 40 oC (direct effect of pressure). The other way is to study the change in viscosity
of the emulsion with respect to temperature in a pressurized environment. Firstly, the results on the direct effect of pressure on the viscosity of
emulsion are discussed followed by the effect of pressurized environment.

60
120
Water-in-Crude Oil Phase Inversion
emulsion Crude Oil-in-
100
102 S-1 Water emulsion

80 170 S-1

Viscosity, cP
340 S-1
60
510 S-1

40 1021 S-1

20

0
0 10 20 30 40 50 60 70 80 90 100
Water Cut, %

(i)

1000

100
Viscosity, cP

80 % WC

60 % WC

40 % WC
10
20 % WC

1
0 200 400 600 800 1000 1200
Shear rate, S -1

(ii)
Fig 5: Effect of (i) water cut and (ii) shear rate on the viscosity of crude oil emulsion at 40 oC and 1 atm in Fann-35 viscometer

Emulsion sample of about 40 ml from the blender is carefully transferred to the cup of Fann-50 viscometer which is maintained at about 40 oC. After
the sample is transferred, the viscometer is operated at about 25 S-1 shear rate and viscosity of emulsion is measured. The pressure is then increased
from atmospheric pressure of about 15 psi to about 900 psi in the step of 100 psi and the viscosity at every step is measured and documented. This data
is shown in Figure 6. During this study, the temperature of emulsion is maintained at 40 oC. No increase in the viscosity is observed with increase in
the pressure from 15 psi to 900 psi at the constant temperature. This shows that the pressure does not directly affect the viscosity of emulsion.

The effect of temperature on the viscosity of emulsion at high pressure is now studied. The experimental conditions at the end of first run (as in Figure
6) are continued, that is, the crude oil sample is now maintained at 900 psi and 40 oC. The temperature of the emulsion sample is then raised from 40
o
C to about 150 oC (a reservoir temperature). The time taken to increase the temperature of the sample from 40 oC to 150 oC is about 140 min. The
pressure is maintained at about 900 psi during this operation. Two different experimental runs are carried out to see the reproducibility of the trends.
Viscosity of emulsion is measured during the temperature rise for these runs. Figure 7 shows the data on the viscosity as a function of temperature at
900 psi for two reproducible trends for the well-ABC crude oil emulsion sample. It is observed that in the presence of pressurized environment the
effect of temperature in reducing the viscosity weakens as shown in Figure 7. The viscosity of the emulsion sample is observed to be around 1000 cP
at 40 oC and 900 psi. As temperature is increased in a pressurized environment of 900 psi, the viscosity values decreased to about 150 cP. No further
decrease is observed for sufficient time, showing that the emulsion is stable at 150 oC and 900 psi. It is well known that as the temperature increases
the viscosity of the emulsion decreases. However, if there would have no pressurized environment of 900 psi at high temperature of 150 oC, it is
expected that the emulsion would have been no longer survive or may attain much lower viscosity than the value of 150 cP. This can be attributed to
high entropy of the emulsion at 150 oC, resulting in breakage of emulsion droplets and evaporation. However, in the presence of pressurized
environment, the entropy of the system is reduced and thus reducing temperature effect, thereby making emulsion stable. Such stable emulsions will
further become highly stable and tighter during flow from well-bore to well surface, as there will be decrease in temperature and phase separation of
light components from the crude oil.

Based on the above analysis on the crude oil composition and rheological studies, it can be concluded that the waxy nature of the crude oil is
responsible for the formation of stable emulsion even in the presence of low level of asphaltene and resins components. High pressure environment of
the reservoir plays an important role in stabilizing emulsion downhole than the crude oil composition alone.

61
10000

Viscosity, cP
1000

at 85-95 oF and 25 S-1


100
0 100 200 300 400 500 600 700 800 900 1000
Pressure, psi

Fig 6: Effect of pressure on the viscosity of well-ABC crude oil emulsion at 40 oC and 25 S-1 in Fann-50 viscometer

10000

1000
Viscosity, cP

Run 1
Run 2

100

10
0 50 100 150 200 250 300 350
Temperature, oF

Fig 7: Effect of temperature on the viscosity of well-ABC crude oil emulsion in pressurized environment of 900 psi in Fann-50 viscometer.

Table 2: Composition of different mixture of chemicals to treat the emulsion of the well-ABC

Sr. No. Chemicals Mixture 1, Mixture 2, Mixture 3,


v/v % v/v % v/v %
1 Xylene (X) 30 15 15
2 Toluene (T) 30 15 15
3 Ethylene glycol mono-butyl ether 10 40 25
(EGMBE, a mutual solvent)
4 Isopropyl alcohol (IPA) 10 10 10
5 Acetic acid (AA) 10 10 10
6 Surfactant 5 5 20
7 Citric acid (CA) 2.5 2.5 2.5
8 Naphthalene (N) 2.5 2.5 2.5

Demulsification of the Waxy Crude Oil Emulsion

An oilfield generally co-produces crude oil and water. In presence of high shear environment if the emulsion of crude oil with water are formed, they
are difficult to flow in the reservoir or well tubing due to their viscous nature and hence require breaking them in-situ to improve production.
However, the task become difficult due to the extreme reservoir conditions and hence the practice of breaking emulsion in the well head separator is
generally employed using gravity settling tanks, cyclone separator, pH adjustment, filtration, membrane separation, etc. All above technique require
high volume separator tools which are difficult to install into deep offshore field.

62
(i) (ii)

(iii) (iv)

(v) (vi)

Fig 8: Effect of various combinations of


chemicals as in Table 2 on the stability
of emulsion (i) toluene, 100 % (ii)
toluene and xylene, 50 % each (iii)
mixture 1 (iv) mixture 2 (v) mixture 3
(vi) xylene, toluene and surfactant,
equal volumes (vii) EGMBE, 100 %

(vii)

63
Additionally, breaking emulsion at wellhead separator does not suffice the goal of high production rate from the oil well. Therefore, chemical
treatment using demulsifier has been common practice to break emulsions near well-bore. It is important to check the efficacy of the demulsifier and
an optimum quantity for the demulsification job. An overdose of demulsifier may aggravate the emulsion or asphaltene precipitation.

Emulsion Treatment

As a part of short term goal to find out the remedy for the emulsion problem of the well-ABC, a treatment is required to be designed. A treatment
design of the mixtures of the chemicals shown in Table 2 is used to break the emulsion in the laboratory and for possible application in the well. Lab
studies are carried out for the effectiveness of these chemicals on the demulsification. Three different compositions of these chemicals are prepared as
in Table 2 and tested with the crude oil emulsion to see any effect on the stability of emulsion. Mixture 1 contains higher percentage of xylene and
toluene, mixture 2 contains higher percentage of ethylene glycol mono-butyl ether (EGMBE) and mixture 3 contains more surfactant as compared to
mixture 1 and mixture 2. Additionally, other combinations are also tried: toluene (100 %), EGMBE (100 %) and a mixture of xylene + toluene (50 %
each), and, xylene + toluene + surfactant (33.33 % each), to treat the emulsion and to see the individual effect of the above combinations on the
stability of emulsions. All the tests carried at normal temperature and pressure conditions. This study will help to understand and recommend suitable
combination/solvents to treat the emulsion based on the effect of the components of the recipe.

An emulsion of the crude oil and injection water in the ratio, 50:50, is created in the lab. The emulsion is observed to be stable at normal atmospheric
conditions. Equal volumes of emulsion (60 ml) and equal volume of the above mixtures (as in table 2) is used to test the emulsion behavior. Figure 8 (i
to vii) shows effect of various mixtures on the stability of emulsions. Toluene is observed as a best solvent to treat the emulsion. Clear separation of
water is observed from the emulsion. The oil components and asphaltenes are observed to be dissolved in the toluene. A good separation of water and
oil is also observed with the combination of toluene and xylene, however, a slight occurrence of emulsion phase is observed at the interface. Mixture 1
is efficient in breaking the emulsion effectively. However, turbidity is observed in the water phase, which may be because of the presence of acids and
alcohols. The presence of acids in mixture 1, in more concentration, can cause the asphaltene to precipitate and can aggravate the emulsion tendency
near well-bore. Other way to pump mixture 1 is to pump xylene and toluene first, followed by pumping the mixture of the rest of solvents. Same
results are obtained in doing so. A result with the application of mixture 2 on the emulsion is quite similar to that of mixture 1. Turbid phase is
observed at the bottom of the test and a dense phase of oil at the top. Higher percentage of EGMBE is not recommended, as it does not observed to be
a good solvent for asphaltene. It just act as agent to sweep the water from emulsion, as is evident from (vii) in Figure 9. Little water getting
precipitated at the bottom of the test can be seen in the figure. Suspensions of heavier particles across the test are also observed. Mixture 3, containing
higher percentage of surfactant than in mixture 1 and 2, is not a good solvent. A loose emulsion is observed to get formed in between the two
separated phases of oil and water. In this case higher surfactant may cause emulsion. A combination of xylene, toluene and surfactant in equal volume
also help to break the emulsion. Here the effect of surfactant might get negated due to more quantity of xylene and toluene in the mixture. However
higher percentage of surfactant could worsen the situation. Based on above test, it is observed that toluene act as a good solvent and can effectively
help break the emulsion. So, only toluene needs to be pumped with light oil as a carrier fluid to treat the emulsion near well-bore region and in the
well tubing.

CONCLUSION

It is observed from above studies that the crude oil sample of well-ABC is capable of forming emulsions with the formation water. It is also observed
from the investigation on the SARA/GC composition analysis of this crude oil that the crude oil shows the presence of more waxy components in the
oil with little asphaltene. Effects of various parameters such as shear rate, water cut, temperature and pressure on the viscosity of crude oil emulsion
were studied individually. The studies on the rheology of emulsion at surface conditions shows a phase inversion point in the water-cut range between
75 and 80 %. The emulsion shows non-Newtonian behavior. The zero-shear viscosity of the emulsion mixture (measured at near zero shear value) at
surface and at HPHT conditions were observed to be between 300-350 cP and 140-160 cP, respectively. It is observed that the magnitudes of
viscosities of the emulsions are higher at surface condition than at HPHT condition. Effect of high pressure and high temperature is studies on the
viscosity of emulsion. It is observed that the pressure does not affect the viscosity of emulsion directly. However, the pressurized environment helps
emulsion to stay stable even at high temperature conditions. The high pressure environment weakens the effect of temperature on breaking the
emulsions. This explains that the emulsions in the reservoir in-situ conditions of high pressure are stable even at the high reservoir temperature.

REFERENCES

Ali, MF, and Alqam, MH (2000) “The role of asphaltene, resins and other solids in the stabilization of water in oil emulsion and its effect on oil
production in Saudi oil field”, Fuel, Vol 79, pp 1309-1316.
Auflem, IH, Kallevik, H, Westvik, A and Sjo¨blom, J (2001) “Influence of pressure and solvency on the separation of water-in-crude-oil emulsions
from the North Sea”, Journal of Petroleum Science and Engineering, Vol 30, pp 1–12.
Aziz, HMA, Darwish, SH, Abdeen, FM (2002) “Downhole emulsion problem, the cause and remedy, Ras Burdan field”, SPE 77847. Presented at SPE
Asia Pacific Oil and Gas Conference and Exhibition, Melbourne, Australia, 8-10 October.
Castro, LU (2001) “Demulsification treatment and removal of in-situ emulsion in heavy-oil reservoirs”, SPE 68852, Presented at SPE Western
Regional Meeting, Bakersfield, California, 26-30 March.
Dutta, B, and Ahamed, HH (2003) “Production improvement by downhole demulsification: A simple and cost effective approach”, SPE 81568.
Presented at SPE 13th Middle East oil show and conference, Bahrain, 5-8 April.
Evdokimov, IN, Eliseev, NY and Eliseev YD (2004) “Effect of asphaltenes on the thermal properties of emulsions encountered in oil recovery
operations”, Fuel, Vol 83, pp 897-903.
Farah, MA, Oliveira, RC, Caldas, JN, and Rajagopal K (2005) “Viscosity of water-in-oil emulsions: Variation with temperature and water volume
fraction”, Journal of Petroleum Science and Engineering, Vol 48, pp 169– 184.

64
Janssen, PH (1998) “Emulsion characteristics of high water-cut oil wells”, SPE 49077, Presented at SPE Annual Technical Conference and Exhibition,
Louisiana, 27-30 September.
Kokal, S, and Al-Juraid, J (1999) “Quantification of various factors affecting emulsion stability: water-cut, Temperature, shear, asphaltene content, de-
emulsifier dosage and mixing different crudes”, SPE 56641, Presented at SPE Annual Technical Conference and Exhibition, Houston, Texas, 3-6
October.
Kokal, S (2005) “Crude oil emulsion: A state of the art review”, SPE Production and Facilities, Vol 20, pp 5-13.
Kokal, S (2006) “An investigative study of potential emulsion problem before field development”, SPE 102856, Presented at SPE Annual Technical
Conference and Exhibition, San Antonio, Texas, 24-27 September.
Li, M, Lin, M, Wu, Z, and Christy, AA (2004) “The influence of NaOH on the stability of paraffinic crude oil emulsion”, Fuel, Vol 84, pp 183-187.
McLean, JD, and Kilpatrick, PK (1997) “Effects of asphaltene solvency on stability of water-in-crude-oil emulsions”, Journal of Colloidal and
Interfacial Science, Vol 189, pp 242–253.
Mehta, SD (2005) “Making and breaking of water in crude oil emulsion”. Master of Science Thesis, Texas A&M University.
Morana, K, Yeunga, A, and Masliyah, J (2006) “The viscoplastic properties of crude oil–water interfaces”, Chemical Engineering Science, Vol 61, pp
6016 – 6028.
Oliveira, RCG, and Gonzalvez, MAL (2005) “Emulsion rheology: Theory vs. field observation”. Offshore Technology Conference.
Quintero, CG, Noïk, C, Dalmazzone, C, and Grossiord, JL (2008) “Modeling and characterization of diluted and concentrated water-in-crude oil
emulsions: comparison with classical behavior”, Rheologica Acta, Vol 47, pp 417–424.
Sansot, JM, Pauly, J, Daridon, JL, and Coutinho, JAP (2005) “Modeling high pressure wax formation in petroleum fields”. AIChE J, 51, 2089-2097.
Schorling, PC, Kesse,l DG, and Rahimian, I (1999) “Influence of the crude oil resin/asphaltene ratio on the stability of oil/water emulsions”, Colloids
and Surfaces A: Physicochemical and Engineering Aspects, Vol 152, pp 95–102.
Sjöblom, J (1996) “Emulsions and emulsion stability”, Marcel Dekker: New York.
Sjoblom, J, Aske, N, Auflem, IH, Brandal, O, Havre, TE, Saether, O, Westvik, A, Johnsen, EE, and Kallevik H (2003) “Our current understanding of
water-in-crude oil emulsions. Recent characterization techniques and high pressure performance”, Advances in Colloid and Interface Science, Vol
100-102, pp 399-473.
Spiecker, PM, Gawrys, KL, Trail, CB, and Kilpatrick, PK (2003) “Effects of petroleum resins on asphaltene aggregation and water-in-oil emulsion
formation”, Colloids and Surfaces A: Physicochemical and Engineering Aspects, Vol 220, pp 9-27.
Tadros, T (2004) “Application of rheology for assessment and prediction of the long-term physical stability of emulsions”, Advances in Colloid and
Interface Science, Vol 108 –109, pp 227–258.
Visintin, RFG, Lockhart, TP, Lapasin, R, and Antona, PD (2008) “Structure of waxy crude oil emulsion gels”, Journal of Non Newtonian Fluid
Mechanics, Vol 149, pp 34-39.
Weiss, J, and McClements, DJ (2000) “Influence of Ostwald ripening on rheology of oil-in-water emulsions containing electrostatically stabilized
droplets”, Langmuir, Vol 16, pp 2145-2150.

65
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Event analysis for offshore riser failure


S.K.Bhattacharyya1, Srinivasan Chandrasekaran2*, and MVS Ram Prasad3
1
Professor and Head, Department of Ocean Engineering, Indian Institute of Technology Madras, Chennai, India
2
Associate Professor, Department of Ocean Engineering, Indian Institute of Technology Madras, Chennai, India
3
Research scholar, Department of Ocean Engineering, IIT Madras
*
corresponding author E-mail: drsekaran@iitm.ac.in

ABSTRACT

Operation experiences in offshore industry show that accidental actions contribute significantly to their failures. Such events are commonly traced
back to errors in design, fabrication or operation. Offshore risers are sections of pipeline rising from the seabed to an offshore installation, extending
as far as the top sides Emergency Shut Down Valve (ESDV); at the seabed end, risers are isolated from the pipeline by Sub Surface Isolation Valve
(SSIV). Risers often contain large inventories of oil or gas at high pressure and have the potential for catastrophic consequences. Majority of riser
leaks occur in the splash zone as the external corrosion and boat impact are dominant in this zone, in conventional jacket platforms. Historical data on
riser leaks are inadequate to predict fatality or damage impacts; so the approach used to estimate risk is to combine modeled consequence zones with
the impact criteria and the initial personnel distribution along with their assumed evacuation response. A case study is presented for an existing
offshore field; to protect company confidentiality, the gas field is designated as ABC gas field which is located in west coast of India in a shallow
water block with water depth ranging between 10-15m. Event analysis for riser failure is carried out by considering parameters specific to leak
scenario namely: i) location of the leak in the riser; ii) size of the leak; iii) orientation of the leak; iv) detection of the leak in order to initiate mitigating
actions; v) isolation of leak by ESDV and SSIV; vi) ignition and possible explosion; vii) weather conditions; viii) distribution of personnel when leak
occurs; and ix) evacuation response.

KEY WORDS: offshore riser failure, Event analysis, annual fatality rate, oil and gas industry

INTRODUCTION

Several platforms may have to be installed to exploit a large oilfield and these are often integrated with a subsea collecting system linked by pipelines
in order to cover a large area. For very deep waters, one solution is the Tension Leg Platform (TLP), which is a buoyant compliant platform,
resembling a huge drilling rig, and is tethered to the seabed by jointed legs kept in tension by controlled ballast adjustments. Alternatively, a subsea
collection system may be linked via a production riser to a Floating, Production, Storage and Offloading (FSPO) vessel that is either a purpose built
ship or a converted tanker or semi-submersible rig where oil is offloaded by a shuttle tanker. Types of common offshore platforms are shown in Fig. 1.

Fig. 1 Types of platforms


Power is generated onboard to drive production equipment and support life. All production systems are constantly monitored for leaks, since both oil
and gas are hazardous substances and extremely flammable. Offshore oil and all gas are brought to shore by pipelines which operate throughout the
year. Pipeline routes should be as short as possible. Slopes that could put stress on unsupported pipe are avoided and seabed sediments are mapped to
identify unstable areas and to see if it will be possible to bury the pipe. Pipeline construction begins onshore, as lengths of pipe are waterproofed with
bitumen and coated with steel-reinforced concrete. This coating weighs down the submarine pipeline even when it is filled with gas. The prepared pipe
lengths are welded together on a barge. As the barge winches forward on its anchor lines, the pipeline drops gently to the seabed, guided by a 'stinger'
(Gerwick, 1999). A typical pipeline installation scheme is shown in Fig. 2.

Fig. 2 Typical pipeline installation scheme

The success of the offshore industry depends on winning and keeping the trust and good will of many stakeholders - employees, suppliers,
shareholders, end customers, communities. Like any relationship, it requires good communication, mutual understanding, honest endeavour and
recognition of shared benefits. The oil and gas industry has brought many benefits to our lives. Oil and gas and their downstream products underpin
modern society, supplying energy to power industry, heat homes, fuel transport to carry goods and people all over the world and are used as raw
materials from which many items in everyday use are derived. The offshore industry employs hundreds of thousands of people and makes a major
contribution to the economy.

SCENERIO IN RISER LEAKS

Majority of riser leaks occur in the splash zone because external corrosion and boat impact are dominant in this zone. Material failures including those
due to corrosion also occur in subsea part because this is often the longest part of the riser and also is difficult to inspect. Failures above the splash
zone occur mainly at the fittings. Where releases occur at deep locations, hydrocarbons diffuse significantly on rising to the sea surface. Typically, it is
assumed that the spread of the hydrocarbon release at the sea surface equals 20% of the water depth. At significant water depths the effective mass
flow rate per unit area when breaking the sea surface is so low that the gas content falls below the Lower Explosive Limit (LEL) if not already below
LEL. Deepwater releases are usually strongly influenced by current effects over the water depth making their eventual location at the sea surface
difficult to predict. Deepwater releases also reduce any potential buoyancy reductions on floating installations. In the risk analysis of offshore
installations of risers and pipelines, the probabilities that are considered are: (a) Subsea: 30%, (b) Splash zone: 50% and (c) Platform: 20%. However,
these values are very sensitive to the decision to exclude leaks from fittings. The probabilities that are considered for leak sizes in risers and pipelines
are: (a) Small (< 20mm): 57%, (b) Medium (20 to 80mm): 15% and (c) Large (> 80mm): 28%. The failure by rupture is taken as 21% and such
failures are always in the ‘Large’ category (Peter and John, 2000).

In the analysis of riser risks, it is normal to assume that the riser Emergency Shut Down Valves (ESDV) at the topsides close quickly enough to
prevent additional flow from the process system significantly adding to the release. Based on comparison with topsides ESDVs, the failure rate of
SSIV has been estimated to be 10−2 per demand. A more pessimistic failure rate, making allowance for the subsea environment and the lack of
experience with large diameter Subs Surface Isolation Valve (SSIV), could be 10−1 per demand. For studies intended to show whether or not SSIVs
are potentially cost effective, it is appropriate to use the lower value of 10−2 per demand. For a given scenario, the consequences must be modelled to
estimate risk, and the consequences of major interest are usually (a) jet fires resulting from ignited gas releases, (b) pool fires and smoke plumes
resulting from ignited oil releases, (c) oil slicks and pollution resulting from unignited oil releases and (d) toxic releases. Gas can be released from
production risers, gas lift or gas injection risers. A subsea release can rise to the surface and engulf part of the facility if it is near the sea surface. This
could introduce gas in areas, such as living areas, where it can be dangerous. Typically the release is isolated by surface riser ESDV at the wellhead or
injection point. SSIV can be used to limit the overall size of release. The time taken to close a wellhead by closing these valves should be short in
order to limit gas releases. The principal effects of Sub Surface Isolation Valve (SSIV) on riser releases are frequency reduction and discharge
reduction. Presence of SSIV makes it easier to isolate and repair the riser, and hence reduces the frequency of riser leaks. For example, the fire
accidents on South Pass 60 Platform B in Gulf Of Mexico (GOM) and Cormorant A in UK might have been prevented if SSIVs had been fitted.
However, there is no data available on probability of frequency reduction. Retrospective installation of SSIV in a riser itself may be a hazardous
activity. If the SSIV is closed when required, it will reduce the inventory available for discharge, and hence reduce the duration of the release. The
reduced size and duration of release reduces the likelihood of the gas cloud finding an ignition source. The ignition probability, however, is very
uncertain. Ignition probability can be significant for full-bore releases in the vicinity of the platform when the probability of finding a source of
ignition is high. In such a case, closure of SSIV will reduce the otherwise high probability of occurrence of a large fire. If the release is ignited, the
SSIV acts to reduce the extent and duration of the fire. Prevention of smoke or fire trapping of personnel inside the accommodation is most significant

68
for full-bore gas riser fires or for oil pool fires. In smaller fires, personnel may still be trapped, but there is a much reduced probability of the living
quarters being impaired before the fire dies down. Prevention of structural collapse is most relevant for fires at sea level, with the potential to cause

structural collapse of the platform or major damage to the side shell of floating installations. Timely closure of SSIV in the burning riser reduces the
probability of the fire burning long enough to cause spread of fire, thereby reduces the probability of failure of other risers.
Considering the significance of the role of risk assessment in offshore risers, a case study is considered in this study for an existing offshore field
producing gas in order to bring out the methods of application of diverse qualitative and quantitative risk assessment and integrity management
methodologies available in the literature. To protect company confidentiality, the gas field is designated as ‘ABC gas field’ in the thesis. It lies in the
west coast of India in a shallow water block with water depth ranging between 10 to 15m. Currently, the ABC field has five offshore jacket platforms
and the produced gas from this field is conveyed to an onshore processing terminal by five pipelines of various sizes (see Fig. 3). Event analysis is
carried out for the riser system of this field and recommendations are made for the integrity management of pipeline system.

Fig. 3 Layout of the ABC field

Risk factors that can cause failure of pipelines can be broadly classified as (a) external interference, (b) corrosion (external and internal), (c) design
defects, (d) material defects, (e) environmental hazards, and (f) human and operational errors. If transported product is corrosive, it results in internal
corrosion of pipelines. In oil and gas industry, injection of corrosion inhibitors into pipelines is carried out to protect pipelines against internal
corrosion. So, any interruption to this corrosion inhibitor injection will also lead to internal corrosion. Probability of internal corrosion is more in 6 O’
clock position of the pipe, other low points, bends and fittings. (DNV RP F101, 1999). Due to abrasion of the transported product, internal erosion
occurs. This may result in deterioration of the inner wall and it becomes a primary target for corrosion. Internal abrasion is generally observed at
bends, tees, valves etc. During design phase of offshore hydrocarbon pipelines, the aspects that should be thoroughly addressed are given in API RP
1111, 1999. If these are not evaluated thoroughly, the design defects result in failure of pipelines by (a)combined bending and tension during
installation and operation, (b) collapse due to external pressure, with the pipe either empty or filled, (c) buckling and collapse due to combined
bending and external pressure, (d) loss of pipeline stability against horizontal or vertical displacements during construction and operation, (e) effects
of thermal expansion and contraction, (f) in place and in service pipeline repair capabilities, (g) fatigue due to hydrodynamic and operational loading
and (h) burst due to net internal pressure.

Pipeline components in a pipeline system should be designed for maximum differential pressure between internal and external pressures to which the
components will be exposed during construction and operation. The maximum differential pressure for a flow line may be due to a shut in pressure
condition. This condition may result from closure of a valve at the production facility without closing valves at the tree, manifold or down hole safety
valve. An important risk that should be evaluated during design phase is the effect of external pressure on all subsea pipeline systems. The significance
of this risk has been demonstrated by buckling of large pipelines subjected to severe bending and external pressure. The design should consider static
loads imposed on the pipeline (DNV 1981). These include the weight of the pipe, coating, appurtenances, attachments; external and internal
hydrostatic pressure and thermal expansion; and the static forces due to bottom subsidence and differential settlement. The weight of submerged
pipeline can be controlled through the combination of the pipe wall thickness and the density and thickness the concrete coating. Weight calculations
should consider stability, both when empty and when full with the fluid to be carried. The weight related forces are of special concern where the
pipeline is not continuously supported, i.e where spans are expected to occur. Spans are also of concern where seismic liquefaction of the supporting
bottom could occur and also where mudslides could occur. Consideration should be given to prevent unacceptably long unsupported lengths by use of
dumped gravel, attached supports, sand bagging, or other suitable means.

69
ENVIRONMENTAL HAZARDS

Earthquakes can liquefy sea bottom sediments supporting a pipeline. As a result, a pipeline may tend to sink, float or move laterally, depending on
specific gravity relative to the liquefied bottom sediment. Hurricanes, cyclones and typhoons can cause high currents and large cyclic wave action,
which together or individually can also cause similar liquefaction. Mudslides, where the moving sediment is effectively liquefied, may subject a
pipeline to large lateral forces (Leynes et al, 2005), thus creating similar effect on a pipeline. In areas of potential mudslides, where the severity of the
slide could cause a tensile pull of the magnitude that might cause damage to a platform or to a subsea connection, break-away connectors should be
considered for protection of the platform or other pipelines (Brillinger, 2003). When conditions are such that an oil spill might result from break away,
the design should provide a built-in check valve to minimize loss of fluid from the pipeline upon breakaway. Sediment transport or scour of
susceptible soils due to bottom currents and/or wave action may result in exposure of a buried pipeline, loss of soil restraint or increase in free spans. It
may not be possible to quantify the risk due to these natural phenomena for a specific offshore pipeline in a given location. If possible, alternative
route may be desirable to avoid a particular zone that is susceptible to these natural phenomena. In the rare conditions where weight coating or
trenching methods may not represent a suitable solution, such as solid rock surface or in shallow water with very high currents, the use of anchors or
pipeline weights may be a viable solution.

RISK ASSESSMENT

TNO Purple book (1999) mentions that risk can be described in different ways: individual risk, societal risk, maximum individual risk, average
individual risk of exposed population, average individual risk of total population and average rate of death. Two popular measures are individual risk
(IR) and societal risk (SR). The former is usually shown on risk contour plot, while the latter is presented with a frequency-number (FN) curve. The
individual risk is defined as the probability of death at any particular location due to all undesired events. It can be expressed as the probability of a
person at a specific location becoming a casualty within a year. For scenarios with risk involving multiple fatalities, the societal risk is defined as the
relationship between the frequency of an incident and the number of resulting casualties. It is usually expressed in the form of a graph of cumulative
frequency (F) of N or more casualties plotted against N (FN curve). Quantitative Risk Assessment (QRA) in offshore industry originated towards end
of 1970s when the methodology and data were captured from nuclear industry. However, Norwegian Petroleum Directorate (NPD) is the organization
that took forward by issuing guidelines for safety evaluation of platform conceptual design (NPD, 1984). These regulations required QRA to be
carried out for all new offshore installations as early as in the conceptual design phase. For many years, Norway was the only country using QRA
systematically. The offshore industry and the authorities in UK were opposing this approach, stating that QRA studies were not the right way to
improve safety. UK authorities, however, came around 10 years later due to Piper Alpha platform accident in 1988. The official inquiry into this
accident recommended that QRA must be introduced into UK legislation (Cullen, 1990). In 1991, Norwegian Petroleum Directorate replaced the 1981
guidelines for risk assessment by ‘Regulations for Risk Analysis’ (NPD, 1990). UK government issued further regulations following the 1992 Safety
Case Regulations (SCR). In the context of this presentation the most important is Prevention of Fire and Explosion and Emergency response
regulations (PFEER, 1995). These regulations present important requirements for active and passive safety systems, as well as emergency
preparedness systems and functions. Next step towards development was in 1998, when blast and fire research carried out for platform topside systems
(BEFETS, 1998). This has focused attention on high blast loads caused by possible gas explosion scenarios on platforms. As a result of this work,
considerable attention is being given to evaluating how explosion scenarios may be included probabilistically in the QRA models. Pipeline operators
that have developed their own risk based approaches include Nova Corporation of Alberta, British Gas and Dow Chemical. There are also many
publications by consulting companies that have developed and applied risk analysis on behalf of pipeline companies. It should be noted that
determining full cost of an incident requires help and support from various groups or departments within the pipeline company.

INTEGRITY MANAGEMENT

Cagno and Caron (1999) proposed a robust Bayesian approach to support the replacement policy of low pressure cast iron pipelines used in
metropolitan gas distribution networks by the assessment of their probability of failure. The analytic hierarchy process has been used as elicitation
method of an expert opinion to determine the distribution of gas pipeline failures. A real world case study is presented in which the company expertise
has been elicited by an adhoc questionnaire combined with historical data by means of Bayesian inference. Walied (1999) presented a probabilistic
procedure to conduct local fracture evaluation of pipelines subjected to multiple cracking in service. Mixed mode fracture mechanics is applied to
multiple cracking cases of two non-coplanar parallel cracks in an infinite elastic solid plate under remote tension. A numerical simulation model is
then used to determine the relation between the proximity of cracks interacting tips and the conditions under which they may not coalesce. Based on
these results, a Monte Carlo simulation model is applied to establish a correlation between the existing population of crack configuration and the
reliability of pipelines affected by multi-site damage. Thus, the relation between failure probability and pipelines serviceability is investigated. It is
concluded that the maintenance strategies used to optimize safety for pipelines could be based on probabilistic analysis that permits risk assessments
associated with the pipelines operating decisions as well as the type and frequency of inspection done to ensure continued safe operation of these lines.

Mazzola (1999) discussed with a case study the accidental scenarios which may possibly occur as a consequence of a pipeline failure and assessed
thermal effects caused by jet fire developing from different rupture sizes. In performing hazard assessments, mathematical modeling techniques with
varying degrees of complexities are adopted. CFD models represent the most accurate approach for determining the characteristics of turbulent flow
fields, including the effects on any impacted structure, and the thermal loading acting on the structure itself. However, they are mathematically
complex; require significant computing resources and also a significant expertise due to the utilization of numerical coefficients. Therefore, these
models are not normally used in routine hazard assessment, where a large number of repetitive calculations must be performed quickly. Cramer et al
(2000) addressed methodology for inspection planning of the pipeline systems and shown how these techniques are being applied by frontrunners in
the offshore oil and gas industry in the inspection planning of pipelines and risers. A detailed case study of risk based inspection planning using the
DNV developed software ORBIT is presented. Dey (2001) presented methodology for selection of pipeline inspection and maintenance strategies for
pipeline monitoring. In the absence of a rational methodology, it requires the entire pipeline to be inspected periodically which wastes time and is

70
expensive. The proposed risk based model reduces the amount of time spent on inspection. The model not only reduces the cost of maintaining
petroleum pipelines, but also suggests an efficient design and operation philosophy, construction method and logical insurance plans. The risk based
model uses Analytical Hierarchy Process (AHP), a multiple attribute decision making technique, to identify factors that influence failure of specific
segments and analyze their effects by determining the probabilities of risk factors. The severity of failure is determined through consequence analysis,
which establishes the effect of failure in terms of cost caused by each risk factor and determines the cumulative effect of failure through probability
analysis. The technique does not totally eliminate subjectivity but improves upon existing inspection methods.

Khan and Haddara (2004) introduced a methodology to develop an optimum risk based maintenance strategy. It comprises three main modules: (a)
risk estimation module, (b) risk evaluation module and (c) maintenance optimization module. A case study was presented to illustrate the
methodology. Human health risk factors were calculated using available failure data, which fits best a lognormal distribution model. Original
calculations indicated that both the societal risk factors and individual risk factors exceeded the acceptable risk criteria. Subsequently, optimal
maintenance intervals were calculated using the maintenance model and a reverse fault tree analysis. It is further verified that that the implementation
of the suggested maintenance and inspection strategy would bring down the risks to acceptable levels. Choi and Bomba (2003) established acceptance
criteria for interpretation of the internal inspection results of subsea pipelines. Acceptable criteria were established with a case study of a 24 inch
diameter, concrete weight coated pipeline that delivers fuel oil to a thermal plant in Ghana. Parameters to be considered in the uncertainties of flaw
sizing include the crack length, depth, orientation of the flaw and number of flaws. Fitness for service of other pipelines can be determined using the
criteria established in this study. In 2001, the federal Office of Pipeline Safety (OPS) promulgated its pipeline Integrity Management Programs (IMP)
rule for hazardous liquid pipelines. Dewolf (2003) examined some parallels and differences between the IMP rules for hazardous liquid pipelines and
the federal Risk Management Program (RMP) regulations for the process industries within a common risk management framework and mentioned that
formal risk assessment is at the heart of both type of programs. He concluded that IMP and RMP are management systems that seek continuous
improvement to change not only what is done, but also how it is done. Also, the success of all these programs depends on replacement of a static
checklist compliance mentality with an active risk management mentality of analysis, formal risk based decision making, continuous review and re-
evaluation in response to changing conditions.

Ale (2002) presented historical development and the present status of the management of risks of major hazardous installations and the transport of
dangerous materials in the Netherlands. Wu-Chein and Apostolakis (1981) examined the risk averse attitude that is included in some risk acceptance
criteria and found that the boundary curve separating acceptable and unacceptable regions does not have to be a straight line on the logarithmic
frequency-consequence space. A curve of variable slope would express the same attitude as long as the slope is less than −1. Schroder and Kauer
(2004) discussed on regulatory requirements related to risk-based inspection and maintenance. Egidi et al (1995) described the ARIPAR project aimed
at the assessment of the major accident risks connected with storage, process and transportation of dangerous substances in the densely populated
Ravenna area in Italy, which includes a large complex of chemical and petrochemical plants and minor industries, essentially distributed around an
important commercial port. The project started by making a complete inventory of fixed installations and transportation activities capable of
provoking major fire, explosion and toxic release events. Then relevant accidents scenarios were developed for single hazard sources. Further,
probabilities were assigned to the events and consequences were evaluated. Finally iso-risk contours and F-N curves were evaluated both for single
sources and for the overall area. This required the development of a particular methodology for analysis of risk and of associated software packages
which allowed examination of the relative importance of the different activities and typologies of materials involved. The methodological approach
and the results have proved to be very useful for the priority-ranking of risk mitigating interventions and physical planning in a complex area.

OBJECTIVES AND SCOPE

Based on the critical literature reviewed above, the present work aims at developing a pipeline integrity management scheme, with focus on inspection
and maintenance planning, for specific shallow water offshore-onshore gas field complex located in Indian west coast based on state-of-the-art risk
assessment methods. The specific objective is to perform event analysis for offshore riser for assessing its risk and acceptability.

PROBLEM DESCRIPTION

The ABC field for gas explorations is situated in the west coast of India. This is a shallow water block with water depth ranging between 10 to15 m.
Currently, the ABC field has 5 offshore platforms. Gas produced in the offshore field is conveyed to onshore processing terminal by 5 pipelines of
various sizes. Management intended to carry out risk assessment of the pipeline system to identify the inspection or maintenance plan required for the
pipelines. The offshore facilities present in the current filed as presented below:

The physical layouts of all 5 unmanned ABC field platforms are more or less similar. Fig. 3 shows lay out of the field. The platforms are all
designated as wellhead platforms, having 6 well slots. Each platform has a dimension of 18m × 10m. There are two primary decks with the cellular
deck at elevation of 15 m and the main deck at elevation 20 m above Mean Sea Level (MSL). Also, there are two more additions of auxiliary decks
with the sump deck at elevation 11 m and the helideck at elevation 25 m for logistic purposes. The cellar deck contains the pig launcher and receiver,
the production manifold and the wellhead panel. The main deck will house the test separator, instrument gas drum, vent knockout drum, chemical
drum storage and injector pumps, vent boom, wellhead water filter and work bench. A sump deck is provided for the closed drain sump and drain
transfer pump. A fresh water storage tank is located beneath the helideck. The following equipments are located on the well platform:(a) production
and test manifolds, (b) pig launcher and receiver, (c) test separator, (d) instruments and utility system, (e) vent scrubber and vent boom, (f) fusible plug
detection system, (g) closed drain system, (h) chemical injection system, (i) solar powered panels for power generation, (j) diesel generator, (k) crane,
(l) helideck, (m) boat landings, (n) work area, and (o) evacuation means (lift rafts, rope ladders and ladder).Various sizes of subsea pipelines
interconnecting the platforms and between platform and the onshore terminal for flow of gas have been provided as given in Table 1.

71
Table 1 Sizes of gas pipelines in ABC field

Pipeline diameter
Sl. No. Pipeline
(mm)

1 X1 to X2 150
2 X2 to X3 200
3 X5 to X4 150
4 X4 to X3 250
5 X3 to OT 300

The existing onshore terminal houses several major facilities. These are: (a) gas compression facilities, (b) one Dew Point Depression (DPD) unit to
meet the consumer gas quality, (c) produced water and waste water treatment disposal system, (d) power generation and distribution system, (e) plant
and instrument air system, (f) fire water system and (g) flare system. The natural gas in this field has no H2S and CO2, i.e., the gas can be termed as
‘sweet gas’. Natural gas from the wells is piped to onshore terminal where the associated water, if any, shall be separated in the slug catcher and knock
out vessel. The gas is then dehydrated and compressed before handing over to customer. The produced water is treated in the existing effluent
treatment plant for removal of oil before its disposal into the sea.

METHODOLOGY

The five main components of the analysis are namely: (a) hazard or failure case identification, (b) failure frequency estimation, (c) consequence
calculations, (d) risk calculation or risk summation and (e) risk assessment using acceptability criteria. Figure 4.1 shows the relationship between steps
and also additional external data requirements. The SAFETI software package structure is designed to mirror this logical approach, and enables the
analyst to concentrate on the site-specific aspects of any study with the more general and repetitive modelling calculations being handled by the
package. This package has been used in the present work.

4.1.1 Collection of Data

In order to undertake this QRA, the information is required about the pipeline network and the facilities. The data requirement can broadly be put in
five categories as (a) plant data, (b) generic failure rate data, (c) population data, (d) meteorological data and (e) ignition source data.The plant data
could be derived from the description of facilities. The generic failure rates data, upon which the present study would be based, are taken from
published data bank sources and fine tuned to reflect the operational experience of the ABC field.

Plant
Data

Generic
Failure Rate Derive Meteorological
Data Failure Data
Cases

Calculate Calculate
Frequencies Consequences

Safety
Management
Factor

Population Calculate Ignition


Data Risks Data

Assess
Risks

Fig. 4 Classical approach to risk analysis

72
FAILURE CASE IDENTIFICATION AND DEFINITION

The first stage is to identify the potential accidents that could result in the release of gas from the risers. This is achieved by a systematic review of the
facilities together with an effective screening process. There is a possibility of failure associated with each mechanical component of the system.
These are generic failures and can be caused by such mechanisms as corrosion, external impact, etc. A small event, such as a leak, may precipitate a
bigger event, thereby causing a larger failure. The range of possible releases for a given component covers a wide spectrum, from a pinhole leak up to
a full bore rupture. It is both time consuming and unnecessary to consider every part of the range. Instead, representative failure cases are generated.
For a given component, these should represent both the range of possible releases and their total frequency. For pipelines, four representative failure
cases are considered namely (a) full bore rupture, (b) 100 mm leak, (c) 25 mm leak and (d) 5 mm leak. For each identified failure case, the appropriate
data required to define that case is input into the SAFETI package. An estimate of the failure frequency is assigned to the failure case, which is based
on the generic failure rate data. When the appropriate inputs are defined, SAFETI calculates the source terms of each release, such as the release rate,
release velocity, release phase and drop size. These source term parameters then become inputs to the consequence modelling. Alternatively, SAFETI
allows these source terms to be input directly.

CONSEQUENCE CALCULATIONS

Using the failure case data developed, the consequence program of the SAFETI package, CONSEQ, undertakes consequence calculations for each
identified failure. CONSEQ initially models the dispersion of the released material. For flammable materials, CONSEQ then proceeds to determine
the effect zones for the various possible outcomes of such a release. A release can ignite close to the source. If a release does not ignite in this way,
and is still flammable, it can be ignited at a number of points downwind if its path is such that it goes across, for example, a road, an area where people
or other ignition sources are present. The particular outcomes modelled depend on the behaviour of the release and the dilution regimes which exist.
This can be quite complex. The CONSEQ program undertakes these calculations for the selected representative meteorological conditions, which are
derived from the annual meteorological conditions in the study area. Consequence models are used to predict the physical behavior of hazardous
incidents. Some models only calculate the effect of a limited number of physical processes, like discharge or radiation effects. More complex models
interlink the various steps in consequence modeling into one package. An overview of the different models available in this software package that are
used in the present work will now be presented. The calculations using the software start upstream of the release point with specified initial conditions
for each scenario. This is followed by discharge modeling to the point where the material emerges into the atmosphere and attains atmospheric
pressure. The Calculations based on dispersion modeling follow, starting with the source terms, possible pool formation etc. based up on Gaussian
dispersion model for buoyant plume dispersion. A definition sketch for discharge and dispersion modeling domains is presented in Fig. 5.
Consequence analysis can be divided into discharge and dispersion modeling, as shown in the Fig. 6. When running consequence models, most of the
physical properties of the material being modeled are input and some data are accessed via an interlinked materials database.

Fluid Pressure = Atmospheric Pressure


Pipeline
Fluid

Orifice

Fluid Flow

Discharge Dispersion
Pipieline Pressure >
Atmospheric Pressure

Fig. 5 Discharge and dispersion of fluid

73
Discharge Models

Dispersion Models

Flammable and Toxic Effect Models

Fig. 6 Consequence modelling flow

The field of consequence modeling for hydrocarbon releases is highly developed, and there are several commercially available computer programs to
model the discharge, dispersion and fire and explosion effects of gases and liquids. Some of these techniques are relatively simple and can be easily
implemented in customized spreadsheets. More complex models are available in stand alone format and also as part of linked software or toolkits. The
final estimation of risk is undertaken within the MPACT program of the SAFETI software package. Each failure case is analyzed to determine its
impact in terms of fatalities. Effect zone information is combined with meteorological, ignition source and population data. Event tree conditional
probabilities leading to a particular outcome and frequency information, extracted from the original failure case description, are used to determine the
level of risk for the specific failure case under consideration. The modeling of flammable impacts can be quite involved. This is because there are
many possible final outcomes from a single release and ignition may occur at several locations. SAFETI allows the analyst to alter the basis upon
which it determines the conditional probabilities within MPACT, thereby altering the computed fatality levels. These variables are stored within the
SAFETI parameter files. MPACT generates the required standard forms of risk measure. It calculates both individual risk at grid points and the
societal group risk of each incident outcome. To calculate societal group risk, the total number of people killed for each release case, event tree
outcome, weather type and wind direction must be calculated. The frequencies of all those combinations contributing to the same number of fatalities
must be summed. SAFETI allows these results to be presented in the form of an FN societal group risk curve. An FN curve is a graph, which plots the
frequency of N or more fatalities per year (F) against the number of fatalities (N). It is conventional that this information is presented on a log-log plot.
The Potential Loss of Life (PLL) combines the FN data to produce an average expected rate of fatalities per year. The PLL is often useful in
comparing the risk between different options.
IMPACT CRITERIA

Where there is line-of-sight between the person and the flame, the main impact is by thermal radiation. The main effects are: (a) burns to exposed skin
and (b) ignition or melting of clothing. Burns are classified in increasing degrees of severity as (a) first degree burns that are superficial giving red,
dry, painful skin, (b) second degree burns that are more than 0.1 mm deep, affecting the epidermis, and forming blisters and (c) third degree burns that
are more than 2 mm deep, affecting the dermis and nerve endings, giving dry skin which has no feeling. The degree of harm caused by thermal
radiation is related to both the intensity of thermal radiation and the time for which the person is exposed to it. The effect of thermal radiation is
defined in widely used data presented in the Tables 2 & 3. Most data on the effects of explosion overpressures on personnel is derived from explosions
of solid phase explosives onshore, in which people were at some distance from the source. The main effects are: (a) direct blast effects of the
overpressure caused by the explosion, causing lung and ear damage which may lead to death (primary effects), (b) impact from missiles projected by
the explosion (secondary effects), (c) impact from collapsing buildings or glass (secondary effects) and (d) whole-body translation due to the blast
wave, resulting in impact with stationary objects, which may cause injury or death (tertiary effects).

Table 2 Effects of Thermal Radiation

Thermal Dose Effects


(kW/m2)4/3 s
125 Threshold of pain

210 Threshold of first degree burns

500 Threshold of second degree burns

1200 Threshold of greater than second degree burns

Threshold of third degree burns


2600
Melting of nylon and polyester fabrics
1500 -3000
Piloted ignition of clothing

74
2500 - 8000
Auto-ignition of everyday clothing
3500 - 12000

Table 3 Human Injury Levels Resulting from Blast Waves in case of Vapour Cloud Explosion

Injury Type Injury Level Peak Overpressure (bar)

Eardrum Failure Threshold 1


50% 3- 4.2
Lung Damage Threshold 2.1-2.4
Lethality Threshold 6.3-8.7
50% 8.7-11.8
Virtually 100% 11.8-16.6

RISK ASSESSMENT AND MEASURES

The final, and most significant, step in the process is the assessment of what the calculated risk levels portray. Risk assessment is a process by which
the results of a risk analysis are used to make judgements, either through relative risk ranking of risk reduction strategies or through comparison with
risk targets or criteria. Where risk criteria have been issued by the regulatory authority, it is possible for interested parties to assess the calculated risk
levels against these criteria. The risk assessment stage determines whether the risks are tolerable, or if risk mitigation measures are required to reduce
the risk to a level, which can be considered to be as low as reasonably practicable (ALARP). The risk levels associated with the facilities are presented
in the following standard forms: (a) Individual risk contours show the geographical distribution of risk to an individual, (b) Societal risk (FN) curves
which show the cumulative frequency (F) distribution of accidents causing different numbers (N) of fatalities. The FN curve therefore indicates
whether the societal risk to the facility is dominated by relatively frequent accidents causing small numbers of fatalities or low frequency accidents
causing many fatalities. The assessment of the acceptability of that risk is left to the judgment and experience of the people undertaking or using the
risk analysis work. The normal approach adopted is to relate the risk measures obtained to acceptable risk criteria.

ABC FIELD CRITERIA

Individual risk Criteria


The following criteria shall apply when judging the tolerability of risk to persons for ABC Field facilities, sites, combined operations or activities. The
same can be graphically expressed as in Fig. 7.
(a) Individual risk to any worker above 10−3 per annum shall be considered intolerable requiring fundamental risk reduction measures.
(b) Individual risk below 10−3 but above 10−6 per annum for any worker shall be considered tolerable if it can be demonstrated that the risks are as
low as reasonably practicable.
(c) Individual risk below 10−6 per annum for any worker shall be considered as broadly acceptable and no further improvements are considered
necessary provided documented control measures are in place and maintained regularly.
(d) Individual risk to any member of the general public as a result of ABC Field businesses activities shall be considered as intolerable if greater than
10−4 per annum, broadly acceptable if less than 10−7 per annum and shall be reduced to as low as reasonably practicable between these limits.

For new facilities and activities, ABC Field shall strive to achieve lower risks compared with that typical for existing facilities, down at least to an
individual risk to any worker of 10−4 per annum, by the appropriate use of best practices including technology and management techniques.

High Risk

Intolerable
Risk

Risk : 10 - 3 Maximum
Tolerable
Criterion
Risk
Tolerable if
ALARP Negligible
Risk: 10 -7 Criterion

Negligible Risk

Low Risk

Fig. 7 ABC Field Individual Risk Criteria

75
Societal Risk Criteria – FN Curves

When considering the risks associated with a major hazard facility, the risk to an individual is not always an adequate measure of total risk; the
number of individuals at risk is also important. Catastrophic incidents with potential multiple fatalities have little influence on the level of individual
risk but have a disproportionate effect on the response of society and impact on company reputation. The concept of societal risk is more complicated
than that for individual risk. A number of factors are involved which make it difficult to determine single value criteria for application to different
situations. These factors include (a) hazards and consequential risks, (b) the nature of the assessment, (c) factors of importance to the company,
government, regulators and authorities, public attitudes and perceptions and aversion to major accidents.
Societal risk is the relationship between frequency of an event and the number of affected people. Societal risk from a major hazard facility can thus be
expressed as the relationship between the number of potential fatalities N following a major accident and the frequency f at which N fatalities are
predicted to occur. The relationship between f and N, and the corresponding relationship involving F, the cumulative frequency of events causing N or
more fatalities, are usually presented graphically on log-log axes. Based on historical major accident experience it is proposed by the UK Health and
Safety Executive that the risk from an accident causing the death of 50 people or more in a single event should be regarded as intolerable if the
frequency is estimated to be more than one in five thousand per annum.

Generic Failure Rate Data

There are several data sources containing varying degrees of information; some are available in their raw data form, whereas some are available only
in summary. The main sources of information are:
(a) UK Gas Transmission Lines (1969-77) from the British Gas Corporation as published by the Health and Safety Executive in their St. Fergus to
Mossmorran NGL pipeline study (HSE, 1978). This data covers some 14,228 km of pipeline and includes leak sizes and causes.
(b) UK Cross Country Pipelines (1963-76), also from the HSE report mentioned above (HSE, 1978). Data from about 2,400 km of pipeline carrying
diverse materials are presented. Unfortunately, however, there is a lack of useful detail in this data set.
(c) CONCAWE - European Oil and Product Lines (1976-84). CONCAWE produces a detailed annual report and a running 5 year analysis of their
19000 km of pipeline. However, these reports express failure size in terms of volume of liquid spilt rather than the dimensions of the hole, so it is
difficult to relate this information to leak size.
(d) US Gas Transmission Lines (1984-2005). Information regarding these lines is collected by the US Department of Transportation Hazardous
Information System from their incident report forms, and is stored in computerised form for ready access. Operators are compelled by statute to
complete both the incident reporting forms and an annual return of pipeline mileage in various classes. In addition, Battelle Columbus Laboratories
analyze this information every year for the American Gas Association and present their results in published reports. The length of pipelines included in
this database is around 294000 km, and it is therefore the largest such database for gas pipelines, by a wide margin.
(e) UK Onshore Operator’s Association (UKOPA) (1962-2004). It presents collaborative pipeline and product loss incident data from onshore Major
Accident Hazard Pipelines (MAHPs) operated by Transco, Shell UK, BP, Huntsman and E-ON UK, covering operating experience up to the end of
2004. The length of pipelines, at the end of 2004, is 21727 km. The total exposure in the period 1952 to the end of 2004 is 654732 km.year (UKOPA,
2005). A comparison of the key failure rate frequencies which may be derived from these data is given in Table 4.

Table 4. Comparison of Failure Rate Frequencies (per 1000 km.years)

Pipeline System Failure USA Gas- UK Gas HSE NGL Risk European Oil and Gas UKOPA
Mode Line Data 1969-77 Study Lines 1974-84 Pipeline Fault
1984-2005 (HSE) (Derived from (CONCAWE Database
Raw Data CONCAWE Data) Unmodified Data) (1962-2004)

Defect 0.12 0.07 0.07 0.16 0.0749


Corrosion 0.069 0.16 0 0.23 0.0505
External Impact 0.017 0.16 0.16 0.22 0.0566
Environmental 0.08 0.007 0.007 0.05 0.0076
Operational/Other 0.023 0.17 0 0.03 0.0734
Total 0.309 0.56 0.23 0.69 0.216

UKOPA database is selected for this study. It has by far the greatest detail, and enables great flexibility of analysis because of failure distribution with
reference to causes. It gives the details in a format readily usable in QRA. The database is designed to reflect the ways in which the UKOPA operators
design, build, operate, inspect and maintain their pipeline systems. Although the pipeline and failure data are extensive, there are pipeline groups (e.g.
large diameter, recently constructed pipelines) for which no failures have occurred. However, it is unreasonable to assume that the failure frequency
for these pipelines is zero. Similarly, further pipeline groups exist for which the historical failure data are not statistically significant. UKOPA
database contains extensive data on pipeline failures and on part-wall damage, allowing prediction of failure frequencies for pipelines for which
inadequate failure data exist. For these reasons, it was chosen as the main source of failure information for the present study.

76
Failure Data Analysis

The total length of Major Accident Hazard Pipelines (MAHPs), above ground, below ground and elevated, in operation at the end of 2004 for all
participating companies (Transco, BP, Shell UK, Huntsman and E-ON UK) is 21727 km. The total exposure in the period 1952 to the end of 2004 is
654732 km.year. Exposure of pipeline before first recorded incident in 1962 is 3740 km.year (included in exposure and incident frequency
calculations). Exposure up to end 2004 of elevated pipeline is 23 km.year. Exposure up to end 2004 of above ground pipeline is 25.7 km.year. There
were 9 out of 172 (5.2%) product loss incidents that resulted in ignition. The overall ignition probability in the present analysis has therefore been
taken as 0.05.Table 5 provides more detail. The overall incident frequency by hole size over the period 1962-2004 is shown in Table 6 and product
loss incidents by cause is shown in Table 7

Table 5 Incidents that resulted in ignition

Affected Component Cause of Fault Hole Diameter Class


Pipe Seam Weld Defect 0-6 mm
Pipe Ground Movement Full Bore and Above
(18” Diameter Pipe)

Pipe Girth Weld Defect 6-20 mm


Pipe Unknown 6-20 mm
Pipe Pipe Defect 0 – 6 mm
Pipe Unknown 40 – 110 mm
Pipe Lightning Strike 0-6 mm
Bend Internal Corrosion 0-6 mm
Bend Pipe Defect 6-20 mm

Table 6 Failure frequency distribution by hole size

Hole Size Class Number of Frequency


Incidents [incidents per 1000 km.year]
Full Bore and Above 7 0.011
110mm – Full Bore 3 0.005
40mm – 110mm 7 0.011
20mm – 40mm 20 0.031
6mm – 20mm 27 0.041
0 – 6mm 106 0.162

Table 7 Products loss incidents by cause

Product Loss Cause No. of Incidents

Girth Weld Defect 33


External Interference 37
Internal Corrosion 2
External Corrosion 31
Unknown 8
Other 40
Pipe Defect 13
Ground Movement 5

77
Seam Weld Defect 3
Total 172

Fig. 8 shows the product loss incident frequency by cause over the period 1962-2004 compared with the frequency over only the last 5 years (2000-
2004). It shows that external interference is one of the main causes of product loss incidents. An overview of the product loss incident frequency by
cause and size of leak in the period 1962 to 2004 is shown in Fig. 9

Fig. 8 Historical pipeline failure frequencies

Fig. 9 Product loss incidents by cause and equivalent hole diameter

CONCLUSIONS

Oil and gas are dominant sources of energy, which are partly produced in a demanding ocean and industrial environment with significant fire and
explosion hazards. Safety of men, environment and assets is therefore a principal concern in this industry. Operational experiences (WOAD, 1996)
show that accidental actions contribute significantly to failures of offshore structures. Such events can be commonly traced back to errors in design,

78
fabrication or operation. To limit the occurrences of undesirable events it is therefore important to avoid errors by conducting risk assessment. It is
also important to carry out risk audits in all life cycle phases of a project. Quantitative risk assessment is a tool to support decisions regarding safety of
systems. The application of risk assessment in the offshore industry has evolved since about 1980 (Moan and Holand, 1981, NPD 1984). The piper
Alpha disaster in 1988 was the direct reason for introducing QRA in UK. This disaster provided tragic confirmation that the major accident predictions
which risk analyses had made were indeed realistic, and that QRA could be useful in trying to reduce the risks. Operators in UK were requested to
evaluate emergency isolation arrangements for subsea pipelines and risers. This concentrated studies on riser hazards and the effect of installing
subsea isolation valves i.e. SSIVs, which isolate the installation from the inventory of the pipeline. Risk analysis methodology is currently applied to
validate offshore facilities at large, as outlined by Vinnem (1999).

Offshore operators are increasingly using quantitative risk assessment techniques as part of their risk management programs. Structured risk
assessment can be used at all stages of a project life cycle. In the design stages, these techniques are valuable tools in ensuring that money is spent
wisely on safety related systems. In the operational stage, QRA can aid the development of procedures. Integrity management system covering issues
such as inspection, emergency planning are crucial to maintain as designed levels of safety and reliability. Audits shall be carried out all through the
operational phase to ensure that risks are not accumulated. The operator is required to develop performance measures for his safety management
system in co-operation with authorities. The authorities will inspect performance against the agreed measures.

REFERENCES

Ale, BJM. (2002), “Risk assessment practices in The Netherlands”, Safety Science, 40, 105-126.
API RP 579 (2000), “Fitness for service”, American Petroleum Institute, first edition, January.
API RP 1111 (1999), “ Design, construction, operation and maintenance of offshore hydrocarbon pipelines (limit state design)”, American Petroleum
Institute, third edition.
API 1160 (2001), “Managing system integrity for hazardous liquid pipelines”, American Petroleum Institute, first edition, November.
Brillinger DR. (2003), “Three environmental probabilistic risk problems”, Statistical Science, 18, 412-421
Choi, HS, Bomba, JG (2003), “Acceptance criteria of defects in undersea pipeline using internal inspection”, Ocean Engineering, 30, 1613-1624.
CONCAWE (2003), “Performance of European cross-country oil pipelines: statistical summary of reported spillages-2001”, OP/STF-1, CONCAWE
Oil Pipeline Management Group.
Cullen, W.D, (1990), “The public inquiry into Piper Alpha disaster”, Department of Energy, HMSO, London.
Dewolf GB (2003), “Process safety management in the pipeline industry: parallels and differences between the pipeline integrity management (IMP)
rule of the office of pipeline safety and the PSM/RMP approach for process facilities”, Journal of Hazardous Materials, 104, 169-192.
Dey, P K. (2004), “Decision support system for inspection and maintenance: A case study of oil pipelines”, IEEE Transactions on Engineering
Management, 51, 47-56.
DNV (1981), “Rules for submarine pipeline systems”, Det Norske Veritas
DNV ( 1999), ”RP F101, Corroded pipelines”, Det Norske Veritas.
DNV (2002), “ Risk management software- User manual for SAFETI”.
Egidi D., F.P. Foraboschi, G. Spadoni and A. Amendola (1995), “The ARIPAR project: analysis of the major accident risks connected with industrial
and transportation activities in the Ravenna area”, Reliability Engineering and System Safety, 49, 75-89.
Gerwick BC. (1999), “Construction of marine and offshore structures”, second edition, CRC press.
Khan FI, M.R.Haddara (2003), “Risk based maintenance (RBM): a quantitative approach for maintenance / inspection scheduling & planning”,
Journal of Loss Prevention in the Process Industries, 16, 561-573.
Leynes RD, W.P.C. Pioquinto and J.A. Caranto (2005), “Landslide hazard assessment and mitigation measures in Philippine geothermal fields”,
Geothermics, 34, 209-221.
Mazzola A. (1999), “Thermal interaction analysis in pipeline systems- a case study”, Journal of Loss Prevention in the Process Industries, 12 ,
495-505.
Moan, T. and I. Holand (1981), “Risk assessment of offshore structures: experiences and principles”, Proceedings of 3rd ICOSSAR, Elsevier,
Amsterdam.
NPD (1984), “Regulations for load carrying structures for extraction or exploitation of petroleum”, The Norwegian Petroleum Directorate, Stavanger.
NPD (1990), “Regulations for risk analysis”, The Norwegian Petroleum Directorate, Stavanger.
PFEER (1995), “Prevention of fires and explosions, and emergency response regulations”, United Kingdom Health Safety and Executive.
Schroder H.C. and R. Kauer (2004), “Regulatory requirements related to risk-based inspection and maintenance”, International Journal of Pressure
Vessels and Piping, 81, 847- 854.
TNO (1999), “Guidelines for quantitative risk analysis”, The Director General of Labour, The Hague, Netherlands
United Kingdom Onshore Pipeline Operators’ Association (2005), “UKOPA pipeline fault database, Pipeline product loss incidents 1962 – 2004”,
4th Report of the UKOPA Fault Database Management Group, Advantica Technologies Ltd.
Vinnem J.E, (1999), “Offshore risk assessment - principles, modeling and applications of QRA studies”, Kluwer academic publishers.

79
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Risk assessment of offshore pipelines


S.K.Bhattacharyya1, Srinivasan Chandrasekaran2*, and MVS Ram Prasad3
1
Professor and Head, Department of Ocean Engineering, Indian Institute of Technology Madras, Chennai, India
2
Associate Professor, Department of Ocean Engineering, Indian Institute of Technology Madras, Chennai, India
3
Research scholar, Department of Ocean Engineering, IIT Madras
*
corresponding author E-mail: drsekaran@iitm.ac.in

ABSTRACT

Oil and gas are dominant source of energy that is partly produced in a demanding ocean and industrial environment with significant fire and explosion
hazards. Pipe lines are considered to be the safest and most economical way of carrying large volume of fluids, especially those which are perceived as
dangerous substances. To limit the occurrence of undesirable events, it is therefore necessary to avoid errors by conducting risk assessment. External
interference, commonly caused by third party activity is one of the main causes of offshore pipe line accidents as they are not contained within a
perimeter controlled by operator like chemical installations. Quantitative risk assessment (QRA) is an important tool to support decisions regarding
safety of systems. Offshore operators are increasingly using quantitative risk assessment as part of their risk assessment programs. Structured risk
assessment can be used at all stages of a project life cycle. In the design stages, these techniques are valuable tools in ensuring that money is spent
wisely on safety related systems; in operational stages, QRA can aid the development of procedures. Considering the significance of role of risk
assessment in offshore pipelines, a case study is presented for an existing offshore field. To protect company confidentiality, the gas field is designated
as ABC gas field which is located in west coast of India in a shallow water block with water depth ranging between 10-15m. Currently, ABC gas field
has five offshore jacket platforms and the produced gas from the field is conveyed to onshore processing terminal by five pipelines of various sizes.
Risk assessment is carried out for the pipeline system of this field and recommendations are made for integrity management of pipeline system.

KEY WORDS: risk assessment, integrity management system, offshore pipeline failure, oil and gas industry.

INTRODUCTION

Large scale geological structures that might hold oil or gas reservoirs are located beneath non-productive rocks, and in addition this is often below the
sea. Most of the oil and gas production platforms in offshore rest on ‘jackets’, which are framed steel structures, fabricated from welded tubular and
are pinned to the sea floor with steel piles. Above the jacket structure are prefabricated units or modules providing accommodation and housing
various facilities including gas turbine generating sets. The drilling rig derrick and the flare stack are located above these modules. Horizontal surfaces
are taken up by store areas, drilling pipe deck and the helicopter pad; a small number of platforms are fabricated from concrete. The most important
functions are drilling, preparing water or gas for injection into the reservoir, processing oil and gas before sending ashore and cleaning the produced
water for disposal into the sea. Power is generated onboard to drive production equipment and support life. All production systems are constantly
monitored for leaks, since both oil and gas are hazardous substances and extremely flammable. On top of a production well there will be a series of
pipes, gauges and valves called the 'Christmas tree’. At this point, crude oil is a hot, frothy, corrosive, high-pressure fluid containing gas, water and
sand. After separation, crude oil is metered and pumped into the pipeline, or stored until sent ashore by tanker. The gas separated from the oil may be
used for fuel or compressed and piped to shore or re-injected into the reservoir. Any liquid dissolved in the gas is removed and then the gas is
compressed, cooled, dehydrated and metered before being piped to shore. Any gas that cannot be used onboard or piped ashore must be burnt in the
platform's flare. A typical jacket platform is shown in Fig. 1. Pipeline construction begins onshore, as lengths of pipe are waterproofed with bitumen
and coated with steel-reinforced concrete. This coating weighs down the submarine pipeline even when it is filled with gas.
Fig. 1 A typical jacket platform

In onshore terminals, crude oil and gas undergo further processing. Any remaining water and gas are removed from oil, which is then stored at the
terminal before transport to refineries. During transportation, great care is taken to avoid or deal effectively with spillage. Business today is much
more than mere economic entities. Successful businesses generate jobs, create new technologies, return profits to shareholders and contribute to the
economic health of the nation. The offshore industry is part of and deeply intertwined with the wider world - a concept which is becoming increasingly
clear as businesses become more global and the world seems to grow ever smaller.

FAILURE IN OFFSHORE PIPE LINES

Offshore pipelines often contain large inventories of oil or gas at high pressures. Although accidental releases from them are rare, they have the
potential for catastrophic consequences and therefore need special attention in their design and analysis. Riser leaks are often critical components in
risk assessment of offshore oil and gas platform complexes and pipeline leaks are often critical components in pollution studies. Risers are sections of
pipeline rising from the seabed to an offshore installation, extending as far as the top sides Emergency Shut Down Valve (ESDV). At the seabed end,
the riser can be isolated from the pipeline by Sub Surface Isolation Valve (SSIV). Risers are mainly vertical in shallow water, but may include
horizontal sections in the cellar deck area. Pipelines are horizontal sections of pipe lying on or under the seabed, used for import or export of
hydrocarbons or other fluids to and from offshore installations and to or from onshore facilities. Spool pieces are horizontal sections of pipe on the sea
bed, typically 100 m long, joining the riser to the pipeline. They are usually constructed to the same specification as risers (ABS 2001). A typical riser
and pipeline configuration is shown in Fig. 2. Before the consequences of a leak from a riser or pipeline can be predicted, it is necessary to specify the
leak scenario, defining parameters such as (a) location of the leak in the riser or pipeline, (b) size of the leak, (c) orientation of the leak, (d) detection
of the leak in order to initiate mitigating actions, (e) isolation of the leak by ESDV and SSIV, (f) ignition and possible explosion, (g) weather
conditions, (h) distribution of personnel when the leak occurs and (i) evacuation response. In reality, a riser or pipeline leak can constitute an infinite
variety of scenarios, ranging from a small pipeline leak requiring repair by divers, to a full bore riser failure which may destroy the installation.

ESDV Process Piping

Cellar Deck

Sea Level

Riser

SSIV Pipeline

Sea Bed

Fig. 2 Typical riser and pipeline configuration

82
In the risk analysis of offshore installations of pipelines, the probabilities considered are: (a) Subsea: 30%, (b) Splash zone: 50% and (c) Platform:
20%. However, these values are very sensitive to the decision to exclude leaks from fittings. The probabilities that are considered for leak sizes in
risers and pipelines are: (a) Small (< 20mm): 57%, (b) Medium (20 to 80mm): 15% and (c) Large (> 80mm): 28%. The failure by rupture is taken as
21% and such failures are always in the ‘Large’ category (Peter and John, 2000).

IMPACT OF PIPELINE LEAKS AND MOTIVATION

The historical data on riser leaks are inadequate to predict fatality or damage impacts, so the approach used in risk estimation is to combine the
modelled consequence zones with the impact criteria and the initial personnel distribution and their assumed evacuation response. For subsea
pipelines, the degree of diffusion of hydrocarbons on rising through the water makes it less likely that releases will have a significant impact on
surface systems. The offshore industry is a major employer of people and makes major contribution to the world economy. Oil, gas and their
downstream products underpin modern society, supplying energy to power industry, heat homes, fuel for transport of goods and people etc. Most
offshore oil and gas are brought to shore by pipelines under high pressure. Accidental releases of hydrocarbons from pipelines can result in
catastrophic consequences. The success of the operating company depends on winning and keeping the trust and goodwill of many stakeholders such
as employees, suppliers, shareholders, end customers etc. Demands are increasing day by day upon pipeline operators to ensure that they proactively
and prudently address the integrity of their systems and assets (HSE 1998). An in depth risk assessment study is the foundation of effective pipeline
integrity program. Integrity management is the process by which pipeline operators ensure long term, safe and reliable operation of their systems in a
profitable way. The target of integrity management is to prevent pipeline failures from having significant impact on public safety, environment and
business operations (Muhlbauer, 2004).

Considering the significance of the role of risk assessment in offshore pipelines, a case study is considered in this study for an existing offshore field
producing gas in order to bring out the methods of application of diverse qualitative and quantitative risk assessment and integrity management
methodologies available in the literature. To protect company confidentiality, the gas field is designated as ‘ABC gas field’ in the thesis. It lies in the
west coast of India in a shallow water block with water depth ranging between 10 to 15m. Currently, the ABC field has five offshore jacket platforms
and the produced gas from this field is conveyed to an onshore processing terminal by five pipelines of various sizes (see Fig. 3). Risk assessment is
carried out for the pipeline system of this field and recommendations are made for the integrity management of pipeline system.

Fig. 3 Layout of the ABC field

A system is called series system if it is in state of failure whenever any of its elements fails. Such systems are often referred to as weakest link
systems. Marine pipelines and risers fall in this category (Bai, 2001).The life cycle cost of an offshore pipeline consists of (a) CAPital EXpenditure
(CAPEX), i.e. initial investment related to procurement and installation costs, (b) OPerating EXpenditure (OPEX), i.e. costs related to maintenance
and (c) losses caused by failures, i.e. a risk related expenditure. Degradation or failure of a pipeline may lead to a shut down of the operation. Operator
or owner wishes to have pipeline with a low initial cost, the highest possible operating margin and an extendable operating period. A life cycle cost
model based on probabilistic economics may be a useful tool to improve the design, inspection and maintenance. The total cost of pipeline is the sum
of CAPEX, OPEX and cost of failures. A target reliability level may then be estimated based on cost optimization, if it is higher than the value
required by legislative requirements (Bai, 2003). Risk factors that can cause failure of pipelines can be broadly classified as (a) external interference,
(b) corrosion (external and internal), (c) design defects, (d) material defects, (e) environmental hazards, and (f) human and operational error. External
interference may be due to (a) dropped objects because of activities on the offshore platform or nearby it, (b) due to dragging or direct impact of
anchors, (c) damages caused by construction operations, shipping operations, fishing operations etc. Brooker (2005) carried out studies on
experimental puncture loads due to external interference of onshore pipelines by excavator equipment. In this study, a series of puncture experiments
of steel pipelines by excavator bucket teeth is presented. Static puncture forces due to a variety of different tooth geometries were obtained for pipes of
4.8 to 6.9 mm wall thicknesses; primarily using small cross-section or sharp tooth types. The results were compared with finite element simulations
and previously derived puncture relationships. The results confirm that pipe wall thickness and bucket tooth dimension are the most important
parameters that influence the puncture load.

The design should also consider stresses induced by impact, vibration due to current induced vortex shedding and other hydrodynamic loading,

83
seismic activity, soil movement and other natural phenomena. All pipeline components such as risers, unsupported free spans, welds, J-lay collars,
buckle arrestors and flex joints must be assessed for fatigue. Potential cyclic loading that can cause fatigue include Vortex Induced Vibrations (VIV),
wave induced hydrodynamic loading and cyclic pressures and thermal expansion loads. The fatigue life of the component is defined as the time it
takes to develop a through wall thickness crack of the component. The design fatigue life shall be at least 10 times the service life for all components.
The effects of thermal expansion and contraction of pipeline system shall be evaluated (Mazzola 1999). When temperature changes are anticipated, the
pipeline approach to a platform or subsea junction should have additional flexibility for expansion and contraction using measures such as slack
curves, pipeline bends and thermal expansion devices. Sufficient measures need to be taken to prevent excessive strains or fatigue damage due to
thermally induced upheaval buckling of buried pipelines or lateral buckling of non-buried pipelines.

RISK ASSESSMENT

EGIG (1992), i.e European Gas Incident Group data, includes the analysis of pipeline incidents that has been collected by a group of eight major gas
transmission system operators in Western Europe from their pipeline system. Hydrocarbon leak and ignition database is developed by the oil
industry’s international exploration and production forum (Exploration and Production forum, 1992). Health and Safety Executive’s PARLOC (1996)
provides useful data on loss of containment for offshore pipelines. UKOPA (United Kingdom Onshore Pipeline Operators’ Association, 2005), a
pipeline fault database, furnishes onshore pipeline product loss incidents from the year 1962 to 2004. SAA HB105 (1998) provides guidelines to
pipeline risk assessment in accordance with AS 2885.1 (2007). Pipeline risk management manual of Muhlbauer (2004) is very useful in identifying
potential risk factors. TNO (Toegepast Natuurweetenschappelijk Onderzoek), Netherlands organization for applied scientific research, played an
important role in the creation of yellow, green, purple and red books containing detailed descriptions of all models and methodologies needed to
perform quantitative safety and risk analysis. These books are viewed worldwide as standard works in their respective fields. The books are now
included in the Hazardous Substances Publication Series published by the Dutch Ministry of Housing, Spatial Planning and the Environment. The
yellow book provides methods for calculating the physical effects caused by the release of hazardous liquids and gases. The green book provides
methods for determining the potential damage to humans and the surrounding area resulting from the release of hazardous substances. The purple
book provides guidelines for quantitative risk analysis. The red book provides methods for calculating the probability of unintended events.

INTEGRITY MANAGEMENT

An inspection and monitoring philosophy is to be established, and this forms the basis for the detailed inspection and monitoring program. Inspection
and monitoring are carried out to ensure safe and reliable operation of the pipeline and riser system (ABS Guide, 2001). For damaged pipes, ASME B
31.4 and B31.8 may be applied to determine whether a damage assessment and repair will be necessary. If a severe damage cannot be repaired
immediately, strength assessment of pipes with damages such as dents, corrosion defects and weld cracks may be performed as defined in Appendix 4
of ABS Guide for building and classing subsea pipeline systems and risers 2001. Non-critical intervention work such as free-span correction,
retrofitting of anode sleds and rock dumping can usually be considered as planned preventive measures. For the localized repair of non-leaking minor
and intermediate pipeline damage, repair clamps may be utilized without the necessity of an emergency shutdown to the pipeline system. For major
pipeline damage, resulting in or likely to result in product leakage, immediate production shutdown and depressurization is invariably required,
allowing the damaged pipe section to be replaced. Non critical repairs in the short term will not jeopardize the safety of the pipeline and hence can
form part of a planned maintenance program such as: (a) corrosion coating repair, (b) weight coating repair, (c) cathodic protection repair, (d) free
span rectification, and (e) installation of engineered back fill (rock dumping).

Fukushima et al (2000) discussed gas pipeline leak detection system using an online simulation method. Leak detection method is based on a dynamic
simulation with wave equations. An industrial application to one of the longest gas pipeline is also presented with its performance information.
Norwegian Petroleum Directorate took initiative to develop a method to assess trends and status for the risk levels in the Norwegian offshore
petroleum industry. Vinnem et al (2006) presented work carried out as part of this initiative, comprising major hazard risk indicators for monitoring of
trends. A method was developed and a pilot study report was issued in 2001, covering the period 1996 to 2000. Annual updates have been performed
since then. The statistical approach is based on recording occurrence of near misses and relevant incidents and results of risk assessments. Evaluation
of safety culture, motivation, communication and perceived risk are also of importance. This has been covered through questionnaire surveys and a
number of interviews, audit and inspection reports as well as accident and incident investigations. De Groot (1982) reviewed the impact of laying and
maintenance of offshore pipes on the marine environment of the North Sea, with an emphasis on the interaction between the offshore oil and gas
industry and the fishing industry. He discussed about natural hazards such as water movement, corrosion, sediment movement and man made hazards
such as anchors, fishing activities etc. He has suggested to base the requirements for each new projected pipeline on a combination of considerations
and to assess the requirements of burial depth, trenching etc. on a case by case basis instead of applying standard requirements.

Leatherdale and Turner (1991) discussed underwater photogrammetry, a technique for inspection and maintenance of platforms and pipelines for the
offshore oil industry. Cameras and flash units are built for the purpose and analytical plotters and computer graphic systems are used for
photogrammetric measurement and analysis of damage, corrosion and weld failures and redesign of underwater structures. Zingaretti and Zanoli
(1998) presented a robust system for the detection and real-time tracking of the submarine pipelines. An active vision system was proposed to predict
changes in the scene, and to direct computational resources to confirm expectations by adapting the process mode dynamically. The system originates
from an image processing algorithm to recognize the pipeline. The accuracy of the algorithm has been enhanced by exploiting the temporal context in
the image sequence. The disturbances on acquired images caused by motion are partially removed by Kalman filter. The filter proves advantageous in
supporting the guidance and control of the ROV and in making the image processing module itself more robust. Fujiyama et al (2004) presented Risk
Based Maintenance (RBM) for steam turbine plants coupled with quick inspection systems. The RBM system utilizes field failure and inspection
database accumulated over 30 years. The failure modes are determined for each component of steam turbine and the failure scenarios are described as
event trees. The probability of failure is expressed in the form of unreliability functions of operation hours or start-up cycles through the cumulative
hazard function method. The posterior unreliability is derived from the field data analysis according to the inspection information. Quick inspection

84
can be conducted using air-cooled bore scope and heat resistant ultrasonic sensors even if turbine is not cooled down sufficiently. Inspection
information comes from degradation and damage measurement. The probabilistic life assessment is carried out using structural analysis and statistical
nature of material properties. The risk function is calculated as the sum product of unreliability functions and expected monetary loss as the
consequence of failure along event trees. The optimum maintenance plan is determined among simulated scenarios described through component
breakdown trees, lifecycle event trees and risk functions. These methods are effective for total condition assessment and economical maintenance of
operating plants. Smith (1995) has discussed risk management in the North Sea offshore industry. On the basis of accumulated experience including
some major accidents, there has been a move away from a rigid, prescriptive approach to setting safety standards. It is recognized that a more flexible,
goal setting approach is more suited to achieve cost-effective solutions to offshore safety. In order to adapt this approach, offshore operators are
increasingly using quantitative risk assessment techniques as part of their risk management programs. Structured risk assessment can be used at all
stages of a project life cycle. In the design stages (concept and detailed design), these techniques are valuable tools in ensuring that money is wisely
spent on safety related systems. In the operational stage, QRA can aid the development of procedures. High quality safety management systems
covering issues such as training, inspection and emergency planning are crucial to maintain ‘as designed’ levels of safety and reliability. Audits of
safety management systems should be carried out all through the operational phase to ensure that risky conditions do not accumulate.

OBJECTIVES AND SCOPE

Based on the critical literature reviewed above, the present work aims at developing a pipeline integrity management scheme for specific shallow
water offshore-onshore gas field complex located in Indian west coast. The specific objective is to perform risk assessment of offshore pipelines to
determine the expected cause of failure and their consequences.

PROBLEM STATEMENT

The ABC field for gas explorations is situated in the west coast of India. This is a shallow water block with water depth ranging between 10 to15 m.
Currently, the ABC field has 5 offshore platforms. Gas produced in the offshore field is conveyed to onshore processing terminal by 5 pipelines of
various sizes. Management intended to carry out risk assessment of the pipeline system to identify the inspection or maintenance plan required for the
pipelines. The offshore facilities present in the current filed as presented below:

The physical layouts of all 5 unmanned ABC field platforms are more or less similar. Fig. 3 shows lay out of the field. The platforms are all
designated as wellhead platforms, having 6 well slots. Each platform has a dimension of 18m × 10m. There are two primary decks with the cellular
deck at elevation of 15 m and the main deck at elevation 20 m above Mean Sea Level (MSL). Also, there are two more additions of auxiliary decks
with the sump deck at elevation 11 m and the helideck at elevation 25 m for logistic purposes. The cellar deck contains the pig launcher and receiver,
the production manifold and the wellhead panel. The main deck will house the test separator, instrument gas drum, vent knockout drum, chemical
drum storage and injector pumps, vent boom, wellhead water filter and work bench. A sump deck is provided for the closed drain sump and drain
transfer pump. A fresh water storage tank is located beneath the helideck. The following equipments are located on the well platform:(a) production
and test manifolds, (b) pig launcher and receiver, (c) test separator, (d) instruments and utility system, (e) vent scrubber and vent boom, (f) fusible plug
detection system, (g) closed drain system, (h) chemical injection system, (i) solar powered panels for power generation, (j) diesel generator, (k) crane,
(l) helideck, (m) boat landings, (n) work area, and (o) evacuation means (lift rafts, rope ladders and ladder).Various sizes of subsea pipelines
interconnecting the platforms and between platform and the onshore terminal for flow of gas have been provided as given in Table 1.

Table 1 Sizes of gas pipelines in ABC field

Pipeline diameter
Sl. No. Pipeline
(mm)
1 X1 to X2 150
2 X2 to X3 200
3 X5 to X4 150
4 X4 to X3 250
5 X3 to OT 300

A stand-by vessel is kept in the field in normal circumstances to fulfill the following functions: (a) prevent unauthorized access to the platform, (b)
prevent collision with platform with passing vessels, (c) providing a temporary work base for crew visiting the platforms, (d) rescue of the men at sea,
(e) emergency evacuation and (f) additionally, small marine crafts are also available in the field to support emergency evacuation, depending upon the
type of work in progress. This field uses one 9-seater helicopter for air transportation of men and materials. It is based at the onshore terminal and is
available round the clock. An offshore well platform in this field is a tripod structure, which has the following facilities: (a) six well slots, (b) a
helideck, a must in view of rough sea and weather conditions in significant part of year, which is suitable for the operation of a BELL–212 helicopter
(or equivalent) and (c) a simplified boat landing facility which would minimize wave loads on jacket while permitting landing of the line boat/small
boat. The boat landing facility is required for any emergency approach and/or evacuation of personnel (for any maintenance) on the platform.

Other facilities include (a) well control panel, (b) inlet gas manifold to divert flow of any well through orifice flow metering system, (c) fusible plug
loop connected to hydraulic shut down panel, (d) DCP skid, (e) ESD/FSD panel/point at suitable locations, (f) pig launcher barrel for platform to
onshore terminal pipeline and pig receiving barrel, (g) solar panel with battery back up sufficient in case of no solar supply for 7 days, (h) liquid trap
for the gas to be used as instrument gas, (i) 3-4 ton capacity pedestal crane for material handling and slick line operations carried as part of reservoir

85
management, (j) all process piping designed for full well shut in pressure, (k) fog horn and navigations lights, (l) life saving and rescue equipment,
such as scramble net, jumping ropes, life raft, life rings, portable fire extinguishers etc., suitably placed, (m) portable generator, (n) Programmable
Logic Controller (PLC) integrated with VHF radio Line of Sight (LOS) lined to onshore control room for control and monitoring of the platform
(since the platform design is for unmanned operations), (p) safety shut down logic and instrumentation as per API recommended practices (RP-14C)
causing the platform to close, when required and (q) remote control system limited for closing of the wing valves in any emergency situation in
offshore or at onshore facilities, adjusting well flow by varying the choke and monitor flow rate production, indicating battery charge/discharge level,
pipeline pressure and temperature, SDV status, navigational light status.

RISK ASSESSMENT METHODOLOGY

The risk assessment methodology adopted in this study follows the Australian code AS 2885.1 (Pipelines - Gas and liquid petroleum - Design and
construction, 1997) along with SAA HB105 (Guide to pipeline risk assessment, 1998). The analysis has three distinct stages, namely, risk
identification, risk evaluation and risk management. The risk identification process is intended to identify all significant threats associated with the
pipeline. This process can be broken down into following four steps:

(a) Location Identification: It provides a basis for the identification of threats that are appropriate to the terrain (offshore or onshore) use and activities
and natural events along the pipeline route. It also identifies locations for consequence analysis in relation to the presence of people and
environmentally sensitive areas. As an example consider the pipeline segment 1 for platform riser and subsea pipeline within 100m radius of platform.
This location has unique risks compared to the pipeline segments such as the subsea pipeline away from the platform and the onshore pipeline or
pipeline within the terminal.

(b) Threat Identification: Threat is an activity or condition with the potential to cause failure or damage to pipeline. A checklist is used to
systematically identify the threats applicable to the pipelines and their locations. Threats could change based on location. For example, the riser
segment has a threat from ship impact as production supply vessels moves in the vicinity of the platform to cater to the operational requirements. The
subsea pipeline segment is less exposed to direct impact whereas it may be more vulnerable to the threat of anchor dragging. The onshore buried
pipeline will have threat of external interference like excavation by third parties.

(c) Existing Protection Measures: The safeguards in place to control or prevent each threat are identified in this analysis. Safeguards include protective
and preventive measures, which can be either physical or procedural. As an example, consider the pipeline segment 1 which has the following existing
protection measures against external corrosion threat incorporated as part of the design: (a) 6mm asphalt enamel coating on pipelines, (b) sacrificial
bracelet anode system for pipelines and (c) an additional anode has been provided for each riser at the splash zone. These existing protection measures
should be accounted for in the risk assessment.

(d) Failure and Hazardous Event Identification: Failure identification involves determination of the potential outcomes from a threat and hazardous
event identification involves determination of the event associated with the threat. As an example, the threat of external corrosion can result in loss of
pipeline wall thickness and ultimately lead to hydrocarbon release.

RISK EVALUATION

Frequency determination

The interpretation of severity and frequency of occurrence of a hazardous event is done in accordance with the industry practice. It is this frequency
which results in loss of integrity that must be categorized. This is the result of combining the frequency of the threat with the probability that the threat
will affect the pipeline and the probability that the result will be a loss of containment or integrity as determined by failure analysis. It is not simply the
frequency of the identified threat. Table 3.1 below provides guidelines for the assignment of frequency category.

Table 2 Guidelines for the assignment of a frequency category

Frequency of Description Nearest numerical


occurrence frequency for guidance
(Per 1000 km per year)
F- frequent Expected to occur at least once per year 1 or greater
E- occasional Expected to occur several times in the life of the pipeline 0.1
D- unlikely Not likely to occur in the life of the pipeline, but is 0.01
possible
C- remote Very unlikely to occur in the life of the pipeline 0.001
B- improbable Examples of this event have occurred historically, but it is 10 –5
not anticipated for the pipeline at this location
A- hypothetical Theoretically possible, but has never occurred to date 10−6 or lower

Severity determination

Some of the major the factors that affect the severity classification are: (a) operating conditions of the pipeline at a specific location, (b) the population
density, (c) type of population (e.g. sensitive ones such as schools, hospitals and retirement homes), (d) presence of environmentally sensitive

86
locations and (e) the potential for the event to escalate (e.g. bushfires and explosions impacting on nearby populations or facilities). Table 3.2 provides
guidelines for the assignment of severity category. This analysis includes both subsea and onshore pipelines where location, people distribution and
environmentally sensitive areas are likely to differ. There are no novel features along the length of the pipelines. The likely distribution of people and
environmentally sensitive areas is considered uniform. By comparison, at the platforms and the onshore terminal there is expected to be a
concentration of threats such as vessels, vehicles, operations and maintenance activities within close proximity to pipeline.

Table 3 Guidelines for the assignment of a severity category

Severity class Description


A- Catastrophic Application only in location, where the number of humans within the range of influence
would result in many fatalities.
B- Major Event causes a few fatalities, major loss of continuity of supply by the pipeline (economic
impact) or major environmental damage
C- severe Event causes hospitalization injuries or restriction of pipeline supply (economic impact)
D-Minor Event causes no injuries and no loss of, or restriction of, pipeline supply.

Based on these factors, the entire pipeline was divided into four locations for risk assessment: (a) location specific: covering the platform risers and
subsea pipelines in the vicinity of the platforms, (b) subsea non location specific: covering the pipelines from the shoreline to the close vicinity of the
platform, (c) non location specific: covering the trenched portion of the pipelines form shore crossing to the onshore processing terminal and (d)
location specific: pipelines within the onshore terminal. Changes made to the pipeline operations or design at later stages in its design life may change
the credibility of the threats perceived at the present time. For each perceived threat, the safeguards already in place are identified first. The safeguards
include protective and preventive measures, which can be either physical or procedural. Failure analysis and hazardous events are combined in the
study to facilitate evaluation of the risks. All hazardous events have been considered for risk evaluation. Results of risk ranking are captured in work
sheets. Risk evaluation can be achieved through the use of the risk matrix as shown in Table 4. Table 5 provides actions required for different risk
classes.

Table 4 Risk matrix

Frequency
Severity class
of occurrence
A -Catastrophic B- Major C- Severe D- Minor
F- frequent H H H I
E-occasional H H I L
D-unlikely H H L L
C- remote H I L L
B-improbable H I L N
A- hypothetical I L N N
Risk rankings: H - high risk, I - intermediate risk, L - low risk, and N - negligible risk

Table 5 Action required for different risk class

Risk class Action required


High (H) Intolerable. Reduce to intermediate class (I) by redesign or reassessment.
Intermediate Repeat risk identification and evaluation.
(I) Determine accuracy and uncertainties of estimation.
Ensure all protective measures and recommendations have been identified so that no
further practicable measures can be applied, i.e. risks are ALARP.
Low (L) Determine management plan for hazardous event to prevent its occurrence and monitor
changes,
which could affect the classification.
Negligible (N) Review at next review interval.

DISCUSSION OF RESULTS

Pipeline segment 1: (subsea pipeline within 100 m radius of the platform)


Fig. 4 indicates the risk picture of pipeline segment 1 before and after implementing the recommendations.

87
Risk Picture - Pipeline Segment 1
Before After

B B
E I

Intermediate
Occassional

Remote
Major

Major

High
Severity Likely Hood Risk Rank

Fig. 4 Risk picture of pipeline segment 1

From the figure it can be noted that the severity is in ‘major’ category for this section. As this segment is in the vicinity of the production platform,
which is a major ignition source, hydrocarbon release in this segment can result in (a) fatalities, as the platforms are manned during day time to take
care of operational requirements and maintenance activities, (b) major loss of continuity of supply by the pipeline which will result in economic
impact to the operating company, (c) major environmental damage from possible major fire accident involving various equipment and flammables on
the production platform. Likelihood is ‘occasional’ before implementation of the recommendations based on industry experience. However, severity
remains ‘major’ even after implementing the recommendations.

RECOMMENDATIONS

Pipeline Integrity Management

Offshore operators are increasingly using QRA techniques as part of their risk management programs. Structured risk assessment can be used at all
stages of a project life cycle. In the design stages, these techniques are valuable tools in ensuring that money is spent wisely on safety related systems.
In the operational stage, QRA can aid the development of procedures that combat and mitigate risk. Integrity management system covering issues such
as inspection, emergency planning etc. are crucial in order to maintain ‘as designed’ levels of safety and reliability. Audits are carried out all through
the operational phase to ensure that risks are not accumulated over time. Though various risks involved in a pipeline system are analysed in this work,
the key for safety of the pipeline system lies in continuously monitoring the risks and taking decisions on various mitigation measures with a holistic
approach. Defining the roles and responsibilities of various personnel involved is also very important. In this section, recommendations are provided,
which basically consists of pipeline integrity management scheme, with a holistic approach on life cycle of pipeline system. The management of the
pipeline integrity consists of four distinct areas of activity, namely, (a) operation and safety systems, (b) pipeline inspection, (c) corrosion
management, (d) modifications management. Within each of these areas an Integrity Engineer (IE) is to be appointed with clear definition of roles
and responsibilities. It is the responsibility of the Integrity Engineer to ensure that all inspection and testing activities are clearly defined and
satisfactorily conducted within the scheduled timescale. IE reports to the Pipeline Manager (PM) who takes the overall responsibility of the system
integrity. The integrity management scheme is based on clear definition of roles, limits of responsibility and understanding risk. The process is
continuous to ensure integrity of the pipelines and minimization of risk to personnel, environment and business.

INTEGRITY ASSURANCE PROCESS

The process for assuring the integrity of the ABC pipelines is based on the integrity assurance cycle. The process comprises each of the following
basic activities with a well developed management process for each activity:(a) define limits for each system, (b) co-ordination of inspection scheme,
(c) definition of monitoring requirements, (d) definition of acceptance and rejection criteria, (e) implementation of monitoring, inspection and testing
schemes, (f) integrity assessment based on the findings, (g) recommendation of remedial actions and (h) update of inspection schemes and tracking of
performance history. For the purpose of assuring technical integrity of the pipelines, the Pipeline Operator is the Overall Responsible Authority
(ORA). To facilitate effective control of the inspection, maintenance, testing and monitoring activities, the pipelines have been divided into sections.
Each of the sections is allocated to IE who has responsibility for performing all of the activities in the integrity assurance loop in respect of their
allocated sections. The PM is appointed by the ORA to approve the schemes developed by each of the IE. The PM is responsible for receiving,
reviewing and communicating the results and assessments obtained from each of the monitoring activities. The PM is responsible for assuring the
technical integrity of the pipeline system to the ORA at any time.

Coordination of Inspection Scheme


Each of the integrity engineers is responsible for coordinating their respective schemes, including appropriate control activities for the sections of the
pipeline allocated to them. The control activities shall be based on risk assessment and industrial Practice. Acceptance or rejection criteria are
developed by the IE, with the approval of the pipeline manager, against which the condition of the pipeline is monitored. Regular review of the
developed schemes are performed to confirm the continuing adequacy and relevant of the control activities. Control activities applicable to operation
of the pipelines include inspection, monitoring, testing and maintenance as appropriate. These activities will be performed to identify potential defects
in a timely fashion and ensure that the pipelines are maintained in efficient working order and good repair, as required by legislation. Pipeline

88
operating parameters are developed for each pipeline, based upon appropriate design codes, the specific design of each pipeline system, its current
condition and operational requirements. These are embedded in the operating procedures. Control activities also monitor the operating parameters of
the pipelines, to ensure that they are not operated beyond their safe operating limits. Change to key parameters of the pipeline, physical or operational,
are subject to the ABC Management of Change process, to ensure the associated risk are reviewed and addressed.

Implementation of Scheme and technical assessment


Each Integrity Engineer ensures that each of the control activities required are performed and reported in a timely manner to ensure continuous
integrity is achieved and shall support the activity (e.g. inspection) throughout. The results of each activity are recorded for tracking purposes by each
IE. The results the inspection, monitoring and control activities are assessed on a timely basis to ensure the continuing fitness for purpose of the
pipelines. The responsible Integrity Engineers shall carry out this in respect of their allocated pipeline sections, and shall maintain a record of the
results together with their assessment. In the event that the results fall outside the acceptance criteria, the results shall be reported to the Pipeline
manager to allow him to assess their significance for continuing operation of the pipeline.

RECOMMENDATIONS

Any findings requiring rectifications/modifications shall be highlighted to the Pipeline manager. The role of any work shall be defined and full
justification provided to the Pipeline manager with an assessment of the impact of such works. Approval of such works is by the Pipeline manager in
liaison with the ABC Management and the Integrity Engineer. Management of the execution of the rectification works is decided by the Pipeline
manager and ABC Management & similarly for the selection of contractor for the work. The Integrity Engineer may be given this task depending
upon the scale of the work, however the role of Integrity Engineer shall remain and they shall participate in a review of the engineering for any such
works, risk assessments and acceptance of the completed work. Summary reports are produced by each IE detailing the activities performed, the
results and their assessments. These reports may be produced following completion of an activity, a series of activities or prescribed intervals and shall
be submitted to the Pipeline manager as formal records. Additionally each IE shall submit a report every 3 months to the pipeline manager
summarizing status of their planned and ongoing works. At approximately three month intervals, a quarterly report is prepared by the pipeline
manager summarizing the works carried out during the period, planned works and main issues arising. This report will be provided to the Overall
Responsible Authority and each of the Integrity Engineers. At least once per year, a summary integrity report is produced by the Pipeline manager
confirming the fitness for purpose of the pipelines. The report shall summarize the work carried out during the year, the main findings, identifies any
outstanding actions and makes appropriate recommendations. Recommendations may include a procedural change, a change to the inspection program
or modification to plant. The Pipeline manager is ultimately responsible for ensuring that the recommendations are reviewed and adopted and
implemented, if required and appropriate. All actions raised against the ABC pipeline system will be tracked using an Action Tracking System, which
shall be maintained by the Pipeline manager.

REVIEW OF SCHEMES

The Pipeline Integrity Management Scheme is reviewed on an annual basis following the technical assessment and provision of recommendations.
The Pipeline manager is responsible for the review of the PIMS and ensures it remains appropriate in the light of each year’s findings. The results of
the review shall be included in the Pipeline manager summary integrity report as a formal record of the review. The Integrity Engineers shall review
annually the risk assessment, control activities and schedules, and if appropriate revise the schemes accordingly. The results of these reviews shall be
included in the summary reports as a formal record of the review. The Pipeline manager is responsible for arranging audits by person not involved in
the day-to-day pipeline integrity management process. Audit is an important aspect of the Pipeline Integrity Management Scheme (PIMS) and shall be
arranged at regular intervals as a means of (a) confirming compliance of PIMS with legislative and company requirements (b) confirming that the
PIMS is effectively implemented (c) providing a basis for continuous improvement. Audits may also be applied following incidents or failures. The
audits and outcomes are reviewed by the pipeline manager and the overall responsible authority. Any actions will be captured in the action tracking
system owned by the pipeline manager.

INTEGRITY MONITORING ACTIVITIES

The integrity of the ABC pipeline system is managed in accordance with this integrity management scheme by division of integrity into four distinct
areas namely; (a) operations and safety systems (b) pipeline inspection (c) corrosion management (d) modifications management. The integrity
engineers appointed to manage integrity in each of these areas have key responsibilities. The integrity engineers, the pipeline manager and the overall
responsible authority are jointly responsible for ensuring compliance with ABC’s technical integrity performance standards the key elements of which
are: (a) process parameters are monitored to ensure operation within specified operating limits (b) excursions shall be recorded and assessed (c)
external and internal deterioration are monitored at levels to meet stated integrity condition parameters and based on informed engineering judgment
(d) surveillance is conducted to identify activities or signs which could lead to failures. The method of determining the frequency of surveys shall be
documented (e) changes in operation, condition, configuration and loadings are reviewed and assessed against design condition and intent (f) critical
valves and associated control systems within the pipeline system (including emergency shutdown valves) shall be tested at intervals to confirm
integrity, availability and reliability. The method of determination of frequency of testing shall be documented. Instrumentation shall be tested at
appropriate intervals to demonstrate required reliability. Any deviation from design specifications or parameters is assessed prior to continued service.
These deviations are recorded and documented.

OPERATIONS AND SAFETY SYSTEMS

Day to day integrity monitoring is achieved through: (a) identifying, understanding and mitigating where practicable, all the risks associated with
Operations (b) monitoring and control of the process parameters (pressure, temperature, flow rates) (c) monitoring, assessment and analysis of the

89
process fluids (composition, H2S, CO2 etc) (d) monitoring and maintenance of all communication and ESD systems (e) ensuring that all monitoring
equipment is regularly maintained and calibrated (f) ensuring that all personnel are competent and have clear procedures and instructions for their
tasks (g) ensuring that all necessary planned maintenance routines are clear, understood, implemented and the results/actions recorded. Pipeline
operations are continually monitored and controlled from the control room at the platforms in accordance with approved operating procedures. This is
in interface with the onshore terminal. The control room monitors the process parameters at the trees, the platforms and at the reception facilities at the
Terminal. Excursions outwith the defined operating envelop are assessed by the control room operators for overall impact to the pipeline. If
unacceptable then the facility can be shut-in from the control room directly or in interface with the onshore terminal. Operations of the pipeline are
also continually monitored via the leak detection system. At each facility and at the platforms the ESD systems are critical both to protection of plant
and personnel. These must be maintained to ensure optimum performance. Work is strictly controlled at the platforms and at each of the offshore
facilities through works control system. The integrity engineer is responsible to ensuring this is done and results reported accordingly or action taken
when appropriate. This will typically include PSVs, ESD valves and their associated control systems.

PIPELINE INSPECTION

Numerous techniques and tools exist to evaluate the integrity of pipelines. Determining which technique and which tool is best requires experience.
Following techniques are considered in developing a pipeline integrity program:(a) coating-damage surveys (b) cathodic protection reviews (c)
hydrotesting (d) magnetic-flux leakage tools (conventional and high-resolution) (e) ultrasonic inspection tools (f) geometry tools (g) inertia-guidance
tools. Application of engineering principles to the data supplied by inspections is crucial for keeping repair costs low. Once data is received from
inspection contractors, it must be related to the operating environment of the pipeline. Factors to watch for include Data quality, Compliance with
specifications & Presentation. Once corrosion damage is detected, it must be determined if the pipeline is structurally sound to operate at the
maximum allowable operating pressure (MAOP). Corrosion damage reduces the capacity of a pipe to contain internal pressure, and if the corrosion is
allowed to proceed, the pipeline may eventually leak or rupture. Pipeline codes in some countries permit an engineering assessment based on a
rigorous fracture mechanics’ analysis. Apart from corrosion, other mechanical damage to pipelines is generally related to dents, gouges, arc burns,
etc., which are caused during construction or operation of the line. In all areas the role is focused on monitoring and control of the external condition
of the pipeline and the associated protection systems such as coatings and CP. Integrity monitoring is achieved by:(a) identifying, understanding and
mitigating against, where applicable, all potential external threats to the pipelines (b) developing risk based inspection and monitoring programs to
monitor for potential threats to the pipeline such as low CP, damaged coatings, etc. (c) developing tracking systems to ensure that trends in the
pipeline condition can be monitored and any potential problems identified (d) review of the third party integrity monitoring and inspection schemes to
ensure that no risk is imported. The risks associated with each section of pipeline are unique and consequently each has its own risk assessment and
risk based inspection and monitoring scheme. The integrity engineer is responsible for developing the inspection schemes and ensuring that the
inspection and monitoring tasks are performed in a timely manner. The inspection requirements are defined by the integrity engineer as are the
anomaly reporting criteria. The detailed inspection task plans are developed by the inspection body. All anomalies are tracked via an anomaly tracking
system to ensure timely close-out. All results are recorded for future reference and tracking purposes.

CONCLUSIONS

Pipelines are considered as the safest and most economical way of carrying very large volume of fluids, especially those which are perceived as
dangerous substances. A pipeline system, compared with road or rail transport, is usually a better controlled system and it largely behaves in a
predictable way. External interference, mostly third party activity, is one of the main causes of oil and natural gas pipeline incidents as pipeline
networks are not contained within a perimeter controlled by operator like chemical installations. Corrosion or material / construction defects are also
common causes of incidents. Qualitative risk is purely a relative measure and depends a good deal on the judgment of experts. The responses of these
experts in the form of a questionnaire and a one-day workshop were used in arriving at the various qualitative measures involved in the assessment.
The recommendations have been suggested with wide consultation with professionals in the industry and in most cases the risk could be improved
from ‘High’ to ‘Intermediate’ or ‘Low’ in all the pipeline segments. Quantitative risk analysis was carried out using industry standard software
considering a comprehensive range of scenarios of hazardous events. For the entire pipeline system, by adopting a host of recommendations, a
reduction of individual risk from 2.4× 10−5 to 1.87×10−6 has been achieved and a reduction of societal risk from 7.75×10−6 to 6.03×10−7 has been
achieved. These reduced risk values represent significant improvement over the base configuration of the gas field complex.

REFERENCES

ABS (2001), “Guide for building and classing subsea pipeline systems and risers”, American Bureau of Shipping.
ASME B 31.4 (1992), “Code for liquid transportation systems for hydrocarbons, liquid petroleum gas, anhydrous ammonia, and alcohols”, American
Society of Mechanical Engineers
ASME B 31.8 (1992), “Code for Gas transmission and Distribution Piping Systems”, American Society of Mechanical Engineers.
Bai Y (2001), “Pipelines and risers”, first edition, Elsevier.
Bai Y (2003), “Marine structural design”, Elsevier.
Brooker D.C. (2005), “Experimental puncture loads for external interference of pipelines by excavator equipment”, International Journal of Pressure
Vessels and Piping, 82, 825-832.
De Groot S.J. (1982), “The impact of laying and maintenance of offshore pipelines on the marine environment and the North Sea fisheries”, Ocean
Management, 8, 1-27.
EGIG (2002), “Gas pipeline incidents, European Gas Pipeline Incident Group”, 5th EGIG report 1970- 2001, EGIG 02.R.0058, Gasunie Technology
and Assets.
Fujiyama K., S. Nagai, Y. Akikuni, T. Fujiwara,K. Furuya, S. Matsumoto, K. Takagi and T. Kawabatta (2004), “Risk based inspection and
maintenance systems for steam turbines”, International Journal of Pressure Vessels and Piping, 81, 825-835.

90
Fukushima K., R. Maeshima, A. Kinoshita, H. Shiraishi and I. Koshijima (2000), “Gas pipeline leak detection system using the online simulation
method”, Computers and Chemical Engineering, 24, 453-456.
Leatherdale J.D. and D.J. Turner (1991), “Operational experience in underwater photogrammetry”, ISPRS Journal of Photogrammetry and Remote
Sensing, 46, 104-112.
Mazzola A. (1999), “Thermal interaction analysis in pipeline systems- a case study”, Journal of Loss Prevention in the Process Industries, 12 , 495-
505.
Muhlbauer W.K. (2004), “Pipeline risk management manual”, Third edition, Elsevier.
PARLOC (1996), “Pipeline and riser loss of containment, The update of loss of containment data for offshore pipelines, No. OTH 551”, United
Kingdom Health and Safety Executive.
Peter K and L. John (2000), “Report on a study of International Pipeline Accidents”, C1 10/00, United Kingdom Health and Safety Executive.
SAA HB105 (1998), “Guide to pipeline risk assessment in accordance with AS 2885.1”, Standards Australia International Ltd.
Smith E.J. (1995), “Risk management in the North Sea offshore industry: history, status and challenges”, Acta Astronautica, 37, 513-523.
TNO (1999), “Guidelines for quantitative risk analysis”, The Director General of Labour, The Hague, Netherlands
Vinnem J.E., T. Aven, T. Husebo, J. Seljelid and O.J. Tveit (2006), “Major hazard risk indicators for monitoring of trends in the Norwegian offshore
petroleum sector”, Reliability Engineering and System Safety, 91, 778-791.
Zingaretti P. and S.M. Zanoli (1998), “Robust real-time detection of underwater pipeline”, Engineering Applications of Artificial Intelligence, 11, 257-
268.

91
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Hazop study for crude oil pipe line


Srinivasan Chandrasekaran1*, Ramesh Babu2, J. and Arshad Ayub, S3
1
Associate Professor, Department of Ocean Engineering, Indian Institute of Technology Madras, Chennai, India
2
Cholamandalam MS Risk Services, Chennai, India
3
Graduate student in Petroleum Engineering, Department of Ocean Engineering, IIT Madras
*
corresponding author E-mail: drsekaran@iitm.ac.in

ABSTRACT

Crude oil pipe lines are mainly used for passing contamination from the sub-surface wells to the separator. These pipe lines made of reinforced
thermosetting resin plastic (RTRP) operate at 600 psi pressure and 212ºF temperature. Under these operating conditions, for a large vessel capacity,
crude oil stored from wells shall become a potential instigator for any chemical hazard. Under such conditions, damage or failure of pipe lines are
predictable using well defined mathematical models by employing HAZOP, even at the design stage itself. The present study discusses such a
structured and systematic method to identify possible hazards, causing damage to public, property and environment. HAZOP study, being a qualitative
method, is capable of modelling the factors that are responsible for causing damage to pipe lines. A case study of AXX Company is considered for the
analysis; for the identified nodes, relevant parameters and guide words are selected that can quantify identified deviations from the design intent.
Appropriate causes, consequences and safeguards are listed and recommended actions are suggested to prevent these identified hazards. The prepared
work sheet, after reviewed by group of Risk Engineers is presented in the paper along with final recommendations. The presented recommendations
and detailed study report shall illustrate the methods to assess and prevent the loss of property or public or environment under similar cases.

KEY WORDS: chemical risk; Hazop; crude oil spill; pipe line failure.

INTRODUCTION

HAZOP (Hazard and Operability) is a systematic study that is widely employed to predict all possible deviations from the given design intent. It
consists of causes, consequences, safeguards and recommendations. Crude oil from the source well pump (ESP) flows to the separator through the
RTRP pipe line. The deviations from this normal behavior may cause damage to public, property and environment. By using P&Ids, flow diagrams
and process description, group of engineers conducted the study to identify the potential hazards in the ensuing process. The objective of periodic
reviewing of the prepared report is to list few recommendations that shall reduce or eliminate the deviations. In recent years, HAZOP study is
becoming increasingly popular; even in few countries, this has been mandatory (for example, in Europe etc) (American Institute of Chemical
Engineers, 2000; Crowl and Louvar 2002). HAZOP technique is a highly disciplined procedure that attempts to identify possible deviations in a
process that shall result in deviations in the design intent. Emphasis is to identify the potential problems; not necessarily solving them. However, for
solution becoming obvious during the study, it is recorded or incorporated parallel while the study is in progress (Skelton 1997). Hazop studies are
necessarily carried out for the following reasons namely: i) to identify any significant hazardous exposures in the local vicinity of the process in
question; ii) to identify exposures that may impact the entire unit; iii) to identify exposures that may affect all plant personnel; as well as iv) to identify
exposures that may affect the surrounding community. It shall also be conducted to ensure adequacy of safety in the process, at design stage itself. The
objectives are: i) that no serious design errors or omissions are made; ii) that there are no obvious procedural problems; and iii) that the design
complies with required codes and standards. The major point to be reemphasized is that the study objectives should be clearly stated, documented, and
agreed upon prior to conducting the study (Lawley 1974; Turney 1990).

OBJECTIVES

In a HAZOP study, team of individuals systematically “brainstorms” the process under review; a set of appropriate guide words are selected to
structure the review. This multidisciplinary team concept allows various viewpoints of team members to stimulate the thinking that shall result in a
creative output. Consequently, a more thorough review is achieved than would occur if members of the team individually reviewed the same process.
Objectives of the present study is focused on three major domains namely: i) to identify hazards related to fire explosion, toxicity etc; ii) to identify
exposures that shall be local, or affect the entire facility or the surrounding community; and iii) to review the design process based on the adopted
safeguards, errors and omissions, procedural problems and compliance with the code regulations (Henly and Kumamoto 1996; Ebling 2000; Khan and
Abbasi 1999). The case study undertaken for illustration is risk assessment of a pipe line carrying crude oil.

PIPE LINE RISK ASSESSMENT

Case of study of AXX company is considered for the analysis. The HAZOP technique was initially developed to analyze chemical process systems,
but has later been extended to other types of systems and also to complex operations and to software systems (Kletz 1983). The Hazop study was
carried out by a team of experts from various disciplines namely inspection, instrumentation, electrical, fire prevention, process engineering,
maintenance and operations as well. The study is conducted by systematically applying guide words with appropriate process parameters to the
various lines and vessels in the plant to identify deviations from the design intent of normal process conditions. The Hazop study followed in the
present paper is focused on the Guide word technique; this technique enables the carry out the study in more systematic and structures manner. It is
therefore customary to use terminologies that are currently in prevalence in the process industry to assess such risks (Bullock et al. 1990). The most
important of these terms are given below:

Intention: The intention defines how the part of the process (being studied) is expected to operate. This can take a number of forms and can be either
descriptive or diagrammatic.
Causes: There are the reasons why deviations might occur. Once a deviation has been shown to have a conceivable or realistic cause, it can be treated
as meaningful.
Consequences: These are the results of the deviations should they occur.
Safeguards: Devices or procedures used for averting or mitigating the consequences.
Node: A section of a process unit generally consisting of piping and vessels that allows the study team to carry out HAZOP in an original fashion.

The guide words are the tools that are used to systematically direct the HAZOP study. It is essential to identify and list these guide words in the start of
the study as they are required by the risk analyzers to judge the potential hazards on the basis of the chosen guide words. They are words or phrases
that, when considered together with a parameter, form a hypothetical deviation for the team to consider. The basic guide words and phrases used in the
present are namely: NO, LESS, MORE, REVERSE, PART OF, AS WELL AS, and OTHER THAN. Parameters considered in the present study are of
two types namely i) general parameters those describe operations more so than physical aspects of the process and ii) specific parameters those
describe the physical aspects of the process. General parameters considered are REACTION, MAINTENANCE, ADDITION, and others those shall
become appropriate for more situations. Process parameters considered are FLOW, PRESSURE, TEMPERATURE, LEVEL, QUANTITY, and
TIME; they describe the process intention of vessels, piping or other equipment in specific terms. All relevant parameters are considered, but the
aforementioned parameters may be appropriate for more situations. Deviations are departures from the design intent that are discovered by
systematically applying the guide word/parameter combinations to study the process. Fig. 1 shows a sample of combination of guide words and
process. A matrix of common deviations is considered during the Hazop review.

 
Fig. 1 Matrix of common deviations
HAZOP WORKSHEETS

Crude oil is flowing from the source well pump (ESP) through RTRP (Reinforced Thermosetting Resin Plastic) pipe line to the separator. This pipe
line is designed at 600 Psig pressure and 2120F temperature. Initially crude is pumped from the source well to the two-phase flow separator using a
pump that is operating at flow rate of 54.51 m3/hr. For improving the flow characteristics, few inhibitors are added to the crude in the process line. PG
&PIC, made available for crude line are examined in detail; the same are not disclosed due to confidential reasons of the conducted study. A drain
vessel is placed in between the well and the separator line to remove the crude when the normal operation is shutdown for maintenance. Vessel is
surrounded by a concrete pit. The vessel consists of LG, LIT as safe guards; gas is sent through a vent to a safe location. While water is draining from
drain valve, crude is pumped to tanker. Different nodes on the pipe line carrying crude oil were investigated for possible hazards. The hazop work
sheets for different nodes are given below. The work sheets show the design intent and deviations, causing potential hazards to the chosen problem.

94
Table 1 Hazop work sheet for Node: Crude oil transit pipeline; Deviation: No/Less flow

Node: Crude oil Transit Pipeline


Deviation: No /Less Flow. Type: Pipe
Document:
Design Conditions/Parameters: Design Pressure=600 psig; design Temperature=2000 F; Maximum Flow Rate=2120 m3/hr
Causes Consequences Safeguards HAZOP Recommendations

1. Pump failure 1. No crude oil at the 1. PIT is available in the 1. Ensure proper maintenance of pump is to be done.
discharge outlet leads to pipeline.
process upset.
2. Scaling inside the 1. Pressure drop increases 1. Mechanism of cleaning
pipe. & hence pump load will of scaling by pig launcher
increase. is available.
3. Leakage in the pipeline 1. Environmental Hazard. 1. OCM is available in the
due to corrosion. 2. Contamination of crude pipeline.
oil. 2. Corrosion allowance is
3. No crude oil is available provided for the pipe line.
at discharge outlet.

4. Ball valve at 8’’-303# 1. Passing of fluid to 8’’- 1. A blind flange set


line kept in open position. 30# line. between two ordinary
flanges are provided.
5. Ball valve connected to 1. The crude oil will rush 1. PG available in the
vacuum tanker kept in to vacuum tanker causing pipeline for indication.
open position. no/less crude oil at the
discharge outlet.
6. MOV 1050 stuck in 1. Possible pressurization 1. PG is available for
closed position. of pipeline in the indication.
downstream of MOV.
2. No crude oil at the
discharge outlet.

7. MOV 1051 kept in 1. Less/No crude oil at the 1. Check valve at MOV
open position. discharge outlet. 1051 is provided.

Table 2 Hazop work sheet for Node: Crude oil transit pipeline; Deviation: More Flow

Node: Crude oil Transit Pipeline


Deviation: More Flow. Type: Pipe
Document:
Design Conditions/Parameters: Design Pressure=600 psig; design Temperature=2000 F; Maximum Flow Rate=2120 m3/hr
Causes Consequences Safeguards HAZOP Recommendations
1. Pump discharge is above its 1. Efficiency of pump will be 1. The pipeline is designed for 1. FIT with high alarm may
3
rated capacity decreased. maximum flow rate of 2120 m /hr be installed.
2. Power loss due to high friction.
2. Reduced delivery head. 1. The crude oil can’t be distributed 1. PG available in the pipeline for
constantly to the downstream. indication.
3.Greater fluid density 1. It will increase the load to pump 1. Flow meter may be
leading to decrease in its efficiency. installed in the pipeline.

95
Table 3 Hazop work sheet for Node: Crude oil transit pipeline; Deviation: Reverse Flow

Node: Crude oil transit pipeline


Deviation: Reverse Flow Type: Pipe
Document:
Design Conditions/Parameters:
Causes Consequences Safeguards HAZOP
Recommendations
1. Leakage through Check Valves. 1. Possible Pressurization 1. A blind flange set between two
Hazard. ordinary flanges is provided.

2. Inadvertent shut down of pump. 2. There will be no pressure in 2. PG & PIT are available in the
the pump. pipeline.

Table 4 Hazop work sheet for Node: Crude oil transit pipeline; Deviation: High pressure
Node: Crude oil Transit Pipeline
Deviation: High Pressure Type: Pipe
Document:
Design Conditions/Parameters:
Causes Consequences Safeguards HAZOP Recommendations
1. Deposition of scales 1. Pressure drop will 1. Pipeline cleaning 1. FIT with high alarm may be provided.
inside the pipelines. increase causing more mechanism by pig
loads to pump launcher is provided.
1. PIT is available.

2. Due to pump 1. Shock losses will be 1. PIT is available. 1. Impeller size or speed may be decreased.
running above its rated maximum causing
value. wastage of pump
power.

3. Due to inadvertent 1. Possible 1. PG is available. 1. SOP to be made to ensure that the MOV s are operated slowly as per
closure of MOV pressurization of 1. Pipeline is designed OEM requirements
(1050) pipeline. to operate for 600psig.

Table 5 Hazop work sheet for Node: Crude oil transit pipeline; Deviation: High temperature

Node: Crude oil transit pipeline


Deviation: High Temperature Type: Pipe
Document:
Design Conditions/Parameters:
Causes Consequences Safeguards HAZOP Recommendations
1. Ambient condition 1. It may contribute to excessive 1. PIT is connected to alarm panel 1. PSV may be installed in the
pressure build up inside the is available in the pipeline. pipeline.
pipeline leading to leak /rupture. 2. The pipe line is designed to
operate at 2000F.
2. Fire situation 1. Fire accident outside the 1.Medium velocity water spray 1. Heat and flame detectors may be
pipeline network may cause the system, Medium velocity water installed.
pipeline burst/rupture. sprinkler system may be provided
3. Internal fires 1. Explosion of pipeline. 1. Line detector and medium
velocity water spray system may be
installed.
4. High temperature crude from 1. It may increase the vapour 1. TI may be installed.
upstream end. pressure of crude causing 2. Expansion with to be
cavitations in the pipeline. provided to the pipeline.
2. Reduces the pump capacity.
3. May lead to severe erosion of
pipeline.
4. The expansion of pipeline due to
high temperature may cause
strained to valve fittings.

96
Table 6 Hazop work sheet for Node: Crude oil transit pipeline; Deviation: Low temperature

Node: Crude oil Transit Pipeline


Deviation: Low Temperature Type: Pipe
Document:
Design Conditions/Parameters:
Causes Consequences Safeguards HAZOP
Recommendations
1. Ambient Temperature 1. The density of crude may increase 1. Proper insulation of pipeline 1. Thickness above
leading to decrease in efficiency of 2. Pipeline is buried underground for critical insulation
pump. most of run thickness may be
provided for the
pipeline.

DISCUSSION OF RESULTS

The completed hazop worksheets analyses the deviation in design intent as given in Tables 1-6. It can be seen from the studies that various safeguards
that are intended to protect the process safety are taken into consideration while analyzing the deviations from the design intent. Despite the presence
of safe guards, identification of many deviations from the design intent shows the probable hazard of the chosen problem. The identified risk shall be
reduced or eliminated by adhering to the recommendations made by the expert team, as shown in the last column of each Hazop work sheet. However,
it can also be seen that all design deviations do not carry an alternative remedial measure (see, for example, Table 3). In such cases, mechanical
methods or computer controlled operations shall ensure avoidance of such deviations. Most commonly, deviations shall be overcome by periodic
inspection of valves, thickness, proper personnel training, and installing some more safe guards like PSV, TI FIT with high alarm; this shall however
reduce the risk but shall not eliminate it completely. The conducted study has few limitations namely: i) success of the review is highly dependent on
the accuracy of process data; ii) appropriate choice of expert team members with relevant technical experience and insight shall justify the arrived
recommendations of the hazop work sheets; as well as iii) the method, although methodical and systematic, shall also lead to tiring experience, making
it difficult to perform over extended period of time; this shall sometimes lead to critical situations called as brain burnout.

CONSLUSIONS

The presented study reviews the case sheet of AXX company’s crude oil cross-country pipe line. Probable hazards are identified on the basis of the
chosen guide words, after examining the process layout thoroughly. The detailed study conducted on the PI diagrams and the process of arriving at the
observed deviations is not presented in the paper as this shall expose the confidential assessment made on the existing project. The project, being of
international importance, the authors shall like to reserve the detailed investigations made on the PI diagrams. Nevertheless, deviations on the design
intent, observed during the study are presented for the most critical node shall find practical application and enhance understanding the proposed
methodology. Recommendations made in the Hazop work sheets shall reduce or eliminate the identified hazards for the chosen case study.

REFERENCES

American Institute of Chemical Engineers (AIChE). (2000), “Guidelines for chemical process quantitative risk analysis”, Second Ed. NY, AIChE.
Bullock, CJ, Mitchel, FR, Skelton, RL (1990), “Operability study applications in selected work field areas: Safety and reliability in the 90s”, Elsevier,
NY.
Chemical Industries Association. (1977), “A Guide to Hazard and Operability studies”, AIChem.
Crowl DA, Louvar, JF. (2002), “Chemical process safety: Fundamentals with applications”, Upper Saddle River, Prentice Hall, NJ.
Ebling CE (2000), “An introduction to reliability and maintainability engineering”, McGraw Hill Publishing Co Ltd, NY.
Henley, EJ, Kumamato H. (1996), “Probabilistic risk assessment and management for engineers and scientists”, IEE Press, NY, Second Ed.
Khan FI, Abbasi SA. (1999), “PROFAT: A user friendly system for probabilistic fault tree analysis”, Process Safety Progress, 18.
Kletz TA. (1983), “HAZOP & HAZAN Notes on the Identification and Assessment of Hazards”, IChemE Rugby, WHA of Process Plant Disasters.
Lawley HG.(1974), “Operability studies and hazard analysis”, Chemical Engineering Progress, 70(4), pp. 45-56.
Skelton, B. (1997), “Process safety analysis: An introduction”, Institute of Chemical Engineers, Warwickshire, U.K.
Turney, RD (1990), “Designing plants for 1990 and beyond: Procedures for control of safety, health and environmental hazards in the design of
chemical plants” Trans IChemE, 68 (B1), pp. 12-16.

97
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

MPD applications ranging from deep water to gas hydrate


Thangarasu K
Department of Petroleum Engineering,
Rajiv Gandhi College of Engineering, Chennai, India,
Email: thangarspe@ gmail.com

ABSTRACT

To meet the ever increasing energy demand petroleum industry is focusing on how to produce more amount of oil from unconventional reserves such as
deep water and unconventional resources such as gas hydrate. To overcome the challenges associated with the deep water reserves exploitations, a new
technique called Managed Pressure Drilling (MPD) is used nowadays. Gas hydrate reservoirs also have the same challenges associated with the deep water
oil reserves as most of the gas hydrate reservoirs lies in the deep water. Narrow drilling window between pore and fracture pressure in deep water is further
reduced in deep water gas hydrate reservoirs. This paper will focus on the particular variation of MPD such as Dual Gradient-Riserless Mud Recovery
(DG-RMR) and its possible application in Deep water gas hydrate reservoirs.

INTRODUCTION

Today in the modern world, energy is the main driver of Earth and its economy. High rate of depletion of petroleum reserves had led us to the peak period
of the world oil production. The scare of the world’s future energy needs including both producing and importing countries paves the way for high energy
demand in the world wide market. Hence the petroleum industry is facing the challenge of how to meet this increased energy demand after all easily
recoverable oil resources has almost depleted. The deep and ultra deep water resources previously labeled with non-drillable attracts the industry majors
with growing technological innovations. Besides that gas trapped in the form of gas hydrate under the deep water regions and permafrost area also
deserves industry focus because of its potential to hasten the peak in petroleum production world wide. But the difficulties associated with these deep
water resources such as high cost of drilling (NPT due to kick\loss cycles, narrow drilling window) need to be removed in order to successfully exploit
these resources. This paper will discuss how the challenges in exploiting the deep water reserves has been overcome by Managed Pressure Drillng (MPD).
This paper also discuss the possibility of using this MPD for Gas hydrate resources.

DEEP WATER CHALLENGES

Drilling in offshore area has been divided into Deep and Ultra deep water regions based on the increasing water depth. As the water depth increases, the
window between the pore pressure and fracture pressure becomes narrower. This is because; in marine environment formation pore pressure is the function
of sea water column. High water column in deep water regions increase the pore pressure due to the high trapment of water in the pore spaces which adds
high amount of hydrostatic pressure. But at the same time, erosion and absence of older sediments reduces the overburden pressure by decreasing the
fracture pressure of the formation. This narrow window between pore and fracture pressure requires high number of different size casing strings to reach
the producing interval which increase the cost of drilling because the increased weight of casings needs fourth or fifth generation floating rigs with high
capacity hook load and storage facilities. Due to the usage of high number of casing strings in the narrow window, it is less possible to reach the well with
the diameter less than what is required. Drilling the top hole area in the shallowest part of the deep water area is one of the problems in the deep water
region. Due to less overburden pressure formation, fracture pressure is low which needs many number of casing strings in the shallower zone itself.
Density fluctuation of drilling fluids in this zone should be avoided to because either it may harm the formation or the drilling rig.Most of the deep water
reservoirs often associated with the lost circulation zones, particularly, the subsalt sections and the reservoirs beneath the depleted zones. Loss of highly
denser mud used in these deep water operations and the associated Non-Productive Time (NPT) affects the economy of drilling considerably.

TECHNOLOGIES AVAILABLE TO DRILL DEEP WATER

Most important problem in deep water region is the close margin between pore and fracture gradient which require very high number of casing strings to
reach the producing interval. These high numbers of casing strings makes the wellbore like telescope and well may reach the producing interval with the
reduced size become unfit. Some of the following technologies seem to play greater role in deep water drilling.
¾ Managed Pressure Drilling (MPD)

¾ Casing while Drilling (CwD)

¾ Solid Expandable Tubular Drilling

Among these available technologies the first one MPD has proved its commercial success in deep water reserves exploitation. Exploitation of deep water
resource alone can not able to meet the increasing world wide demand. Effective ways to produce Gas hydrate, an unconventional resource, is the need of
these days.

MANAGED PRESSURE DRILLING

Managed Pressure Drilling is an adaptive drilling process used to precisely control the annular pressure profile throughout the wellbore. Circulation system
,is designed in a closed way such that well is not open to the atmosphere by using Rotating Control Devices (RCD) and Choke Manifold, thus the wellbore
annular pressure can be controlled by back pressure. Objectives of MPD are to reduce drilling risks and costs in narrow pressure margins, manage the
pressure profile, avoid fluid influx and contain incidental flow and to reduce the casing strings. There is no single way to achieve all those objectives;
several variations of MPD are available to achieve these objectives.

Advantages of MPD:
¾ Deeper open holes with few mud density changes

¾ Improved well control process with less NPT

¾ Less number of casing strings to Total Depth (TD)

¾ High rate of penetration with the use of less weight Mud

¾ No formation influx into the wellbore

Variations of MPD:
MPD is classified into four categories based on how it reduces the NPT of the drilling program.
1. Constant Bottom Hole Pressure (CBHP)

2. Dual Gradient Drilling (DG)

3. Controlled Mud Cap Drilling (CMC)

4. Return Flow Control Drilling (RFC)

Critical Analysis of Variations of MPD:

Variations Methods Mechanism Kick/loss Casing depth Drilling rate Critical


control selection parameters
Constant DAPC Maintain the Bottom Reduce No gain in Increases as Particle settling
Bottom Hole Hole Pressure (BHP) considerably reduction of it uses low velocity of mud, BP
Pressure by adjusting the back casing string weight mud
(CBHP) pressure

CCS Maintain the BHP by Reduces No gain in Increases as Pump rate, AFP,
maintaining Annular considerably reduction of it uses low annulus diameter
Friction Pressure casing string weight mud
(AFP)
Mud Cap PMCD Heavy weight mud Lost Reduces the Increases Viscosity of Cap
Drilling (MCD) and back pressure circulation is casing mud, BP, volume of
used to drilling ahead controlled numbers LCZ
of Lost Circulation Z not
eliminated
CMCD Maintain the BHP by Avoid lost Not Increases Mud level in the
adjusting mud level in circulation considerable marine riser, and
marine riser density of mud

100
Dual Gradient DG Reduces the No effect No effort Not Density of injected
Drilling (DG) Hydrostatic Head by considerable fluid
injection of lighter
weight fluid
RDG Uses two different No effect Reduces Not Mud lift pump,
gradients to make the considerably considerable density of mud in
effort as if the rig is in the riser
sea floor
Return Flow Divert the gas influx No no no Efficiency of
Control (RFC) safely to the flare diverter
section

GAS HYDRATES- PROMISING UNCONVENTIONAL RESOURCES

Gas hydrates are natural gas bearing ice like material which occurs in deep marine continental shelves and permafrost area. Following table illustrate the
potential of world wide gas hydrate deposit. Due to the low price of natural gas in the market these gas hydrates are considered as uneconomical to exploit.
But the peak in world oil production along with technological improvements like Gas to Liquid (GTL) conversion using fisher-Tropch process increased
the demand of drilling gas hydrates. Around twice the amount of Unconventional and conventional carbon energy is present in the form gas hydrate,
approximately 2.2 × 1016 pounds. Important property of gas hydrates is change in specific volume during the transition to gas hydrate from Free State and
vice versa. This may happen in reservoir, wellbore or surface. Gas Hydrate Stability Zone (GHSZ) is lies between 328 ft to 1640 ft below the ocean floor
and from 1640 ft to 6560 ft below the permafrost zone.

Table 2: World Gas Reserves

S.No Resource Type Quantity

1 Continental hydrates 5,000 – 12,000,000 *1012

2 Oceanic hydrates 30,000 – 49,100,000 *1012

3 Conventional 320,222 *1012

Gas Hydrate Challenges:

Similar to the deep water resources, these gas hydrates also posses unique set of challenges. Narrow margin between pore and fracture pressure in deep
water area and narrow margin between Hydrate dissociation pressure and Fracture pressure complicates the drilling. Some of the notable problems in Gas
hydrate drilling are as follow
¾ Collapse of well due to gas disassociation from the sediments

¾ Surface hole instability

¾ Plugging of well bore annulus if the well bore pressure reduces below the GHSZ

¾ Dissociation of gas in marine riser

¾ Dissociation of gas in upper part of the wellbore

¾ Plugging of marine risers due to the cooling effect of seawater

¾ Plug formation in choke and kill line which harm the well circulation

Need of MPD for Gas Hydrate:

Successful drilling in these hydrate reservoirs require innovative methods to overcome the difficulties associated with it. Instead of maintaining the bottom
hole pressure between formation pore pressure and fracture pressure, drilling in gas hydrate reservoir require to it to be balanced between hydrate
dissociation pressure and fracture pressure. This makes the drilling window narrow in these zones and further narrower in deep water. Application of
conventional or under balance drilling technology in these reservoirs may lead to the hydrate formation in the wellbore due to the upward decrease of
pressure in the wellbore. Pressure in the wellbore in both this method of drilling solely depends on mud weight say hydrostatic weight. The hydrostatic
mud weight increase with the depth and as the wellbore is open to atmosphere, rock cuttings from the hydrate reservoirs faces the drop in pressure while
travelling to the surface. Once it reaches the GHSZ of pressure and temperature, hydrate formation in the well bore is inevitable. Drilling in narrow
drilling window to avoid and control the gas dissociation from the reservoir and avoiding the hydrate formation in wellbore can be accomplished by the
MPD method.

101
Fig 1: wellbore pressure profile without BP

Fig 2: wellbore pressure profile with BP

The upward decrease of pressure in both the conventional and underbalanced drilling is compensated by the application back pressure so that hydrate
formation in the wellbore is avoided.
Application of conventional overbalance drilling technique to these gas hydrate reservoirs failed to maintain the wellbore geometry. As the formation is
overbalanced during drilling, any small increase in pressure while trip in process may harm the formation and leads to the loss of mud. This conventional
method requires large number of casing to compensate for the possibility of lost circulation and shallow gas flow.
Underbalanced drilling also failed to drill these wells successfully due to the following factors. Either increase in temperature or drop in pressure below the
GHSZ will lead to the gas dissociation from the sediments. This dissociation of sediments leads to the rapid collapse of wellbore in the reservoir section.
Further the liberated gas along with the cuttings again starts to form hydrate in the riser due to the drop in temperature from seabed to sea surface. This
hydrate plugs the riser and upper well head.

Dual Gradient-Riserless Mud Recovery (DG-RMR):

A particular variation of MPD such as dual gradient drilling along with Riserless Mud Recovery seems to be effective to overcome these problems. A
distinct and most important variation of dual gradient MPD is Riserless Mud Recovery. Pressure in the drill pipe at well head is balanced with sea water
hydrostatic pressure so that there is two gradient n the pressure regime, one from the surface to seabed which is equal to seawater hydrostatic head another
one from seabed to the bottom hole by the mud weight. This dual gradient is the alternative to the previous pump & Dump method of drilling used in the
deep water. This dual gradient variation reduces the casing strings as the true hydrostatic head in the wellbore starts at the seabed.

102
Fig 3: Application of RMR-DG to reduce Casing strings

By using the subsea BOP and RCD cuttings along with the mud is diverted into a Subsea mud pump and than to a return line. Rate of this subsea mud
pump is controlled to be less than the pumping speed of the mud pump. Back pressure generated by this method helps to avoid both the drop and increase
in the pressure of the wellbore. Thus by eliminate/ reduces the NPT and reduces the drilling time.
In deep water temperature is very low compared to the subsurface reservoir condition and if the pressure in the return line is still high, hydrates may form
in the return line. One of the possible solution to prevent the hydrate formation in the return line is to inject hydrate inhibitors in the return line but this is
limited if the cuttings are to be disposed in open system.
As the upper riser part is opened to the atmosphere, drilled cuttings experience a pressure drop here. Some times when the pressure drop is very low gas
will dissociate from the cuttings and may reach the rig floor. If it reaches the drilling floor it may cause fire there so that it should be diverted away from
the rig floor. Association of RCF method of MPD with this installation diverts the drilling mud to the flare line if it contains gas. But the possibility of gas
dissociation in the upper return line is very less.

Case Study: DG-RMR in Sakhalin Island


Field: Blavery Neftogez, Sakhlin Island, BP joint venture operation
Objectives: Zero discharge into sea due to environmental regulation, Shallow gas control, Reduce drilling time.
Solution handled: By using DG-RMR method, top hole section was drilled with full specification mud weight so full gauge hole was drilled at the top. To
avoid risk in gas zone mud weight is lowered and back pressure was applied..
Result: drill time or more than 9 days reduced due to not drilling a pilot hole, not using the marine riser, eliminating the shallow gas by maintaining
required pressure in the wellbore
This case study shows the success of DG-RMR in deep water drilling to eliminate NPT. So it can also be applied to the deep water gas hydrate as this also
have the similar problems of deep water oil reservoirs.

LIMITATION
¾ This paper focus only about how to use MPD in deep water hydrate reservoir.
¾ No effort was put to describe the mathematical relationship between GHSZ and BHP.
¾ No effort was put for the design of hydrate inhibitor valve in the return line.

CRITICAL ANALYSIS

¾ This proposed study is based in the predictions of higher demand for gas from hydrates in the future.

103
¾ With the current market price of natural gas it is difficult to implement this technology in the deep water unless the hydrate reservoir has large
volume of gas.

¾ Even through this method reduce drilling cost by eliminating the NPT associated with kick/loss cycle, reduces the casing string, it still requires
hike in gas price to implement worldwide.

¾ Proactive installation of Inhibitor injector valve and RCF add the cost of the rig but it is needed to handle the emergency.

CONCLUSION

Today in the peak period of world oil production, E&P industry is facing the challenge to meet the increasing energy demand. This triggered the industry
concentration to develop the deep water oil wells with the technological advancement. Particularly, MPD variations play the immense role in developing
these deep water resources which are previously labeled as Un-Economical to drill.
MPD method, a collection of different variation eliminates the NPT in drilling, and reduces the casing strings. But to meet the increasing energy demand
more oil and gas production from the unconventional resources is required. DG-RMR variation of MPD can be used to drill gas hydrate reservoir, a
promising alternative unconventional resource.
For the successful implementation of this DG-RMR, gas prices need to be hiked but this method of drilling to produce gas hydrate will definitely play a
major role in meeting the ever increasing energy demand. Conclusion is case speficic – what sort of generalities you observe or suggest, for a theoretical
suggestion or paper like this you can never suggest real technical solutions, you need to make it general. Limitations of this paper is not highlighted at all?
Not to the alignment requirement of papers.

REFERENCES:
Greg myers. “Ultra-Deepwater Riserless Mud Circulation With Dual Gradient Drilling”, Scientific Drilling, no. 6, July 2008.
Neil forrest, “Subsea Equipment For Deep Water Drilling Using Dual Gradient Mud System” SPE/IADC 67707.
Erdem tercan. “Managed Pressure Drilling Techniques, Equipment & Applications”, a thesis submitted to the graduate school of natural and applied
sciences of middle east technical university.
Don Hannegan. “The time Has Come To Develop Riserless Mud Recovery Technology’s Deepwater Capabilities”, D r i l l i n g c o n t r a c t or,
September/October 2006.
George Medley. “Distinct Variations of Managed Pressure Drilling Exhibit Application Potential”, March 2006, Vol. 227, No. 3.
Ken Smith. “MPD Helps to Make Problems Disappear “, d r i l l i n g c o n t r a c t o r, September/October 2006.
Afolabi Ayoola Amodu. “Drilling Through Gas Hydrate Formations: Possible Problems And Suggested Solutions”. Texas A&M University, August 2008.
Mari Oma Engevik. “Risk Assessment of Underbalanced and Managed Pressure Drilling Operations”, Trondheim June 8, 2007.
Don Hannegan, P.E. “MPD – Uniquely Applicable to Methane Hydrate Drilling”, SPE/IADC -91560.
Børre Fossil, “Managed Pressure Drilling for Subsea Applications; Well Control Challenges in Deep Waters.”, SPE/IADC 91633.
Paul Scott. “Pushing the Limits of Riserless Deepwater Drilling”, AADE-06-DF-HO-30.
Harsh Gupta, “Gas Hydrates: A Potential Source Of Energy From The Oceans”, Intergovernmental Oceanographic Commission (IOC), June 25, 2003.
Todd, Richard J. “Managed Pressure And/Or Temperature Drilling System And Method”, United States Patent Application 20090236144 , 09/24/2009.
Franklin, Lindsay J. “Gas Hydrates Drilling Procedure”, United States Patent 4422513, 12/27/1983.

104
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Design of new-generation drill ship with ice operating capabilities


Bo-ram Kim1*, Tae-wan Kim2 and Seung-chul Kang3
1*&3
Graduate student,2Professor, Department of Naval Architecture and Ocean Engineering,
Seoul National University, Seoul, Republic of Korea.
E-mail: 1*kbr0376@snu.ac.kr, 2taewan@snu.ac.kr, 3kangsch@snu.ac.kr

ABSTRACT

Many resources like below oceans covered with ice, and till today remain relatively unexplored. Furthermore, ice covered oceans hold the key for
many of earth climate's secrets. For example, how do the sea ice coverage and the sea water properties change? How do plants and animals survive
under the most extreme conditions of the earth? Which information of past climate change can be read from the sediments at the sea-floor and how can
the future climate changes can be predicted? In order to explore vital resources and understand earth climate better, a drillship that can operate in the
cool summers and freezing winters of the ice covered oceans and that can drill deep into the sea floor is needed. The research on design of drill ship
with ice operating capabilities is rare in the existing literature. In this paper, we explore a multi-functional design of a drillship for deep ocean basins
that also support climate/environmental research facilities. The design features incorporate advanced technological facilities like dynamic positioning
in closed sea-ice cover, satellite navigation and ice-management. Our design investigation aims for deep drilling rig, which enables sampling of the
ocean floor and sub-sea up to 5,000m – 7,500m water depth and 1,000m – 1,500m penetration. The on-board equipmentis accommodated in a flexible
manner and infrastructure (in particular, winches, cranes, etc.), free deck-space and separate protected deck areas are optimized. Finally, we present
show that the design approach can be efficiently used to design a drill ship that meets the specifications of the highest ice-class for polar icebreakers.

INTRODUCTION

In order to explore more resources in arctic deep water, conceptual design of drill ship with ice operating have been performed. In this paper we show
historical background of offshore oil industry and marine routes in the Arctic Ocean. In the following, some points to be considered to design arctic
drillship will be suggested.

OFFSHORE OIL INDUSTRY

History of the offshore oil industry


The modern petroleum development was started in 1859 when E.L. Drake succeed to develop the 21m deep well at Pennsylvania, U.S. After that time
onshore oil wells were developed widely.Petroleum development area was extended to the sea when offshore oil wells were started to be developed in
Summerland, California, U.S. in the early 1900s.

Offshore plant
For inshore oil well, the jack-up rigs (Fig 2) have beeninstalledwidely. Jack-up rig is consisted with hull and number of legs, and can be put up to a
maximum depth about 200m. After jack-up rig completing the wells, jacket works to produce oil (Fig 3). Jacket is a fixed-producing unit which has
truss substructure and upper producing plant.
Fig 1Offshore at Summerland, CaliforniaError! Reference source not found.]

Fig 2 Jack-up Rig

Fig 3 Jacket

Development of shallow oil wells were going to finish, offshore oil development has been extended deeper and deeper. And floating equipment such
as semisubmersible rig or drillship was developed. Semisubmersible rig is a floating drilling rig which has minimal water area to improve motion
response. Semisubmersible is divided into semisubmersible drilling rig and semisubmersible production unit by usage. Fig 4 shows traditional
semisubmersible drilling rig and Fig 5 shows semisubmersible production unit. Semisubmersible has two pontoons under waterline and four to six legs

106
which connect pontoons and deck. By reducing water area, semisubmersible can get good motion response, but the deadweight has to be decreased.
Thus semisubmersible drilling rig need another supply boat to load weights and semisubmersible production rig also requires extra crude oil carrier to
storagethe crude oil which semisubmersible production rig produced.

Fig 4 Semisubmersible Drilling RigError! Reference source not found.]

Fig 5 Semisubmersible Production RigError! Reference source not found.]

Drillship, as shown at Fig 6, has ship hull with a large drilling derrick. Compare with the semisubmersible, drillship has enough deadweight because of
the large water area. Although the drillship has bad motion response, it can be improved by using dynamic positioning system which consists of
several azimuth thrusters.

107
Fig 6 Drillship –‘Stena DrillMax’Error! Reference source not found.]

ARRANGEMENT OF DRILLSHIP

Drillship is divided into the upper structure for drilling and hull. At the middle of the drillship, there is a large hole which is called ‘moon pool’ and
drilling derrick is mounted on it. At ahead of derrick, there is drilling pipe rack. And at astern of derrick, riser rack is located. There are knuckle boom
cranes to handling each drilling pipe and riser. At ahead of riser rack, there is accommodation for the sailors living. And the front of the
accommodation, there is helideck for helicopter landing there. To control movement and posture of drillship, the azimuth thruster propulsion is
equipped at the bottom of the hull.

Fig 7Arrangement of drillship

ARCTIC OCEAN MARINE ROUTES AND ARCTIC DRILLSHIP

Arctic Ocean marine routes


There are two marine routes in Arctic Ocean. One of them is Northwest Passage, and the other one is northern sea route. Around the 16th century
people believe that they can sail into the Pacific from North Atlantic near Greenland. John Dais tried to find Northwest Passage three times in 1585,
1586 and 1587 in the time of Queen Elizabeth I, but the route was never pioneered.
Northern sea route follows coast of northern Russia from the Atlantic to the Pacific. In 1553 Spanish explorer, SebastianoCaboto almost succeededto
find the route, but he ended his way at the Ob River. The orange line of Fig 8 shows Northwest Passage and the blue line of Fig 8 shows northern sea
route.

108
Fig 8 Arctic ocean marine routesError! Reference source not found.]

Although the advantage of short distance, Northwest Passage and northern sea route is unusable route because of ice cap. However today’s global
warming makes ice cap melted, so it is possible to pass the route. In fact, a Canadian cargo ship ‘MV Camilla Desgagnés’ succeed to make a voyage
from the Atlantic to the Pacific in November 2008Error! Reference source not found.].

Fig 9 represents the current route and Northwest Passage which connect Rotterdam and San Francisco. Current route is 15,000km distant, but
Northwest Passage is 11,000km distant which is shorter than 25% of current route. Fig 10 represent the current route and northern sea route which
connect Rotterdam and Yokohama, Japan. Current route is 21,000km distant, but northern sea route is 12,700km.

Fig 9 Northwest Passage

109
Fig 10Northern Sea Route

Arctic drillship
The opening of Arctic Ocean marine routes makes oil well development in Polar Regions become increasingly active.Previously introduced in Fig
6‘Stena DrillMax’ is an Arctic drillship which can operate in minus 40Ԩ, significant wave height 16m and 41m/s wind. This Arctic drillship has no
ability to break ice, so another ice breaker is needed to form a fleet. As can be seen in Fig 11 and Fig 12, the nuclear icebreaker breaks the ice first, and
diesel propulsion icebreaker breaks twice. After then drillship can pass through ice.

Fig 11 Ice Breakers and Arctic DrillshipError! Reference source not found.]

110
Fig 12 Ice Breakers and Arctic Drillship Error! Reference source not found.]

In recent years, a drillship which has ice breaking ability without the additional support of any icebreaker. Fig 13 shows ‘Vidar Viking’ which was
originally built as icebreaker but converted to drillship with ice breaking ability. Although ‘Vidar Viking’ has drilling derrick and helideck but it is not
designed for drilling originally. So it is not suitable for commercial drilling.

Fig 13Vidar VikingError! Reference source not found.]

111
However ‘Stena DrillMax ICE’(Fig 14) or ‘AURORA BOREALIES’(Fig 15) which are now in conceptual design stage have enough ice breaking
ability to be used for commercial drilling.

Fig 14 Stena DrillMAX ICEError! Reference source not found.]

Fig 15 AURORA BOREALIESError! Reference source not found.]

Because of the hazard environment, the following items should be considered to design Arctic drillship.
- Hull Strengthening for Ice class ship,
- Green Water Protection,
- Anti-icing/De-icing equipment,
- Powerful thruster and Thruster Canister.

Hull strengthening for Iceclass ship

IACS(International Association of Classification Societies) established hull strengthening rule for Iceclass ship. In that rule, ship hull divided into
bow(B), bow intermediate(BI), midbody(M) and stern(S). And each of them is divided into icebelt(i), lower(l) and bottom(b)Error! Reference
source not found.]. All of each region has different hull strength.

112
Fig 16 Hull Strengthening for IceclassshipError! Reference source not found.]
Green Water Protection

North Sea, which belongs to Arctic Ocean, has average temperature minus 20Ԩ,significant wave height 16m and 41m/s wind.
Due to the high significant wave height, the greenwater effect may be occurs. The greenwater effect is occurs by waves over the foredeck and can
make equipment damaged which installed on the deck. So to avoid damage of greenwater effect, some special structure should be designed. Fig 18
shows a kind of that structure, greenwater protection wall.

Fig 17Greenwater

Fig 18 Greenwater protection wall

Anti-icing/De-icing equipment

Due to low temperature in Polar Regions ice can be formed on the deck of ship as Fig 19shows. In this situation workers cannot move on the deck and
the drilling operation on the deck is almost impossible. So anti-icing or de-icing technology should be adapted to Artic drillship.

Fig. 20 shows stairs and escape way which has anti-icing and de-icing function.

113
Fig 19 Ice accumulations on deck

Fig. 20 Heated Traced stairs (upper) and escape ways (lower)

Powerful thruster and Thruster Canister


When drilling operation, drillship should keep the position without moving. To positioning one position in spite of wave influence some special
propulsion system which can rotate 360 degree is needed. Fig 21 shows that special propulsion called azimuth thruster. ‘Stena DrillMax Ice’ in Fig 14
is designed to install six sets of 5.5kW Azimuth thruster UUC505.

Fig 21Azimuth thruster

114
If some problemshaveoccurredwith propulsion system of drillship, generally the drillship should go to the repair shipyard for repairs. But it takes too
long time to move from Arctic Ocean to shore where repair shipyard located. Long delay time of drillship means very much financial loss. To avoid
this situation nowadays thruster canister as shown in
Fig 22 and Fig 23 is installed to drillship. By using thruster canister, thruster can be moved to the deck of drillship and fixed. The loss of time is
decreased impressively.

Fig 22 Thruster Canister - AppearanceError! Reference source not found.]

115
Fig 23Thruster Canister - DraftError! Reference source not found.]

CONCLUSION

In this paper we show historical background of offshore oil industry, and new marine routes in the Arctic Ocean which is able to be used by the effect
of global warming. The opening of new marine routes in the Arctic Ocean means that offshore oil well region can be extended to the Arctic region. To
operate in Arctic hazard region, Arctic drillship should be having special equipment as mentioned in this paper.

Nowadays the development of drillship with ice operating capabilities is in progress, conceptual design stage. Because of the depletion of oil well on
normal region, arctic drillship should be developed soon. The development of arctic drillship makes big benefits to oil companies and shipyards.

ACKNOWLEDGEMENTS

This study was performed as part of the research project grant number 2009-00812190 and supported by the Ministry of Education, Science and
Technology, government of Republic of Korea. We acknowledge and appreciate the support provided.

REFERENCES

AURORA BOREALIES, http://www.eri-aurora-borealis.eu


BassTech, http://www.bass-tech.com
European Space Agency, http://www.esa.int/
G.H. Eldridge, “Nature and Science on the Pacific Coast,” Pacific Committee of the American Association for the Advancement of
Science, 1915.
Hyundai Heavy Industry, Inc., http://www.hhi.co.kr
IACS, Requirements concerning POLAR CLASS, IACS Req. 2007
Kathryn Moran, Jan Backman, and John W. Farrell, “Deepwater drilling in the Arctic Ocean’s permanent sea ice,”Proceeding of the Integrated Ocean
Drilling Program, Vol. 302. doi:10.2204/iodp.proc.302.106.2006
Petrobras, http://www.petrobras.com.br
Stena Drilling, http://www.stena-drilling.com
Subrata K. Chakrabarti, Handbook of offshore engineering, Elsvier, 2005.
The CBC News, “1st commercial ship sails through Northwest Passage,” 2008. 11. 28.
Thrustmaster of Texas, Inc., http://www.thrustmastertexas.com
United Oil and Gas Consortium Management Corp, http://unoilgas.org

116
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Deepwater Drilling Platforms


Reetanshu Kumar1, Shiva Dayal1, Rahul Ranakoti1 and Astha Gupta1
1
Applied Pertroleum Engineering, University of Petroleum and Energy Studies,
Dehradun Uttarakhand (India)
E-mails: ranakoti_rahul@yahoo.com/ kreetanshu@gmail.com/shivasrivastava90@gmail.com

ABSTRACT

The paper will present the importance of offshore hydrocarbon exploitation, the basic steps in the development process and the introduction of the
major structural concepts (jacket-based, FPSO, GBS-based, TLP, floating) for deepwater drilling. Floating production storage and offloading (FPSO)
vessels offer a cost-effective field development solution, especially in deepwater areas lacking an adequate pipeline network. Most FPSOs are
permanently moored, i.e. the complete system is designed to withstand any kind of extreme environment at the field location. FPSOs that can be
quickly disconnected from their moorings and risers have also been designed and deployed. The key feature of this type of disconnectable FPSO is
that it can be disconnect and so avoid dangerous environmental conditions such as icebergs, hurricanes in the Gulf of Mexico and typhoons in the
South China Sea. In this paper, the concept of disconnectable FPSOs for deepwater field development is also presented. Key technologies and their
engineering analyses are highlighted. The merits and demerits of disconnectable vs permanent FPSOs are then evaluated. The paper concludes that
both permanent and disconnectable FPSOs are versatile floating systems and their selection depends on safety, technological, cost and operational
considerations.
Over the years many types of platforms have been used for different purposes under various conditions. Structures that float near the water surface-
Recent development include
• Tension Leg platforms
• Semi Submersible
• Spar
• Ship shaped vessel (FPSO)

These are widely used these days especially in deepwater drilling.


Fixed structures that extend to the Seabed are.
• Steel Jacket
• Concrete gravity Structure
• Compliant Tower

All environmental phenomena of importance should be considered before setting up a platform such as wind, waves, currents, tides, ice loads and
earthquake.
Tensile strength is the key mechanical properties for strength design of structures. The materials used are required to have satisfactory weldability and
fracture toughness that is satisfactory for the intended application environment (temperature). Fatigue and corrosion characteristics are also important
material properties.

This paper describes the design of a Floating Production Storage and Offloading system (FPSO) for the deepwaters as well as the conditions that affect
its stability and production capability.

INTRODUCTION

With offshore oil reserves conservatively estimated at some 160,000 barrels and some 490 trillion cubic feet of gas, the exploration for oil has
extended to all corners of the globe and probe deeper into the vast expense of the oceans. Exploration is now going on off the coast of almost all
countries and over 40 countries now have production facilities in place or undergoing construction. Today offshore production accounts for about 20%
of the world’s oil production, which 10 years ago, was just 3% of the total oil production.
As the search for oil leads to exploration in deeper waters and rougher seas, the semi-submersible, some with displacement of over 30,000 tonnes are
required and fewer jack-up units will be used. In recent years, drillships which are dynamically positioned without the use of anchors have become
increasingly popular as they can carry large quantities of supplies and travel to the remote regions of the seas with little difficulty.
Thus technology and skills are now clearly available to drill for oil at considerably greater depths than is currently being done and at a cost quite
similar to that of an onshore drilling rig. The advent and necessity of offshore drilling on the sea floor promises to be very exciting, challenging and
rewarding venture for a technical community and will involve a major investment of human and financial resources of many companies and nations
around the world.

TYPES OF OFFSHORE PLATFORMS

1, 2) conventional fixed platforms; 3) compliant tower; 4, 5) vertically moored tension leg and mini-tension leg platform; 6) Spar ; 7,8) Semi-
submersibles ; 9) Floating production, storage, and offloading facility; 10) sub-sea completion and tie-back to host facility.[6]

Fixed platforms

1. Fixed Jacketed Structure:


¾ Design: production; steel framed tubular structure attached to sea bed with piles which are driven into the seafloor (legs act as a
guiding device or “jacket” for the piles); constructed in sections and transported to the site; design lifetime of 10-25 years.
¾ Depth: shallow… 500 m or less
¾ Advantages: (1) support large deck loads, (2) may be constructed in sections & transported, (3) large field, long term production
(supports a large number of wells), (4) piles result in good stability, (5) little effect from seafloor scour
¾ Disadvantages: (1) costs increase exponentially with depth, (2) high initial & maintenance costs, (3) not reusable, (4) steel
structural members subject to corrosion
2. Gravity Structure:
¾ Design: production; large reinforced concrete bottom mounted structure which uses its weight to resist environmental
loads, not attached to the bottom with piles
¾ Depth: medium… up to 350 m
¾ Advantages: (1) support large deck loads, (2) possible reuse, (3) construction and testing may be completed before floating
and towing to site, (4) large field, long term production (supports a large number of wells), (5) may have large storage
capacity, (6) more tolerant to overloading & sea water exposure than steel jacketed platforms
¾ Disadvantages: (1) cost increases exponentially with depth, (2) foundation settlement, (3) subject to seafloor scour, (4) may
require more steel than steel jacket structures
3. Complaint tower(Guyed Tower):
¾ Design: small field production; slender truss-steel structure supported by a “spud-can” foundation and held upright by multiple wire or
chain guy lines which are held in place by clump weights and anchor piles; clump weights may rise off the seafloor in extreme
conditions and provide added restoring force.
¾ Depth: medium… 200-600 m
¾ Advantages: (1) low cost (lower than steel jacket), (2) good stability - guy lines and clump weights give added restoring force, (3)
possible reuse
¾ Disadvantages: (1) high maintenance costs, (2) small fields only, (3) cost increases exponentially with depth, (4) difficult mooring

118
Movable platforms

1. Tension Leg Platform:


¾ Design: production; semisubmersible tethered to seafloor with vertical anchor lines (cables or pipes) maintained in tension by the
excess buoyancy of the platform
¾ Depth: shallow to deep… 120 - 1500 m
¾ Advantages: (1) mobile & reusable, (2) stable - minimal vertical motion, (3) low cost increase with depth increase, (4) deep water
capability, (5) low maintenance cost
¾ Disadvantages: (1) high initial cost, (2) high subsea cost, (3) fatigue of tension legs, (4) difficult maintenance of subsea systems, (5)
little or no storage
2. Semisubmersible :
¾ Design: exploratory and production; floating structure; towed to the site, ballasted and moored (anchored); large vertical columns
connected to large pontoons, the columns support the deck structure and equipment
¾ Depth: shallow to medium…90 - 1000 m
¾ Advantages: (1) mobile with high transit speed (~10 kts), (2) stable - minimal response to wave action, (3) large deck area.
¾ Disadvantages: (1) high initial & operating costs, (2) limited deck load (low reserve buoyancy), (3) structural fatigue, (4)
expensive to move large distances, (5) limited dry-docking facilities available, (6) difficult to handle mooring systems and land BOP
stack & riser in rough seas
3. FPSO:

The design basis of this FPSO has been developed with the intent of enabling the unit to be easily redeployable in different locations, either as an
Early Production System (EPS) for large fields, or a Full Field Development on marginal or medium size fields (30-80,000 bopd).
Different technologies are compared and their design and operational limitations identified, for riser design (flexible risers, hybrid riser towers, top
tensions risers, single well riser topers, permanent and disconnectable) and station keeping design, (permanent or disconnectable, dynamic positioning
or anchor leg moorings, polyester or wire rope mooring systems). Those aspects of the design which require technology new to the Gulf of mexico,
will be highlighted and comparable applications of the technology elsewhere in the world, if applicable, identified.
There has been an increased interest recently in the application of FPSOs for the development of deepwater fields in the Gulf of Mexico. Damage to
offshore facilities and the pipeline infrastructure during recent hurricane seasons, has made the benefits of the FPSO concept more apparent, with its
large onboard oil storage capacity and capability to offload via shuttle tankers or pipelines, and its ability to disconnect from risers/moorings and avoid
the hurricane conditions design case. Understanding the issues relative to the use of FPSOs in the Gulf of Mexico and developing a design, which
utilizes the proven technology of FPSOs elsewhere in the world is key to applying this concept successfully in this region.
So far there have been no FPSOs installed in the US portion of Gulf of Mexico even though the concept has been increasingly selected for field
developments elsewhere in the world. What are the key issues when reviewing the suitability of a FPSOs for this region, and what has changed
recently to affect the selection process? The damage caused to the offshore units and the pipeline system by hurricane Ivan, Katrina and Rita, plus the
continuing exploration success in the ultra deep water beyond the continental shelf, are two of the main drivers towards the current interests in FPSOs
in the Gulf of Mexico.
The ability of the FPSOs to be designed to disconnect and sail away prior to the arrival of a hurricane , and the fact it is not dependent o n the pipeline
network to transport oil, make it an inherently more robust and safe design concept. Recent increases in the hurricane wind speed and wave height
design criteria , by the operating companies, further enhances the benefits of the FPSOs concept , particularly if the FPSO is disconnectable to avoid
the hurricane condition.

119
Special design issues for GoM FPSO

There are special issues to be considered when developing the design and schedule of the FPSO and hence the price. Individually these issues are not
unique to the USA GoM, but the combination of all of them in one region is
• Regulatory and environmental requirements.
• Vessel selection and refurbishment for design life extension.
• Riser and station keeping/mooring design for hurricane and loop current conditions in deepwater.
• Top sides design including associated gas disposal.
• Oil import and export facilities.
• Shuttle tanker requirements.

Vessel selection and refurbishment

The size of tanker for the FPSO depends on the crude oil storage volume requirement and the deck space required for the facilities layout. The crude
oil storage volume requirement is established by the following variables
• Size of oil offloading parcel size (shuttle tanker size capacity).
• The practical frequency of shuttle tanker berthing and offloading operations at the FPSO.
• The oil production rat e of the field and the topsides processing capacity.
• The preferred storage capacity contingency in order to maintain full production even if the shuttle tanker is delayed for some reason
(weather, operation al problems, or offloading terminal delays).

With a maximum shuttle tanker size of 350,000 bbls, and an assumed production rate of 60-80,000 BOPD, a minimum FPSO oil storage capacity of
500-600,000 bbls, would be suitable for a remote GoM deepwater location. This storage volume compares to the capacity of an Aframax size
tanker(80,000- 120.000 dwt).
Riser and station keeping
When selecting the station keeping system for an FPSO in the GoM the main challenges are
• Hurricane environment
• Loop current
• Unpredictable seabed conditions( e.g. deep ravines)
• Sub-sea architecture

A consideration applicable to the GoM region for permanent moored FPSOs would be its ability to weathervane in a hurricane environment. A study
conducted by SBM to access the survivability of an FPSO in the unlikely event of a category five hurricane passing over a FPSO, concluded the
following
• Even though the wind direction could change by almost 180degree in less than 2 hours in the path of the hurricane, higher mooring line
tension has not been observed in the eye of the hurricane, due to reduced wave height and wind speed. Instead the extreme mooring line
tension design is governed by the responses caused by the hurricane fringe because of higher wave height, whilst maximum wind speed
from the time series are almost the same for both locations and current speed is far less than 1min/sec most of the time.
• Transient analysis has shown that during the rapid change in wind direction, vessel can potentially be exposed to beam sea. Although this
does not correspond with the highest tension in mooring legs, it can lead to green water impact, which should be further investigated.

The disconnectable turret mooring system is the most common station keeping system used for FPSOs located in hurricane prone regions of the world
(e.g. NW Australia and SE China). Total operating experience for these FPSOs is over 100 years, with no major incidents. In addition to their use in
hurricane regions, disconnectable turret moored FPSOs have been used in iceberg prone regions such as eastern Canada. Of the 20 disconnectable
FPSOs installed worldwide, 6 have had external turrets and the remainder internal turrets. For deepwater areas such as the GoM, an internal turret will
be selected due to high vertical loads on the turret from the risers and mooring.

Station keeping capabilities

Once the preferred type of station keeping system has been determined the primary considerations for the station keeping design encompass the
following two aspects:
• Limiting the offsets to within the acceptable limits for the riser design.
• Withstanding the environmental condition for system survivability.

For the benefit of flexible riser designs the FPSO offset is normally limited to 6-8% of the water depth for intact survival condition and approximately
8-10% of the water depth for the broken line case.
For ultra deep water, greater than 1000m water depth, where there are size and pressure rating limits for flexible risers, a different riser solution will be
required, such as riser tower. It has been assumed during this study that the offset limitations for a single or multiple riser tower would be similar to

120
those values stated above for free hanging catenary risers.
Setting a limit on the environment depends on the site – specific conditions and on the operational philosophy of the operator.
For a disconnectable turret there are additional design issues to be considered. Since the main advantage of the disconnectable FPSO is its ability to
avoid the harshest weather conditions, the selection of the environmental conditions for disconnecting is key to optimizing the dezign. Reaching the
balance between minimizing the design criteria to reduce cost, and minimizing the outstation environmental design condition to increase the FPSO up
time.
• FPSO to remain connected during 100 years winter storm and 100 year loop current.
• FPSO to disconnect from mooring system as and when the predicted storm conditions are expected to exceed 100year return winter storm.
• The disconnected mooring and riser system to withstand 100 year hurricane environment (safety factors those applicable to an extreme
loading condition) and to survive the thousand year hurricane as a special check often required by some operator.

Mooring system selection

For deep water environments the mooring aspect has to be accounted in totality. That is the riser, offloading and mooring systems design cannot be
isolated from one another.
For the mooring line configuration, two options are currently available; a chain + wire combination or a chain +polyester combination. A study
performed by SBM found that the choice and cost effectiveness depend on the environment and water depth. For disconnectable systems, the riser
buoy buoyancy requirement could be a consideration , with the lighter polyester rope reducing the size of the buoy required.
For deepwater mooring, both taut and semi taut system could be considered. Mooring line combination can either combination of polyester and chain
or all steel (chain and wire rope). It is SBMs experience that for water depth greater than 100 meters and subjected to severe environmental conditions,
as in the case for GoM , it is economical to select a taut or semi taut log system employing chain and polyester. Some of the key findings of an internal
study performed by SBM for application of polyester mooring systems include:
• Considerable experience has been gained since the first installation of the polyester mooring lines in the Campos Basin in 1997. Since then it
has been employed in GoM also.
• Polyester benefits increase with water depth. In deeper waters to minimize the riser buoy buoyancy requirements, a steel system has to be
designed with spring buoys because of increased restoring capacity requirements and increased horizontal and vertical loads.
• The indirect cost benefits associated with polyester taut log system are saving due to reduced sub – sea footprint area and reduced vertical
loads for the floater.
• Higher flexibility with respect to sub-sea layout and more cost effective riser solutions.
• If general guidelines are followed, there is no reason to assume that it is more difficult or more expensive to install polyester system than a
steel system.
• In ultra deep water, the low suspended weight of polyester will make handling easier thus minimizing the installation cost.

Additional operational costs are expected for polyester systems. These include
• The requirement for monitoring the line tensions.
• Re-tensioning the system even though the latest findings are challenging the requirement for re-tensioning.
• Regular ROV inspection.
• Removal and testing of inserts etc.

Turret location

In order for a turret – moored vessel to weathervane, the location of the turret with respect to the vessel mid-ship is critical. The farther the turret
location from the vessel mid-ship the better the weathervaning capability. The downside is higher vessel motions which may impact the behavior of
riser system.
A sensitivity analysis is critical to arrive at an optimum location for the turret. The analysis should be performed using the selected tanker and the
mooring and riser configurations. The following points should be considered while deciding on the preferred turret location
• Vessel motions and weathervaning requirements.
• Riser dynamics.
• Mooring loads
• Storage volume requirements
• Vessel structural compatibility vis-a –vis location of transverse and longitudinal bulk heads.
• Location of accommodation
• Client preference

For the internal turret FPSO, depending on the turret diameter, the preferred turret location is either at the fore peak compartment or at the first cargo
tank aft of the collision bulkhead.

121
Turret arrangement
Two types of turret systems are feasible for deepwater development, namely an internal turret and an external turret. The merits and drawbacks of both
the systems are discussed below

External turret mooring system

Its main advantages are


• Single and the cheapest weathervaning mooring system.
• Ease to integrate into the vessel or barge structure.
• Easy to impact
• Easy hock-up of anchor legs and risers

And the disadvantages are


• Require a cantilever to avoid risk of anchor legs/hull interference.
• Bow trimming may be required for conversion. For new builds bow could be custom designed.
• Limited number of risers although SBM has detailed design for external turrets capable to accommodate vertical loads of more than 5000 t
for permanently moored system.For a disconnectable system. the number of risers would be limited due to higher rating loads.
• Higher motions owing to the distance between the turret axis and the vessel mid-ship.
• Higher pull loads during the rating of a disconnectable turret.

Internal turret mooring system

The main advantages are:


• Well protected
• Can accept large number of risers (75 risers for Petrobras P53 project)and large vertical loads for permanently moored system. For
disconnectable system the number of risers would be limited by the buoyancy restrictions of the disconnected buoy.
• No interference issues between anchor legs and the hull.
• Reduced vessel motions at top of riser.

But the disadvantages are:


• More complex hull integration.
• Submerged components.
• Loss of cargo space or volume.
• Higher cost.

For a disconnectable FPSO, even though both external and internal turret designs exist, for the deepwater GoM, an internal disconnectable system
would be preferred.

Turret bearing selection

The various options available include:


• Single piece 3 race roller bearing.
• Segmented 3 race roller bearing.
• Bogie wheel arrangement.

Both single piece and segmented roller bearing could be used for the turret arrangement. Though both types are proven to give satisfactory
results, segmented roller bearings are normally used only for very large diameter turrets (>8 m and <14 m) and are usually more complex in
design. Ensuring the alignment among various segments is a tedious task.They are also more sensitive to hull flexural motions than single
piece roller bearing.Therefore, for large diameter bearings(>8 m ), the preferred choice would be bogie wheel arrangement rather than
segmented bearing.

Anchor selection:

Suction piles, Vertically Loaded Anchors (VLA), Suction Embedded Plate Anchors (SEPLA), or torpedo anchors (Petrobars Patent) could
be considered for GoM. Knowing the water depth, soft soil conditions and the large loads, the installation of driven/grouted piles would be
expensive for deepwater applications. Similarly, considering the requirement for proof loading of the anchors, portable drag in the future,
deepwater, and the benefits of a turn monitoring system, drag anchors are not suitable for deepwater applications.

122
Riser design consideration in deep waters

Some of the special issues facing the GoM deepwater system design and selection include:

• Riser sizes.
Available riser sizes will be a limitation for deepwater application and particularly for flexible risers.
• Fluid (content) temperature and pressure.
A characterstic of deepwater wells in the GoM region is its high wellhead pressure.This in combination with the water depth will
have limitaion on the available flexible riser sizes.For steel risers, the penalty would be on wall thickness and thus on the weight.
• Loop current.
This is a phenomenon specific to the GoM region. Even though loop current may not be a major concern for flexible risers, it
would be a challenge for steel risers. Votex Induced Vibration (VIV) to be carefully considered in the design.
• Hurricane.
Beyond 1000 m water depth, flexible riser design and installation becomes complex and expensive. Cheaper and reliable
alternative could be steel catemaries or lazy wave risers. Steel risers are, however more prone to fatigue damage than its flexible
counter parts, especially for FPSO type of surface structure. For flexible risers due to the inherent damping present within the
structure, fatigue may not be a problem. Some of the riser design options available for the GoM include:
• Flexible risers
a)simple category
b)lazy wave configuration utilizing buoyancy modules.
• Steel risers
a)simple category
b)lazy wave configuration utilizing buoyancy modules.
• Single Line Offset Riser(SLOB)
• Hybrid Riser Tower(HRT)

Free hanging caternary risers(flexible and steel) could impose considerable loading on the turret and the requirement of disconnectable buoy sizes,
such an arrangement could influence the mooring system design and an integrated approach needs to be accounted for. If the risers cannot be placed
symmetrically with respect to the turret, an unsymmetrical mooring line configuration (by varying the number of lines on each side or by varying the
line orientation or by varying the line pre-tensions) needs to be considered.
SLOR and riser tower options utilize a combination of concentrated subsurface buoyancy and jumper hoses to connect to the FPSO. The steel riser
tower is isolated from the vessel motions, but interference between risers and mooring lines shall be evaluated. Safety installation considerations
(presence of sub surface buoyancy) could play a major consideration in the selection. The tower behavior in the disconnected mode subjected to
hurricane environment to be carefully accounted for in the design.
Each option has its own pros and cons. Final selection shall be based on field specific requirements, operational and installation constraints, over all
cost and operator preference.

Topside Design and Associated Gas Disposal:

There are no major issues associated with the design of the topsides facilities for an FPSO for the GoM, with the exception of the disposal of the
associated gas from the oil production. In many deepwater areas of the world where FPSO use is prevalent and there is not yet any gas export system,
injection of the gas is acceptable. In the GoM this may be permitted if necessary for reservoir pressure maintenance but not for gas disposal.
Therefore for fields in the GoM where gas production is expected to be minimal, and would not justify a deepwater pipeline, consideration has been
given to other gas export system (e.g. gas to liquid conversion).

Oil Import and Export Facilities:

A deepwater FPSO could act as a host platform for the export of crude and gas in additional to its basic function as an FPSO. Any import (either fuel
gas or crude oil) to the FPSO will be via risers. But for export, dual options are available, namely:
i)Tandem offloading via shuttle/off-take tankers
ii)Pipeline

The major advantages of tandem offloading include:


i)FPSO production is not affected by the offloading activities as the produced crude is stored in the vessel tanks
ii)Could be a reliable and cheaper alternative to pipeline export

For the pipeline option, the primary advantages are:


i)No requirement to temporary store the produced oil in the vessel cargo tanks

123
ii) Does not need to depend on off-take vessels

A commonly perceived concern for tandem offloading from a turret moored FPSO is the associated downtime. Industry experience worldwide has
been that unlike for a spread moored FPSO, with proper crude oil storage capacity in the FPSO and the planning of the shuttle tanker operations ,the
downtime due to weather is minimal. The primary reasons include:
i)Due to inherent weathervaning capability, the FPSO will always be aligned with the predominant force direction pertaining to the prevailing
environmental combination. Sideways (quartering, beam etc.) environmental conditions are therefore, not an issue.
ii)For tandem offloading, due to the reasons stated above , the approach for the off-take tanker will be against the prevailing environment, which
is the safe approach. Interference from the mooring lines or risers is limited.

CONCLUSION

The FPSO is a well proven technology for the development of remote deepwater fields, even though it has not been utilized yet in the US GoM.
The particular combination of pipeline infrastructure established regulatory procedures, HPHT reservoirs, and extreme Matocean conditions in the
USA GoM are unique in the world however, none of these issues on their own are unique. Experience has been gained elsewhere in the world with
FPSOs addressing all of these issues, and by applying this experience to field developments in the GoM, an FPSO can be successfully utilized there as
well.
A design which address the issues, combining proven technology from other projects around the world, is described. The proposed design would
be an Aframax sized tanker with a disconnectable, internal turret moored with a polyester rope system to suction piles. Produced oil would be
offloaded to a tandem moored shuttle tanker and associated gas would be exported via a gas pipeline.
The turret moored FPSO was selected over DP due to the requirement far a novel new technology to provide a rapid emergency disconnect in
the event of a DP drift off or drive off failure scenario, and a lower OPEX.
A disconnectable turret mooring was selected over a permanent turret mooring since it is an inherently safer and more robust design, avoiding
the hurricane event. The technology of disconnectable turrets is well established with twenty units currently operating, with over 100 years cumulative
operating experience in hurricane (typhoon/cyclones) regions of the world.

ACKNOWLEDGEMENT

The authors would like to thank the concerned faculty members of University of petroleum and energy studies under whose able guidance
and support this paper on DEEPWATER DRILLING PLATFORMS could be completed.

REFERENCES

Luo, Y. and Schuurmans, S. “FPSOs for deepwater Gulf of Mexico Applications”


Maclure, D. and Walters, D. “Freestanding Risers in the Gulf of Mexico-A Unique Solution for Challenging Field Development Configuration”
Michael Frank J.,”Offshore Drilling Technology”

124
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

NUMERICAL STUDIES ON DEEPWATER DRY TREE SEMISUBMESIBLE


R. Sundaravadivelu, N. Srinivasan and V. P. K. Kaundinya
Department of Ocean Engineering,
Indian Institute of Technology, Madras, Chennai 600036, India

GENERAL

The deep water offshore platforms (>300m) is expected to grow from 3% in 2003 to10% in 2020. The different types of floating deepwater platforms
used in the oil and gas industry are Tension leg platform (TLP), Semi-submersible, Spar and FPSO. The semi submersible is generally used for wet
tree operation and is not suitable for East Coast in the Bay of Bengal since the 100 year design waves are of 20m significant height and 14s period.
Dry tree support semi-submersible Platform (DTSP) is designed to overcome the critical disadvantages of both the SPAR and the Semi-submersibles
concepts. SPAR, with large deck payload, suffers from disadvantages in design, fabrication, transportation and installation that are highly cost-
controlling factors in the selection for a deepwater production application. SPAR installation cost is significant and with severe environment, the risers
are affected by fatigue type of failure. Hence, a dry transportation with self-erection capability is looked with a semi-submersible concept with
telescopic keel tank. Truss-pontoons are used to reduce the heave motion of the platform during severe storm. The new vessel technology is suitable
for dry-tree support production operation in deepwater in the east coast of India. It is also sized with extra deck pay loads where the drilling modules
can be accommodated and the drilling and production can be achieved from one platform. This reduces the cost of the entire field development in the
east coast of India in the deepwater. The platform can withstand 100 yrs return storm with acceptable heave, pitch and surge motions and highly stable
and reliable to operate with moderate storms. Self installation features and mobility and reuse of the platform in other locations are added advantages
of this platform concept. The risers, moorings and umbilical hanging loads are significant in a deepwater and could be supported by the extra
buoyancy in the keel tank and thus the deck pay load is not affected by the riser, mooring and umbilical pull down loads.

LITERATURE REVIEW

Halkyard et al (2002) studied the deep draft semisubmersible with a retractable heave plate. This paper has demonstrated the feasibility of the DPS
2001-4 for application in the deep waters of the GOM. The structure itself is a combination of a typical semisubmersible top hull and the lower
truss/heave plate used on truss spars. Global responses were calculated and were found to be superior to typical semisubmersibles.

Nagan Srinivasan et al (2006) studied the hydrodynamic behavior of the truss pontoon semisubmersible and separated flow damping. The
experimental results shown in this paper explained the hydrodynamic behaviors of the truss pontoon semisubmersible. John Murray et al (2007)
studied the hydrodynamics analysis of a conceptual dry tree semisubmersible for drilling and production platforms. The relative areas of the column
and pontoon were varied to demonstrate the global effects on the hydrodynamic forces acting in these structural components while the area of the
heave plate was kept constant.

DRY TREE TRUSS PONTOON SEMI SUBMERSIBLE.

DTSP is a column stabilized floating unit and the columns are connected by simple truss structure all around. The conventional shell pontoon is
replaced by the truss-pontoon. The Fig-1 shows the 3D diagram of DTSP with description of different structural units.
Fig 1 3D View of the DTSP

The purpose of the telescopic keel tanks is:

1. To provide required mass balance to bring the overall cg of the vessel down. To provide considerable amount of added mass and separated
flow damping and to bring down the heave natural frequency and thus the heave motion.

2. To support columns as a base to facilitate the fabrication process in the fabrication yard and to support the entire weight of the vessel with
required deck load and free float with its minimum draft such that load on/out to the barge and self installations are feasible.

3. To support predominant load of the riser pretension and mooring loads.

Table1. Details of DTSP

PARTICULAR DETAILS
Displacement (T) 80605

Draft (m) 79.9

Free Board (m) 15.47

Column Width (m) 16.764

Column Length (m) 19.815

Column Height (m) 53.35

Keel-tank Length (m) 85.4

Keel-tank Breadth (m) 85.4

Keel-tank Depth (m) 2.8

Column Center to Center (Longitudinal) (m) 64.081

Column Center to Center (Transverse) (m) 60.97


Deck Pay Load 17150 T

126 

 
NUMERICAL INVESTIGATION

The numerical investigation is carried out using WAMIT. The discretization of the structure is shown in Fig.2. The radiation damping obtained from
will not be realistic due to the large viscous damping associated with keel plate. Hence a physical model study is carried out and the heave damping of
9.05% and pitch & roll damping of 10.66 % obtained using the free oscillation study is used in the numerical study.

Fig.2. Discretization

The  numerical study is carried out for head sea (0deg), beam sea(90deg) and quarter sea(45deg) and the results are shown in Fig.3 to Fig.8. The
results indicate that the heave RAO is less than 0.1 for wave period up to 20s, and the maximum heave RAO of 1.6 is observed at the heave
natural period of about 30s.

 
 

Fig.3 Surge RAO

127 

 
Fig.4 Sway RAO

Fig.5 Heave RAO

128 

 
Fig.6 Roll RAO

Fig.7 Pitch RAO

129 

 
CONCLUSIONS

• From the model test results, the damping of the system is comparatively higher than the conventional semisubmersible. The damping values
are 9.05 % and 10.66 % with respect to critical damping in heave and pitch motions respectively, where as for semisubmersible the damping
is in the range of 2 to 3 %. The damping of the system is higher in the structure when it is subjected to wave excitation, since the amplitude
of the free oscillation test is less than the response of the DTSP subjected to wave excitation.
• The heave motion and heave motion of DTSP is within the acceptable limit of +/- 5m in operating condition to accommodate the dry-tree
facility in deck. Since, the DTSP structure has the high natural period than the storm wave period in deepwater region; it is suitable for
production in the extreme sea condition.
• It is possible to use the DTSP structure like TLP and SPAR for drilling, and production using dry-tree facility on the deck.

REFERENCES
Halkyard J., J. Chao, P. Abbott, CSO Aker; J. Dagleish, Ocean Energy; H. Banon, BP and K. Thiagarajan (2002), “A Deep Draft Semisubmersible
with a Retractable Heave Plate”, OTC -14304.
John Murray, Arcandra Tahar and Chan K. Yang (2007), “Hydrodynamics of Dry Tree Semisubmersibles”, ISOPE – 2007, JSC-491.
Nagan Srinivasan, Subrata Chakrabarti and R. Radha (2006) “Response analysis of a truss –pontoon semi- submersible with heaves plates”,Journal of
Offshore Mechanics and Arctic Engineering,128,100-107, OMAE2005-67522.

130 

 
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

A New Pipe Handling Approach for Deepwater Slim Drillship


Alex Kah Keong Tan1, Asbjorn Mortensen2, Sudhan Vincent3 and Abhijit D. Dimble4
Department of Drilling Equipment,
Keppel Offshore & Marine Technology Centre Pte Ltd, Singapore
E-mails: 1alex.tan@keppelom.com,2asbjorn.mortensen@keppelom.com,3sudhan.vincent@keppelom.com,4abhijit.dimble@keppelom.com

ABSTRACT

This paper outlines a proposal for a new pipe handling approach for a deepwater slim drillship optimized for re-entry to existing subsea wells. The
proposed slim drillship is designed with a new pipe handling solution to reduce downtime for pipe handling. This new pipe handling approach is also
tailored to facilitate various specialized well intervention operations which results in switch over time between different well intervention operations
being reduced significantly. The ship is designed to be able to drill wells up to 22500 ft below the drillfloor at water depth up to 7500 ft and intended
to provide drilling, well intervention and subsea operation service in the North Sea, Mexican Gulf, West African, Asian and Brazilian waters.

KEY WORDS: Pipe handling; slim drillship; well intervention; subsea operation.

INTRODUCTION

Driven by the increasing world demand for energy, the offshore oil drilling business, which is considered well established in shallow waters, is set for
significant development towards oil exploration and production in deeper waters. Oil drilling activities in water depth up to 550 ft can be carried out
by jack-ups and fixed platform rigs while drilling in deeper waters require floaters such as drillships or semi-submersibles. Deepwater offshore oil
production involves drilling subsea wells, which is an area of the oil business with rapid development of new technology. Subsea well intervention is
a necessary work operation in order to avoid declining output from a subsea well. The emerging oil demand has pushed oil companies to keep pace
and maintain or increase production levels from existing fields. According to consultants Infield Systems and Douglas Westwood, there will be more
than 5500 subsea wells by the end of 2010 and this number is still likely to accelerate in years to come. Most deepwater drill ships and semis are
made for exploration drilling, and needs a lot of setup time when changing between various intervention modes requiring going into a pressurized
well. Efficient deepwater well intervention vessels are therefore necessary with the purpose of reducing intervention time and enhance cost
effectiveness for well intervention.

The main challenge in the design of an intervention class vessel is to minimize downtime when arranging the intervention and drilling equipment. A
ship shaped hull is also considered important since intervention operations in general are less time consuming than drilling a whole well, so the vessel
must be able to move quickly between locations. Design of a drilling vessel starts with considering BOP handling in the moonpool area, in close
relation to planning tubular handling to an elevated drillfloor above the moonpool. The height difference which is necessary between the BOP
handling deck and the drillfloor drives how cranes are arranged and in general the weight distribution of equipment and structures. The height of
drillfloor and size on cranes drives the stability of the ship, requiring it to be longer and wider to avoid it to become unstable. The shape of a ship hull
forces tubular handling to occur along two longitudinal axes. By definition, an intervention ship should be smaller and more lightweight than a full
size drilling vessel, since its flexibility is less than the well established bigger drillships and semis. This means the financial risk of an intervention
vessel is higher for an owner, since currently the market is smaller for intervention than full bore deepwater well drilling. The selling factor for an
intervention ship is cost to perform an intervention operation, compared to the gain in oil production.

An intervention ship should by definition have access directly to the well filled with hydrocarbons through its subsea production tree and through the
inside of its production liner, without filling the well with mud. The slim drilling vessel should also by definition be able to drill to new targets inside
an existing well, meaning it needs a mud system as well as a marine riser for fluid return to the vessel. The mud system must be able to switch
effortlessly between oil based mud, water based mud, and clean production fluids, since time to clean tanks between intervention jobs shall be
minimized. For our vessel, we have chosen a nominal high pressure riser with internal diameter 11”, since most subsea wellheads has bore limitations
between 5” up to 9”, in order to drill through its Xmas tree to save time.

In this proposed slim drillship design, the prime technology driver is a new conceptual pipe handling approach. This enables the slim drillship to
reduce significantly the downtime for tubular handling during drilling operations via faster racking speed and offline stand building of tubulars. In
addition, the lower elevation of the setback and the corresponding stand building deck allows a more efficient horizontal pipe handling, compared to
catwalk decks on the same level as the drillfloor. The use of slanting catwalk chutes is not considered an option, due to its track record with accidents
and dropped objects, and less efficiency than horizontal feeding. The new pipe handling approach with dual horizontal catwalks from forward and aft
direction, is also aimed at handling third party well intervention equipment skids in order to reduce turnaround time between drilling and
intervention. The layout of the drillship is maximized by this new pipe handling solution to generate extended areas for third party well intervention
equipments and it offers the equipments direct access to the well center ensuring rapid turnover between various tools going into the well.

Fig 1: Overview of the slim drillship

The slim drillship uses a high pressure (HP) riser configuration as illustrated in Fig 2. The HP riser is connected to the subsea wellhead via a Seabed
BOP containing a duplicated Emergency Shear BOP, a hang-off RAM and a Lower Marine Riser Package (LMRP) for quick riser disconnection.
Further up near the vessel’s hull, a submerged Surface BOP is terminating the high pressure channel to the well, with annular preventer and pipe
rams, and choke and kill connection. The advantage of using HP riser is that it enables the use of standard intervention equipment for dry
intervention. Such equipment and procedures are commonly used on fixed platform rigs, meaning the coiled tubing (CT) BOP and wireline (WL)
BOP is located on the drillfloor.

132
Fig 2: HP riser configuration

PIPE HANDLING CONCEPT

The new pipe handling solution revolves around a new KOMtech design automated vertical pipe handling machine, the Rotaset Racker, housed on a
stand building deck which is a lowered deck at 4.5m below the drillfloor to carry out offline stand building via powered mouseholes. An A-shaped
box type telescopic mast is developed to complement the Rotaset Racker for tubular transfer to the well center. Forward open deck area on the
starboard side of the vessel and under deck section are set aside for tubular storage while the aft poop deck area is used for riser storage. Horizontal
tubular feeding machines for moving tubulars and risers from their storage location at forward and aft of the vessel respectively towards the drillfloor
are also integrated so that the complete pipe handling system can work seamlessly together.

Horizontal Pipe Handling


Horizontal tubular feeding is by means of a forward drill pipe catwalk machine for transportation of drill pipes/casings and an aft riser catwalk
machine for transportation of risers from storage location to and from the stand building deck and drillfloor respectively. The catwalk machine is also
optimized for the transportation of well intervention equipments which will be discussed in the next section.

Tubular Storage
Drill pipes and casings storage are located forward of the moonpool on the main deck starboard side of the vessel with combined capacity for 400
number of 9-5/8 casing joints total length 16800 ft or 1560 number of 3-1/2” slim size drill pipe joints total length 45200 ft. Table 1 shows the
capacity of the tubular storage for different tubular size. Additional storage area for drill pipes is allocated in the pipe hold below the main deck with

133
capacity to hold up to 780 number of 5” standard size drill pipe joints total length 22600 ft. This arrangement for tubular storage with pipe hold
below deck not only enhances tubular payload which makes the slim drillship less dependent on tubular supply vessels/infrastructure reducing
downtime, but it also maximizes the limited deck space on the slim drillship creating available large deck space on the forward port side of the vessel
for a laid down area for cargo as well as a ROV hangar by which ROVs are activated for subsea operations.

Table 1: Capacity of tubular storage


Tubular Storage Location Tubular Size No. of Joints Total Length
Main Deck 3 ½ ” Drill Pipes (5” TJ) 1560 45200 ft
2 x (W=4.5m x H=3 m) or 5” Drill pipes (7” TJ) 780 22600 ft
or 9-5/8” Casing 400 16800 ft
or 13-3/8” Casing 270 11200 ft
Pipe Hold 3 ½ ” Drill Pipes (5” TJ) 1560 45200 ft
(W=14 m x H=2.5 m) or 5” Drill pipes (7” TJ) 780 22600 ft

Pipe Craneage
The vessel comes with three major knuckle boom cranes and a gantry crane. A tubular handling overhead crane is also installed inside the pipe hold
for handling single tubular via two vertical scissor lift tables to be raised onto the main deck eliminating manual guiding through the pipe hatch by
operators done traditionally with crane hoisting. This improves speed and safety of the procedure. The tubulars are transferred to the drill pipe
catwalk machine by a knuckle crane with bundle pipe gripper for fast feeding of tubular to the drillfloor. This crane is located forward end of the
catwalk machine on top of the ROV hangar so as to also access tubulars from the pipe hold hatch. A gantry crane installed over the riser storage area
is designated to handle the riser onto the riser catwalk machine at the aft of the moonpool. Two other knuckle boom pedestal type deck cranes are
also installed at mid ship and ship aft for transferring drill pipes/casings and risers respectively to/from quay or tubular supply vessel.

Fig 3: Tubular storage on open deck starboard side with tubular craneage and horizontal tubular feeding machine

Horizontal to Vertical Pipe Handling


A V-door arm, referred to as the V-door machine, picks up single tubulars from the forward catwalk and raises them to vertical orientation, in order
to put them into one of two vertical mouseholes. The catwalk machine/ conveyor has a tail-in arm to assist the lower end of the tubular, in order to
put it safely into the mousehole. For further sequence to build stands, see following section about this.

Vertical Pipe Handling Machine


The vertical pipe handling machine, the Rotaset Racker, is a new racking machine design which claims to be the world’s fastest racking system. It
achieves a racking speed of about 50-60 stands an hour as given by the thread make- and- break cycle time and the drawworks hoisting speed,

134
meeting the industry’s expectation and overcoming the shortcoming of most vertical racking machine able to rack about 30-40 stands per hour. The
Rotaset Racker is a fully automated vertical racking system with two rotating setbacks and a separate two- arm column racker. Its tripping sequence
is divided into 2 parallel activities to improve racking speed: One activity for internal handling inside setback and to drillfloor, and a second activity
to handle the stand from setback vicinity to its destination. The setback area is located on the lowered stand building deck. Its storage capacity
includes sections with belly board for slim pipes (3 ½”), sections for double casing stand building, as well as normal triple stands, heavyweights and
collars. Casing stand building is performed in parallel with running casing doubles into the well, in order to reduce the size of the setback. This not
only allows the slim drillship to gain an advantage with a lower Center of Gravity (COG) and hence improves stability, but also maximizes the
available space on the drillfloor. The lower elevation of the stand building deck not only permits the forward knuckle boom crane to reach the main
deck tubular storage with ease, but also improves safety and workability during offline stand building due to its relative far distance from the well
center (compared to traditional mono derricks with offline stand building).

The fingerboards have a double barrier giving it redundancy to prevent stands from falling out. This cylindrical and rotating fingerboard is placed
inside a stationary ring to prevent the stand escaping from the fingerboard by accident. An open slot which is normally covered by another rotating
ring allows the column racker to pick the outermost stand.

Fig 4: View of Rotaset Racker on the stand building deck Fig 5: Setback area

Offline Stand Building


The slim drillship has a separate platform for offline stand building which is a prime contributor to drilling efficiency. Offline stand building is
carried out on the stand building deck which has two powered mouseholes between the two rotating setbacks of the Rotaset Racker. The V-door
machine transits tubular into the first mousehole where a hydraulic iron roughneck or a casing tong connects the two tubulars. As the mouseholes do
not penetrate the moonpool, offline stand building is possible at the same time as HP riser is established in the moonpool. When the triple stand is
finished, the V- door machine and a dedicated tail- in arm brings the stand to the drillfloor. The column racker transfers the completed stand to the
setbacks or directly to the well. A dedicated storage finger is catered for each casing diameter and for each setback, so the casing string is build to
doubles on-the-run, i.e. in parallel, while run into the well from the other setback. This arrangement enables offline stand building of tubulars without
interfering with drillfloor operations allowing different operations to be carried out simultaneously and thus increase the efficiency of the slim
drillship.

135
A-Shaped Mast
A box type mast with an A-shape base is designed to integrate with the pipe handling system in order to support the top drive and at the same time to
handle stands seamlessly from the rotating setbacks at the stand building deck to the well center on the drillfloor. A main feature is that its crown
block extends outside the mast structure, enabling crane access to the well center. The mast is constructed with internally reinforced steel plate sheets
in order to take down the integrated torque resulting from the protruding crown block. Compared to a truss work derrick, this compact box type plate
mast occupies minimal footprint on the drillfloor. The well center line can be accessed through an opening from behind the mast by the racking arms
for tripping-in or tripping-out when the dolly system supporting the drilling torque from the top drive is in extended position.

The telescopic design of the mast reduces the COG of the ship when it is retracted improving vessel stability. Fig 5 shows the telescopic design of the
mast to allow the ship to lower the crown block and sail through the Bosporus Straits and the Panama Canal. The telescopic collapse movement is
controlled by four cylinder based winches and this design enables the ship to move fast between arctic locations and tropical waters without passing
south of South America.

73m above sea level

56.7m above sea level

Fig 6: Box type telescopic A-shaped plate mast

The new pipe handling approach utilizes the limited space available on the slim drillship to full potential. This concept allows for efficient, fast and
safe pipe handling for drilling operations on a small vessel. In addition, this pipe handling solution on the slim drillship from tubular storage
arrangement, pipe handling machines to box type plate mast design is developed in mind for facilitating subsea well intervention operations.

136
OPTIMIZING SUBSEA WELL INTERVENTION WITH PIPE HANDLING SOLUTION

Subsea well intervention and slim drilling has become increasingly important in the last decade for improving the hydrocarbons recovery in terms of
production flow and reservoir drainage percentage in aging wells. Traditionally, subsea well intervention used to be expensive and uneconomical due
to time consuming setup for intervention and the use of costly deepwater drilling units. In recent years, these operations have been made cheaper with
new technology such as the use of HP intervention riser to improve access to subsea wellheads. The challenge in the next decade will be in
developing and meeting the demand for more efficient subsea well intervention systems.

The slim drillship is designed for re-entry into existing subsea wells to perform subsea well intervention operations through a HP riser string between
the ship and the wellhead. The efficiency of the subsea well intervention operations performed by the slim drillship is enhanced with the new pipe
handling approach which eases the workflow for subsea well intervention operations reducing turnaround timings between various operations. The
fast mode change between various services from drilling with jointed pipe to subsea well intervention operations and between different intervention
operations will optimize the total operational capability of the slim drillship. This allows for a completely different as well as larger operational
window compared to traditional drilling and well intervention units.

The pipe handling approach of the slim drillship facilitates subsea well intervention operations by integrating the handling of the third party well
intervention equipments as well as maximizing the layout of the ship for accommodating these equipments for fast mode change.

Optimizing Layout
The streamline pipe handling solution creates an extended drillfloor outside the wind wall to accommodate third party well intervention equipments
such as coiled tubing injector head, wireline control unit, wireline reels, wireline frequency drives and a wireline well control unit for storage of
wireline equipments and tools. The close vicinity of these equipments and their reserved positions eliminate the change over time between operations
which is a burden on time for standard drilling units. The tubular storage arrangement allows for available deck space to be also created on the poop
deck (Fig 7) just below aft of the extended drillfloor for coiled tubing reels with spooler unit, coiled tubing power pack and control unit, and other
coiled tubing tools and equipments.

Fig 7: Drillfloor and nearby main deck, stand building deck and poop deck layout

The proposed layout is incorporated for coiled tubing, wireline as well as chemical stimulation directly to the wellhead as they are standard in many
drilling and well intervention programs today. The coiled tubing and wireline equipments can be integrated permanently on the assigned location on
the slim drillship to minimize the handling of such equipments if they are temporary and thus enabling the vessel to be “ready for operation” almost
immediately upon arriving at offshore location eliminating any downtime which is common for traditional drilling/intervention units. However,
depending on the chartering contract on operational requirements of the vessel, the coiled tubing injector head and wireline equipments on the open
area of the extended drillfloor can be replaced with other third party equipment for other subsea services if the need arises. The open area on the
extended drillfloor is multi-purpose by design as it is devised with a direct access to well center line.

With the compact box type plate mast replacing the traditional truss work type derrick on the slim drillship, the layout is not only enhanced by the
availability of more drillfloor space, but large equipments can also be skidded or hoisted by cranes directly to the well center as there is no lattice

137
type derrick structure around the mast hoisting system and thus there is no V-door limitation for access.

Integrated Skidding System on Extended Drillfloor


A skidding system is integrated with the riser catwalk rails which double up as skid rails. The riser catwalk rails are customized to allow big frames
housing well intervention equipments to be skidded by push/pull skid tractors. Skidding beams are also integrated transversely across the extended
drillfloor (refer to Fig 8) for transverse skidding. A skid tractor with integrated control stand and power unit is mechanically attached to the frame
which acts as a skid base. The skid tractor employs hydraulically activated clamp/wedge lock to secure one end to the skidding beams and
extending/retracting hydraulic cylinders to create push/pull motion at the other end. The hydraulic cylinders and clamp/wedge lock work in tandem to
allow the skid tractor to travel along the skidding beam pushing/pulling the skid base in the same direction along the skidding beams. At the
intersection point between the riser catwalk rails and the skidding beams, slots are created on the catwalk rails and an external jacking system is
implemented. The jacking system is designated to lift the skid base onto the higher extruded riser catwalk rails from the skidding beams and vice
versa. Another skid tractor which travels in the longitudinal direction along the riser catwalk rails will take over the task of pushing/pulling the skid
base towards/from the well center along the longitudinal direction.

Optimizing Coiled Tubing Operations


Coiled Tubing is a repair and maintenance application involving running a continuous tubing which is rolled onto a reel down to the production zone
through an injector head which pushes the coil through a coiled tubing BOP into the well against the well pressure. Coiled tubing offers several
advantages over other well intervention methods. This method allows for pumping of liquid regardless of position and direction of travel for
performing mechanical operations downhole. Furthermore, with the installation of an electrical conductor or hydraulic conduit, complex intervention
techniques can also be performed safely.

The new pipe handling approach optimizes the switch over process from/to drilling or other subsea well interventions to/from coiled tubing
operation. The coiled tubing injector head, together with the pressure control equipments including the coiled tubing BOP, is housed within a guide
frame which doubles up as a skid base for safe skidding from/to its assigned parking position on the starboard side of the extended drillfloor to/from
the well center. With the horizontal pipe handling solution optimized to handle skids for coiled tubing operations, traditional critical lifting and
handling of the coiled tubing injector head with cranes or winches and manriding can be eliminated. This enhances and safety and speed of the switch
over process to/from coiled tubing operations.

Fig 8: Skidding direction on the extended drillfloor

138
Optimizing Wireline Operations

Wireline operation is carried out using single-strand (slickline) or multistrand (braided line) wire or cable for intervention activities in the well.
Typical monitoring includes running a probe through the pressurized well at the end of a wire to measure properties of the well surroundings, such as
oil, water and gas content. Another class of wireline applications involves running tools into the well, energized by wire pull or electric power
through an e-line. Such wireline intervention operations are performed when the well is pressurized by hydrocarbons, as opposed to stabilized with
mud when drilling ahead.

Similar to coiled tubing operations, the new pipe handling approach optimizes the switch over process from/to wireline operations reducing turn over
downtime. Wireline well control equipments such as BOP, grease injector head, auxiliary/logging tools, fishing tools and radioactive/explosive
materials are contained within a base frame which will be skidded transversely on the skidding beams from/to the assigned parking position at the
starboard side of the vessel and longitudinally on the riser catwalk rails to/from the well center via mechanically attached skid tractors.

With the skidding system integrated to the horizontal riser handling system, complex lifting and handling of wireline well control equipments to/from
the well center from/to the assigned parking position on the extended drillfloor can be eliminated to improve operational safety and speed for the
switch over process to/from wireline operations. Furthermore, the Rotaset Racker can be used as storage for and directly handle wireline lubricators,
which have pipe resembling ends, to the well center via the racking arms doing away with manual handling of wireline lubricators which otherwise
need to be laid down on main deck storage area.

Optimizing Plugging & Abandonment

Plugging and abandonment of subsea well is performed by setting a series of cement plugs in the wellbore with integrity or inflow test conducted at
each plugging stage to confirm the isolation of hydrocarbon bearing formation. This subsea operation is normally carried out when logs conclude that
there is inadequate hydrocarbon potential to complete the well or after production operations have exhausted the aging reservoir of hydrocarbons.

Special equipments which are required for plugging operation like cementing heads, and pulling of casing sections with cutter tools and hydraulic
hammers, can be stored in the pipe stand for easy access to the well center during mode change to plugging and abandonment. These tools are bulky,
but can be mounted with drill pipe extensions to fit their height to become standard triple stand height. The special racking fingers on the drillfloor
can be used to store such bulky tools, avoiding time consuming operations to lay them down.

Optimizing Other Subsea & Drilling Operations

Bottom Hole Assembly (BHA): BHA is the lower portion of the drill string which must provide force for the drill bit to break the rock and provide the
driller with directional control of the well. The assembly consists of several specialized devices with complex electronics for measurements and
logging while drilling, as well as stabilizers and reaming tools with larger diameters. The drillfloor special setback area has special racking fingers for
storage of these bulky and heavy collars.

Managed Pressure Drilling (MPD): Managed Pressure Drilling requires the use of a HP riser to put the mud return under pressure in order to enable
the use of lightweight mud. The mud back pressure may easily be changed as demanded by pressure losses during circulation or tripping operation,
or varying pore pressure. The HP slick joint above the surface BOP has a Rotating Control Head (RCH) for MPD connection to a MPD manifold in
the moonpool area as well as a quick connector to establish a HP line (blafro telescopic/ locked joint) to CT and WL BOPs on the drillfloor. Under
the Surface BOP annular preventer, a backup MPD connection exists to be used when the pumps are off-line during connections and tripping
operations when the drill string is stripping through the RCH. If requested by the drilling operator or oil company, the RCH joint may also located on
the drillfloor, for ease of access for maintenance of critical seals. A surface mounted RCH comes with a skid to be installed in the rotary, in order to
give access for a trackless iron roughneck in high elevation. The distance between the drillfloor installed RCH and surface BOP gives ample room to
the entire BHA, when stripping out of the well during tripping operation.

It is important when drilling MPD to measure the balance between incoming and outgoing mud in order to detect kicks early and to adjust any mud
imbalance quickly. It is also equally important to remove gas from the mud return which is exposed in a downhole atmosphere containing
hydrocarbons. The slim drillship is proposed with a mud cleaning system from upstart company Cubility which employs vacuum technology to
remove gas and a rotating screen in the shaker to separate cuttings. The advantage of this setup is also less height for the mud return system which
fits a slim drillship design, as a separate pump level is not necessary under the settling tanks. The mud return flow is calculated from the RPM of
fixed displacement return pumps, so any kicks resulting from gas drawn into the mud flow can be detected very early and accurately which is an
advantage for the drilling monitoring system. The mud back pressure can also be controlled in real time to prevent the mud from damaging the wall
properties in the production zone of the well.

139
Fig 9: Mud cleaning system (with integrated degasser) layout under the drillfloor

CONCLUSION

The offshore oil drilling and production business is experiencing a rapid technological and market advancement. New techniques for improved
recovery rates from established oil wells, including drilling horizontal sections to new targets inside the reservoir, requires new surface equipment on
the drilling vessel. It is therefore an increasing demand for vessels which are efficient and capable of performing various drilling and intervention
operations as well as carrying out the different operations simultaneously. The slim drillship is designed to perform wireline operations, coil tubing
operations, through tubing slim drilling with jointed drill pipe and managed pressure drilling in water depths of 7500ft and wells depths of up to
22500ft below the drillfloor within a HP riser. At 145m long and 30.2m beam, the slim drillship is a fraction of the size of most high performance
deepwater drillships currently available in the market which translates to a lower construction and operating cost for the drilling operators. A new
pipe handling approach was necessary for the vessel of such class and it is subsequently developed to integrate with the operational capabilities of the
slim drillship. The optimization of the new pipe handling approach to facilitate drilling and subsea intervention operations including optimizing the
layout for permanent integration of third party well intervention equipments and implementation of drillfloor skidding systems ensures that switching
between various operations is simple, fast and safe without any major and critical procedures required. Preparation for mode change including
clearing of deck should therefore be managed within 1 to 3 hours compared to a standard semi-submersible which might require downtime of up to 3
or 4 days. Looking ahead, oil becomes harder to find, meaning increase in number of subsea wellheads and requirement in improved oil recovery
rates. This translates to increased demand for such specialized and more efficient subsea intervention vessels.

REFERENCES

G. Humprey, “A Slim hole Drilling & Intervention Drillship utilising Expandable and MPD/UBD technologies”, IADC International Deepwater
Drilling Conference & Exhibition, Rio de Janeiro Brazil, March 2010.
A. Mortensen, Y.J. King, “Rotaset Racker Vertical Pipe Handling Machine”, KOMtech Technology Review (2009), pp 46-52.
A. Mortensen, Alex Tan, Ricky Ng, Abhijit Dimble, Sudhan Vincent, Tom Ostvold, “A Conceptual Development of a Slim Drill Ship”, IADC/SPE
Asia Pacific Drilling Technology Conference and Exhibition, Ho Chi Minh City, Vietnam, 1–3 November 2010.

140
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Treated Fly Ash: A Drilling Fluid


Chandan Guria, Akhilendra Kumar Pathak and Tarkeswar Kumar*
Department of Petroleum Engineering
Indian School of Mines, Dhanbad 826 004

ABSTRACT

Acceptability of fly ash (collected from National Thermal Power Corporation, Noida, India: Field No. 08) as a drilling fluid has been studied. A float
and sink test has been carried out for classification of fly ash to reduce its polydisperseity. Depending on the specific gravity of solution used in float
and sink test, different fractions of fly ash are obtained with specific particle sizes. Laser beam particle size analyzer is used to determine the particle
size and its distribution. Rheological parameters, mainly, apparent viscosity, plastic viscosity, Bingham yield, and gel strength of aqueous suspension
of treated fly ash have determined using a standard Fann VG meter (Model 35). Based on Newtonian behaviour of aqueous suspension of treated fly
ash, shear rate is calculated. Effects of particle size, particle size distribution and solid loading on rheology have been studied. Results of the treated
fly ash suspensions have also been compared with the standard aqueous bentonite suspensions.

KEW WORDS: Fly ash, Float and sink test, Drilling fluid, Particle size, Particle size distribution, Solid loading

INTRODUCTION

Shape, hardness, Fly ash is a pozzolanic material and it is an admixture of several mineral oxides. Coal fired thermal power plants produces huge
amount of fly ash and generates world’s largest quantity of industrial solid wastes (Giere et al., 2003), which creates severe waste disposal problem
and recognized as an environmental pollutant. Because of environmental problems posed by fly ash, a good deal of work on fly ash has been under
taken worldwide. Utilization of fly ash as a resource material has been studied extensively in many areas such as extraction of valuable minerals, water
pollution control, production of ceramic products, agriculture, building materials, paint and plastic industries. Usually, the utilization of fly ash is
determined by their physical and chemical properties which include fineness, particle size, particle size distribution, specific surface area, particle
freeze-thaw resistance, and activity in aqueous suspension etc. Many investigators have been carried out towards effective utilization of fly ash with
understanding of potential environmental pollution and health impacts associated with the disposal of fly ash by land filling. Details of the utilization
of fly ash in the above mentioned areas have been reviewed extensively by Ahmaruzzaman (2010). In addition to the above mentioned
applications, fly ash also finds a potential application as a foamable drilling fluid for deep water offshore well operations (Totten, et al., 1997), and
stabilization of drilling fluid wastes to avoid ground water pollution (Deeley and Laguros, 1987; Thompson, 1994). About 53.0 % of India’s energy
supply is based on coal and shall be so for few the next few decades (Friedman and Schaffer, 2009). There are about 82 coal fired thermal power
stations producing approximately 110 million tonnes of fly ash per annum in the Country (Kumar et al., 2007). Till date, nearly 38.0 % of fly ash has
been successfully utilized in the Country at preset (which is reported by National Thermal power Corporation, India), in various fields including land
filling, cement making, concrete products making particularly bricks, blocks and tiles, road making, and mines backfilling. In the present study,
acceptability of fly ash as a drilling fluid has been studied. For this fly ash has been classified into various size fractions based on specific gravity of
fly ash. A float and sink tests have conducted to classify the fly ash samples. Particle size and its distribution are analysed and rheological studies have
been carried out using the aqueous suspension of classified fly ash samples. Standard Fann VG meter is used to obtain the shear stress from dial
reading in degrees of the bob neglecting the small end effect at the bottom and shear from rotational velocities in rpm of the rotor after Newtonian
fluid approximation. Effects of particle size, particle size distribution and solid loading on rheology are also carried out and the results are also
compared with aqueous suspension of bentonite.

EXPERIMENT

Materials
Raw fly ash was collected from National Thermal Power Corporation, Noida, India, (Field No. 08) and specific gravity, particle size range, and total
evaporable moisture content were determined as 2.18, 10 - 100 µm, 1.55 % respectively. Loss on ignition was measured by burning the sample in
muffle furnace at 8000C for 3 hours and found to be 3.0 %. The chemical composition of fly ash as oxides in weight percent was SiO2: 59.23, Al2O3:
32.16, Fe2O3: 5.03, CaO: 0.9; and TiO2: 2.59. Zinc chloride (specific gravity: 2.91, Assay: 99.9 %) supplied by SD Fine-Chem Ltd., Mumbai, India
was used to prepare aqueous solution for float and sink test. Bentonite (Montmorillonite) powder supplied by Sigma-Aldrich (Kolkata) was used as
reference sample for rheological analysis. The typical specification of bentonite powder used for this studies are: loss on drying – 3.0 % and pH
– 10.0 (4.0 g in 200 ml water). Underground raw water with total hardness (as calcium carbonate): 680 mg/l and total dissolved solids: 955 mg/l
without pre-treatment was directly used for preparation of aqueous suspension of bentonite, virgin fly ash and treated fly ash.

METHOD

Classification of fly ash

Float and sink tests were carried out for classification of fly ash with different particle sizes. Raw fly ash were collected from power plant site and
sieved through 200 mesh to reject naturally occurring fines. Sieved fly ash sample has a size range of 10 to 100 µm with average specific gravity of
2.18 were subjected to float-sink tests with three specific gravity solutions for classification of fly ash with desired specific gravity. Silica being a high
specific gravity material, float and sink test will help to reduce undesirable silica content in the fly ash and is very helpful for drilling fluid application
to reduce corrosion of kelley, drill pipe, drill collar, drill bit, mud pump, and etc., due to abrasion. Aqueous zinc chloride solution is used a solution
medium for separation of fly ash. Aqueous zinc chloride solutions of 1.5, 1.8 and 2.0 specific gravity were prepared to carry out float and sink tests.
To prepared 1.5, 1.8 and 2.0 specific gravity aqueous zinc chloride solution, 188.0 g, 265.0 g and 358.0 g of solid zinc chloride solution were added in
100 ml water respectively. Time required for settling the sink depends on the specific gravity of the aqueous solution and it is found to be 2 – 12
minutes within the range of specific gravity studied. Amount of float separated from the solution also depends on the solution specific gravity and it
was found to be 0.28 g of dry float/g of virgin dry fly ash, 0.35 g of dry float/g of virgin dry fly ash and 0.40 g of dry float/g of virgin dry fly ash for
1.5, 1.8 and 2.0 specific gravity zinc chloride solutions respectively. Collected float was mixed with distilled water and separated using centrifuge. The
process was repeated for several times to contaminated remove zinc chloride and finally float was washed with methanol to remove the residual zinc
chloride. Washed float was dried in oven at 100 oC for 48 hours and dried fly ash was then used for rheological studies.

Rheology

For rheological studies (mainly apparent viscosity, plastic viscosity, Bingham yield and gel strength), aqueous suspension of treated fly ash with
different specific gravity and bentonite was placed in a rotational coaxial-cylindrical type Fann VG meter (Model 35). The rotor of the meter is driven
by an electric motor with six standard rotational speeds (600, 300, 200, 100, 6 and 3 rpm). These rotational speeds are measured in rpm are the
measure of shear rate whereas corresponding dial readings in degree measure the applied shear stress. Details of the Fann VG meter and its operations
are given in the standard text (Bourgoyne et al., 1986) and are not discussed here.

Analysis

Particle size distribution of treated fly ash

Virgin fly ash and treated fly ash with different specific gravity zinc chloride solution (i.e., 1.5, 1.8 and 2.0) are analysed for particle size and particle
size distribution (which includes differential distribution, oversize cumulative distribution and undersize cumulative distribution). The
size analysis is carried out using a laser beam particle size analyzer (Mastersizer S Ver. 2.19, Malvern Instruments Ltd., Malvern, U.K.). In this
method diffracted light produced when a laser beam passes through a dispersion of particles in liquid. The angle of diffraction increases as particle
size decreases, so that this method is particularly good for measuring sizes between 1000 to 0.001 µm. Following system details are used for analysis:
range lens - 300 mm, beam length - 2.40 mm, presentation - 3RHD, particle refractive index - (1.9285, 0.1000), dispersant refractive index - 1.3300,
analysis model – polydisperse, and distribution type – volume, particle specific gravity – 2.2. Particle size analysis is obtained in the form of
differential, cumulative undersize and cumulative oversize distribution. Details of the analysis were reported in Table 1.

Raw water analysis

Underground local raw water was used for rheological studies of bentonite, virgin fly ash and treated fly ash suspension. A typical analysis of the
underground raw water is as follows: Total hardness (as Calcium carbonate) – 680 mg/l, total dissolved solids – 955 mg/l, Iron (as Fe) - 0.5 mg/l,
chloride (as Cl) – 100 mg/l and others (Nitrate, fluoride, mercury, arsenic, lead and pesticides) – nil.

Equations for shear stress and shear rate estimation


Common laboratory Fann VG meter is a concentric cylinder rotational viscometer with six rotational speeds (Model 35) and is commonly used to
measure the rheological properties of drilling fluid which includes plastic viscosity, apparent viscosity, yield stress, initial gel strength and 10 minutes
gel strength. In this viscometer, bob (inner cylinder) is stationary and rotor sleeve or cup (outer cylinder) rotates at a constant speed. The dimension of
the rotor sleeve and the bob are specified in API RP 13B. In the region of annulus, the fluid moves in a circulatory pattern. Due to the presence of
gravitational and centrifugal force in the rotational viscometer, pressure (p) depends on z (height) and r (radius). Reasonable assumptions for the
velocity and pressure are: Under these assumptions, all the terms in the equation of

142 
 
continuity are zero (Bird et al., 2002) and the generalized component equation of motion in cylindrical coordinate system in terms of τ simplifies to the
following equations (Bird et al., 2002)
r-component:

(a)
θ-component:

(b)
z-component:

(c) (1)
where is the θ-component momentum flux (shear stress) in the positive direction of r and ρ = density of the working fluid.
Equation (1a) tells how the centrifugal force affects pressure and shear stress distribution in the annulus of the rotational viscometer. Equation (1b)
gives indirect information about the velocity distribution without assuming any type of fluid flow. Equation (1c) gives the effect of gravity (g) on
pressure (hydrostatic effect). In the present study, generalized θ-component equation of motion is considered for analysis of fluid flow in the annulus.
Now, θ-component equation of motion is integrated with respect to r giving
(2)
On differentiating the above equation is rearranged into the following convenient form:

(3)
Above differential equation demonstrates the variation of shear stress with radius of the concentric rotational viscometer. In a rotational viscometer,
fluid shear rate is function of radius and fluid velocity. At a given radius (r), fluid velocity (v) is related to the angular velocity (ω) by
(4)
Thus the change in velocity with radius (i.e., shear rate, ) is given by

(5)
First term in the right hand side of Eq. (5) describes the fluid layers are not slipping past one another but moving as solid plug (i.e., no slip condition is
applicable to the fluid layer which is immediate to the rotor sleeve) whereas the second term represents the shear rate due to slippage between two
successive layers. Considering only slippage between two successive layers and combining Eqs. (3) and (5) following equation is obtained

(6)
At steady state, when the rotor sleeve is rotating at constant angular velocity ω2 and the bob held at motionless (i.e., ω1 = 0), the torque applied by the
torsion spring at the bob must be equal but opposite in direction to the torque applied to the rotor sleeve. Integrating Eq. (6) within corresponding
limits, the following equation is obtained

(7)
Equation (2) is applicable to any point in the annulus and written in the following form

(8)
Now substituting upper limit of the integral of Eq. (7), i.e., in terms of , one may obtain the following integral with modified limits i.e.,

(9)
Float and sink test is carried out using virgin fly ash to classify the particles into small particle size and narrow
particle Above equation is quite general and applicable for Newtonian fluid, i.e., where µ is fluid viscosity and simplifying Eq. (9), shear
rate at the inner wall is obtained as

(10)
Above equation is convenient to express in terms of Fann VG meter rpm reading and is given by

143 
 
(11)
and substituting appropriate dimensions of bob and rotor [i.e., bob radius (r1) = 17.245 mm and rotor radius (r2) = 18.415 mm], one may obtain
. Shear stress at the inner wall neglecting end effect at the bottom of the bob is given by

(12)
Substituting appropriate dimensions of the Fann VG meter (i.e., bob radius = r1 = 17.245 mm, rotor radius = 18.415 mm, height of immersed cylinder
= h = 38.0 mm and spring constant = k = 3.87×10-5 N.m/scale unit), one can also obtain the wall shear stress at the bob wall in SI unit (i.e., Pascal) and
is given by , where θ = Fann VG meter dial reading. Therefore, shear rate and stress can easily be calculated using Newtonian fluid
approximation. Similarly, apparent viscosity, cp (µaF), plastic viscosity, cp (µp) and Bingham yield point, N/m2 (Yb) were calculated from the
following relations respectively (Monicard, 1982):

(13)

Results and discussion

Size distribution. This test is conducted using aqueous zinc chloride suspension to reduce polydispersity of the fly ash samples. After float and sink
test using different specific gravity aqueous suspension (i.e., 1.5, 1.8 and 2.0), float samples are obtained and analysed for particle size and particle
size distribution. Particle size distribution (differential, cumulative undersize and cumulative oversize) for virgin fly ash and treated fly ash samples
(Specific gravity: 1.6, 1.8 and 2.0) are shown in Figs. 1a-1d and the details of the distribution statistics for these fly ash samples are also given in
Table 1. It is observed that the least specific gravity fly ash float (i.e., specific gravity with 1.5) gives lowest particle size with lowest span value,
which indicates the narrow distribution with reduced polydispersity.

In order to use these treated fly ash samples as a drilling fluid, several rheological tests have been performed with the aqueous fly ash suspension using
a standard Fann VG meter (Model 35) and all the rheological results are compared with aqueous bentonite suspensions with different solid loading.
Dial reading of bob (in degree) and rotational speed of rotor (in rpm) for all the treated fly ash samples and bentonite suspensions with different
particle size and solid loading are noted and shown in Figs. 2a – 2d. Apparent viscosity is an indication of total pressure required to flow a fluid
whereas yield point and plastic viscosity represents the force require to initiate the flow of fluid and additional force required to flow the suspension
respectively. Numerical values of these rheological parameters give an idea about the behaviour of the bentonite and treated fly ash suspensions under
drilling condition. Corresponding to Figs. 2a – 2d, apparent viscosity, cp (µaF), plastic viscosity, cp (µp) and Bingham yield point, N/m2 (Yb) are
calculated using Eq. (13) and the results for all solid loadings are reported in Table 2. It is noted that Eq. (13) is based on Newtonian fluid flow
approximation. It is noted that rheological behaviour of bentonite suspension and treated fly ash suspension with lower specific gravity (i.e., specific
gravity: 1.5) gives a comparable rheological properties. It is also observed that Bingham yield point in all cases are very high as compare to plastic
viscosity, which indicates that shear rate of thinning will predominate at higher shear rate. This property is desirable drilling which will help to clean
the bit efficiently. The higher yield point/plastic viscosity ratio also indicates the shear thinning mud which is desirable for drilling fluid as it sets to a
gel and helps to suspend the cuttings when circulation is stopped. It is also evident that the gel will also break quickly to a thin fluid when it is agitated
(Gray and Darley, 1981).

Initial gel strength (i.e., gel at 10 seconds) and final gel strength (i.e., 10 minutes gel) for all the suspensions are also reported in Table 2 and
comparable results are obtained for bentonite and treated fly ash suspensions. Similar gel strength for all the samples indicates the comparable solid
holding capacity of the bentonite as well as treated fly ash suspensions. It is also noted that the aqueous suspension of bentonite and fly ash exhibit low
gel strength with the increase in solid content in the suspension, which indicates the less clay content is very much effective for drilling fluid.
Optimum clay loading is also very low (usually 3 % weight by volume) for most of the oil fields (Park et al., 1960; Elward-Berry and Darby, 1997).
Maintaining a constant solid loading in the treated fly suspension, reducing the particle size of the solid phase leads to an increase in the number of
particles in the system. Total surface area of the particles is also increased with increased number of particles. Therefore, for a given solid loading,
viscosity of the suspension will increases with decrease in particle size. Similarly, with narrow particle size distribution (low span value), viscosity
usually increases due to reduced polydispersivity and packing efficiency, which increases particles-solvent interaction.

The effect of this change in particle size and size distribution on apparent viscosity, plastic viscosity and Bingham yield point of the suspension are
also shown in Table 2. In the present study, it is observed that apparent viscosity, plastic viscosity and Bingham yield point of the suspension
increases with the reduction in particle size. Similar phenomena are also noticed for particle size with narrow distribution. Least average surface mean
diameter (i.e., 2.75 µm) particles with narrow particle size distribution (i.e., lowest span of 0.931) is obtained for the least specific gravity float sample
(i.e. 1.5) which results an improved apparent viscosity, plastic viscosity and Bingham yield point over other specific gravity samples.

144 
 
As solid loading increases, apparent viscosity, plastic viscosity and Bingham yield point are also increases. This is mainly due higher number of
particles results in more particle-particle interactions which increases the resistance to flow causing increase of viscosity (Krieger-Dougherty, 1959).
Above results show that particle size, particle size distribution and solid loading can be used to manipulate apparent viscosity, plastic viscosity and
Bingham yield.

CONCLUSIONS

For potential use of fly ash as a drilling fluid, classification of fly ash is carried out using float and sink test. Different specific gravity float samples
are obtained with different average particle size and particle size distribution. Low specific gravity float gives smaller average particle size with
narrow particle size distribution whereas higher specific gravity float gives higher average particle size and broad particle size distribution.
Rheological studies have been carried out in a standard Fann VG meter with the different specific gravity float samples assuming Newtonian fluid
flow approximation. Particle size, particle size distribution and solid loading have a major role to control the suspension rheology considerably.
Smaller particle size, narrow particle distribution and large solid loading improve apparent viscosity, plastic viscosity and Bingham yield point
considerably and similar results are obtained as compared to bentonite suspension. Comparable values of gel strength of treated fly ash and bentonite
suspensions are also an indication of comparable solid holding capacity. Rheological analysis shows that treated fly ash with smaller particle size,
narrow particle size distribution and high solid loading is acceptable for drilling fluid applications.

ACKNOWLEDGEMENT

Partial financial support from Indian School of Mines, Dhanbad [through minor research project grant 2009/MRP/PE/Acad] is gratefully
acknowledged. Help from Institute of Drilling Technology, Oil and Natural Gas Corporation, Dehradun, India for particle size analysis is gratefully
acknowledged.

REFERENCES

Ahmaruzzaman, M., 2010. A review on the utilization of fly ash, Progress in Energy and combustion Sci. 36, 327–363.
Bird, B., Stewart, W., Lightfoot, E., 2002. Transport Phenomena, Second Edition. John Wiley & Sons, Inc, New York.
Bourgoyne Jr., A. T., Millheim, K.K., Chenevert, M.E., Young Jr. F.S., 1986. Applied Drilling Engineering, Soc. Pet. Eng., Richardson, Tx.
Deeley, G.M., Laguros, L.W., 1987. Stabilization of drilling fluid waste with fly ash, Mat. Res. Soc. Symp. Proc., 86, 77-85.
Elward-Berry, J., Darby, J.B., 1997. Rheologically stable, nontoxic, high temperature, water based drilling fluid. SPE Drill. Complet., 158– 162
(September).
Giere R., Carleton L.E., Lumpkin G.R., 2003. Micro- and nanochemistry of fly ash from a coal-fired power plant, Am Mineral, 88, 1853-1865.
Gray, G.R., Darley, H.C.H., 1981. Composition and Properties of Drilling and Completion Fluids, 5th ed., Gulf Professional Publishing Company,
Butterworth-Heinemann.
Krieger I.M., Dougherty T.J., 1959. Concentration Dependence of the Viscosity of Suspensions Trans. Soc. Rheol. 3, 137-152.
Kumar, V., Zacharia, K.A., Sharma, P., April 2007. Fly ash Utilization: Indian scenario & case studies, http://www.tifac.org.in/news/flyindia.htm.
Monicard, R., 1982. Drilling Mud and Cement Slurry Rheology Manual, Gulf Publishing Company, Editions Technip, Paris.
Park, A., Scott Jr., P.P., Lummus, J.L., 1960. Maintaining low solids drilling fluids. Oil Gas J., 81– 84 (May).
Friedman, C., Schaffer, T., 2009. India’s Energy Options: Coal and Beyond, South Asia Monitor, Number 132 (August).
http://csis.org/files/publication/sam_132.pdf
Thompson, L.F., 1994. Drilling fluids waste minimization and stabilization using polymer technology. Paper SPE 29196 Presented at the SPE Eastern
Regional Meeting, Charleston, West Virginia.
Totten, P.L., Griffith, J.E., King, B.L., Foamable drilling fluid, EP0761798, 1997. http://www.freepatentsonline.com/EP0761798.pdf

Table 1 Particle size distribution statistics of virgin and treated fly ash

Description Virgin fly ash: sp. Treated fly ash: float Treated fly ash: Treated fly ash:
gr. 2.2 sp. gr. 1.5 float sp. gr. 1.8 float sp. gr. 2.0

D 101 (µm) 4.85 1.17 1.25 1.26

D25 (µm) 10.25 2.08 2.62 2.99

D502 (µm) 18.22 16.92 5.62 4.72

                                                            
1
Diameter of the particles below 10 % volume of under size cumulative distribution 
2
 Diameter of the particles below 50 % volume of under size cumulative distribution 

145 
 
D75 (µm) 29.54 3.795 11.01 8.47

D903 (µm) 45.61 16.92 22.8 14.47

Span4 2.237 0.931 2.799 3.834

Volume mean 11.42 8.0 8.95 10.39


Diameter (µm)
Surface mean 3.25 2.75 2.99 3.21
Diameter (µm)
Number mean 0.94 0.98 0.97 0.93
Diameter (µm)

Table 2 Rheological parameters of bentonite and treated fly ash suspensions

% loading (g/100 ml Apparent Plastic Bingham Gel0 Gel10


raw water) viscosity viscosity yield point
(µaF, cp) (µp, cp) (Yb, N/m2) (N/m2) (N/m2)

Bentonite
2 3.0 0.5 5.25 4590.0 0.50 0.75
3 4.5 1.5 6.74 1530.0 0.50 1.0
4 7.5 2.5 11.23 1530.0 0.50 1.0
5 8.75 2.1 14.33 2720.0 0.50 1.0
6 11.5 3.0 18.47 2380.0 0.50 2.0
Treated fly ash suspension (sp. gr. 2.0)5
2 1.1 0.4 1.60 1275.0 0.375 0.50
3 1.5 0.75 1.87 510.0 0.50 0.75
4 1.9 0.8 2.59 892.5 0.50 0.75
5 2.2 0.9 3.04 963.3 0.50 0.75
6 2.5 1.0 3.49 1020.0 0.50 1.00
Treated fly ash suspension (sp. gr. 1.8)1
2 3.3 1.5 4.24 680.0 0.375 0.50
3 4.0 1.5 5.74 1190.0 0.50 0.75
4 4.5 1.5 6.74 1530.0 0.50 0.75
5 5.6 2.3 7.73 953.5 0.50 1.00
6 6.5 2.5 9.23 1122.0 0.50 1.00
Treated fly ash suspension (sp. gr. 1.5)1
2 4.8 1.5 7.24 1700.0 0.50 0.50
3 5.5 2.0 7.98 1275.0 0.50 0.75
4 6.5 2.0 9.98 1785.0 0.50 0.75
5 7.3 1.0 12.99 5865.0 0.50 1.00

                                                            
3
Diameter of the particles below 90 % volume of under size cumulative distribution 

4
and represents the breadth of the particle distribution 

5
Washed and dried fly ash float obtained after float and sink experiment 

146 
 
6 9.0 2.0 14.98 3060.0 0.50 1.50

Legend for figures


Fig. 1 Particle size distribution of virgin and treated fly ash of different specific gravity
Fig. 2 Variation of dial reading with rotation for the bentonite, virgin fly ash and treated fly ash suspensions in Fann VG meter
Fig. 3 Variation of wall shear stress at the bob (τ) with shear rate at the bob (γ’) for virgin and treated fly ash samples (Fig. 2)

a. Virgin fly ash

b. Float sp. gr. 1.5

147 
 
c. Float sp. gr. 1.8

d. Float sp. gr. 2.0

Fig. 1

148 
 
30
Raw water bentonite suspension

2.0 g/100 ml

25 3.0 g/100 ml

4.0 g/100 ml

5.0 g/100 ml

6.0 g/100 ml
20

Dial reading, degree


15

10

0
0 200 400 600
Rotation, rpm

6
Raw water flyash suspension: sp. gr. = 2.0

2.0 g/100 ml
5
3.0 g/100 ml

4.0 g/100 ml

5.0 g/100 ml

4 6.0 g/100 ml
Dial reading, degree

0
0 200 400 600
Rotation, rpm
 
 

149 
 
14
Raw water flyash suspension: sp. gr. = 1.8

2.0 g/100 ml
12
3.0 g/100 ml

4.0 g/100 ml

10 5.0 g/100 ml

6.0 g/100 ml

Dial reading, degree 8

0
0 200 400 600
Rotation, rpm

20
Raw water flyash suspension: sp. gr. = 1.5

2.0 g/100 ml

3.0 g/100 ml
16
4.0 g/100 ml

5.0 g/100 ml

6.0 g/100 ml
Dial reading, degree

12

0
0 200 400 600
Rotation, rpm

Fig. 2

150 
 
14
Raw water bentonite suspension
12

10

Shear stress (τ), pas


8

4 2.0 g/100 ml water


3.0 g/100 ml water
4.0 g/100 ml water
2 5.0 g/100 ml water
6.0 g/100 ml water

0
0 500 1000 1500
Shear rate (γ), s-1
 
3
Raw water fly ash suspension sp. gr. = 2.0

3
Shear stress (τ), pas

1
2.0 g/100 ml water
3.0 g/100 ml water
4.0 g/100 ml water
1
5.0 g/100 ml water
6.0 g/100 ml water

0
0 500 1000 1500
Shear rate (γ), s-1
 
 
 

151 
 
8

Raw water fly ash suspension sp. gr. = 1.8

Shear stress (τ), pas


4

2.0 g/100 ml water


2
3.0 g/100 ml water
4.0 g/100 ml water
5.0 g/100 ml water
6.0 g/100 ml water

0
0 500 1000 1500
Shear rate (γ), s-1
 
12
Raw water fly ash suspension sp. gr. = 1.5

10
Shear stress (τ), pas

4
2.0 g/100 ml water

3.0 g/100 ml water

4.0 g/100 ml water


2
5.0 g/100 ml water

6.0 g/100 ml water

0
0 500 1000 1500
Shear rate (γ), s-1
 
Fig. 3

152 
 
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Study the effect of polymers on the rheological properties and filtration properties of water
based drilling fluids
Vikas Mahto, A.K. Pathak and V.P. Sharma
Department of Petroleum Engineering,
Indian School of Mines, Dhanbad, Jharkhand, India
Email: vmahto@yahoo.com;akhilendra_patahk@yahoo.co.in;vps_ismpe@hotmail.com

ABSTRACT

The bentonite clay is most commonly used for preparation of water based drilling fluids. The organic polymers are added in bentonite water
suspension to improve its rheological properties and to control filtration loss properties. In this work, the usefulness of Indian bentonite clays in oil
well drilling fluids has been studied. The physicochemical properties like cation exchange capacity, swelling index and yield of clay were determined
using standard procedures and after studying the effect rheological properties of organic polymers on bentonite water suspensions, some drilling fluids
were prepared. The developed drilling fluid combinations have shown best rheological properties and fluid loss control at optimum concentration of
polymers which is required for better performance in oil well drilling

KEY WORDS: Bentonite, Rheology, Polymer, Gel Strength, Filtration Loss

INTRODUCTION

Bentonite clay is the cheapest raw material for the preparation of drilling fluids. It contains montmorillonite and it is added in fresh water to increase
the hole cleaning capability, to reduce water seepage or filtration into permeable formation, to form a thin filter cake of low permeability, to promote
hole stability in poorly cemented formations and to avoid or overcome loss of circulation (Chilingarian and Vorabutr, 1983; Caenn and Chillingar,
1996)

The high clay solids in drilling fluid greatly reduce the rate of penetration (Gatlin, 1960) because solid particle cushion between bit tooth and rock so
that clean sharp impact is not achieved. An increase in solid content increases viscosity and yield point which in turn increases system frictional losses
and thereby reduce the pressure drop (and velocity) which is applied across the bit and thus chip clearance time is increased which ultimately
decreases penetration rate. Thus low bentonite content sufficient enough to achieve desired gel strength is desirable for control of total solids. Since
bentonite clay is unable to provide satisfactory rheological properties and filtration properties, polymers are added to achieve the desired result

The partially hydrolyzed polyacrylamide (PHPA) is used as clay extender, either dry-mixed into clay or added to the rig to a low-bentonite mud (Fig
1). It is a linear copolymer of acrylate (anionic) and acrylamide (nonionic) monomers. The ratio of acrylic acid to acrylamide groups on the polymer
chain can be varies depending upon molecular weight (Lummus and Azar, 1986). A concentration of approximately 10 to 30% acrylate groups
provides optimal anionic characteristics for most drilling applications.

Fig 1: Structure of Partially Hydrolyzed Polyacrylamide


The PHPA in fresh water, salt water and sea water was found to be non-toxic (Orszulik, 1997). It is primarily used for shale control. PHPA reduces
uncontrolled build of colloidal solids, control of viscosity and provide high penetration rate. The temperature stability of PHPA is 149oC (Lummus and
Azar, 1986; Bruton et. al, 2000).

The xanthan gum is a high-molecular weight natural heteropolysaccharide is produced by the bacterial fermentation of the microorganism
Xanthomonas campestris (Navarrete, 2000; Khan et al, 2003). The chemical structure of xanthan gum is shown in Figure 2. The polymer is slightly
anionic in nature but shows good compatibility with electrolyte solutions with tolerance for salinity and fair tolerance for hardness ions. One of the
most important functional properties of xanthan gum is its superior ability to control aqueous fluid rheology. Water solutions of xanthan gum are more
pseudoplastic than other polysaccharide products.

Fig 2: Structure of Xanthan Gum

Xanthan gums major application in drilling fluids is as thickener/viscosifier or more precisely as a suspending agent. It shows higher suspending
ability of xanthan gum than that of any polymer currently used in drilling fluids. The exceptional suspending ability of xanthan gum at low
concentrations favors its use whenever transportation costs are high. Although Xanthan gum is not a filtration-control agent, it is compatible with
filtration reducing substances such as Bentonite and CMC.

The gaur gum is derived from the seed of gaur plant (Fig 3). It is dirty white to yellow powder It is a branched copolymer and its molecular weight is
in the order of 2.2 x 105(Parija et. al, 2001). It provides very high viscosity at lower concencntration. Structurally, the backbone of polymer is
composed of mannose unit and every other mannose unit has a galactose branch bonded to it (Chatterji and Borchardt, 1981). Chemically it is a
galactomannan composed of linear chains of (1-4) –b-D-mannopyranosyl units with D-galactopyranosyl units attached by (1-6) linkages. There are no
ions present in the polymer structure thus the polymer is termed as nonioinic.

Fig 3: Structure of Guar Gum

The Carboxymethylcellulose is derivatized cellulosic polymers which have been modified to impart water salability to the cellulose molecule.
Carboxymethylcellulose usually referred to as “CMC”, is sold as the sodium salt. It should be called as “Sodium Carboxymethylcellulose”. The

154
polymer is manufactured by the reaction of monochloroacetic acid or sodium monochloroacetate with sodium-hydroxidetreated cellulose.

R (OH) 3+ClCH2CO2H+2NaOH = R (OH) 2OCH2CO2Na +NaCl+2H2O

Fig 4: Structure of Carboxymethyl Cellulose

It is used primarily to control fluid-loss control (Hughes and Zones, 1990). To make CMC, OH groups on the glucose rings of cellulose are ether-
linked to carboxymethyl (-OCH2-COO-) negative charged groups. Each glucose ring has three OH groups capable of reaction having three degree-of
substitution. Degree of substitution determines water solubility and negativity of the polymer, which influences a CMC's effectiveness as a mud
additive. Drilling grade CMC used in muds typically has degree-of-substitution around 0.80 to 0.96. The viscosity depends largely on the molecular
weight of the starting cellulose material. Sodium Carboxymethylcellulose is an anionic polymer and is adsorbed on clays. Filtration is sharply reduced
by low concentration of CMC and especially by the higher molecular weight products. The low-viscosity grade CMC is used for filtration reduction of
heavily weighted muds. The medium-viscosity grade is used in muds having the common range of solids contents. The CMC suspensions are shear-
thinning; they have high apparent viscosities at very low shear rates. The apparent viscosity decreases with rise in temperature. The effectiveness of
CMC in reducing filtration and raising viscosity decreases as salt concentration increases. Some calcium containing muds are thinned by small
addition of CMC. CMC is co-precipitated along calcium and magnesium by raising pH as is observed with starch. The temperature stability of CMC is
121oC and is not subjected to bacterial degradation (Lummus and Azar, 1986).

The polyanionic cellulose (PAC) is used as fluid loss reducer for fresh water and salt water muds. But it also acts as viscosifier in these systems (Fig
5). It is available in two types (high or low viscosity grade) both of which impart same degree of fluid loss control but different degree of viscosity.
The temperature stability of polyanionic cellulose is 1490C and is not subjected to bacterial degradation (Lummus and azar, 1982; Plank, 1992; Bruton
et al, 2000).

Fig 5: Structure of Polyanionic Cellulose

The substantial quantity of starch is used in mud to control fluid loss in water muds ranging from freshwater to saturated-salt to high-pH lime muds.
Starches have thermal stability to about 250°F [121°C]. They are subject to bacterial attack unless protected by high salinity or bactericide. Starches
are carbohydrates of a general formula (C6H10O5)*n and are derived from corn, wheat, oats, rice, potatoes, yucca and similar plants and vegetables
(Gray and Darley, 1981). They consist of about 27% linear polymers (amylose) and about 73% branched- polymers (amylopectin).

Fig 6: Structure of Amylose

The two polymers are intertwined within starch granules. Granules are insoluble in cold water, but soaking in hot water or under steam pressure
ruptures their covering and the polymers hydrate into a colloidal suspension. This product is a pre-gelatinized starch and has been used in muds for
many years (Chiliangarian and Vorabutr, 1983). Amylose (Fig 6) and amylopectin (Fig 7) are nonionic polymers that do not interact with electrolytes.
The use of starch typically causes a minimal increase in viscous while effectively controlling fluid loss.

155
Fig 7: Structure of amylopectin

The selection of drilling fluid systems depends upon the type of geological formations, depth, temperature and the type of wells like vertical,
horizontal, multilateral, ERD or drain holes. The types and composition of the additives in the drilling fluids varies depending upon the situations and
conditions for drilling operations. For the optimization of the composition of the drilling fluids, thorough study of the effect of additives on the drilling
fluid systems is essential. This paper presents the effect of organic polymers on general behavior of the water based drilling fluids which will help us
for the proper optimization of the properties of clay based drilling fluid systems.

EXPERIMENTAL

Materials

The raw bentonite clay sample used for this work was obtained from Kutch region, Gujarat, India. The partially hydrolyzed polyacrylamide, xanthan
gum, gaur gum, polyanionic cellulose, carboxy methyl cellulose and starch were procured from Indian oil field.

Experimental Procedures

Initially, bentonite water suspensions were then prepared and rheological properties like apparent viscosity, plastic viscosity, initial gel strength,
and 10 min. gel strength and filtration properties were measured using Fan V-G meter 35SA model and standard API filtration Test. These properties
were again measured by adding organic polymers and most favorable mud systems were prepared. The developed drilling fluid combinations were
hot rolled in a roller oven for 16 hours to measure its temperature stability.

The apparent viscosity, plastic viscosity, yield point were calculated from 300 rpm and 600 rpm readings of Fan V-G meter using standard
procedure for field testing of drilling fluids(Recommended Practice, 1988)

RESULTS AND DISCUSSIONS

The rheological properties and filtration properties of bentonite water system at different different clay concentrations are shown in Table 1. The bentonite
clay consists of superimposed layers composed of two Si-O tetrahedral sheets framing an Al-O-OH octahedral sheet. In aqueous dispersion, water can
penetrate into the interlayer space and cause swelling of bentonite. The behavior of bentonite is complex due to unisometric clay particles exposing
different crystal faces, on which an electrical double layer can develop, which differs both in sign and the magnitude of the total surface potential (Singh
and Sharma, 1997). Consequently the rheological characteristics in these suspensions are achieved by clay particle-clay particle and clay water suspension.
The filtration properties of the clay water suspension decreases with the increase in clay concentration and the reason for this may also be the same.

Table 1. Effect of conc. of bentonite clay on rheological properties and filtration properties of Bentonite- water suspension

Conc. Of Apparent Plastic Viscosity Yield point Gel in Gel 10 API Filter Loss
bentonite Viscosity (cp) (N/m2) (N/m2) (N/m2) (ml)
(g/l) (cp)
10 1.75 1 0.75 0.375 0.5 85
20 2 1 1 0.5 0.75 76
30 2.5 1 1.5 0.5 1 50

156
40 3 1 1.75 0.5 2 44
50 4 2 2 0.5 2.5 42.5

Table 2 summarizes the rheological properties and filtration properties of 3 %( w/v) or 30 gm/l bentonite concentration in water with this clay. Since the
solids content should be less for effective control of total solids in drilling fluid , 3% (w/v) bentonite water suspension was chosen as base mud , which is
normal concentration used in oil fields for drilling purposes(Park et al, 1962; EIward-Berry and Darby, 1997) .

Table 2. Rheological properties of 3 % bentonite clay-water suspension

Sl. No. Properties Unit Observed Values


1 Apparent Viscosity cp 2.5
2 Plastic Viscosity cp 1
3 Yield Point N/m2 1
4 Initial Gel Strength N/m2 0.5
5 Final Gel Strength N/m2 1
6 API Filter Loss ml 50

The rheological properties of bentonite at 3% conc. require further improvement. Thus, organic polymers were added to 3% clay water suspension in
order to improve its rheological properties. Each of these polymers at different concentrations was added to 3% Bentonite-water suspension to obtain
most suitable water based drilling fluids.

Effects of PHPA on the rheology of drilling fluid

The effect of PHPA on the rheology of 3% water based drilling fluid is shown in Table No. 3. The apparent viscosity shows almost sharp increasing
trend with increase in concentration of PHPA. Plastic viscosity shows almost steady increasing trend as concentration of PHPA increases.

Table 3. Effect of conc. of PHPA on rheological properties of 3% Bentonite- water suspension

Conc. of PHPA Apparent Plastic Viscosity Yield point Gel in Gel 10


(w/v)% Viscosity (cp) (N/m2) (N/m2) (N/m2)
(cp)
0.1 12.5 8 4.5 2 4
0.2 17 10 7 2.5 6
0.3 22.5 14 8.5 3.5 7
0.4 25 13 12 4 9
0.5 28.5 14 14.5 4.5 10

Effects of Xanthan Gum on the rheology of drilling fluid:

The effect of xanthan gum on the rheology of 3% water based drilling fluid is shown in Table No. 4. The apparent viscosity shows almost sharp
increasing trend with increase in concentration of xanthan gum. Plastic viscosity shows almost steady increasing trend as concentration of xanthan
gum increases.

Table 4. Effect of conc. of Xanthan Gum on rheological properties of 3% Bentonite- water suspension

Conc. of Xanthan Apparent Plastic Viscosity Yield point Gel in Gel 10


Gum Viscosity (cp) (N/m2) (N/m2) (N/m2)
(w/v)% (cp)
0.1 9.5 5 4.5 2.5 5
0.2 15 7 8 5.5 7
0.3 34 8 26 14.5 17.5
0.4 44.5 16 28.5 16.5 19.5
0.5 67 22 45 21.5 24

Effect of Gaur Gum

The effect of Gaur gum on rheological properties is shown in Table 5. It is seen that apparent viscosity, plastic viscosity, yield point, initial gel
strength and 10 minute gel strength increased with increase in the concentration of guar gum. The reason for this is the hydrogen bonding of hydroxyl
group of Gaur gum with clay.

157
Table 5. Effect of conc. of Gaur Gum on rheological properties of 3% Bentonite- water suspension

Conc. of Guar Apparent Plastic Viscosity Yield point Gel in Gel 10


gum Viscosity (cp) (N/m2) (N/m2) (N/m2)
(w/v)% (cp)
0.05 6.5 3.5 3 1.5 2.5
0.1 11.5 6 5.5 2.0 3
0.15 20 9 11 4.5 5.5
0.2 25 10 15 6 7
0.25 28.5 10 18.5 6.5 7

Effects of CMC on the rheology of drilling fluid:

The effect of CMC on the rheology of 3% water based drilling fluid is shown in Table No. 6, It can be concluded that plastic viscosity, yield point
and 10 min gel strength are showing increasing trend on increasing concentration of CMC. The increasing trend is coming due to carboxy-group in
CMC, which causes dispersibility of water. This causes increase in friction between different particles. This results in increased required shearing
stress to induce shear rate and thus yield point increases. The increase in viscosity can be due to the increased repulsion charges of Na+ ions of CMC.

Table 6. Effect of conc. of CMC on rheological properties of 3% Bentonite- water suspension

Conc. of CMC Apparent Plastic Viscosity Yield point Gel in Gel 10


(w/v)% Viscosity (cp) (N/m2) (N/m2) (N/m2)
(cp)
0.25 5 4 2 1 3
0.5 8 6 4 1.5 4.5
0.75 10 7 6 2 7
1.0 13 8 10 3 9
1.25 41 20 21 9 30

Effect of Polyanionic Cellulose

Table 7 shows that viscosity (apparent, plastic), yield point, yield point/ plastic viscosity ratio and gel strengths increases with increase in polyanionic
cellulose concentration. The reason for this is that adsorption and flocculation occurs via hydrogen bonding between solid surfaces and hydroxyl group
on the polymer

Table 7. Effect of conc. of polyanionic cellulose on rheological properties of 3% Bentonite- water suspension

Conc. of PAC Apparent Plastic Viscosity Yield point Gel in Gel 10


(w/v)% Viscosity (cp) (N/m2) (N/m2) (N/m2)
(cp)
0.1 11.5 7 4.5 1.75 7
0.2 16 10 6 2.5 12.5
0.3 26 14 12 3.5 15
0.4 33 17 16 6 24
0.5 41 20 21 9 30

Effects of Starch on the rheology of drilling fluid:

Table 8 shows that viscosity (apparent, plastic), yield point, and gel strengths increases very slightly with increase in starch concentration. The starch
cannot be used as the visosifier due to its poor rheological behavior. It is mainly used for the control of fluid loss.

Table 8. Effect of conc. of Starch on rheological properties of 3% Bentonite- water suspension

Conc. of Starch Apparent Plastic Viscosity Yield point Gel in Gel 10


(w/v)% Viscosity (cp) (N/m2) (N/m2) (N/m2)
(cp)
0.2 4.5 2 2.5 2 3
0.4 4.5 2 2.5 2.5 4
0.6 5 2 3 3 4.5
0.8 5 2 3 2.5 4.5
1.0 5.5 2 3.5 3 4.5

158
Development of water based drilling fluids

The apparent viscosity, plastic viscosity and yield point and gel strengths increase with increase in the concentration of organic polymers with 3%
bentonite as evident from Tables 1-8. From these studies the six most favorable drilling fluids were developed by varying the compositions. Table 9
summarizes the three favorable drilling fluid system and their compositions are as follow:

Mud A: 3% Bentonite, 0.1% Polyanionic Cellulose and 0.1% Gaur gum

Mud B: 3% Bentonite, PHPA 0.25 %, 0.2% Polyanionic Cellulose and 0.2%CMC

Mud C: 3% Bentonite, 0.4% Polyanionic Cellulose and 0.2% Xanthan gum

Mud D: 3% Bentonite, 0.2% Polyanionic Cellulose, 0.2% CMC and 0.2% Gaur gum

Mud E: 3% Bentonite, 0.2% Polyanionic Cellulose, 0.2% CMC and 0.2% Xanthan gum

Mud F: 3% Bentonite, 0.2% Polyanionic Cellulose, 0.4% Starch, and 0.2% Xanthan gum

Table 9. Properties of developed drilling fluid systems at ambient temperature

Sample Apparent Plastic Viscosity Yield point Gel in Gel 10 API Filter Loss
Viscosity (cp) (N/m2) (N/m2) (N/m2)
(cp)
Mud A 21 10 11 3.5 6.5 12.5
Mud B 57.5 30 27.5 20 22 15
Mud C 20 13 7 5 12.5 18
Mud D 61.5 28 33.5 19 22 13
Mud E 66 29 37 30 33 14
Mud F 33 18 12 15 10 14

Table 10. Properties of developed drilling fluid systems after aging for 16 hours at 75oC.

Sample Apparent Plastic Viscosity Yield point Gel in Gel 10 API Filter Loss
Viscosity (cp) (N/m2) (N/m2) (N/m2)
(cp)
Mud A 21.5 11 11.5 3.5 7 12
Mud B 60 35 25 20 22 15
Mud C 21 14 7 5 12.5 18
Mud D 62 28 34 19 22 13.5
Mud E 66 29 36 30 33 14
Mud F 34 18 16 15 10 14

These combinations were also tested in hot roller oven for 16 hours at 75oC. The rheological properties and API filter loss before and after hot rolling
of each drilling fluid combination is also favorable which is shown in Table 10. Thus these drilling fluid systems were stable at 75oC.

CONCLUSIONS

The following results can be concluded from this experimental work:

(i) The Indian bentonite clay under study are quite suitable for the development of water based drilling fluids

(ii) The combination of organic polymers and Indian bentonite clay gives favourable rheological properties and optimum fluid loss.

(iii) The temperature stability of developed drilling fluids are 750C

REFERENCES

Bruton, J Mercer, R. and Paul, D (2000). “The application of new generation CaCl2 mud systems in the Deepwater GOM” SPE paper no. 59051,
IADC/SPE Drilling Conference, New Orleans.
Caenn, R and Chillingar, GV (1996). “Drilling fluids: State of the art,” J. Petrol. Sci. Eng., vol 14, pp. 221-230

159
Chatterji, J and Borchardt, JK (1981). “Application of water soluble polymers in the oil field. J. Petrol. Technol. November, pp. 2042-2054
Chilingarian, GV and Vorabutr, P (1983). “Drilling and Drilling Fluids”, 1st ed. Elsevier Science Publishers, Amsterdam, Netherlands, pp 50
Elward-Berry, J and Darby, JB (1997). “Rheologically stable, nontoxic, high temperature water based drilling fluids,” SPE Drilling Completion,
September pp. 158-162
Gatlin, C (1960). “Petroleum Engineering: Drilling and Well Completions”, Prentice Hall Englewood Cliffs, New Jersey, pp 125
Gray, G R. and Darley, HCH (1981). “Composition and Properties of Oil Well Fluids”, 4th ed. Gulf Publishing Company, Houston, Texas, pp. 495
Hughes, TL and Jones, TGJ (1990). “The chemical composition of CMC and its relationship to the rheology and fluid loss of drilling fluids”, SPE
paper no. 20000, IADC/SPE Drilling Conference, Houston, Texas.
Khan, R, Kuru, E Tremblay, B and Saasen, A (2003). “An Investigation of Formation Damage Characteristics of Xanthan Gum Solutions Used for
Drilling, Drill-In, Spacer Fluids, and Coiled Tubing Applications”, Canadian International Petroleum Conference, Calgary, Alberta
Lummus, JC and Azar, JJ (1986). “Drilling fluids optimization – A practical field approach”, Pennwell Publishing Company, Tulsa, Oklahoma, pp.
112,160
Navarrete, RC, Seheult, JM and Coffey, MD (2000). “New Bio-Polymers for Drilling, Drill-in, Completions, Spacer Fluids and Coiled Tubing
Applications” SPE Paper No-62790, IADC/SPE Asia Pacific Drilling Technology, Kuala Lumpur, Malaysia
Orszulik, ST (1997).” Environmental Technology in the Oil Industry”, Blackie Academic & Professional, London, UK.
Parija, S, Misra M and Mohanti, AK (2001). “Studies of natural gum adhesives: an overview,” J. Macromolecular Sci.-Polymer Rev., C41 (3), pp.
175-197
Park, A, Scott Jr., PP and Lummus, JL (1960). “Maintaining low solids drilling fluids,” Oil Gas J., May, pp. 81-84
Plank, JP (1992). “Water based muds using synthetic polymers developed for high temperature drilling,” Oil Gas J., 90, pp. 40-45.
Recommended Practice (1988). Standard procedure for field testing drilling fluids, 12th ed. Recommended Practice, Vol 13 B (RP 13B), API,
Washington, USA.
Singh, PK and Sharma VP (1997). “Effects of electrolytes on zeta potential of beneficiated Indian bentonites” J. Sci. Ind. Res., 56, pp. 281-287

160
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Conceptual development of a computer model for rheological analysis of oil based


drilling fluids at HT/HP
S. Nainar1, N. Willenbacher2 and R. Sharma3
1-3
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
2
Institut für Mechanische Verfahrenstechnik und Mechanik
Karlsruher Institut für Technologie, Germany
E-mails: 1subhashini.nainar@gmail.com, 2norbert.willenbacher@kit.edu, 3rajivatri@iitm.ac.in,

ABSTRACT

Normally oil based drilling fluids are non–Newtonian fluids. Hence, flow properties of fluid cannot be defined by a single viscosity value. Therefore,
rheological characterization is imperative for complex fluid model analyses. Proper application of drilling fluids in actual reservoir condition
demands that the downhole pressure be taken in to account in evaluating the rheology of oil mud. In this regard computer analyses provide the plots
indicating the variation of each rheological flow characteristics with change in value of parameter in the drilling fluid. In a computer model
regression analyses can be performed on the experimental results obtained to analyze the flow behavior of oil based drilling fluid under high pressure
(HP) and high temperature (HT) conditions. Additionally, from the computer model optimum quantity of different additives can be easily computed
with efficient analyses. In the present paper, we discuss the conceptual development of a computer model for rheological analysis of oil based drilling
fluids at HT/HP.

KEY WORDS: Fluids, Rheological analyses, non-linear multivariate regression analyses, apparent viscosity, plastic viscosity, yield point, power-
law plastic viscosity, power-law index, Bingham model, Herschel Bulkley model, Newtonian, Non-Newtonian, emulsion, power law, plastic
viscosity.

INTRODUCTION

Rheology is the study of science and movement of matter. Normally oil based drilling fluids are non–Newtonian fluids. Flow properties of fluid
cannot be defined by a single viscosity value. Therefore, rheological characterization is imperative for complex fluid model analyses. Proper
application of drilling fluids in actual reservoir condition demands that the downhole pressure be taken in to account in evaluating the rheology of oil
mud. The important aspects of rheology can be listed as:

• Rheology of mud is mostly shear history dependent and the extent of this effect is defined by GEL10. At higher temperatures the gel strengths
will be thermally activated.
• Application of pressure modifies the behavioral characteristic of mud. Rheological parameters of the fluid are determined with Rheometer.
• At high pressure and high temperature solids do not have a constant density.
• The volume variation of the emulsion is the linear combination of oil and brine volumes.
• The volume variation of solids is negligible.
• Rheological properties will be extremely dependant on the microstructure of the composition.
• Oil mud produces high rates of penetration at high temperature and pressure.
• Filtration rates in oil based mud are minimized by the addition of emulsified water.

Oil-based drilling fluids show significant rheological properties dependence on temperature and pressure. The well control incident rate in HP/HT
wells is much high which makes it imperative to use oil mud. The world wide locations of HP/HT wells are shown in Fig. 1. Geologically, HP/HT
wells are concentrated in the US Gulf Coast and the North Sea, where reservoir pressures of 16,000 psi and bottom-hole temperatures of 350°F are
very much common. In Asia, in some nations like Thailand, more than 45 high-temperature wells have been drilled successfully. And, in future
because of the movement of exploration in deeper offshore waters the interest and importance in HP/HT wells are expected to grow fast. Oil based
muds are desired in HP/HT wells. The oil mud shows suitable rheological properties of hole cleaning, cutting lifting and friction pressure loses in
annulus. Furthermore, thermal stability in deep, high-temperature wells, increased lubricity in deviated offshore wells, and hole stability in thick,
water-sensitive shales are the advantages of oil-based muds. The viscosity and density of oil mud varies with temperature and pressure, and oil-based
muds show significant density dependence on temperature and pressure. The oil-based muds are compressible, and their densities can vary in
proportion to the amount of solids, oil, and water in the oil/mud mixture. In general density decreases with an increase in temperature and increases
with an increase in pressure, and the rates of change depend on the amount of water and solids in the oil/mud mixture. Furthermore, oil muds
compress under pressure and expand with temperature while drilling deep hot wells. The visco-elastic behavior of the fluids with pressure is
dependent on the presence of clay in the fluid. The Mud weight at down-hole temperature and pressure can be significantly different from those
measured at the surface in HT/HP deep wells.

Fig 1: The world wide locations of HP/HT wells.

In this regard computer analyses provide the plots indicating the variation of each rheological flow characteristics with change in value of parameter
in the drilling fluid. In a computer model regression analyses can be performed on the experimental results obtained to analyze the flow behavior of
oil based drilling fluid under high pressure (HP) and high temperature (HT) conditions. Additionally, from the computer model optimum quantity of
different additives can be easily computed with efficient analyses. It is in this environment that the master’s thesis of the first author of the present
paper has been planned. The present reported work is part of her planned thesis, and in our work (including master’s thesis), broadly we have the
following objectives:

ƒ To experimentally study the flow behavior of oil based mud.


ƒ To develop equations that explores the relationships of temperature and pressure effects on the rheological properties of drilling fluid.
ƒ Based on the relationships obtained, the correct selection of rheological model to describe drilling fluids rheology can be made for hydraulic
calculations during drilling operations.
ƒ The relationships are to be investigated that are able to model the pressure and temperature behavior of the apparent viscosity, plastic viscosity
and yield point using the method of regressive analysis.
ƒ To characterize constants and correlation coefficients for various drilling fluids.
ƒ The characterized constants, temperature and pressure values should be substituted into expression obtained, and the rheological parameters at
down-hole could be obtained.
ƒ Finally selecting the best of additive, base fluid and their optimum composition in oil based mud can be determined.

In the past, traditionally, numerical programs have been used by the major drilling fluid companies for use in drilling fluid rheological analysis.
These programs have typically inputs like shear stress data from two fixed shear rates and calculated Bingham and power-law parameters, plastic
viscosity, yield point, power-law plastic viscosity, yield point, power-law index, and consistency coefficient, etc. However, with the recent advent of
powerful algorithms in computational science, and inexpensive and efficient computing facilities more sophisticated computational models can be
developed for the rheological modeling of drilling fluids. In the lack of a computer model, the extent of rheological analysis is limited to some simple
calculation of the Bingham model parameters typically expressed as plastic viscosity and yield point. In the past, the power-law model has been used
to model drilling fluids and the two power-law constants have been determined by either an electronic or slide rule type calculator.
However, our larger aim is too broad to be addressed in one paper. Therefore, in this paper we restrict ourselves to the conceptual development of a
computer model for rheological analysis of oil based drilling fluids at HT/HP. This paper will discuss the considerations in developing such
programs with particular reference to drilling fluids and the conventional oil field viscometer. Also, we shall show typical output from such a
program.

162
The remaining of this paper is organized as follows: Section 2 describes the brief review of literature; Section 3 presents the basic design of drilling
fluids at HP/HT; Section 4 describes the computer model briefly, and Section 5 concludes the paper with identifications of future areas of research
and application. Some technical details about the terms used in the present paper and design examples are given in Appendix 1.

BRIEF REVIEW OF LITERATURE

Maxey (2009) studied non aqueous fluid rheology at elevated temperature and pressure. The important observations from his paper are as follows:
• Information on the visco-elastic and yielding nature of these fluids is beyond the reach of viscometer.
• The flow profiles, yield stress, and visco-elasticity were examined using rheometer.
• Structural properties of the fluids can be obtained through the gel strength tests.
• Increasing the pressure serves to increase rheological properties (such as plastic viscosity).
• Clay based fluid exhibited only minimal changes in the flow curve with temperature, with the greatest differences exhibited at decreased
viscosity at high shear rates.
• As temperature increased, Clay free fluid exhibited a fairly consistent decrease in viscosity over the test range and is temperature sensitive.
• Without the organophilic clay, Clay free fluid exhibited a lower overall viscosity than did fluid with clay. The visco-elastic behavior of the
fluids with pressure was dependant on the presence of clay in the fluid.
• Organophilic clay-containing fluid exhibited modest decreases in yield stress at increased pressures.
• Increasing the pressure would cause the fluid to behave more like an elastic solid and exhibiting less structural dominance with the observed
decrease in yield stress.
• Clay free fluid’s yield behavior has no dependence on pressure while Fluid with clay shows a small but definite trend.
• The decrease in overall structural dominance matches well with the observed decrease in yield stress for fluid with clay under increased
pressure.
• Clay free fluid exhibits a much smaller yield stress than is found in fluid with clay. The overall structural dominance of fluid with clay is seen to
be higher than clay free fluid but the loss modulus for both fluids is roughly the same.
• With clay, a consistent decrease in overall structure of the fluid was observed with increasing pressure.
• Maximum strain within the LVR (linear viscoelastic region) for Clay free fluid was about half that of fluid with clay.

Tiu et al. (2006) presented a comprehensive study on yielding behavior of visco-plastic materials. The important observations from their paper are as
follows:
• Complex fluids exhibiting a yield stress have extremely variable shear properties ranging from elasto-plastic behavior below the yield stress to
viscous behavior beyond that.
• The techniques developed for measuring yield stress can be categorized into two groups: indirect and direct methods. Direct method obtain the
yield stress based either on dynamic or static principles.
• The vane torsion method, widely accepted direct method. The yield stress is calculated from the peak torque measurement and the vane
dimensions.
• Indirect methods involve the fitting of the shear stress-shear rate data to rheological models such as the Bingham, Casson and Herschel-Bulkley
models, etc. The yield stress is determined by the extrapolation of the flow curve at low shear rates to zero shear rate. The results obtained using
indirect methods are very sensitive to the models used for fitting the rheological data.
• Yield stress obtained by the direct methods is more reliable than the indirect methods.
• Quantitative understanding of this progressive transition in rheological behavior from linear elastic response to non-linear elasto-plastic
deformation is necessary to understand the creep behavior of a complex material.
• The yield stress increases steadily with increasing solid concentration, while the slope of the stress plateau decreases with increasing
concentration. This enhances the particle-particle interactions and the gel would become harder, tougher and more brittle.
• At lower stress values, the shear strain appears to be a unique function of the shear stress, and is independent of the rate of stress increment. At
large stress values, the shear stress is a unique function of strain rate, which represents a fluid behavior.
• The transition from solid to liquid is not abrupt, but goes through a region of a stress plateau. The extent of the stress plateau reflects the
breaking process of the internal structure of the material. A large slope implies that the breaking process occurs in a wide range of stress; hence,
the structure of the material exhibits ductile-type failure. A small slope reflects a brittle-type failure of the material structure.
• The deformation of a material during creep can be classified into three regions. There is an initial elastic response when stress is applied to the
sample, followed by a creep process. After a certain time, the creep rate increases slowly at first followed by a sharp upturn, indicating that some
viscous component was involved. A steady strain rate is finally developed, corresponding to a viscous flow condition in the material.
• Yielding process of a material is time-dependent. A prolonged creep time can cause failure of the material structure at lower stresses, even at
well below the conventional yield stress value.
• The decrease in strain recovery ratio from 100 to 0 % represents a change of the materials’ behavior from elastic solid to viscous liquid.
• The lower critical stress represents the transition from purely elastic to plastic behavior; and the upper critical stress represents the transition of
plastic to viscous flow. The two stresses obtained from recovery ratio can be interpreted as the limits of solid-like behavior and liquid-like
behavior, representing the two yield stresses, static and dynamic.
• The structure failure only occurs once the accumulated strain reaches a certain value. Therefore, using a critical strain to characterize structure
failure or yielding of visco-plastic materials may be more appropriate than a yield stress.

163
Skalli et al. (2006) presented a study on surface and core wetting effects of surfactants in oil-based drilling fluids. The important observations from
their paper are as follows:
• Surfactants are used in oil-based drilling fluids to emulsify water and to ensure that cuttings are wetted by oil.
• Increasing concentration of surfactant produced decreases the rates of water imbibitions.
• Surfactants reduce the system free energy by lowering the energy of the interface (calculated as area * surface tension).
• Polyaminated fatty acids and mixtures of polyaminated fatty acids and oxidized tall oil are two types of surfactants.
• Addition of surfactant to crude oil had less effect than sequential exposure of surfaces first to crude oil, then to surfactant solutions.
• Changes in wetting toward more oil-wet conditions were much more dramatic if surfaces were exposed to the emulsifier after they had been
treated with crude oil.

Nguyen (2006) discussed the problem of predicting dynamic Barite sag in oil based drilling fluids. The important observations from their paper are as
follows:
• More sag problems are encountered with oil based mud (OBMs) than with water base mud (WBMs).
• Sag problems are more in OBMs because they are generally more viscous than WBM, and efforts are made to reduce viscosity by minimizing
additives used for suspending barite.
• Invert base mud do not develop gel strengths in the same manner as WBMs. As the result, the frequency of problem associated with barite sag is
higher with OBMs, compared to water-based systems.
• Barite sag could lead to major drilling problems such as lost circulation, well-control difficulties, poor cement jobs, and stuck pipe.
• Annular velocity, drill pipe rotation, rheology of drilling fluids and inclination angle are important parameters affecting dynamic sag.
• Sag is decreased due to increased flow and pipe rotation.
• Preventing dynamic sag is more difficult than preventing static barite sag.
• The onset of dynamic barite sag occurs at a very low shear rates and in deviated well-bore using invert-emulsion mud.
• Ultra-low shear rate viscosity appears to have a greater influence on dynamic barite sag than mud weight.
• Sag tendency was significantly higher in deviated wells than in vertical ones.
• If calcium chloride (CaCl2) salt is replaced with a selection of other salts in invert-emulsion fluids, the sag stability could be affected
significantly.
• Viscosity levels should be within the limits of the “Prevention Window” which are indicative of a low probability for dynamic barite sag. The
preventive window is shown in Fig. 2.
• At high velocity, particularly when turbulence was observed, sag beds were removed and incorporated back into the mud.
• Organoclays in OBM will help reduce the problem of barite sag.
• Thinning of OBM can increase the problems of barite sag.
• Surfactant in required amount is used to thin the OBM.

Fig 2: The preventive window for dynamic barite sag.

Berthezene et al. (1999) presented a comprehensive study on Methane solubility in synthetic oil-based drilling mud. The important observations from
their paper are as follows:
• Predicting methane solubility is of crucial importance for predicting gas eruption risks when drilling.
• This is particularly the case of offshore deep drilling areas for controlling gas hydrate formation.
• When drilling in high pressure and high temperature conditions, the interactions with the reservoir fluids can significantly modify the properties
of the mud.
• At higher pressures, the methane and the oil are completely miscible.
• The increased solubility of gases in the mud creates a risk of gas kicks or of gas hydrates formation at the surface or in the mud column.
• Gas solubility in water is very small, gas solubility data in oil are equivalent to the solubility in the complete drilling fluid.
• The heavier the hydrocarbon liquid, the higher the pressure needed for a given solubility.
• The interactions between the mud filtrate and the reservoir fluid can lead to solid deposits such as asphaltenes, paraffins or salts inducing severe
formation damage.
• At very high pressure, a critical point is observed above which complete miscibility occurs. This critical pressure increases quickly with the
number of carbon atoms.
• The presence of an ester group in the base oil results in a polar moment that could reduce the total solubility, especially at high pressure.

164
Bailey and Weir (1998) discussed investigation of methods for direct rheological model parameter estimation. The important observations from their
paper are as follows:
• Modeling accuracy is affected by the choice of rotational viscometer readings used to calculate these parameters.
• It provides the necessary expressions for direct parameter solution.
• Modeling accuracy has an impact on characterization of drilling fluid’s non-Newtonian behavior over all expected shear rates encountered in the
circulating system.
• The main problem in using the models is increase of complexity in solving model parameters and the uncertainty associated with the selection of
the appropriate rotational viscometer readings to be used to solve their parameters.
• Parameter evaluation based on nonlinear least squares will deliver better general fluid characterization.
• Readings taken from any coaxial cylindrical rotational viscometer requires some compensation and correction in order to deliver more
representative non-Newtonian shear rates.
• The direct solution approach, however, establishes fixed parameters based on measured data and are not subject to variation due to nonlinear
fitting.
• Direct parameter solution refers to the method of rheological model parameter solution using unspecified rotational viscometer readings to solve
a set of explicit and or implicit equations derived from the rheological model itself.
• Different input rotational viscometer readings were found to result in non-convergence in some of the models tested indicating a limited
robustness in the direct solution approach.
• Relative model performance was analyzed by comparing their residual mean square (RMS) responses against a large set of 414 different
rotational viscometer reading sets comprising of eight dial readings in each set.
• Conventional readings of 600 and 300 rpm do not serve Laminar flow which is more likely to occur at lower shear rates.
• 600/30 rpm has the smallest median RMS value of all combinations tested and is distinctly more accurate than results obtained using
conventional 600 rpm and 300 rpm readings.
• Models must possess the smallest maximum RMS and the smallest median RMS value and lower- and inter-quartile spreads.
8 8
• A two-parameter model will have 28 ( C2 ) possible reading combinations while three-parameter model will have 56 ( C3 ). Possible reading
8
combinations and 70 ( C4 ) possible reading combinations for a four parameters model.
• Flow equations derived directly from the rheological model are suitable for laminar flow only.
• Modeling accuracy decreases as the value of the in situ shear rate differs more and more from those shear rates used to determine rheological
model parameters.
• The non-Newtonian fluid being characterized at these high rotations will be in the region of turbulent flow due to the high shearing present.
• Laminar flow which is more likely to occur at lower shear rates is often poorly served by concentrating the rheological parameter solution using
high viscometer rotation readings.
• All two-parameter models yield 100% solution success, higher-order models showed less consistency.
• The Herschel and Bulkley model demonstrated a 100% success in providing solution.
• The non-Newtonian fluid being characterized at these high rotations will be in the region of turbulent flow due to the high shearing present.

Caenn and Chillingar (1996) have reported a comprehensive treatise on “Drilling fluids: State of the art”. In their book, from the viewpoint of the
proposed thesis, the important observations are as follows:
• Drilling fluids are classified into air based mud, water based mud and oil based mud.
• Classification of mud is based on type of base fluid used in preparation of drilling fluid.
• Most air-drilled wells are relatively shallow in hard, competent formations.
• Drilling fluids are used while drilling as it serves various purposes.
• A mud with a sufficient amount of polyol behaves more like an oil mud in its wetting characteristics.
• Polyols greatly improves the lubricity; shale stability and formation return permeabilities of the mud. It prevents problems of gas hydrates and
bit balling.
• Polyol group includes glycols, glycerol, and pentaerythritol and also such products as trimethylolethane, trimethylolpropane, 1, 2, 6-hexanetriol,
sorbitol, inositol, and poly (vinyl alcohol).
• Polymers in water based mud may be natural, modified or synthetic.
• All of the diesel replacements either eliminate or minimize the aromatic content, thereby making the material non-toxic or less toxic.
• New generation mud will have the advantages of oil mud but the handling and disposal characteristics of water based mud.
• New generation mud is mainly based upon synthetic fluids, e.g. esters, ethers, polyalphaolefins, glycols, glycerines and glucosides.
• Oil mud prevents ‘Bit Balling’, provides high lubricity, enables hole cleaning and suspension but has many disadvantages.
• Oil-based mud, in general, is easier to change into turbulence, required for good hole cleaning than water-based mud.
• The ability to rapidly increase the low-shear rate viscosity and gel strengths to high levels is easier in water-base Mud than in oil-based mud.
• In new generation mud, the base fluid is synthetic.
• Biodegradability and bioaccumulation characteristics can be improved in new generation mud.
• Esters, polyalphaolefins, ethers, and detergent alkylates are used to replace diesel or mineral oil for synthetic oil mud.
• Synthetic oil mud will have the advantages of oil mud but with the handling and disposal characteristics of water mud.

Hamida et al. (2004) discussed the problem of rheological characterization of aqueous waxy hull-less barley (WHB) solutions. The important
observations from their paper are as follows:

165
• WHB performed reasonably well as a viscosifying agent, especially in saline, high temperature environments and in high pH environments.
• They are non-starch polysaccharides are composed of mixed linked glucose polymers.
• It is rich in soluble fiber components, most notably β − glucan which provides unique viscosity properties in high temperature, high pH
environments.
• β − glucan may be effectively extracted from barley at elevated temperatures.
• NaOH has also been found to yield high percentages of β − glucan in cereals.
• Since WHB is biodegradable, such an additive is considered environmentally friendly.
• WHB produces a relatively low yield point mud before and after aging.
• The cost of WHB is on the average 3 times higher than the cost of bentonite.
• WHB can be applied as an additive to oil-based mud since β-glucan can be extracted effectively using solvents.
• In saline solution WHB clearly outperforms bentonite fluid and it requires no additive to remain stable.
• Low concentration WHB- water mixture behaves almost like Newtonian. At higher concentrations WHB–water mixture exhibits shear thinning
characteristics.
• The viscosity remained stable over a wide range of pH (2–10) and decreased with increasing temperature.
• Plastic viscosity rises with an increase in WHB concentration.
• WHB with polymers can further increase the yield point after aging.
• The rheology of bentonite–fresh water mud with increasing WHB fraction shifts from Bingham-plastic to power law.
• Gel strength does not change with an increase in WHB concentration.
• The plastic viscosity increased dramatically at higher WHB concentrations.
• The yield point is marginally improved after adding WHB.
• Increasing WHB concentration did improve the apparent viscosity of aged bentonite–fresh water mud

Based upon our review of literature the major limitations of the existing literature can be listed as:
• Oil based fluid has not received proper attention in the past because research was concentrated on shallow water reservoirs.
• Rheological characterization of drilling fluid is still difficult because of limitations and expenses in conducting experiments at high temperature
and pressure.
• There is an inherent non-linear complexity present in flow behavior of drilling fluid. These need to be explored with computational modeling.
• Some flow instabilities could be visualized in the behavior of the mud at higher pressure applications to the mud at times of minimum wall slip.
The instabilities could be assumed to be as a result of instrumental error.

In our proposed thesis we plan to address the above mentioned limitations. And, in the present paper, some of our results in the forward direction
have been reported.

BASIC DESIGN AND TESTING OF DRILLING FLUIDS AT HP/HT

A. Basic fluid design

The hole cleaning and control of formation pressures are critical functions of the drilling fluid. These are achieved when the mud
pressure is sufficient to hold back pore pressure and the viscosity is sufficient to transport cuttings from the bit to surface. Viscosity
must also be sufficient to hold mud solids in suspension. The conventional calculations of down-hole pressure, which assume constant
drilling fluid properties, are both practical for day-to-day use and accurate enough for routine wells. The down-hole static pressures
are easy to calculate from mud weight measured at surface, while additional pressures due to circulation can be calculated using
established relationships between pump rate and drilling fluid rheological properties. Errors that result from ignoring variations in mud
properties are small in relatively shallow wells. However, in deeper offshore wells these are significant. In these settings, the mud
engineers should concentrate on formulating drilling fluid properties for maximum rates of penetration and optimal condition of the
hole. The reservoir formations can commonly withstand moderate overpressure before being fractured, which permits mud engineers
to add a comfortable safety margin when weighting the mud. However, mud properties vary with down-hole pressure and temperature,
affecting the accuracy of both surface measurements and down-hole estimations of mud weight and viscosity. In HPHT wells these
variations can be significant because of the limited safety margins available. The ability to predict these effects is critical to the
successful drilling of HP/HT wells. Though may be small but serious errors in computing the drilling fluid pressure at the reservoir
result from ignoring uncertainties due to either temperature or fluid properties.

It is important that the selection between drill with oil or water-base mud is made before detailed analysis of the mud properties is
attempted. Water-base muds (WBMs) include conventional bentonite muds and polymer systems, while oil-base mud can be an invert
emulsion, all-oil system, or synthetic oil-base. Normally, each has advantages in cost, environmental impact and drilling performance.
The stability (rheological) and fluid-loss control of oil-base muds at high temperatures is a clear advantage in HP/HT wells. The oil
based muds are stable to at least 450°F [230°C] in 16-hour laboratory tests. However, oil-base muds do have some significant
disadvantages under HP/HT conditions. E.g., the solubility of gas in the base fluid that makes kick detection more difficult. A gas
influx while drilling will go into solution and remain there without causing a significant increase in mud volume until it is relatively

166
near the surface. On coming out of solution, the gas rapidly increases in volume, requiring quick reactions to control the well. Also,
the thermal expansion of oil-base muds is higher than water-base muds, which can lead to pressurization of the annulus. The
conventional advantages and disadvantages of oil-base mud also apply to HP/HT wells, e.g. good protection against differential
sticking, reservoir protection and well-bore stability in shales, claystones and salt. Because of their severe environmental impact oil
(diesel) based muds are slowly losing their application domains to bio-degradable synthetic oil-base mud. The bio-degradable
synthetic oil-base mud has low toxicity that reduces the potential for environmental damage during its use. Recent research results
have shown its successful use up to temperatures 250°C so far. The important design issues of drilling fluids for HP/HT are listed in
Table 1. We have designed our drilling fluid for HP/HT. The composition of the drilling fluid is listed in Table 2.

Table 1: The important design issues of drilling fluids for HP/HT.

Property of the drilling fluid Desired performance in HP/HT wells


Plastic viscosity As low as it can be. Reason: To minimize ECD.
Yield stress and gel strength Intermediately sufficient. Reason: It should be sufficient to prevent sag, but should not be high as to cause
gelation or high surge and swab pressures.
HP/HT fluid loss As low as it can be. Reason: To prevent formation damage and risk of differential sticking.
HP/HT rheology Intermediately sufficient. Reason: in order to control sag, gelation and ECD.
Compressibility Intermediately sufficient. Reason: To account for down-hole pressures and minimize ECD.
Stability to contaminants Intermediately sufficient. Reason: In order to be stable in the presence of gas, brine and cement.
Gas solubility Intermediately sufficient. Reason: To avoid kick.
Stability to aging Intermediately high. Reason: To ensure that the properties do not change over time under either static or
dynamic conditions
Solids tolerance Intermediately high. Reason: To ensure insensitivity towards drilling solids.
Weighting Intermediately high. Reason: To ensure that it is being able to be weighted up rapidly if a kick
is taken.

Table 2: The composition of the designed drilling fluid.

S. no Components Description Quantity(g) % weight Specific gravity Volume (in


cc)

1 Diesel (sp.gr=.85) Grade 2-D S15 196.775 42.6 .85 231.5


2 Water (sp.gr=1) Distilled 63.2 13.67 1 63.2
3 DFL Lube OBM 2 .43 .89 2.247
4 DWA VERSAWET 2 .43 1 2
5 VG-69 organophilic clay - 6.45 1.4 1.57 4.12
viscosifier and gelling agent

6 Lime ≥96 %, powdered (Ca(OH)2 ) 2 .43 2.24 .89

7 Calcium chloride ≥98 %, dehydrated, powder 22.3 4.83 2.15 10.37


8 Barite Barium sulphate 167.3 36.21 2.88 58.1
Total quantity: 462.025 g, total volume: 372.427 cc, and density of fluid = 1.32g/cc.

Brief test results

In the present work, three types of rheometers have been used: capillary rheometer, rotational rheometer and physica rotational rheometer. The fluid
described in Table 2 has been tested under different conditions. The analysis of the designed fluid is done at different temperatures such as room,
40.8, 52.2 degree C. The combined results from both capillary and rotational Rheometer are presented briefly.

167
Fig 3: The apparent viscosity variation with respect to apparent shear rate in rotational rheometer at different temperatures.

From Fig 3, we can see that the variation in the viscosity at the low shear rate with the temperature is not remarkable due to the active participation of
Vander Waal forces between the emulsified particles. The decrease in the viscosity with the increase in the temperature is observed. From Fig. 5, we
can see that the viscosity variation with the temperature at the high shear rate is remarkable as at high shear rate the Vander Waal forces are
superseded by the temperature effects. It can be seen that with the temperature increase there is a drop in the viscosity values. To determine the wall
slip and the exact gap thickness to eliminate it, the test has been performed on sample placed in the rotational Rheometer with different gap thickness
starting with .5mm and going up to 2.5mm.It is found from the graph that the sample tested with lower gap thickness possesses low viscosity than the
larger gap thickness .This suggests the existence of wall slip .Wall slip reduces the shear rate and viscosity and give rise to large velocity gradient in a
very thin layer close to wall boundary. Fig. 6 is indicative of the presence of the wall slip. When used with plates of less gap thickness, it can be
clearly seen that the plots with high gap thickness almost merge with other indicating that the influence of the wall slip is less at high gap thickness. It
is clear in Fig. 7 that the viscosity of the sample is tending to merge at higher gap thickness. Wall slip is negligible beyond 2.5mm.Thus we can
perform all the experiment with the sample in the Rheometer with the gap thickness of 3 mm. We compute the true viscosity from the measurements
in rotational rheometer (Mooney correction), and the true viscosity is computed after calculating the true strain.

Fig 4: The apparent viscosity variation with respect to apparent shear rate in capillary rheometer at different temperatures.

Fig. 5: The shear stress variation with respect to apparent shear rate in plate plate geometry Rotational rheometer for different gap thickness
(Mooney correction).

168
Fig 6: The torque variation with respect to apparent shear rate in plate plate geometry rotational rheometer for different gap thickness (Mooney
correction).

Fig. 7: The viscosity variation with respect to wall shear stress in plate plate geometry rotational rheometer for different gap thickness
(Mooney correction).

Fig 8: The apparent strain variation with respect to (1/gap) in plate plate geometry of rotational rheometer (Modified Mooney).

169
Fig 9: The true viscosity variation with respect to wall shear stress in plate plate geometry of rotational rheometer.

The true viscosity from the slope of the apparent strain v/s (1/gap2) plots corresponding to different wall shear stress values has been computed and
listed in Table 3. From Fig. 8, the slope for the individual plot corresponding to a shear stress is found, e.g. slope = 2 * slip velocity, and true shear
rate = apparent shear rate - 2 * slip velocity / gap height. From Fig. 9, it is clear that the viscosity reduces with an increase in the shear stress and the
shear rate. The viscosity reduces till it becomes constant which is a representative of Herschel Bulkeley model. The rise at the bottom of the graph in
the above plot is indicative of shear thinning. The above plot consists of the true viscosity which is computed from the true shear rate and wall shear
stress.

Table 3: The computed true viscosity from the slope of the apparent strain v/s (1/gap2) plots corresponding to different wall shear stress values.

The couette geometry of the rotational rheometer is used in the measurement of the yield stress of the test fluid. If a tangent is drawn to the plateau
then the line would intersect on the stress axis at a stress of about 0.25 Pa. 0.25 Pa is the yield stress. At about 0.25 Pa yield stress the fluid breaks its
gel structure and starts flowing as the free flowing continuous fluid. When the applied stress is less than critical stress, the yield stress, they do not
flow but deform plastically like a solid with definite strain recovery upon the removal of stress but when the yield stress is exceeded the fluid flows
like a truly viscous material with finite viscosity. Drilling fluids exhibit the non-Newtonian, shear thinning and visco-plastic behavior. They also
exhibit the time dependent behavior by showing an increase in viscosity with time once the stress is removed and decrease in viscosity with time
under constant stress. This property of drilling fluids is known as thixotropic. From Fig. 11, the yield stress can be identified as 0.5 Pa because the in
the enlarged plot it can be seen that the viscous flow begins at 0.5 Pa once the gel strength is over come by the application of yield stress. Fig. 12
demonstrates the presence of the yield stress as the lower stress limit at which the viscosity tends to infinity. We can infer from the Fig. 12 that the
stress at which the viscosity tends to infinity is 0.5 Pa. Thus we could describe that 0.5 Pa is the yield stress. Fig. 13 describes the variation of the
strain rate with the increase in the wall shear stress. The change in the slope can be witnessed beyond the slight disturbance .The point where the
slope change corresponds to the yield stress. The value of the yield stress is 0.5 pas.

170
Fig 10: The wall shear stress variation with respect to apparent strain rate in couette geometry of rotational rheometer.

Fig 11: The wall shear stress variation with respect to apparent strain rate in plate plate geometry of rotational rheometer with direct data
extrapolation.

Fig 12: The apparent viscosity variation with respect to wall shear stress in plate plate geometry of rotational rheometer with logarithmic plot of
apparent viscosity and wall shear stress.

171
Fig 13: The apparent shear rate variation with respect to wall hsear stress in plate plate geometry of rotational rheometer by
interpretation from strain rate and shear stress graph.

BRIEF DESCRIPTION ABOUT THE COMPUTER MODEL

The experimental studies are costly and time consuming. Furthermore, HP/HT conditions are difficult in laboratory settings and upper limits of the
testing ranges in the present HP/HT testing facilities are limited because of hardware settings. This makes predictive and forecasting tools essential in
HP/HT testing. Even after assuming that the HP/HT tests are complex nature and have highly non-linear characteristics in test results, we can explore
simulation of the test results with modern soft computing tools. And, with significant research into the simulation of HP/HT test parameters, not only
we can get the trends right but also the exact behavior over the short span in the advanced ranges of HP/HT.

We present a method in which a combined methodology driven by neural networks, fuzzy logics, and genetic algorithms (Haykin (1998), Ross
(2000), and Chambers (2004)) is used to develop a predictive and forecasting model for HP/HT test results. We use neural networks (NN) to predict
testing conditions at some future desired span of time (i.e. in our applications five steps). In neural networks the back propagation neural networks
are utilized. The output from the neural network model is input to a Weighted Fuzzy Petri Net based decision modeler (WFPNBDM). Another input
to the WFPNBDM is from the expert’s analysis dealing with the impact analysis of physical, chemical, and other non-specified conditions at a future
date. The WFPNBDM is designed using fundamentals from the theory of Petri nets, and genetic algorithms (GA). The present forecasting model is
shown in Fig. 14. To briefly show how the present forecasting model works; let us assume that at time t we want to predict some parameter from
HP/HT test results at some future value of time t + i + j . This time period (i.e. t + i + j - t ) is user specified, dependent upon the time structure of
the systems. The desired and relevant data is observed at time t , then the neural networks produce the outputs ptu+ i , ptv+ i , ……. ptw+ i , which are the
inputs to the WFPNBDM. The other input to the WFPNBDM are inputs from expert’s analysis at t + i . The WFPNBDM contains a base of decision
rules in a linguistic manner, and these are called fuzzy rules. The inputs to the WFPNBDM are translated into linguistic terms via a procedure of
fuzzification, and then these are processed using the fuzzy rules, fuzzy logic inference. In the next step, via a defuzzification and firing procedure a
quantitative prediction of the selected HP/HT test parameter is made at time t + i + j . We investigate the forecasting of few chosen HP/HT test
parameters. As shown in the Fig. 1, the starting point is a series of neural networks that are used to forecast each of the HP/HT test variables in the
chosen range. These networks are indicated as NN u - NN v - NN w where u + v + w is the number of economic variable that are selected for the

chosen range. Each of these neural networks has a number of inputs x(u1,,or...,vn,uor,orw nv ,or nw ),t which would decide the structure of the neural networks.
The inputs to the networks are at a time t , and they correspond to the values of those variables at time t on which HP/HT test variables for chosen
) corresponds to the values of the HP/HT test variables at time t + i where i refers to the
u ,or v ,or w
range depend. The output of these networks (i.e. p t +i
future period of time for which the forecasting is done. These serve as inputs to the WFPNBDM along with inputs from analysis of the experts. The
WFPNBDM follows an algorithm which then results in the final values of HP/HT test variables.

Normally, a neural network (i.e. precisely artificial neural network) is a layered, and hierarchical structure, and it consist of input layers, output
layers, and hidden layers which are located between these input and output layers. Each of these layers has a number of processing elements that are
called neurons (or nodes/units). The neurons are interconnected via signal paths with weights; receive their inputs either from outside of the network
via neurons in the input layers, or from other neurons via neurons in the hidden layers or output layers. After receiving the inputs the neurons
compute their outputs by their transfer or activation functions, and these outputs are sent either to other neurons in the networks or to the output of
the network layers as per the demands of network structure. We use feed forward neural networks (Haykins (1998)) that are trained by the supervised
learning back propagation algorithm. In the feed forward neural network, we use the ‘gradient descent’ technique (Haykins (1998)) to optimize the
weights of the neural network in a backward, and layer-by-layer manner so that the vector of the neural network outputs for a vector of training
inputs achieves the desired outputs. The feed forward neural network is shown in Fig. 15 (Haykins (1998)). It has been observed that more complex a
system becomes traditional mathematical methods are less able to model it; whereas humans can describe and deal with a complex situation with
more ease using natural language and imprecise thinking (ROSS (2004)). This ability of natural language and imprecise thinking motivated the

172
‘fuzzy sets’, and ‘fuzzy logic’. As per fuzzy set theory an element can belong to set 1 and set not-1 at the same time, and this allows ambivalence in
fuzzy set theory. Fuzzy set theory is a response to the insufficiency of Boolean algebra in which only ‘0’ and ‘1’ are recognized. In fuzzy set theory
everything is a matter of degree. Fuzzy logic is a set of reasoning with fuzzy sets, fuzzy truths, fuzzy connectives, and fuzzy rules of inference.
Fuzzy sets tries to emulate human reasoning in a natural, logical, systematic, and mathematical way. Therefore, they are able to mimic human
thinking and decisions very well. In the world of business, most of the information that is essential for decision making is imprecise, and ambiguous
(i.e. for example market penetration possibility is expressed as likely, very likely, and unlikely, etc.). As any other competitive business, the shipping,
and shipbuilding markets are full with fuzzy information, and fuzzy reasoning. It is appropriate to use fuzzy sets whenever we have to deal with
ambiguity, vagueness and ambivalence in mathematical or conceptual models of empirical phenomena.

Fuzzy logic models linguistic terms mathematically, and maps the numerical (i.e. crisp) onto verbal (i.e. fuzzy), and vice versa. Therefore it builds a
bridge between the two sets of completely different domains or reach (i.e. world). Because of this fuzzy logic is better suited to model and analyze
economic and social systems. A fuzzy logic based system involves four major operations: fuzzification, activation of fuzzy rules, fuzzy inference,
and defuzzification and firing process. Normally, the inputs and output variables of the fuzzy system are called fuzzy variables, and the collection of
all sets to which a fuzzy variable can belong is set called the universe of discourse. The universe of discourse is covered by overlapping fuzzy sets
that are defined by fuzzy membership (i.e. truth value) functions. The degree of membership (i.e. truth value) that an element belongs to a certain
fuzzy set is determined by the corresponding fuzzy membership function. In general fuzzy sets are marked with intrinsic linguistic terms or words.
The crisp numerical inputs are converted into linguistic terms (i.e. fuzzy sets) along with their truth values based on the fuzzy membership functions
via fuzzification. We use trapezoidal fuzzy membership functions for decision variables. The fuzzy rule bank is a collection of fuzzy rules which
contain a series of linguistic expressions. Normally, fuzzy rules are in the form of IF − THEN statements that interpret the linguistic input
information obtained from the process of fuzzification, and then provide linguistic output information.

Following Ross (2004), and assuming that there is a fuzzy production rule Ri appearing in the knowledge base of a rule based system which is
shown as follows:
(
Ri :IFd j and d k THEN CF µi , and ) (1)

IF X 1 is A1 AND X m is Am THEN Y is C , (2)


where d j ( X 1 is A1 ), and d k ( X m is Am ) are propositions, µi is a fuzzy number defined in the universe of discourse [ 0,1] indicating the
certainty factor value of the rule Ri , and the weights of the propositions d j , and d k are ŵ j and ŵk respectively. Additionally, in Equations (1)

and (2) ŵ j and ŵk are fuzzy numbers defined in the universe of discourse [ 0,1] , and Ai and C are fuzzy sets (i.e. fuzzy values). Assuming that

the fuzzy truth value of the proposition d j is tv j , where tv j is a fuzzy number defined in the universe of discourse [ 0,1] . Then, the fuzzy truth

value of the proposition d k is equal to tv j ⊗ µi , where “ ⊗ ” is the multiplication operator of fuzzy numbers. Equations (1) and (2) are very basic
and all other types of fuzzy rules (i.e. for example rules involving OR , NOT , LOGICAL OPERATORS like IF − BUT − THEN , and multiple
consequents, etc.) can be re-converted into the forms of Equation (1) and (2). The fuzzy rule bank is expressed as a linguistic matrix for numerical
implementation. If the linguistic input information is available then the fuzzy rules are determined as per the linguistic fuzzy rule matrix. In our
applications we assume four different basic fuzzy rules for HP/HT test variables. The logical and numerical operations are implemented on the
activated fuzzy rules via the process of inference, to determine their levels of activations. The levels of activations are determined according to the
truth values of the fuzzy sets bounded by the activated fuzzy rules, and result into clipped output fuzzy sets. In this work, we use ‘correlation-
minimum inference’ (i.e. simply cutting off the top of the fuzzy set, and using only that area of the set which lies below the user specified current
degree of belief (DoB). The low belief in a fuzzy concept implies that the true discrete value lies outside the supremum region. The supremum region
expands as belief decreases from maximum to minimum (i.e. for implementation 1 – 0) until it equals the support of the fuzzy set. Since, in our
applications, we wish to retain the impact of a set even if its degree of belief (DoB) is decreasing, and assume that the percent area of the set is
notably higher than its DoB always; therefore correlation-minimum inferencing is better suited for our applications. The output fuzzy sets (i.e.
clipped fuzzy sets) are demystified and converted into the crisp and numerical outputs via the process of defuzzification and firing. In this work, we
use ‘area-centroid weighting’ (i.e. simply cutting off the top of the fuzzy set, and using only that area of the set which lies below the user specified
current degree of belief (DoB) for defuzzification and firing. In our applications we use symmetrical fuzzy sets (i.e. trapezoids), and the centroid
position is independent of DoB and inferencing Method. The centroid is always located at the mid-point of the supremum range. And, this allows the
supremum position (i.e. average position) to be used instead of calculating the centroid on each cycle through the controller which is computationally
expensive. Furthermore, we consider the fuzzy set as a singleton. After deciding the input and output variables of the system to be modeled the
Weighted Fuzzy Petri Net based decision modeler (WFPNBDM) is designed. Following, Chen (2002), and using the fundamentals of Petri nets and
fuzzy Petri nets, we use a generalized Weighted Fuzzy Petri Net structure (WFPNS), as,
WFPN = ( P,T ,D,I ,O, f ,α , β ,W ) , (3)
where,
P = { p1 ,..., pn } is a finite set of variables, (4)
T = {t1 ,...,tm } is a finite set of transitions, (5)
D = {d1 ,...,d n } is a finite set of propositions, (6)
P ∩T ∩ D = φ, P = D , (7)

173
I : T → P ∞ is the input function, a mapping from transitions to sets of variables, (8)
O : T → P ∞ is the output function, a mapping from transitions to sets of variables, (9)
f : T → fuzzy numbers in [0,1] is an association function, a mapping from transitions to fuzzy numbers defined in the universe of discourse
[0,1] , (10)
α : T → fuzzy numbers in [0,1] is an association function, a mapping from variables to fuzzy numbers defined in the universe of discourse
[0,1] , (11)
β : P → D fuzzy numbers in [0,1] is an association function, a bijective mapping from variables to propositions,
(16)
W : P → fuzzy numbers is an association function, a mapping from variables to fuzzy numbers. (12)

The design of WFPNBDM requires that the number of fuzzy sets, their associated membership functions for the input and output variables,
transitions for the input and output variables, associated weights for the input and output variables, and the specifications of the fuzzy rules are
determined efficiently. The WFPNBDM is designed so that it can model the system efficiently. The complexity of a WFPNBDM increases with the
unknown parameters. As mentioned previously, we use two layers of WFPNBDMs. It means that in the first layer there are three WFPNBDMs for
HP/HT test variables and in second layer there is one WFPNBDM for desired second stage HP/HT test variables. Genetic algorithms (GA) do not
require that the objective function is continuous, twice differentiable, and mathematically well formulated. Also, GAs operates on a domain (i.e.
population) of potential solutions (i.e. individuals, or chromosomes) at each step of iteration and do not evolve a single solution. Additionally, GAs
takes a pre-conceived random search strategy and ‘a priori’ knowledge about the searching space where the solution will lie is not required. Since,
GAs operate on a domain of potential solutions and search randomly, so they may miss the local optimal solutions, [12]. Because, of this limitation
GAs are better suited for higher dimensional, ill defined, and nonlinear problems. To automate the simultaneous identification of both the fuzzy
membership functions and fuzzy rules, we use genetic algorithm (GA) driven optimization. The details of the algorithm can be found in Chambers
(2000).

u
x1,t Input from expert’s analysis for conditions at time
Neural t +i .
x2u,t network
ptu+ i
( NN u ) Crisp input
for the
xnuu ,t Weighted Crisp output
Fuzzy Petri Fuzzification D and F process from the
Net based Weighted Fuzzy
v
decision Activated Fuzzy Petri Net based
x1,t modeler sets and Clipped Fuzzy
Neural decision
v activated truth membership
x 2 ,t network value functions
modeler
ptv+ i
Inputs for HP/HT test parameters

( NN v )
Fuzzy rule Activated Fuzzy Fuzzy
xnvv ,t
Prediction of HP/HT test parameters

rules
bank inference
w1 Neural Weighted Fuzzy Petri Net based Market
x1,t decisions
network decision modeler
x2w,t1 ptw+1i yt + i + j
( NN w1 ) (Two layers)
Basic nomenclature
xnww1
1
,t ƒ x(u1,...,n
,or v ,or w
,or n ,or n ) ,t
= input parameters to neural
u v w
w2
x1,t Neural networks at time t .
network ptw+2i
( NN w2 ) ƒ ptu+,or
i
v ,or w
= prediction of neural networks at time
w2
x nw ,t t +i .
2
ƒ yt + i + j = output produced by WFPNBDM at time
w3
x1,t Neural
network ptw+3i t +i+ j .
x w3 ( NN w3 ) ƒ D and F process: Defuzzification, and firing
nw ,t
3 process.
Fig 14: The structure of the integrated forecasting system.

174
Output target Output
input layer hidden layer hidden layer output layer output layer
input neurons weights neurons weights neurons vector OVO target
vector
vector ∑ / f( ) ∑ Error
IVP calculation OVT
between
OVO and
∑ / f( ) ∑ OVT

Back-
propagation
∑ / f( ) ∑ learning
algorithm
W1 W2
Weights adjustment

Fig 15: The feed forward neural network.

Once the HP/HT test parameters have been predicted, our idea is use them to optimally design mud rheology and hydraulics. Our proposed
optimization procedure has been derived from inputs from various industries (e.g. Halliburton USA, Dowell Tunisia and Baker Hughes, USA.) and
the optimization model is shown in Fig. 16 and Fig 17.

Fig 16: The optimization model for drilling fluid properties (1 - low conditioning requirements, 2 - low ECD and risk of fracture, 3 - high rate of
penetration, 4 - minimized differential sticking, 5 - minimized sagging, 6 - good hole cleaning, 7 - maximum prevention of kicks, 8 - good hole
circulation, stability and washouts minimized, and 9 - increased hydraulic horsepower at bit), adapted and derived from inputs from various
industries (e.g. Halliburton USA, Dowell Tunisia and Baker Hughes, USA).

175
Fig 17: The optimization model for drilling mud design, adapted and derived from inputs from various industries
(e.g. Halliburton USA, Dowell Tunisia and Baker Hughes, USA)

CONCLUSIONS

In this paper, we have presented a conceptual development of a computer model for rheological analysis of oil based drilling fluids at HT/HP. Our
presented model is an integrated predictive forecasting model that is designed around a rheological characterization. The predictive forecasting model
presented in this work uses a combined methodology driven by neural networks, Weighted Fuzzy Petri Net based logics, and genetic algorithms, and
the integrated predictive forecasting model will be used to simulate the HP/HT tests. The application of neural networks allows the detection and
learning of nonlinear relationships amongst the various HP/HT test variables and parameters. The application of Weighted Fuzzy Petri Net based
logics and genetic algorithms allow the incorporation of expertly defined parameters. Hence, the HP/HT test simulation with integrated predictive
forecasting model can be representative of real world HP/HT test situations to a large degree.

However, in this work we have neither attempted a generalized simulation model for HP/HT tests, nor we have considered all the parameters in the
integrated predictive forecasting model. A more detailed integrated predictive forecasting model considering specific parameters relevant to one
particular test fluid can be an obvious extension of the present work. Also, a detailed sensitivity analysis of the integrated predictive forecasting
model incorporating different parametric values of HP/HT test variables has not been done. Each of the ideas presented in this work can be further
explored ‘in-detail’ to develop a comprehensive long-range HP/HT test simulation. Our future work shall go in this direction, and currently this is
under investigation.

ACKNOWLEDGEMENTS

The first author is supported by the DAAD (Deutscher Akademischer Austasch Dienst) scholarship for the academic year 2010-2011, and she is
presently DAAD fellow at Karlsruher Institut für Technologie, Germany. The experimental work is being done at Institut für Mechanische
Verfahrenstechnik und Mechanik, KIT, Germany. The authors thank all the staff members of Institut für Mechanische Verfahrenstechnik und
Mechanik, KIT, Germany for their help, support and encouragement in the experimental studies. Some of the data used in the present work has been
supplied by our industrial research collaborators (e.g. Halliburton USA, Dowell Tunisia and Baker Hughes, USA). The authors gratefully
acknowledge their support and encouragement for the present research.

REFERENCES

Bailey, W. J. and Weir, I. S. (1998), “Investigation of methods for direct rheological model parameter estimation”, Journal of Petroleum Science
and Engineering, Vol 21, pp. 1–13.
Berthezene, N., De Hemptinne, J.-C., Audibert, A. and Argillier, J.-F. (1999), “Methane solubility in synthetic oil-based drilling muds”, Journal
of Petroleum Science and Engineering, Vol. 23, pp. 71–81.
Caenn, R. and Chillingar, G. V. (1996), “Drilling fluids: State of the art”, Journal of Petroleum Science and Engineering, Vol. 14, pp. 22l-230.
Chambers, L. D. (2000), “The Practical Handbook of Genetic Algorithms: Applications”, 2nd edition, Chapman and Hall/CRC, USA, pp 67–78.
Chen, S.-M. (2002), “Weighted Fuzzy Reasoning Using Weighted Fuzzy Petri Nets”, IEEE Transactions on Knowledge and Data Engineering,
Volume 14, no. 2, pp 386-387.
Hamida, T., Kuru, E. and Pickard, M. (2004), “Rheological characteristics of aqueous waxy hull-less barley (WHB) solutions”, Journal of
Petroleum Science and Engineering, Vol. 69, pp. 163-173.
Haykin, S. (1998), “Neural Networks: A Comprehensive Foundation”, 2nd edition, Prentice Hall, USA, pp 55–67.
Maxey, J. (2009), “Non aqueous fluid rheology at elevated temperature and pressure”, Society of Petroleum Engineers, USA.
Nguyen, Q. D. and Boger, D. V. (2010), “Measuring the flow properties of yield stress fluids”, Annu. Rev. Fluid Mech.
Peterson, J. L. (1981), “Petri Nets Theory and Modeling of Systems”, Prentice Hall, Englewood Cliffs, N.J., USA, 1981.
Ross, T. J. (2004), “Fuzzy Logic with Engineering Applications”, 2nd edition, John Wiley and Sons, USA, pp 132–147.
Skalli, L., Buckley, J. S., Zhang, Y. and Morrow, N. R. (2006), “Surface and core wetting effects of surfactants in oil-based drilling fluids”,
Journal of Petroleum Science and Engineering, Vol. 52, pp. 253–260.
Tiu, C., Guo, J. and Uhlherr, P. H. T (2006) “Yielding Behaviour of Viscoplastic Materials”, Vol 12, pp. 653- 662.

176
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Numerical Modeling of Fractured Hydrocarbon Reservoirs: An Overview


G. Suresh Kumar
Associate Professor, Petroleum Engineering Program, Department of Ocean Engineering,
Indian Institute of Technology - Madras, Chennai-36, India
E-mail : gskumar@iitm.ac.in

ABSTRACT

This paper provides an overview on the basics of conventional reservoir modeling in general followed by modeling requirements associated with
fractured hydrocarbon reservoirs. General mathematical issues involved in solving a set of partial differential equations used to describe the migration
of hydrocarbons in the subsurface system have been discussed in detail under conventional modeling, while reservoir simulation modeling addresses
the modeling concepts pertaining to reservoir simulation of fractured hydrocarbon reservoirs.

KEY WORDS: Modeling; Fractured hydrocarbon reservoir; Conventional modeling; and Reservoir simulation modeling

INTRODUCTION

Almost all hydrocarbon reservoirs are affected by fractures (Stearns and Friedman, 1972), which may either enhance or suppress the reservoir
permeability (Tillman, 1983; Lorenz et al., 1988). Thus, it is widely acknowledged that fractures have an important role in the filtration of
hydrocarbons in the reservoir although the main pore volume belongs to the low permeability rock-matrix with micro-fractures. The oil recovery
mechanisms from such low-permeability rock matrix are essentially determined by the multi-phase reservoir fluid properties namely wettability and
capillarity. As a consequence, the ability to numerically model multi-phase fluid flow in a fractured hydrocarbon reservoir, characterized by a low-
permeability rock matrix, combined with, a high permeability fracture system is extremely important in the estimation of production rate of oil and/or
gas. Therefore the knowledge of multi-phase fluid flow within the high permeability fracture, and its associated fluid mass transfer at the fracture-
matrix interface is very critical in assessing the sweep efficiency of the reservoir. At the outset, the most striking hydrogeologic feature of a fractured
hydrocarbon reservoir is the overwhelming variability of the fundamental reservoir rock properties namely porosity and permeability. In fractured
reservoirs, the rocks themselves do not transmit the hydrocarbon fluids but it is the high permeability fractures that form preferential pathways through
the basically impervious rock-matrix (Gustafson and Krasny, 1993). These fractures commonly control the hydraulic responses of hydrocarbon
reservoirs, and eventually determine the production rate of hydrocarbon oil and/or gas. In order to characterize such fractured reservoirs, and, to better
understand the fluid flow and its associated processes, the data on the distribution of fracture aperture length, fracture aperture width, fracture spacing,
fracture density, fracture orientation and fracture connectivity are required, in addition to the details on the flow characteristics, asperities, degree of
channeling and the effect of fracture skin, for modeling purposes. It is to be noted that all the above details may not remain a constant over a period of
time, while the details on its variation with increasing in situ stress, may be needed for modeling purposes. In view of such reservoir rock
complexities, the concept of Representative Elementary Volume (REV) for the characterization of hydrocarbon reservoirs is extremely difficult. In
general, fractured hydrocarbon reservoirs are conceptualized to consist of two homogeneous systems of markedly varying porosity (dual-porosity) and
permeability (dual-permeability). In such systems, ensuring the continuity of mass fluxes at the fracture-matrix interface is very critical, as the length
scale of fracture aperture thickness is extremely small, when compared with the length scale of fracture length, and as the fluid flow velocities within
the high permeability fractures are orders of magnitude higher with respect to the fluid flow velocities in its associated low-permeability rock matrix.
For example, a finite thickness fracture aperture is represented as a single line in the case of one-dimensional finite difference modeling. Such
numerical models do not precisely estimate the pressure drop or absolute permeability distribution of single phase fluid flow, along the flow direction,
when either the cell width along the fracture (Dx) or the time step (Dt) is significantly increased.

The general equation used to describe the fluid flow in a coupled fracture-matrix system is Navier-Stokes Equation (NSE), which considers mass and
momentum conservation over the fracture aperture thickness (Brush and Thompson, 2003). The NSEs form a non-linear system of partial differential
equations that are difficult to solve for undulated fracture walls. In order to convert this complex non-linear system of PDEs into a simpler linear
system, NSE is converted into Stokes equation by assuming the inertial forces within the high permeability fracture to be much less than the pressure
and viscous forces. Further, if the variability of undulated fracture walls is assumed to be gradual (then the component of fluid velocity perpendicular
to the direction of fluid flow will be approximately zero), and also, if the fracture walls are assumed to be parallel to the flow direction and
perpendicular to the vertical axis, then, the three-dimensional flow field can be approximated with a two-dimensional description, which eventually
leads to Local Cubic Law (LCL). Later, the effect of fracture surface roughness was incorporated by Singhal and Gupta (1999). Thus, Local Cubic law
has been well established to describe the single phase fluid flow in smooth parallel walled fractures (Witherspoon et al., 1980). In addition, a set of
well established non-linear PDEs for wetting and non-wetting fluid phases to describe the multi-phase fluid flow in fractures are available. It is to be
noted that a new driving force is introduced in the form of capillary force at the immiscible fluid interface in the case of multi-phase fluid flow, while
the pressure forces due to the pressure differences and gravitational forces only drives the single phase fluid flow (Tatomir, 2007). However, in such
multi-phase systems, the non-linear dependence of the saturation on the capillary pressures and on the relative permeabilites complicates the numerical
solution under varied saturation, depending on (a) pressure formulation (having pressures as unknowns); (b) pressure-saturation formulation (having
the pressure of the fluid with the highest affinity and the saturation of the other phase as unknowns); or (c) saturation formulation (having the phase
saturations as unknowns). For example, the numerical results of Pruess and Tsang (1989, 1990) showed a strong phase interference, with the sum of
effective permeabilities to be much less than unity at intermediate saturations. Further, atleast the author is not clear whether the oil production results
from spontaneous imbibition of water into low-permeability matrix blocks from the high permeability fractures, followed by a subsequent
displacement of the expelled oil through the fractures, or due to the presence of capillary continuity within the low permeability rock matrix that leads
to the oil expulsion by viscous displacement, and, numerical modeling for simultaneous happenings of these two processes becomes a challenging
task. The problem leads to further complexity for multi-phase hydrocarbon fluid flow, as the wetting phase of the hydrocarbons along the fracture
walls significantly affects the fluid mass transfer at the fracture-matrix interface, and numerical modeling of multi-phase fluid flow across fractures
between low-permeability reservoir rock-matrix yields a complex coupled partial differential equations. In addition, the conventional assumption of
symmetricity applied for a coupled fracture-matrix dual-porosity system may not be always valid as the imbibitions and drainage characteristics alter
the mass transfer coefficient over spatially and temporally. On the other hand, Bonnet et al., (2001) used a fractal (a set without a characteristic length
scale) model to describe the fractured reservoirs, which cover a whole range fracture length scales. However, in such models, the capturing of mass
exchange between high permeability fracture and low permeability rock matrix becomes difficult to model numerically. In particular, the models for
petroleum reservoir simulation have provided reasonably reliable solutions to the very complex petroleum reservoir conditions, when it comes to the
question of prediction, although material balance approach provides good results at history matching production performance. Numerical simulators
can be comfortably referred to as multi-dimensional, multi-phase transient material balance programs. As against analogy and volumetric methods of
reserve estimation, numerical simulations require a large amount of data to predict the reservoir performance (say, production rate, oil recovery rate)
under a variety of primary/secondary/tertiary producing/recovery mechanisms. Thus, the numerical simulators has the ability to model the reservoir
development conditions, well before the field implementation, and eventually, assists in deducing an optimal rate of hydrocarbon production/recovery.
In essence, reservoir simulation acts as a greatest artifact in petroleum reservoir management aspects as the numerical simulation results help to
maximize the economic recovery of hydrocarbons, while the same helps to minimize the operation expenses. The passion towards reservoir simulation
builds up as the knowledge on mathematics, physics, reservoir engineering and computer programming enhances. The main advantage of reservoir
simulation being, a model can be run several times without any appreciable expense, and that too over a short period of time, while a real time field
performance conducted in-situ can be produced only once with a heavy initial investment. However, it is to be noted that any reservoir numerical
simulation requires a tedious task of processing a huge amount of data required to be fed as input, in order to achieve a higher resolution as well as
precise predictions. The general procedure towards reservoir simulation modeling involves (a) building up of a reservoir model at a local scale from
the available data, and scaling up the same to a large scale simulation model using computational grids; (b) selection of reservoir development and
operation strategies; (c) modeling techniques; and finally, (d) predicting the reservoir performance. Thus, the purpose of the present paper is to review
the basics of conventional reservoir modeling in general, followed by the basics of reservoir simulation modeling with its associated complexities in a
fractured hydrocarbon reservoir.

Conventional Modeling

During a mathematical modeling process, two sets of information are essential for correct and accurate simulation of hydrocarbon fluid flow. First,
there is the number of parameters (variables) involved in the process; second, there is accurate knowledge of the interrelationships (no matter how
complex and nonlinear) between different parameters. In a conventional reservoir modeling, the major physical events are related to the action and
interaction of various physical/chemical/biological processes, representing the number of parameters involved. On the other hand, the mostly non-
linear interrelationships between these different parameters necessitate the numerical solution of the associated partial differential equations as
analytical solutions often do not exist for the same. However, such numerical methods introduce an approximation resulting from truncation errors and
subsequently emphasis on grasping the concepts of stability, convergence, and consistency. Thus the major differences from considerations and
procedures for constructing numerical models of a conventional hydrocarbon reservoir system are in the mathematical approximations that arise in
representing different contributing physical processes, the strength of the coupling among theses processes, the input data and finally on the scale
effects (Oran and Boris, 2001). In general, a model is a simplified representation of a real system. A conceptual model is a hypothesis for how a
system or process operates. This hypothesis can be expressed quantitatively as a mathematical model. Mathematical models are abstractions that
represent processes as equations, physical properties as constants or coefficients in the equations, and measures of state or potential in the system as
variables. Most of the reservoir simulation models in use today are deterministic mathematical models. Deterministic models are based on
conservation of mass, momentum, and energy and describe cause and effect relations. The underlying assumption is that given a high degree of
understanding of the processes by which stresses on a system produce subsequent responses in that system, the system’s response to any set of stresses
can be predetermined, even if the magnitude of the new stresses falls outside of the range of historically observed stresses (Konikow, 1996; Konikow
and Reilly, 1998).

Deterministic reservoir models generally require the solution of partial differential equations. Exact solutions can often be obtained analytically, but
analytical models require that the parameters and boundaries be highly idealised. Numerical methods yield approximate solutions to the governing
equation (or equations) through the discretisation of space and time. Within the discretised problem domain, the variable internal properties,

178
boundaries, and stresses of the system are approximated. Deterministic, distributed-parameter, numerical models can relax the rigid idealised
conditions of analytical models or lumped-parameter models, and they can therefore be more realistic and flexible for simulating field conditions (if
applied properly). The number and types of equations to be solved are determined by the concepts of the dominant governing processes. The
coefficients of the equations are the parameters that are measures of the properties, boundaries, and stresses of the system; the dependent variables of
the equations are the measures of the state of the system and are mathematically determined by the solution of the equations. When a numerical
algorithm is implemented in a computer code to solve one or more partial differential equations, the resulting computer code can be considered a
generic model. When the grid dimensions, boundary conditions, and other parameters are specified in an application of a generic model to represent a
particular geographical area, the resulting computer program is a site-specific model. The ability of generic models to solve the governing equations
accurately is typically demonstrated by example applications to simplified problems. This does not guarantee a similar level of accuracy when the
model is applied to a complex field problem. If the user of a model is unaware of or ignores the details of the numerical method, including the
derivative approximations, the scale of discretisation, and the matrix solution techniques, significant errors can be introduced and remain undetected.

The partial differential equations describing hydrocarbon fluid flow can be solved mathematically using either analytical solutions or numerical
solutions. The advantages of an analytical solution, when it is possible to apply one, are that it usually provides an exact solution to the governing
equation and is often relatively simple and efficient to obtain. In general, obtaining the exact analytical solution to the partial differential equation
requires that the properties and boundaries of the flow system be highly and perhaps unrealistically idealised. For most field problems, the
mathematical benefits of obtaining an exact analytical solution are probably outweighed by the errors introduced by the simplifying assumptions of the
complex field environment that are required to apply the analytical model. Alternatively, for problems where the simplified analytical models are
inadequate, the partial differential equations can be approximated numerically. In so doing, the continuous variables are replaced with discrete
variables that are defined at grid blocks (or nodes). Thus, the continuous differential equation, which defines the spatial/temporal distribution of fluid
flow everywhere in the system, is replaced by a finite number of algebraic equations that defines the fluid pressure at specific points. This system of
algebraic equations generally is solved using matrix techniques. This approach constitutes a numerical model. The equations describing multiphase
hydrocarbon fluid flow processes are second-order differential equations, which can be classified on the basis of their mathematical properties. There
are basically three types of second-order differential equations, which are parabolic, elliptic, and hyperbolic (Peaceman, 1977). Such equations can be
classified and distinguished based on the nature and magnitude of the coefficients of the equation. This is important because the numerical methods for
the solution of each type should be considered and developed separately for optimal accuracy and efficiency in the solution algorithm.
Two major classes of numerical methods have come to be well accepted for solving the hydrocarbon fluid flow equation. These are the finite-
difference methods and the finite-element methods. Each of these two major classes of numerical methods includes a variety of subclasses and
implementation alternatives. Both of these numerical approaches require that the area of interest be subdivided by a grid into a number of smaller
subareas (cells or elements) that are associated with node points (either at the centres of peripheries of the subareas). Finite-difference methods
approximate the first derivatives in the partial differential equations as difference quotients (the differences between values of variables at adjacent
nodes, both in space and time, with respect to the interval between those adjacent nodes). There are several advanced text books that focus primarily
on finite-difference methods (Peaceman 1977; Remson et al. 1971; Von Rosenberg 1969). Finite-element methods use assumed functions of the
dependent variables and parameters to evaluate equivalent integral formulations of the partial differential equations. In both numerical approaches, the
discretisation of the space and time dimensions allows the continuous boundary-value problem for the solution of the partial differential equation to be
reduced to the simultaneous solution of a set of algebraic equations. These equations can then be solved using either iterative or direct matrix methods.
Each approach has advantages and disadvantages, but there is very few hydrocarbon fluid flow problems for which either is clearly superior. In
general, the finite-difference methods are simpler conceptually and mathematically, and are easier to program for a computer. They are typically keyed
to a relatively simple, rectangular grid, which also eases data entry tasks. Finite-element methods generally require the use of more sophisticated
mathematics but, for some problems, may be more accurate numerically than standard finite-difference methods. A major advantage of the finite-
element methods is the flexibility of the finite-element grid, which allows a close spatial approximation of irregular boundaries of the reservoir
interfaces and (or) of parameter zones within the reservoir when they are considered. However, the construction and specification of an input data set
is much more difficult for an irregular finite-element grid than for a regular rectangular finite-difference grid. Thus, the use of a model preprocessor,
which includes a mesh generator and a scheme to efficiently number the nodes and elements of the mesh and to specify the spatial co-ordinates of each
node, is valuable to effectively utilize the advantageous features of a finite-element model. The numerical methods that work best for parabolic partial
differential equations are not best for solving hyperbolic equations, and vice versa. Thus, no one numerical method or simulation model will be ideal
for the entire spectrum of compositional oil model problems likely to be encountered in the field. Although finite-difference and finite-element models
are commonly applied to hydrocarbon fluid flow problems, other types of numerical methods have also been applied to such problems, including
method of characteristics, particle tracking, random walk, and Eulerian-Lagrangian methods. The numerical errors can generally be reduced by using a
finer discretization (either time steps or spatial grid).
The partial differential equations describing the hydrocarbon fluid flow include terms representing derivatives of continuous variables. Finite-
difference methods are based on the approximation of these derivatives (or slopes of curves) by discrete linear changes over small discrete intervals of
space or time. If the intervals are sufficiently small, then all of the linear increments will represent a good approximation of the true curvilinear
surface. However, it is to be noted that a certain error is involved in approximating the derivatives by finite-differences, but this error will generally
decrease as the cell width along x or y-directions is given smaller and smaller values. This error is called a “truncation error” because the replacement
of a derivative by a difference quotient is equivalent to using a truncated Taylor series, so that the exact solution of a difference equation differs from
the solution of the corresponding differential equation (Peaceman 1977). Also, it may not be possible to achieve an “exact” solution of the difference
equation because of limits of precision in storing numbers in a digital computer. In solving a large set of difference equations, many arithmetic
operations are performed, and round-off errors may sometimes accumulate.The finite-element method (FEM) is a numerical analysis technique for
obtaining approximate solutions to a wide variety of problems in physics and engineering. The method was originally applied to structural mechanics
but is now used in all fields of continuum mechanics. Huebner (1975) describes four different approaches to formulate the finite-element method for a
problem, which are: the direct approach, the variational approach, the weighted residual approach, and the energy balance approach (Huebner 1975).

179
The finite-element method (FEM) uses a concept of "piecewise approximation". The domain of the problem, that is the extent of the river to be
simulated, is divided into a set of elements or pieces. In theory, the elements can be of different shapes and sizes. Most FEM computer programs use
one shape element, most commonly either triangular or quadrilateral elements. Point values of the dependent variable (for example, depth, pressure, or
concentration) are calculated at nodes, which are the corners or vertices of the elements, and a simple equation is used to describe the value of the
dependent variable within the element. This simple equation is called a basis function and each node that is part of an element has an associated basis
function. The simplest basis functions that are usually used are linear functions. The solution to the differential equation for flow or transport is
approximated by a set of elements in which the dependent variable only varies linearly within the element, but the entire set of elements approximates
the complex distribution of depth or concentration.
The finite-difference and finite-element approximations lead to an algebraic equation for each node point. The set of algebraic equations may be
solved numerically by one of two basic methods: direct or iterative. In direct methods, a sequence of operations is performed only once to solve the
matrix equation, providing a solution that is exact, except for machine round-off error. Iterative methods arrive at a solution by a process of successive
approximation. They involve making an initial guess at the solution, then improving this guess by some iterative process until an error criterion is
satisfied. Therefore, in these techniques, convergence and the rate of convergence are of concern.

Direct methods (solution by determinants/solution by successive elimination of the unknowns/solution by matrix inversion) have two main
disadvantages. The first problem is one of computer resource requirements, including large storage (memory) requirements and long computation
times for large problems. The matrix is sparse (contains many zero values) and to minimise computational effort, several techniques have been
proposed. However, for finite-difference and finite-element methods, storage requirements may still prove to be unavoidably large for three-
dimensional problems. The second problem with direct methods is round-off error. Because many arithmetic operations are performed, round-off
errors can accumulate for certain types of matrices.
Iterative schemes avoid the need for storing large matrices, which make them attractive for solving problems with many unknowns. Numerous
schemes have been developed; a few of the more commonly used ones include successive over-relaxation methods, iterative alternating-direction
implicit procedure, and the strongly implicit procedure. Because iterative methods start with an initial estimate for the solution, the efficiency of the
method depends somewhat on this initial guess. To speed up the iterative process, relaxation and acceleration factors are used. Unfortunately, the
definition of best values for these factors commonly is problem dependent. In addition, iterative approaches require that an error tolerance be specified
to stop the iterative process. An optimal value for the tolerance, which is used to evaluate when the iterative calculations have converged on a solution,
may also be problem dependent. If the tolerance is set too large, then the iterations may stop before adequate numerical accuracy is achieved. If the
tolerance is set too small, then the iterative process may consume excessive computational resources in striving for numerical precision that may be
orders of magnitude smaller than the precision of the field data, or the iterative process may even fail to converge. More recently, a semi-iterative
method, or class of methods, known as conjugate-gradient methods, has gained popularity. One advantage of the conjugate-gradient method is that it
does not require the use or specification of iteration parameters, thereby eliminating this partly subjective procedure.To obtain a unique solution of a
partial differential equation corresponding to a given physical process, additional information about the physical state of the process is required. This
information is supplied by boundary and initial conditions. For steady-state problems, only boundary conditions are required, whereas for transient
problems, boundary and initial conditions must be specified.
Mathematically, the boundary conditions include the geometry of the boundary and the values of the dependent variable or its derivative normal to the
boundary. In physical terms, for reservoir flow model applications, the boundary conditions are generally of three types: (a) specified value (pressure),
(b) flux (corresponding to a specified gradient of pressure), or (c) mixed boundary condition, in which the flux across a boundary is related to both the
normal derivative and the value. The initial conditions are simply the values of the dependent variable specified everywhere inside the boundary at the
start of the simulation. Normally, the initial conditions are specified to be a steady-state solution. Regarding model design and development, the first
step in model design and application is to define the nature of the problem and evaluate the purpose of the model. Although this may seem obvious, it
is an important first step that is sometimes overlooked in a hasty effort to take action. This step is closely linked with the formulation of a conceptual
model, which again is required prior to development of a simulation model. In formulating a conceptual model, the analyst must evaluate which
processes are significant in the system being investigated for the particular problem at hand. Some processes may be important to consider at one
spatial or temporal scale of study, but negligible or irrelevant at another scale. The analyst must similarly decide on the appropriate dimensionality for
the numerical model. Regarding model calibration, uncertainty in parameters logically leads to a lack of confidence in the interpretations and
predictions that are based on a model analysis, unless it can be demonstrated that the model is a reasonably accurate representation of the real system.
The objective of the calibration procedure is to minimise differences between the observed data and calculated values. Usually, the model is
considered calibrated when it reproduces historical data within some acceptable level of accuracy. Because of the large number of variables in the set
of simultaneous equations represented in a model, calibration will not yield a unique set of parameters. Where the match is poor, it suggests (i) an
error in the conceptual model, (ii) an error in the numerical solution, or (iii) a poor set of parameter values. It may not be possible to distinguish among
the several sources of error (Konikow and Bredehoeft 1992). Even when the match to historical data is good, the model may still fail to predict future
responses accurately, especially under a new or extended set of stresses than experienced during the calibration period.
Numerical methods in general yield approximate solutions to the governing equations. There are a number of possible sources of numerical error in
the solution. If the modeller is aware of the source and nature of these errors, they can control them and interpret the results in light of them. The
sensitivity of the solution to the grid design (or spacing), time-step criteria, nature and placement of boundary conditions, and other numerical
parameters should also be evaluated, even if an inverse or regression modelling approach has been used. This is frequently overlooked, but failure to
do so may cause critical design flaws to remain undetected.

Model calibration may be viewed as an evolutionary process in which successive adjustments and modifications to the model are based on the results
of previous simulations. If a model is to be used for prediction relating to a problem or system that is of continuing interest or significance to society,
then field monitoring should continue and the model should be periodically postaudited, or recalibrated, to incorporate new information, such as
changes in imposed stresses or revisions in the assumed conceptual model. A postaudit offers a means to evaluate the nature and magnitude of

180
predictive errors, which may itself lead to a large increase in the understanding of the system and in the value of a subsequently revised model.
Revised predictions can then be made with greater reliability. The criteria for labelling a model as validated are inherently subjective. In practice,
validation is attempted through the same process that is typically and more correctly identified as calibration, that is, by comparing calculations with
field or laboratory measurements. However, the non-uniqueness of model solutions means that a good comparison can be achieved with an inadequate
or erroneous model.

RESERVOIR SIMULATION MODELING

Reservoir simulation has been used an artifact since the beginning of petroleum engineering in the 1930’s. Until 1960, the concept of reservoir
simulation was largely associated with analytical models, zero dimensional material balances, and one-dimensional Buckley-Leverett calculations
(Coats, 1982). Later, with the advent of computer programs, it was feasible to (a) enhance the dimensionality (2D/3D) of the problem; (b) account the
transient nature of fluid flow as against simple steady state conditions; (c) consider multi-phase multi-component fluid flow; and (d) to address the
challenges posed by the heterogeneity, and such models were able to address the dominant reservoir problems such as pressure depletion/maintenance,
which more or less addresses the issues related with primary recovery. And, further improvement was observed during 1970’s, when the concept of
reservoir simulation started addressing challenging issues pertaining to secondary and tertiary recoveries such as water flooding, chemical flooding
and thermal flooding. The art of developing numerical simulation model for a petroleum reservoir to assess its performance refers to the building and
development of a numerical model resulting from a well defined mathematical model and deeply grasped conceptual model. A model can simply be
referred to as any representation of a real system. Since, a real petroleum reservoir system on a field scale consists of a complex interaction between
various physical, chemical, and biological processes, the term “simplified model” becomes relatively biased based on the concerned modeler’s
individual specialization (say, Chemist / Geologist / Drilling Engineer / Reservoir Engineer), and eventually, the numerical modeler has to face the
challenge of its associated uncertainty and risk, depending on the gap that exists between the simplified model and the complex reality.

Any modeling consists of developing three main stages, which can be referred to as (a) Conceptual Modeling; (b) Mathematical Modeling; and (c)
Numerical Modeling. A conceptual model involves virtually visualizing the complex physical/chemical behavior of a deeply buried hydrocarbon
reservoir with a broader perspective, involving knowledge from multi-disciplines. During conceptualization, the modeler needs to visualize various
reservoir rock and fluid properties, and eventually, needs a deeper understanding on its associated rock-fluid interaction. Conceptual model essentially
anticipates a full range of possible reservoir system responses, and eventually, paves a strong foundation to the future reservoir management actions. A
mathematical model follows the conceptual model. Any mathematical model involves the translation of various physical/chemical/biological reservoir
processes that were conceptualized into an equivalent mathematics. A mathematical model can be as simple as a simple algebraic equation, while the
same can be as complex as a set of coupled equations containing higher order hyperbolic partial differential equations. It is to be noted here that while
a conceptual modeling does not provide any barrier to virtually visualize the complex reservoir system, the art of mathematical modeling needs a
strong mathematical understanding to translate the visualized processes. Thus, a modeler without a strong mathematical foundation, but with a very
good physical understanding of a complex reservoir system may fail to provide a reliable mathematical model or a simplified mathematical model
with a huge risk and uncertainty associated with it. Mathematical modeling involves either an empirical approach based on data analyses, or a
mechanistic approach evolving from mathematical descriptions based on theory. It also projects the time factor indicating whether the system is static
(time independent) or dynamic (time dependent). The mathematical model can be homogeneous for a deterministic system, which do not address the
data variability, while the same can be stochastic, which essentially addresses the data variability and its associated uncertainty. It is to be ensured that
the developed mathematical models are tested and verified for scenarios similar to the reservoir conditions. When a simplified mathematical model is
developed due to insufficient mathematical understanding, then the assumptions involved and the limitations associated with it needs to be portrayed
precisely. It is to be noted that in a petroleum reservoir, if the deduced set of initial and/or boundary conditions together with the multi-phase multi-
component partial differential equations cannot be prescribed capriciously, then the respective model should be tested for Existence (atleast one
solution to the problem exists), and Uniqueness (atmost one solution to the problem exists), and eventually, makes the problem mathematically
meaningful. Thus, it is always not possible to solve a set of coupled non-linear partial differential equations, particularly with time dependent
boundary conditions (the depth to Gas-Oil Contact, Water-oil Contact), which needs a continuous dependence on boundary and/or initial value. It is
also possible to have complicated reservoir geometry and/or boundaries. Under such circumstances, a numerical model needs to be developed to
address such issues. A numerical model estimates solution to a complex mathematical model that expresses the conservation of mass, momentum
and/or energy through numerical approximations. For example, in the case of Finite-Difference technique, either ordinary or partial derivatives are
approximated using appropriate numerical approaches, and the resulting set of algebraic equations are grouped into a single matrix and solved. It is to
be noted here that the finite-difference technique provide the solution only at a finite number of discrete points of the solution domain, while the value
of the dependent variable remains unknown between any two nodal points, and eventually, an extrapolation is needed. However, sometimes, a
numerical model can be used to verify the correctness of an analytical solution with the introduction of improved numerical modeling techniques such
as Finite Volume Technique. In this technique, the derivatives are not approximated by the difference quotients but are considered in their integral
form.

Any numerical model essentially involves solving the parabolic, hyperbolic and/or elliptic types of mathematical equations. The parabolic type
equations are associated with time marching solution domain, and they need to satisfy the von Neumann stability criteria based on whether the scheme
opted is implicit/explicit or Crank-Nicholson. The hyperbolic type equations play a crucial role at discontinuities such as interface between high
permeability fracture and low permeability rock-matrix, typically, encountered in a fractured reservoir. However, Lax and Wendraff (1960) have
shown that such discontinuous weak solutions can be computed without any special treatment of the discontinuity if a conservative form of differential
equation is adopted. The elliptic type of equations is independent of time, without any directional bias. The elliptic type of equations is generally
referred to as boundary value problems associated with a closed solution domain. In general, any reservoir model would consist of a combination of
parabolic/hyperbolic/elliptic partial differential equations based on the problem description. Such equations can be solved either as a one-dimensional /
two-dimensional or three-dimensional model depending on the focus of the project. For example, gravity drainage of hydrocarbons could be simulated
using a one-dimensional model in the vertical direction, and the corresponding field performance may be projected approximately. However, it is to be
noted that any 1-D model may not be able to capture the well behavior locally. In the absence of any distinct layering of soils and/or rocks, the impact

181
of different well patterns may be studied using an areal 2-D model, while a cross-sectional 2-D model is used for layered reservoirs. However, in the
case of a naturally fractured reservoir, where the details on the fracture network, connectivity of fracture networks and fluid-rock interactions are
required to assess the reservoir performance, a volumetric 3-D model is used. The above mentioned models can be solved either using fully explicit or
by Implicit Pressure and Explicit Saturation (IMPES) formulations. However, the issue of stability needs to be addressed explicitly in an IMPES
formulation. The main disadvantage of using a 3-D model is that it requires a large CPU time to run the model due to the large and sparse coefficient
matrix encountered generally. In addition to Cartesian co-ordinates, coning models are used in polar co-ordinates to study the individual well effects.

The above model needs input data at each node in the case of finite-difference / finite element or at each sub-cell in the case of finite-volume include
the details on (a) reservoir geometry (length/thickness/width/depth); (b) reservoir rock properties (porosity and single phase absolute permeability as a
constant mean value, while multi-phase relative permeability as function of saturation); and (c) reservoir fluid properties (reservoir pressure and
gas/oil/water saturation as a constant mean value, while density/viscosity/compressibility/Gas-Oil-Water formation volume factors/Gas-Oil
ratio/Solution Gas-Oil ratio as functions of pressure; and capillary pressure as a function of saturation). In addition, the details on the well location as
well as the injection and production rates through/from the well are required. In general, the hydrocarbon reservoir system is characterized by (a)
generally unknown geology; (b) poor understanding of fluid dynamics; and (c) thermodynamics complexities. In addition, since the data pertaining to
reservoir simulation is deduced from different sources like core data, thin sections, seismic data, well log, well test, etc.., it may not be feasible to reply
on these data as its range varies between several orders of magnitude (10-5 to 108 cm). Hence, it is necessary to build atleast a couple of models on a
smaller scale describing the geology of the reservoir (geological models) and the same needs to be upscaled to a larger scale, i.e., on a computational
grid with least errors. Then, proper initial and boundary conditions needs to be established followed by the selection of development and operating
strategies. Then, multi-component multiphase fluid flow describing the hydrocarbon fluid flow can be solved and the reservoir performance can be
predicted. The uncertainty associated with the models can also be estimated either by minimizing or maximizing the objective functions.

CONCLUSION

This paper has made an attempt to provide few basics on the conventional modeling and the modeling complexities pertaining to fractured
hydrocarbon reservoirs. Though this overview is limited to a discussion on the basics of the computational techniques and its associated issues
pertaining to fractured hydrocarbon reservoirs, the proper application of the selected codes for the concerned computational technique is as important,
if not more important, than code selection. A code, no matter how well suited to a particular application, could give erroneous and highly misleading
results if used improperly or with incomplete or erroneous input data. Conversely, even a code with very limited capabilities, or a code used at a site
which has not been well characterized, can give very useful results if used intelligently and with a full appreciation of the limitations of the code and
the input data.

REFERENCES

Bonnet, E., Bour O., Odling N., Main I., Berkowitz B., Davy P., and Cowie P. (2001). Scaling of fracture systems in geological media, Rev. Geophys.,
vol. 39, pp. 347-383.
Brush D. and Thomson N.R. (2003). Fluid flow in synthetic rough-walled fractures: Navier-Stokes, Stokes, and local cubic law simulations”; Water
Resources Research, vol. 39(4), pp. 1085.
Coats, K. H. (1982) Reservoir Simulation: State of the Art, Journal of Petroleum Technology, Distinguished Author Series, SPE 10020, pp. 1633-
1642.
Gustafson, G., and Krasny, J. (1993). Crystalline Rock Aquifers: Their Occurrence, Use and Importance, IAH Memories, vol. XXIV, Part 1, pp 3-20.
Huebner, K.H. (1975). The Finite Element Method for Engineers. John Wiley & Sons, New York: 500 pp.
Konikow, L.F. (1996). Numerical models of groundwater flow and transport. In: Manual on Mathematical Models in Isotope Hydrogeology,
International Atomic Energy Agency Rept. IAEA-TECDOC-910, Vienna, Austria: 59-112.
Konikow, L.F., Reilly, T.E. (1998). Groundwater Modelling. In: The Handbook of Groundwater Engineering [J.W. Delleur, ed.], CRC Press, Boca
Raton 20:1-20.40.
Konikow, L.F., Bredehoeft, J.D. (1992). Ground-water models cannot be validated.
Advances in Water Resources 15 (1): 75-83.
Lax, P. D., and b. Wendroff. (1960) System of Conservation Laws, Communications on Pure and Applied Mathematics, vol. 13, p. 217.
Lorenz, J. C., Warpinski, N. R., Teufel, L. W., Branagan, P. T., Sattler, A. R., and Northtrop, D. A. (1988). Results of the Multi-well Experiment: In-
situ Stresses, Natural Fractures, and other Controls on Reservoirs, EOS, Transaction American Geophysical Union, vol. 69 (35), pp. 817-826.
Peaceman, D.W. (1977). Fundamentals of Numerical Reservoir Simulation. Elsevier, Amsterdam:176 pp.
Pruess, K. and Tsang, Y. W. (1989). On relative permeability of rough-walled fractures, Proc. Fourteenth Workshop on Geothermal Reservoir
Engineering, Stanford Univ., Stanford, CA, pp. 123-130.
Pruess, K. and Tsang, Y. W. (1990). On two-phase relative permeability and capillary pressure in rough-walled rock fractures, Water Resources
Research, vol. 26(9), pp. 1915-1926.
Oran, E. S., and Boris, J. P. (2001). Numerical simulation of reactive flow. 2nd edition. Cambridge University Press, 529 pp.
Remson, I., Hornberger, G.M., Molz, F.J. (1971). Numerical Methods in Subsurface Hydrology. Wiley, New York: 389 pp.
Singhal B.B.S. and Gupta R. P. (1999). Applied Hydrogeology of Fractured Rocks”, Kluwer Academic Publishers, Dordrecht, Boston, London.
Stearns, D. W., and Friedman, M. (1972). Reservoirs in Fractured Rocks, American association of Petroleum Geologists, Memoir, vol 16, pp 83-106.
Tatomir, A. B. (2007) Numerical Investigations of Flow Through Fractured Porous media, Masters Thesis, WAREM, Stuttgart University, 79 pages.
Tillman, J. E. (1983). Exploration for Reservoirs with Fractured Enhanced Permeability, Oil & Gas Journal, vol. 81, pp. 165-180.
Witherspoon, P. A., J. S. W. Wang., K. Iwai., and J. E. Gale. (1980). Validity of Cubic Law for Fluid Flow in a Deformable Rock Fracture, Water
Resources Research, vol. 16(6), pp. 1016-1024.
Von Rosenberg, D.U. (1969). Methods for the Numerical Solution of Partial Differential Equations, Elsevier, New York: 128 pp.

182
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Role of reservoir fluid sampling and phase behavior studies in reservoir management
Anil Kumar and Chaman Lal
KDMIPE, Kaulagarh Road ONGC Dehradun
*email: anilkb15@yahoo.co.in

ABSTRACT

Representative reservoir fluid sampling is most important to obtain accurate PVT parameters. With more and more exploration, appraisal and
development activities in marginal fields, accurate fluid characterization becomes more critical. The PVT properties of reservoir fluid are essential
for efficient reservoir management through the life of a reservoir from discovery to abandonment. Inaccurate fluid characterization often leads to
high uncertainties in estimation of in place volume, prediction of recovery, characterization of reservoir and optimization of production &
development strategies. The main objective of a successful sampling is to obtain representative samples having all the components of hydrocarbons
present in the reservoir fluid in the reservoir. The success of sampling of a reservoir fluid depends on the employment of correct sampling procedures
and well conditioning before and during sampling. Though reservoir fluid sampling procedures and PVT experimental techniques are well
documented but no industry guide lines are in place for sampling from different types of reservoir. There are two general methods of sampling-
surface and subsurface sampling. However within these, there exist different methods that can be more applicable to particular type of reservoir than
to another. In several case studies, it has been observed from the results of PVT analysis of sub surface/surface sample of different wells that the
various PVT parameters determined are not accurate because of sampling method employed was not correct. The results of analysis have indicated
that only certain procedure of sampling is correct for a particular type of reservoir to get the accurate PVT results. The paper presents a discussion on
better understanding of different sampling methods, concepts behind selecting the sampling procedure and impact of sampling procedure on the
outcome of results of PVT analysis by using some field examples.

INTRODUCTION

Properties of reservoir fluid PVT are critical for efficient reservoir management throughout the life of reservoir from discovery to abandonment.
Reliable PVT properties of in situ fluids are essential for the determination of in place volumes and recovery factor and key input for reservoir
simulations. Fluid characterization and distribution within the reservoir help in defining reservoir continuity and communication among various
zones. Success of obtaining the quality data by PVT analysis depends on the representative reservoir fluid samples. The objective of reservoir fluid
sampling is to collect a sample that represents the fluid in the reservoir at the time of sampling. Ensuring that the sample is representative of the
reservoir fluids at actual reservoir conditions is a difficult task to achieve.

The purpose of PVT analysis is to obtain reliable high quality data for reservoir evaluation and development. The PVT data requirement depends on
the fluid type and the expected development and production strategies. For example, extra heavy oils require customized PVT cells and experimental
procedures to accelerate the time needed for attaining equilibrium conditions because of the slow gas liberation. On the other hand, more complex
near critical fluids and miscible gas injection processes need special PVT tests and precise measurements techniques to record the complex phase
behavior exhibited by these fluids. The basic PVT analysis comprised constant composition expansion, differential liberation, and multistage
separator tests to describe the oil PVT behavior as a function of reservoir and surface conditions. Fluid properties may be position, structure or time
dependent, requiring collection of multiple samples at different locations at different times in the reservoir. Whether or not the sample is
representative is also affected by how the sample is collected, transferred and transported from reservoir to the laboratory. To plan sample in a field
it is essential to understand the reservoir first. The reservoirs can be classified in two main categories viz., Oil Reservoir and Gas Reservoir.

OIL RESERVOIR

Oil reservoirs can be divided in to two categories:- ordinary oil reservoirs and near critical oil reservoirs. The ordinary oil reservoirs are sometimes
called black oil reservoirs. There is no sharp dividing line between an ordinary and near critical oil. Fluid composition and reservoir temperature
greatly influence the behavior of reservoir fluid. It is very difficult for an engineer to decide the type of reservoir at the time of initial production
testing of an exploratory well. However certain thumb rules are there when an engineer can make out the possible type of reservoir. Ordinary oils are
generally having a GOR up to approximately 2000 ft3/bbl or 360 m3/m3 and oil gravity up to 450 API.

Near critical oils are often referred to as volatile oils also. The reservoir fluid in this case exhibit the properties of oil existing in the reservoir at a
temperature near its critical temperature. These properties include a high shrinkage immediately below bubble point pressure. In some cases this
shrinkage can be as low as 45% of the hydrocarbon pore space within 10 psi below the bubble point pressure. Near critical oils generally have GOR
approximately 2000-3000 ft3/bbl or 360-540 m3/m3 and oil gravity above 400 API.

GAS RESERVOIR

Gas reservoir can be divided in to three categories: - Gas condensate, wet gas and dry gas
In case of gas condensate reservoir generally single phase (gas) exist at reservoir conditions. As reservoir pressure declines during production,
retrograde condensation takes place in the reservoir thus making the reservoir in two phases. In case of wet gas the fluid in reservoir exist as gas
throughout the reduction in reservoir pressure. However, at separator conditions at surface some liquid is formed because of presence of some higher
components of hydrocarbon. Since there is no liquid formation inside the reservoir throughout its production life but liquid is formed at surface
condition such gas reservoirs are called wet gas reservoir. In dry gas reservoir the fluid does not contain enough of heavier hydrocarbon molecules to
form liquid inside the reservoir as well as at surface during production and such reservoirs are classified as dry gas reservoir. Out of these reservoirs,
change in phase in reservoir during production occurs only in gas condensate reservoir and as such are important to study for its phase behavior.

SAMPLING METHODS

Two techniques are employed in general practice in obtaining the reservoir fluid samples. Samples may be obtained by surface sampling or sub
surface sampling. Whichever method is employed it is imperative that the samples be representative of the reservoir fluids. They should be obtained
during early life of a reservoir since preferential flow of one phase makes it more difficult to reproduce accurately a reservoir fluid from a reservoir
when appreciable quantity have been produced at pressures below saturation pressure.

Subsurface Sampling

During the early life of an oil reservoir the flowing phase is usually liquid. In oil reservoir, a representative sample of this flowing phase is usually
best collected by sub surface sampling. However, in order to obtain a representative sub surface sample, a well must be properly conditioned. The
conditioning of a well for a bottom hole sampling depends upon the drawdown of the reservoir pressure (the difference between shut-in reservoir
pressure and flowing bottom hole pressure). Finally following general procedures are recommended for conditioning a well in preparation for
subsurface sampling.

• The well should be flowed sufficiently to recover any contaminating material which might have resulted from drilling, perforation and
completion.
• The sample must be taken at a pressure as near the existing pressure of the oil or flowing phase in the reservoir at the drainage limit as
possible. The sample must be taken at a time when the oil has been subjected to a minimum reduction in pressure in its movement in to the
well bore.
• The rate should then be reduced as low as possible and flow the well for about 24 hrs.
• Finally close the well for about 24 hours.
• Following the well conditioning, but prior to sampling, a measurement of bottom hole pressure and temperature should be made in order to
correlate laboratory data with reservoir data. A pressure gradient should be determined to locate the water level in tubing if exist and decide
the depth at which sample may be collected.

Surface Sampling

Surface sample for PVT analysis is collected at separator. Success of surface sampling depends largely on stabilization of well producing at a
constant rate while holding a constant pressure and temperature on the separator. A fairly low rate, which will minimize the pressure draw down, in
and around the well bore to prevent retrograde condensation in the reservoir rock in a gas well and evolution of solution gas in an oil well is
recommended. The following procedures are recommended for conditioning the well in preparation for obtaining separator sample at surface.

• The well should be flowed sufficiently to clean out any foreign oil in case of flowing of two or more wells together or contaminating
material if well is not fully clean.
• In case of more than one well are flowing through separator, flow of well which is to be sampled should be diverted through a test separator
and it should be flowed for sufficient time to flush out line and separator already filled with the multi well fluid to have a representative
fluid samples.
• The well should then be flowed at a relatively low rate generally until liquid and gas production has stabilized.
• The flowing gas-oil ratio should be measured accurately as this value will be used as the basis for recombination of the separator oil and
gas samples in the laboratory.
• After satisfactory gas-oil ratio measurement, the sample of oil and gas are taken from the high pressure separator or first sage separator.
• In case of multistage separator, sampling should be done only from first stage as all the components of oil and gas are available at this
separator. Whereas at second stage separator, oil reaches after removing the separated gas at first stage thus losing the components of gas
separated at first stage.

184
For gas condensate reservoirs subsurface samples may not be representative enough of the reservoir fluid. In condensate reservoir enough condensate
may be precipitated to build up a continuous phase and flow with vapor leaving behind condensate. A normal bottom hole sample obtained from a
shut in well in this type of reservoir would ordinarily contain only the liquid that has separated from the vapor in the tubing and would have little or
no relation to the physical characteristics of the phase that is actually flowing through the reservoir rock. Therefore it s recommended that reservoirs
which are considered to be condensate reservoirs and those where there is a doubt as to whether the reservoir is an oil or condensate, be sampled
from separator at surface.

These are some standard methods of samplings but it is a big task for a reservoir engineer to decide which method would be most suitable for a
particular type of reservoir. If an oil reservoir is saturated or reservoir pressure is very close to bubble point pressure then the subsurface sample
would not be representative. This can be interpreted with the field examples where subsurface sample gave erratic results of PVT analysis because of
employing incorrect method of sampling.

Case Study-1

PVT analysis of subsurface sample of well A from one of southern and well B from one of western Indian fields indicated both the reservoir as under
saturated whereas both the reservoirs are saturated having gas cap. The details of wells and summary of PVT results are presented in table-1 & 2

Table-1, General well information

Sl. No. General Information Well A Well B


1 Perforated Interval 2569-2572, 2577-2580 3204-3206
2 Reservoir Pressure Kg/cm2 206.7 280
3 Reservoir Temperature (0F) 240 279
4 Oil Rate m3/d 25 10
5 Gas Rate m3/d 22000 22780
6 Field Gas Oil Ratio (GOR) 890 2283
7 Depth of Sampling 2560 3100

Table-2, Brief of PVT data generated

Sl. No. PVT Parameters Well A Well B


1 Saturation pressure Kg/cm2 200.38 154.67
2 Flash oil FVF 1.4704 2.02
3 Flash GOR 138 246.5
4 Differential oil FVF 1.4253 2.0089
5 Solution GOR 135 237
6 Shrinkage factor 0.7016 0.4977
7 API gravity 41.6 45.46

In both the wells, the results of PVT analysis showed that saturation/bubble point pressure is less than reservoir pressure indicating both the fields as
under saturated though in reality both the fields are having gas cap and are saturated reservoirs. The reason for getting the bubble point pressure less
than reservoir pressure is that the sample was trapped from bottom after flowing the well. While flowing, the wells were subjected to some draw
down resulting in to release of some dissolved gas. Now the bottom hole sample trapped was having less dissolved gas than actually it should have at
reservoir pressure and the fluid was thus under saturated. In fact in such cases the bubble point pressure would be near to flowing pressure at which
the wells were flowing just before or during sampling. It is therefore, concluded that if the reservoir is just saturated or having gas cap should be
sampled at surface only and the gas & oil collected separately should be recombined at the field GOR in laboratory to make it representative and
obtain accurate PVT results.

Now question is how to know whether the reservoir is saturated and is having gas cap if there is no indication of gas cap in well log data of an
exploratory well with no previous history. In that case field GOR is the best indicator to decide whether to go for sub surface sample. For example, a
reservoir with a depth of about 2000 m and having field GOR more than 300 m3/m3 may be saturated and sub surface sampling may lead to erratic
results.

Case Study-2

Some reservoirs are very tight and it need very high draw down even for a flow through a smallest possible choke. Very high pressure drop during
flow may cross bubble point pressure and such reservoirs are also recommended for surface sampling as subsurface sample would give wrong bubble
point pressure. In one of the example, it was observed in the results of PVT analysis of an under saturated reservoir which was flowing at less then
bubble point pressure and sampled at sub surface that the bubble point pressure was more than actual bubble point pressure and solution GOR was
more than field GOR. In this reservoir, sample was collected in flowing condition. Since the well was flowing in two phases at the time of sampling
as the flowing pressure was less than bubble point pressure, sub surface sample was mixture of gas and oil with more gas. This sample was made in
single phase by applying pressure in laboratory resulting higher solution GOR and bubble point pressure than actual.

185
This can be shown in an example of a field of western on land basin where an exploratory well C producing oil at the rate of about 20 m3/d with a
very less gas (rate could not be measured). The testing results clearly indicated the reservoir to be under saturated. However the sub surface sample
was collected in this well in flowing conditions for PVT analysis. The results of PVT analysis of the sample indicated the reservoir to be saturated.
The details of well and results of analysis is presented in table-3.

Table-3, General well information & Brief of PVT data generated

Sl No General Information & PVT Well C


Parameters
1 Perforated Interval (m) 1856.5-1861
2 Reservoir Pressure Kg/cm2 179.5
3 Reservoir Temperature (0F) 156.2
4 Oil Rate m3/d 20
5 Gas Rate m3/d -
6 Field Gas Oil Ratio (GOR) -
7 Flowing Bottom Hole pressure Kg/cm2 100.1
8 Depth of Sampling (m) 1820
9 Saturation pressure Kg/cm2 179
10 Flash oil FVF 1.85
11 Flash GOR 253
12 Differential oil FVF 1.64
13 Solution GOR 223
14 Shrinkage factor 0.6076
15 API gravity 49.15

The erratic results of PVT analysis of sample from well C are because of incorrect method of sampling. The well is having reservoir pressure about
179.5 Kg/cm2 and flowing pressure 100.1 Kg/cm2 indicating a very high draw down of about 79 Kg/cm2. The sample was collected from bottom in
flowing condition and it appears that the bubble point pressure had crossed because of very high draw down. Thus the reservoir was flowing in two
phases at the time of sampling with the result trapping of oil and free gas. In this case, gas was more than actual dissolved gas making the sample
saturated at reservoir pressure and temperature. As mentioned earlier the well should be flowed through a very small choke to give minimum possible
draw down and if it is not possible because of reservoir characteristic, the surface sampling should be planned in order to avoid erratic PVT results.
But in this case it was not followed causing the wrong interpretation of reservoir based on PVT analysis.

CONCLUSIONS

1. Sampling of reservoir fluid posed significant challenges. Representative reservoir fluid is most essential to have accurate PVT parameters
as the fluid characterization strongly affects in place volume, recovery factor and development plan of a reservoir.
2. Fluid sampling program should be made after having a good understanding of reservoir by going through various available data like well
log, reservoir pressure and producing field GOR etc.
3. Additionally the latest technologies available for mud logging, formation testing and fluid evaluation are brought to the well site to reduce
subsurface uncertainties as well as reduce the decision time.
4. Emphasis should be given to samples collected after proper well conditioning and in case of bottom hole sample the well should be flown
with minimum drawdown to collect the sample above bubble point pressure.
5. Surface oil and gas samples must be taken simultaneously or close together at first stage separator in case of multistage separation.

REFERENCES

Burcik, E. J. (1957), “Properties of Petroleum Reservoir Fluids”, John Wily & Sons, INC. London
Dake, L. P. (1978), “Fundamentals of Reservoir Engineering”, Elsevier Scientific Publishing Company New York
Dodson, C. R., Goodwill, d. and Mayer, E. H. (1953). “Application of Laboratory PVT Data to Reservoir Engineering Problems”, JPT, 287-298
Mccain, W. D. Jr. (1993), “Chemical Composition Determines Behavior of Reservoir Fluids”, Pet. Eng. Intl, 18
Reudelhuber, F. O. (1957), “Sampling Procedures for oil Reservoir Fluids”, JPT, 15-18
Standing, M. B. (1977), “Volumetric and Phase Behavior of Oil Field Hydrocarbon Systems”, Millet Print Inc. Dallas Texas.
Honarpour, M. M., Nagarajan, N. R., and Sampath, K., (2006), “Rock/Fluid Characterization and their Integration-Implication on Reservoir
management”, JPT 58(9).
Nagarajan, N. R., Honarpour, M. M., and Sampath, K., (2007), “Reservoir Fluid Sampling and Characterization-Key to Efficient Reservoir
Management”, JPT, SPE 103501.

186
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

An assessment of non linear AVA inversion for reservoir parameter estimation using two global
optimization methods
Rajesh R. Nair1, Yudhvir Singh2 and Suresh Ch. Kandpal3
1,2
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
2
Bhaba Atomic research Centre, Mumbai.
2
E-mails: 1rajeshnair@iitm.ac.in, yudhvir_vijay@rediffmail.com, 3sur.kandpal@gmail.com

ABSTRACT

We assess the performance of two optimization methods (genetic algorithm and fuzzy clustering means) for reservoir parameter estimation by
evaluating their degree of variance using AVA inversion scheme. AVA provides the subsurface reservoir parameters ie compressional velocity,
shear velocity and density of various layers. The problem is inherent in the non linear nature of the forward model, thus making this inversion quite
intricate. During inversion, error obtained in the first layer percolates down and carries on cumulatively to the underlying layers and hence solution
may not converge with desired error bound. We demonstrate here using two optimization methods, how these problems can be effectively addressed.
Methods based on local optimization fails and did not provide the acceptable fitness as obtained using least square inversion scheme alone. The
robust and proficient genetic algorithm and cooperative fuzzy clustering means yields promising estimates of reservoir parameters within 0.3 to 3
percent error.

KEY WORDS: AVA Inversion, reflection coefficient, Fuzzy logic, cooperative algorithm, cluster analysis, Genetic algorithm

INTRODUCTION
The AVA (amplitude variation with angle) response is effectively used to determine the reservoir parameters like compressional velocity (VP), Shear
velocity (VS), density (ρ) across an interface, which is governed by Zeopperitz equation (Zeoppritz, 1919) and is given in concise matrix formed by
Aki and Richards (2002). However the exact mathematical equation of the Zeopperitz equation is exceptionally long, thus making it difficult to
perceive how reflection amplitude varies if a rock property is changed slightly. To give an intuitive idea of how these amplitudes are related to
various physical parameters, a number of approximations to the Zeopperitz equation have been made by several scientists (Aki and Richards, 1980;
Shuey, 1985), which are valid for small angle of incidence and low impedance contrast. Since the Zeopperitz equation is non linear function of model
parameters, the approximation to the exact equation causes an error in determining the model parameters. So, to exploit the full information
contained in the AVA data, we need to use the exact equation. Demirbag and Costain (1993) show that the exact densities are quite complicated to
determine as the equation always talks of the relative density and the convergence of VS is also not also good for small offsets. Hence to take up this
issue, we use a non linear multi-paramter inversion technique initially with a fuzzy c means method and then inverted VP using travel time inversion
and estimate Vs and density from AVA inversion using a Genetic algorithm (GA) optimization method.

The reflection coefficient at any given angle of incidence is completely determined by the density, P wave and S wave velocities of each medium.
Therefore the major task in AVA analysis is to restore relative amplitudes in seismic pre stack records before extraction and interpretation of the
amplitude information. In exploration geophysics, we rarely deal with simple isolated interfaces. However we must begin our understanding of offset
dependent reflectivity with the partitioning of energy at such an interface (Castagna and Backus, 1993; Ostrander, 1984). The angles of incident,
reflected and transmitted rays synchronous at the boundary and defined by Snell’s law. The matrix representation of Zeopperitz equation is visibly
used describes reflection coefficient as a function of incident angle and thus can be used to create a forward model for the reflection coefficient data.
The expression for calculating the offset from angle of incidence takes care of the ray bending (Saian and Kaila, 1994) at different interfaces and
hence gives accurate results for calculated travel time. We therefore assess the performance of genetic algorithm based on travel time inversion and
also co-operative least squares and Fuzzy c- means clustering (Nair and Kandpal, 2010) on a synthetic model to investigate whether any noticeable
improvement is obtained.
AVA METHOD AND FUZZY C-MEAN OPTIMIZATION

Here we want to demonstrate the applicability and performance of the present inversion approach ‘Fuzzy c-means clustering’, by linking the three
parameters compressional velocity, shear velocity and density with a least squares AVA inversion .We perform a non linear AVA inversion for
estimation of Vp, Vs and ρ and ‘ε’ (error function). The basic assumptions made are that the earth has approximately horizontal layers at each
common depth point, and that each layer is described by the three parameters VP, VS and ρ (Zoeppritz, 1919; Aki and Richards, 2002). We compute
the reflection coefficient for each angle and at each layer boundary. A synthetic offset seismic gather can be calculated by convolving the reflection
coefficients with predetermined wavelets. A simple illustration of this process is given in Fig. 1(a).The seismic source wavelet, used, is a Ricker
wavelet generally used in seismic modelling (Hoversten et al., 2006). These synthetic data are compared with the observed data to form a misfit
function. We then perturb each parameter of the solution vector in turn to form a new earth model, and generate new synthetic data, which are
compared with the observation again. The process is repeated until the root mean square error function between the observed and the synthetic data is
minimized. To reduce the non-uniqueness problem, the inversion algorithm is constrained by bounding the solution within a range of physical
solutions.

Using the simple least squares optimization for a multi-parameter situation may not be right choice, to deal with the non uniqueness and unstabilities.
To be effective, the available data sets should be linked during the model generation process, for example by a cooperative inversion. It is
understood that, the cluster analysis, realises the information exchange between the individual models and data, but it cannot completely overcome
the problems of inversion in several cases (Paasche and Tronicke 2007).We examine here, whether one can resolve the density parameter estimation
using Fuzzy c-means at least to a first order.

Fuzzy c-means (FCM) is a data clustering technique wherein each data point belongs to cluster to some degree that is specified by a membership grade.
FCM play a roll of partitioning arbitrary n vectors into c fuzzy groups, also it finds a cluster center for each group such that it iteratively minimize
objective function (Hoeppner et al. 1999), that represents the distance from any given data point to a cluster center weighted by the data point’s
membership grade. Well-known fact about FCM indicates that FCM employs Fuzzy partitioning such that a data point can belong to several groups with
the degree of membership grades between zero and one. The optimum number of clusters c is computed by minimizing the normalized classification
entropy (NCE) (Lines et al., 1988; Paasche et al., 2006). Fuzzy c-means (FCM) is based on minimizing the objective function (Bezdek, 1981).

FCM Cluster Analysis with Damped Least Squares Technique: A union of conventional data set inversion algorithm with FCM technique provides
the sole theme of co-operative cluster inversion. To ensure information exchange between different geophysical data sets and models during the
inversion process, at each iteration FCM cluster analysis is applied. First of all we define the n different geophysical data sets (VP, VS and ρ), set the
initial number of clusters to Cmin=2 and guess a reasonable number of clusters Cmax. Subsequently, equally discretized initial parameter models are
generated using uniform squared grid cells for all data sets. Then FCM cluster analysis is performed on these n individual models. Following this,
membership information is used to calculate new model parameters for a cluster centre value weighted with the degree of membership of model cell
to the c clusters at each model cell. Now conventional regularized least squares single input data inversion is performed using new parameter models
as input models. In the damped least squares technique, we minimize the residual that is defined as the sum of squares of the differences between the
idealized model response and observed data.

As the optimum value of c is usually not known a priori, we have to repeat the procedure using several reasonable values for c. The final-model with
the optimum c is selected from these possible solutions by visually analyzing the calculated NCE and rms errors values. We seek the model with
minimum complexity which concurrently minimizes the NCE values and the root mean square errors (RMSE). Fig. 1 (b) shows the flowchart
depicting the sequence of FCM clustering adapted for AVA Inversion. We have used least squares inverted output as the starting model for Vp and
Vs, while using the Fuzzy clustering scheme.

Synthetic Data: We use a convolution modelling technique, to generate synthetic data. The input model for our synthetic example consists of five
heterogeneous layer boundaries with acoustic impedance contrasts. The chosen model is typical for Gas Hydrates in a sedimentary setting, where the
parameters tend to cluster within certain ideologies. The Vs, Vp and density profiles corresponds to published elastic properties of sediments
containing hydrates (Dvorkin et al., 1999). Aki and Richards Equation (Florence et al., 2004; Hoversten et al., 2006) is used to compute the
reflectivity profile for angles of 1 to 40 degrees corresponding to the angle range of the field data. The synthetic data are added with a random noise
component of 25% with respect to signal to make it a true data situation. Fig. 1 (c) shows the forward modelled synthetic data, used as an input for
the inversion scheme.

Results of the standard inversion with least square optimization: The non linear inversion does not find the exact parameter values but the results
are satisfactory. When adding noise to the data (S/N=4), the unconstrained non linear inversion is not destabilized as the linear inversion. We
conclude from the synthetic modeling that the non linear inversion is more robust than the linear inversion. The non linear inversion is little affected
by the noise level in the data. Fig. 2 (a-c) shows the initial guess model (three dotted lines), grey indicates the well log guided one and other two
arbitrary guess model (sky blue and violet dotted lines) again used for Fuzzy c-means inversion on either side of well log guided guess model. Least
squares inversion with initial guess model (shown as sky blue and violet dotted lines) is shown with same colours. The misfit of Vp, Vs and density
and layer boundary fitness is higher based on least square inversion alone, with better fit obtained for the well log constrained initial guess in contrast
with arbitrary lower and higher guess used in the inversion scheme.

188
Fig. 1. (a) Workflow for nonlinear AVA inversion, (b) Flow chart showing the sequence of steps in the present scheme of fuzzy c-mean clustering
adapted for AVA inversion, (c) Forward modeled synthetic data used for fuzzy c-means clustering.

0 0 0
(b) (c)
(a)
-10 -10 -10
Depth (m)

Depth (m)
Depth (m)

-20 -20 -20

-30 -30 -30

-40 -40 -40

1600 2000 2400 2800 3200 3600 1000 1200 1400 1600 1800 2000 2050 2100 2150 2200 2250 2300
V p (km/s) Vs (Km/s) Density (Kg/m3)

Fig. 2. Results of standard inversion with optimization using least squares minimization scheme: Green line indicates true model, red line indicates
inverted model and dashed line indicates the initial guess model. Here a,b,c shows the initial guess model (three dotted lines), grey indicates the
well log guided one and other two arbitrary guess model (sky blue and violet dotted lines) again used for Fuzzy c-means inversion on either side of
well log guided guess model. Least squares inversion with initial guess model (shown as sky blue and violet dotted lines) are shown with same
colours.

Results based on Fuzzy c-means clustering approach: We test the non linear inversion algorithm on a gas hydrate model, first by simple inversion
based on least squares optimization scheme with homogenous layering as a starting model, and then by means of Fuzzy c-means applied using the
cooperative fuzzy c-means clustering inversion.

The output from the least squares inversion is quite off from the true model. Therefore to avoid more searching the empirical relationship derived
from Hoversten et al. ( 2006), is used as a constraint for the starting model for all three parameters. Fig. 3 (a-c) shows the fuzzy inversion shown as
pink and grey colours of two other least square inputs for higher and lower values respectively with respect to Fig. 2 (a-c). The acceptable fitness
percentage is not achieved in all cases (with three initial guess models) with least squares inversion alone. While Fuzzy c-means gives best fitness for

189
well log guided initial model and its least squares input, the Fuzzy c- means fitness is not largely affected even if we use arbitrary guess model
within some defined limits from the well log guided model. This shows the ability of the clustering algorithm in retrieving model parameters.

0 0 0 0
(a) (b) (c)

-10 -10 -10 -10

Depth (m)
Depth (m)

Depth (m)
-20 -20 -20 -20

-30 -30 -30 -30

-40 -40 -40 -40

1600 2000 2400 2800 3200 3600 1000 1200 1400 1600 1800 2000 2050 2100 2150 2200 2250 2300
V p (km/s) Vs (Km/s) Density (Kg/m3)
0.03 0.1 0.3
(d) (e) (f)
0.08

0.02 0.2
RMSE
NCE

0.06

NCE
0.04
0.01 0.1
0.02

0 0
0
0 20 40 60
0 2 4 6 8 Iterations 0 10 20 30 40 50 60
Iterations
Number of clusters

Fig. 3: (a) to (c), Fuzzy c-means inversion after least squares inverted results are given as input model. The blue line indicates inversion using Fuzzy
c-means optimization scheme, red standard inversion with least squares optimization scheme and green line indicates the true model. Here a,b,c
shows the Fuzzy inversion shown as pink and grey colours of two other least square inputs for higher and lower values respectively (Fig. 2 a-c). (d)
Indicates the minimum clusters for a lowest normalized classification entropy (e) RMSE error shows the parameters are obtained, when the error
goes to minimum (f) NCE values again indicates that results are obtained when the NCE values also goes to minimum.

AVA METHOD AND GENTIC ALGORITHM OPTIMIZATION

To show the effectiveness of the method, we preferred to apply the method to synthetic data where the model to be estimated is known. We have
considered a 2D three layered model (shown in Fig. 4(b)), corresponding to different layers, some reflection travel times at different offsets are
analytically computed using equation and for each interface the p-wave reflection coefficients have been calculated using Zeopeeitz equation. Both
these data contaminated by some random noises have been inverted to determine different degrees of random noises along with the corresponding
best fitted model. A reduced time scale with a reduction velocity of 6000 m/s has been used through out to present all the reflection times in a
compact form.
Genetic algorithm are global optimum methods that apply some known principles of biological evolution (eg.survival of the fittest, mutation and
crossovers etc) to find the best possible solutions for a given problem. Being global optimum search methods, these algorithms explore the whole
space of parameters. This feature makes evolutionary algorithms appropriate to solve the geophysical inverse problems, which on account of the non
uniqueness, may have several local optima (Sen and Stoffa, 1995; Stoffa and Sen, 1991). Different steps in a Genetic algorithm are:

a. Initialization of population
b. Fitness Evaluation
c. Selection of fit individuals
d. Biological operations
e. Fitness Evaluation
f. Termination condition examination
g. Termination

The different steps in the GA are shown in Fig. 4(a).

190
Fig. 4. (a)Different steps in the Genetic Algorithm (GA) inversion. (b) Synthetic 3-layered Earth model assumed for GA inversion.

Here to solve afore mentioned inverse problems, we have used Binary coded classical genetic algorithm using elitist approach. The methods used in
different steps have been elaborated below briefly:

Initialization: The initial step in employing a GA is to design a coding scheme that represents the possible solution to a given problem. We assume
that we know the physical parameters controlling the observed geophysical response. We randomly generate each of these physical parameters and
make them the initial model parameters. These are then searched in combination with all the other parameters to find a solution that best agrees with
the observed geophysical data. In the simple binary coding scheme, these random model parameters are coded to binary domain. Each bit
corresponds to a gene which can take a value of zero or one and each individual in the population is completely described by its bit string called a
chromosome.
In a GA, the coding scheme for each model parameter can be different. This means the search space of each model parameter can be independently
defined and its resolution independently specified. Thus the basic concept of coding the physical parameters using a binary representation limits the
search space, defines the resolution of each model parameter, and limits the number of acceptable models.

Forward model and Fitness function: A forward model is created which takes model parameters as an input and does a simulation to generate the
response. After initialization, the different chromosomes are fed into the forward model and a simulation profile is created. This profile is matched
against the observed data to calculate the amount of misfit or RMS error. The misfit function is mentioned below:

1 N
f (C ) = rms(C ) = ∑ (d i,obs − d i,calc ( xi , C )) 2
N i =1
(1)

We wish to minimize this misfit function else we can construct a different function which we can maximize in order to minimize misfit. This
function is called as fitness function and can be formulated as follows:
1
Fitness = (2)
1 + misfit

Mutation: This operator causes a global as well as local movement of a chromosome in the search space. This operation is done by randomly
selecting a site and flipping the binaty value at that site.

The mutation probability (Pm) is kept very low, because it can destroy good solutions if done on a wrong site.

Crossover: This operation greatly accelerates search early in the evolution of the population. As the name suggests, two chromosomes exchange
information to produce another two daughter chromosomes. The chromosomes thus formed contain characteristics of both the parents. To perform
this operation a random site is generated and the chromosome strings are exchanged about it. Crossover probability is kept relatively high, as it
accelerates the convergence.

191
Selection Operator: Once the fitness of each individual model in the population is determined, selection is employed. .This operator pairs individual
models for reproduction based on their fitness values. Models with higher fitness values are more likely to get selected than models with lower fitness
values. This is the analog of inheritance since fitter models will be selected for reproduction more often. The probability of selection can be directly
related to the fitness value or some function of the fitness (Goldberg, 1989). It is also possible to arbitrarily eliminate or select models which are
below or above the specified fitness thresholds. Some of the different methods of selection are: Roulette wheel selection, stochastic remainder
selection, Tournament selection (Goldberg and Deb, 1991).

Elitism: While performing the selection, some of the models having best fitness values are deterministically selected and copied to the next
generation. These individuals are called as Elites and the procedure is known as elitism. Elitism when incorporated greatly accelerates the
convergence and makes the GA fast but if done in excess may result in premature convergence.

The major advantage of GA is their flexibility and robustness as Global search methods. They are “weak methods” which do not use gradient
information and make relatively few assumptions about the problem being solved. GAs can quickly scan a vast solution set. Bad proposals do not
affect the end solution negatively as they are simply discarded. The inductive nature of the GA means that it doesn’t have to know any rules of the
problem, it works by its own internal rules. This is very useful for complex or loosely defined problems. An efficient optimization algorithm must use
two techniques: Exploration to investigate new and unknown areas in the search space, and Exploration to make use of knowledge found at point
previously visited to help find better points. GA being such an algorithm proves its worth. So, GA’s can deal with highly nonlinear problems and
non-differentiable functions as well as functions with multiple local minima or maxima. They are readily amenable to parallel implementation, which
renders them unable in real-time.

The primary drawback of genetic algorithms results from their flexibility. The designer has to come up with encoding schemes that allow the GA to
take advantage of the underlying building blocks. One has to make sure the evaluation function assigns meaningful fitness measures to the GA. It is
not always clear how the evaluation function can formulated for the GA to produce an optimal solution. GAs are also computationally extensive but
today’s high technological advancement in computations makes this point irrelevant. Sometimes convergence is a problem if the mutation and
crossover constants and type of selection are not properly set.

Methodology (GA): The method described here, uses a GA based approach for performing the inversion. An attempt has been made to exploit all the
above mentioned advantages of genetic algorithm to solve the problem. As mentioned above that due to the complexity and non-uniqueness of the
Zoeppritz equation, it is difficult to invert all six parameters across a reflector. To overcome this problem, we can adopt the layer stripping method.
But in stripping method, some error is generated at every interface that increases and accumulates as we go down to invert the parameters of
underlying layers.

In this approach we first invert the Travel time response in order to get P-wave velocity and thickness of the strata. The algorithm initializes by
assuming a population of layered earth models (number of layers is given as an input by the user) and randomly assigning values of VP and t
(thickness) to each layer. As this system works on all the layer parameters as a whole and inverts them simultaneously, the problem due to layer
stripping simply vanishes. The error what so ever generated gets disseminated into all the inverted parameters. As a result of this, the error per
parameter gets minimal. Experimental results show that convergence of parameters from the travel time data is considered for low offset voz. up to
3000 m.

Module-1 of the algorithm computes the P-wave velocity and thickness of various layers. This VP and thickness hence calculated are given as an
input to the module-2 that uses a similar genetic algorithm for the inversion of seismic reflection coefficient. Constraining the VP increases the
changes of the convergence. Results show that beyond certain offset the inversion of reflection coefficient data yields unique parameters (density
ratios and s-wave velocities of different layers) and the error or misfit is within the permissible bounds. During this inversion the important thing to
note is that, the layer thickness is also given as an input to the module-2 for the conversion of incident angle to offset considering ray bending. Hence
thickness also plays a role of partial constraint. The quality of inversion is substantially improved.

Application on Synthetic data: To show the effectiveness of the proposed method, it is always preferable to apply the technique to synthetic data
where the model to be estimated is known. We have considered a 2-D, three – layered model, as shown in Fig. 4(b). Corresponding to different
layers, some reflection travel time at different offset are analytically and for each interface the P-wave reflection coefficients have bee calculated
using Zoeppritz equations.

Both these set of data, contaminated by random noises have been inverted to determine the model parameters. The synthetic data of travel-time and
reflection coefficient tainted by different degrees of random noises along with the corresponding best fitted model are shown in Figs. 6 and 7. In Fig.
6 the Max-offset is 3000m, and Figs. 6(a,c,e) are the reduced travel time curve for different layers with 0%, 1% and 3% noise level respectively. The
Figs. 6(b,d,f) are the reflection coefficient curves for different interfaces with 0%, 1% and 3% noise level respectively. In Fig. 7 the Max-offset is

192
4000m, and Figs. 7(a,c,e) are the reduced travel time curve for different layers with 0%, 1% and 3% noise level respectively. The Figs. 7(b,d,f) are
the reflection coefficient curves for different interfaces with 0%, 1% and 3% noise level respectively. A reduced time scale with a reduction velocity
of 6000 m/s has been used throughout to present all the reflection times in a compact form.

The Genetic Algorithm used here has following attributes:


Population size: 60
Mutation constant: 0.02
Crossover constant: 0.8

The Genetic Algorithm was applied through 2 diferent modules, one each for travel time and AVO inversion. Module-1 and 2 consisted of various
sub-modules namely initializer, ADC (n-bit analog to digital converter), DAC, fitness calculator, crossover and mutation applier, selection operator,
forward response generator, angle to offset converter, TIR-handler (takes care of total internal reflections taking place). All these sub-modules
controlled from a main central function and execute the inversion. The initial population and the population after 100 iterations of GA are shown in
Fig. 5 (a) and Fig. (b) respectively. Note that here fitness is calculated on a scale of 0-1.

Fig. 5 (a) initial population and its fitness. (b) Population and its fitness after 100 iterations

Fig.6. The Max-offset is 3000m. (a,c,e) Reduced travel time curves for different layers with 0%, 1% and 3% level of noise. (b,d,f) Reflection

193
coefficient curves for different offset with 0%, 1% and 3% level of noise.

Fig.7. The Max-offset is 4000m. (a,c,e) Reduced travel time curves for different layers with 0%, 1% and 3% level of noise. (b,d,f) Reflection
coefficient curves for different offset with 0%, 1% and 3% level of noise.

RESULT AND DISCUSSION

When a cooperative inversion procedure (least square and fuzzy c-mean) is adopted, first with the starting model as a homogenous model, then we
found that the cooperative Fuzzy c-means gives a better fit to the model than the simple inversion with least squares, and further with least squares
inverted output as a starting model, the convergence is better, with NCE and RMSE values going on to minimum. The results in Fig. 3 based on
Fuzzy c-means are close to the results of the true model. When one uses the joint inversion, in such a way that the output from one scheme is used as
input to the other scheme, then the model will be better refined (Trietel and Lines, 2001). We agree that our scheme works much better when layer
boundaries are well defined. As it is shown in Fig. 2, with poor model fit for single inversion and a better model fit for cooperative inversion as
shown in Fig. 3, when the starting model is the least squares inverted model. The results are affected by the noise and the algorithm provides good fit
if the input with well log guided guess is used. The collated table (Table 1) is obtained after inversion results shown in Fig. 4 (a-c) and Fig. 6 (a,-c) is
used to compare the uncertainity in the least square inversion and Fuzzy c-means. Least squares inversion alone shows that generally Vp with an
error of 14-25 %, Vs shows 14-20 % error and density (rho) of 3 % only. Well log constrained least square inversion is better and most acceptable in
comparison with the lower and higher guess values. Fuzzy c-means inversion shows the best acceptable percentage error 0.68-1.74 % for VP, VS and
density (ρ).

Table 1: Collated table comparing the uncertainty in the inversion schemes (co-operative fuzzy c-means) for the three parameters.

Percentage Error in
Initial Guess VP VS ρ
Least squares (lq) alone
Lower value 21.38 16.20 3.53
Median value (Well log constraint) 14.45 14.71 3.83
Higher value 25.13 19.28 3.63
Fuzzy c-means with least squares (lq) as input
Lower value derived lq 5.57 3.31 1.83
Well log value derived lq 1.74 2.04 0.68
Higher value derived lq 7.00 4.62 1.17

194
In the second approach (GA), we used a very sophisticated and robust GA based approach for constrained AVA inversion using travel time. Travel
time inversion using GA gives outstanding results for estimating P-wave velocities and corresponding layer thickness of multilayered subsurface. It
handles vigilantly the non-unique nature of the Zoeppritz equation. This method gives almost a perfect fit for the synthetic data without noise for
offset to depth ratio above 2. The total error generated while inverting all the parameters is 0.069%, 0.223%, 0.39% for 1st, 2nd, and 3rd layers
respectively. Table 2 shows that the error enhances with increase in noise. But the error generated is again minuscule and within the permissible
bounds. For 4000m offset, maximum error generated is 2.1% for 1% noise and 2.35% for 3% contamination. On comparing results, experiments
reveal that both non-uniqueness and error start decreasing with increase in offset.

Table 2. Calculated values for VP, VS, density and thickness of different layers with Genetic Algorithm and for different offsets. The effect of noise
levels can be seen form the table.

Max – offset 3000m Max –offset 4000m


Layer VP VS ρ Thickness %error VP VS ρ Thickness %error
When noise level 0%
1 1501.5 9.7 1000 1497.3 0.35 1502 1.3 1000 1500 0.069
2 1683.6 404.34 1203.43 302 0.72 1700.7 401.3 1197.35 302.34 0.223
3 2219.2 608.29 1252.64 158.79 0.88 2211.9 602.45 1248.25 148.79 0.391
When noise level 1%
1 1509.8 21 1000 1493.2 0.9 1504.9 18 1000 1491.8 0.45
2 1718.8 394.65 1208.43 329.69 1.69 1677.1 402.15 1205.6 303.71 0.94
3 2267.6 608.33 1255.23 157.42 2.07 2123.0 604.53 1254.38 147.85 2.10
When noise level 3%
1 1505 72 1000 1497.3 2.0 1501 7.00 1000 1497.3 0.3
2 1607.5 406.06 1207.01 292.75 3.2 1740.5 410.16 1211.49 318.75 2.24
3 2316 594.4 1256.01 185.27 3.88 2269 606.45 1244.57 168.36 2.35

CONCLUSION

We have reassessed the nature of model fitness and stability issue for AVA based scheme in inverting compressional, shear velocity and density. We
used a cooperative Fuzzy c-means cluster analysis combined with least squares scheme; reinforce the local character of the least squares technique
by finding a local minimum of the error function and the well data allows constraining the solution to a region of the model space where the global
minimum of the error function is located. Our results demonstrate a successful methodology to resolve the three parameters well. We conclude from
the synthetic modelling that the cooperative Fuzzy c-means non linear inversion is more robust than the conventional single optimization scheme
algorithms for inversion.

The second inversion method presented here inverts the reflection coefficient by taking constraints from travel time inversion using Genetic
Algorithms. This technique is capable of estimating all the elastic parameters (VP, VS and ρ) and thickness of multilayered earth in one go. The
feasibility of this method is demonstrated on a generalized synthetic model contaminated with different degrees of noises. From the obtained result
we conclude that the method exhibits exceptionally good outcome if offset to depth ratio is above 2. Even for low offset to depth ratios the method
performs fairly well. This method not only converges fast, but also inverts data with minimal error possible.

REFERENCES

Aki, K, and Richards, PG (1980), “Quantitative seismology: theory and methods (Eds.)”, W.H.Freeman and Co., New York.
Aki, K, and Richards, PG (2002), “Quantitative Seismology”, 2nd edition, University Science Books.
Bezdek, J (1981), “Pattern Recognition with Fuzzy Objective Function Algorithms”, Plenum Press, New York and London.
Castagna, JP, Batzle, ML, and Kan, TK (1993), “Rock physics: the link between rock properties and AVO response. In: J.P. Castagna and M.M.
Backus, Editors, Offset-Dependent Reflectivity-Theory and Practice of AVO Analysis”, SEG, Tulsa, pp 135-171.
Demirbag, E, Courh, C, and Costain, JC (1993), “inversion of P-wave AVO, In: Castagna, J.P., and Backus, M.M. (Eds.), Offset-dependent
reflectivity – Theory and practice of AVO analysis”, Society of Exploration Geophysics, pp 287-302.
Dvorkin J, Prasad, M, Sakai, A, and Lavaie, D (1999), “ Elasticity of marine sediments. Geophysical Research Letter, Vol 26, pp 1781-1784.
Florence C, Rector, J, and Hoversten, M (2004), “Study of gas hydrates in the deep-sea Gulf of Mexico from seismic data”, The Leading Edge,
Vol 23, pp 4356-4369.
Goldberg, DE, and Deb, K (1991), “A comparison of selection schemes used in genetic algorithms”, Foundation of Genetic Algorithm, Vol I,
pp 69-93.
Goldberg, DE (1989), “Genetic Algorithms in Search, Optimization and Machine Learning”, Wesley Pub.Co.
Hoeppener, F, Klawonn, F, Kruse, R, and Runkler, T (1999), “ Fuzzy cluster analysis: Methods of classification, data analysis and image
recognition”, John Wiley and Sons, Inc.
Hoversten et al. (2006), “Direct reservoir parameter estimation using joint inversion of marine seismic AVA and CSEM data,” Geophysics., Vol 71,

195
No 3, pp C1-C13.
Lines, LR, Schulz, AK, and Trietel, S (1988), “Cooperative inversion of geophysical data,” Geophysics., Vol 53, pp 8-20.
Nair, RR, Kandpal, SC (2010). “Effect of co-operative fuzzy c-means clustering on estimates of three parameters AVA inversion,” Journal of Earth
System Science, Vol 119, No 2, pp 137-145.
Ostrander, WJ (1984), “Plane-wave reflection coefficients for gas sands at nonnormal angles of incidence”, Geophysics, Vol 49, pp 1637-1649.
Paasche, H, and Tronicke, J (2007), “Cooperative inversion of 2D Geophysical data sets: A zonal approach based on Fuzzy c-means cluster
analysis”, Geophysics, Vol 72, No 3, pp A35-A39.
Paasche, H, Tronicke, J, Holliger, K, Green, AG, and Maurer, HR, (2006), “Integration of diverse physical-property models: Subsurface zonation
and Petrophysical parameter estimation based on Fuzzy c-means cluster analyses”, Geophysics, Vol 71, No 3, pp H33-H44.
Sain, K and Kaila, KL (1994), “Direct calculation of interval velocities and layer thicknesses from seismic wide angle reflection times”, Geophysical
Journal International, Vol 125, pp 30-38.
Sain, K, and Stoffa, P (1995), “Global optimization methodology in geophysical inversion”, Elsevier Publication. Tokyo.
Shuey, RT (1985), “A simplification of Zoeppritz equations”, Geophysics, Vol 50, pp 609-614.
Stoffa, PL, and Sen, MK (1991), “Nonlinear multiparameter optimization using Genetic Algorithms: Inversion of plane wave seismograms”,
Geophysics, Vol 56, pp 1624-1638.
Trietel, S, and Lines, L (2001), “ Past, Present and future of geophysical inversion-A new millennium analysis”, Geophysics, Vol 66, No 1, pp 21-
24.
Zoeppritz, K (1919), “On the reflection and penetration of seismic waves through unstable layers”, Goettinger Nachr (ed) Erdbebenwellen VIII B, pp
66-84.

196
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Gas Hydrate stability and saturation estimation on Cascadian Margin using ODP Legs 889 & 891:
inferred from petrophysics modeling
Yudhvir Singh1, Ramakrushna Das2 and Rajesh R.Nair3
1,3
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
2
Department of Geology and Geophysics,
Indian Institute of Technology Kharagpur, India
E-mails: 1yudhvir_vijay@rediffmail.com, 2 happyiitkgp@gmail.com, 3rajeshnair@iitm.ac.in

ABSTRACT

We investigate an effective medium model for the elastic modulii of high porosity ocean bottom sediment. The elastic constants of the dry sediment
frame depend on porosity, elastic modulii of the solid phase, and effective pressure. The elastic modulii of saturated sediment are calculated from
those of the dry frame using Gassmann’s equation. The elastic modulii of marine sediment are calculated as a function of porosity, mineralogy,
effective pressure and pore fluid compressibility. The result shows that the shear elastic modulus is more sensitive than the bulk elastic modulus to
the effective pressure. We applied the model to the Cascadian margin leg 146 ODP site 889 and ODP site 891 and observe the gas hydrate and free
gas presence in the site. The hydrate stability zone of site 889 is 224 mbsf and in case of site 891 is 239 mbsf. The corresponding volume of methane
gas in the reservoir ranges from 130 to 415 billion cubic feet for ODP site 889 and 210 to 514 billion cubic feet for ODP site 891.

KEY WORDS: Gas Hydrate, Effective Medium Modeling, Cascadian Margin Leg 146.

INTRODUCTION

Gas hydrates are clathrates of water molecules holding the natural gas (primarily methane). Bottom simulating reflectors (BSRs); (Shipley et al.,
1979) are the first observed indicator for presence of gas hydrate in seismic profile and have opposite reflection polarity than the sea bed. The
negative polarity of this reflection indicates that it results from a decrease in acoustic impedance with depth. There are some other indicators
including plunging BSRs, velocity amplitude pull-downs, double BSRs, blanking and frequency attenuation etc (Thakur, 2010). BSR occurs
approximately at the depth where the base of the gas hydrate stability zone is predicted based on thermodynamic equilibria (Tucholke et al., 1977;
Shipley et al., 1979; Hyndman and Spence, 1992). Because of the temperature dependence of hydrate stability, the depth of the BSR provides a
means of mapping the thermal gradient and heat flux in the overlying sediment (Davis et al., 1990). Although some details of the seismic reflection
properties are not yet fully understood, it appears that the strength and the characteristics of the BSR is determined by the presence of free gas below
the gas hydrate zone (Paull et al., 1996). The presence of free gas represents a very large change in seismic velocity, and therefore produces a very
strong and sharply defined reflection. Theoretical models (Xu and Ruppel, 1999) and synthetic studies (Wood and Ruppel, 2000) indicate that the
BSR is not a necessary condition for the presence of hydrate, as it only occurs when there is free gas beneath the distinct gas hydrate phase boundary.
Indeed, sediments containing gas hydrate have been recovered from areas where there is no BSR (Mathews and Von Huene, 1985).

In an attempt to calibrate the geophysical constrained parameters related to BSR with the direct measurement, the Ocean Drilling Program (ODP)
sampled various BSR horizons on the continental slopes around the Pacific Rim (e.g. Peru, Chile, Costa Rica, Oregon/Washington, Japan) and on the
passive US Atlantic margin (Blake Ridge) also with the aim of understanding gas hydrate characteristics, distribution and concentration in
continental margin settings. According to Kvenvolden (1993), the worldwide estimates for the amount of gas, most of which is methane, in natural
gas hydrates ranges from 1.1 × 10 5 to 2.7 × 10 8 tcf for oceanic deposits and from 5.0 × 10 2 to 1.2 × 106 tcf for permafrost formations. The estimates
are based primarily on the theoretical conditions of hydrate stability rather than direct measurements of the presence and concentrations of hydrate
deposits. USGS has estimated that for the entire United States, the mean in-place volume of gas associated with natural gas hydrates is 3.2 ×105 tcf.
For example, based on the recovery of gas hydrates in pressurized core barrels at the Arco-Exxon N.W. Eileen State No. 2 wildcat well on the North
Slope of Alaska, it has been estimated that about 1.1 × 10 2 m3 of methane is present in gas hydrates in the region of the Prudhoe Bay and Kuparuk
River oil fields (Collett, 1993). The worldwide amount of methane carbon in gas hydrate may be equivalent to about twice the amount of carbon that
resides in all fossil fuels on Earth and also an amount of methane that could be as much as 3000 times the amount in the present atmosphere (USGS,
1998).

METHODOLOGY

Estimation of Elastic Moduli of Mineral Composition: To relate the elastic moduli of sediment to porosity, pore fluid compressibility, lithology, and
effective pressure we first model the elastic moduli of the dry rock frame. At zero porosity the bulk and shear moduli of rock are K and G
respectively. Accounting for mineralogy, the elastic moduli of solid phase are calculated from bulk and shear moduli of the individual mineral
constituents using Hill`s (Hill, 1952) average formula.

1⎡ m ⎛ m f ⎞ ⎤
−1

K = ⎢ ∑ fi K i + ⎜ ∑ i ⎟ ⎥ (1)
2 ⎢ i =1 ⎝ i = 1 K i ⎠ ⎥⎦

1⎡ m ⎛ m f ⎞ ⎤
−1
(2)
G= ⎢ ∑ f i Gi + ⎜ ∑ i ⎟ ⎥
2 ⎢ i =1 ⎝ i =1 Gi ⎠ ⎥⎦

where m is the number of the constituents, ƒi is the volumetric fraction of the ith constituent in the solid phase, Ki and Gi are the bulk and shear moduli
of the ith constituent.
The difference between the lithostatic and hydrostatic pressure is known as the effective pressure. The effective pressure is calculated by the formula:

P = (1 − φ )( ρs − ρ f ) gD (3)

Where φ is porosity, ρs is the bulk density of sediment, ρf is the density of pore fluid, g is acceleration due to gravity; and D is the depth below sea
floor.

Calculation of Elastic Moduli of Sediment: Then calculate the effective bulk moduli(KHM) and shear moduli(GHM) of the dry sediment by using the
Hertz-Mindlin contact theory at porosity φc as below:

1
⎡ G 2 n 2 (1 − φc ) 2 ⎤3 (4)
K HM = ⎢ p⎥
⎣ 18π (1 − v)
2 2

1
5 − 4v ⎡ 3G 2 n 2 (1 − φc ) 2 ⎤ 3 (5)
GHM = ⎢ p⎥
5(2 − v) ⎣ 2π 2 (1 − v) 2 ⎦

where P is effective pressure, K and G are the bulk and shear moduli of the solid phase, ν is the Poission`s ratio of the solid phase and n is the
average number of grain contact in sphere pack. The Poission`s ratio ν is calculated by the formulas below:

ν=
( 3K − 2G ) (6)
2 ( 3K + G )

Calculation of Elastic Moduli of Marine Sediment with Dry Frame:


Case I: Porosity below critical porosity: For the rock whose porosity is below critical porosity φc (φc=36% - 40%, Nur et al., 1998), this effective
medium model connects two end points in the modulus porosity plane, one at zero porosity where the rock elastic moduli are those of the solid phase
and the other at φc where rock is modeled as a random pack of identical spheres (Fig. 1).

Fig. 1. Hashin-Shtrikman (Hashin and Shtrikman, 1963) arrangement of sphere pack, solid and void. From left to right pure solid at zero porosity,
pore solid enveloped by the sphere pack at critical porosity, void enveloped by sphere pack phase and void at 100% porosity. These two end points in
the porosity mocluli plane can be connected with the curves. That have algebraic expressions of the upper or lower Hashin-Shtrikman (1963) bound
(bulk and shear moduli) for the mixture of two components like the pure solid phase and the phase that is the sphere pack. We are modeling

198
unconsolidated sediment so the lower bound is most appropriate. Therefore we are looking for the softest arrangement of the two end point material.
In Fig. 1 an imaginary physical realization of this arrangement is shown. Here the softest component like the sphere packs envelopes the stiffest
component like the solid. Now we calculate the bulk and shear elastic moduli of dry frame where the porosity φ < φc. We know that at porosity φ the
concentration of the pure solid phase in the rock is 1- φ/ φc and that of the sphere pack phase is φ/φc. Now the bulk modulus (Kdry) and shear modulus
(Gdry) of dry frame are calculated by the mathematical formula as below:

−1
⎡ ⎤
⎢ φ / φc (1 − φ ) / φc ⎥ (7)
K dry =⎢ + ⎥ − 3 GHM
4
4 4
⎢ K HM + GHM GHM ⎥
⎣ 3 3 ⎦

−1
⎡ φ / φc (1 − φ ) / φc ⎤
G dry = ⎢ + ⎥ −Z (8)
⎣ K HM + Z Z ⎦

GHM ⎡ 9 K HM + 8GHM ⎤
where Z= ⎢ ⎥
6 ⎣ K HM + 2GHM ⎦

Case II: Porosity above critical porosity: In this model the porosity is higher than the critical porosity. In this model for the high porosity marine
sediment, the low porosity end point is at φc in Fig. 1 and the elastic moduli are calculated from Eq. 1. Marine sediment is unlikely to look like a pack
of identical spheres. The role of this pack is only to provide the end point elastic moduli for building the desired effective medium mode. The other
end point (Fig. 1) at100% porosity is void of zero rigidity (i.e. without pore fluid). Our intention is to model sediment with a non-zero stiffness of the
dry frame. This is why we hypothesize that the appropriate elastic moduli come from the upper Hashin-Shtrikman (1963) bound. Now we calculate
the bulk modulus (Kdry) and shear modulus (Gdry) of dry frame where the porosity is φ > φc, the concentration of the void phase is (φ - φc)/(1- φc) and
that of the sphere-pack phase is (1- φ)/(1- φc). Then the elastic moduli of the effective dry frame are given by the mathematical formula as below:
−1
⎡ ⎤
⎢ (1 − φ ) /(1 − φc ) (φ − φc ) /(1 − φc ) ⎥ (9)
K dry =⎢ + ⎥ − 3 GHM
4
4 4
⎢ K HM + GHM GHM ⎥
⎣ 3 3 ⎦

−1
⎡ (1 − φ ) /(1 − φc ) (φ − φc ) /(1 − φc ) ⎤
G dry =⎢ + ⎥ −Z (10)
⎣ K HM + Z Z ⎦
Calculation of elastic moduli of marine saturated sediment: The subsequent saturated sediment elastic bulk (Ksat) and shear (Gsat) moduli are
calculated using the Gassmann equation (Gassmann, 1951):

φ K dry − (1 + φ ) K f K dry / K + K f
K sat = K (11)
(1 − φ ) K f + φ K − K f K dry / K

Gsat = Gdry (12)

where Kƒ is the bulk modulus of the pore fluid which, at saturation Sw, is the isostress average of those of water (Kw) and gas (Kg) and the bulk
modulus of the pore fluid is given by the formula as below.

−1
⎡S (1 − S w ) ⎤ (13)
Kf = ⎢ w + ⎥
K
⎣⎢ w K g ⎦⎥

Calculation of elastic Velocity of marine saturated sediment: The elastic wave velocity is calculated from the elastic moduli and bulk density ρB as
from Eqs. 14 and 15. The different parameters used in the study are listed in Table 1.

4 (14)
V p = ( K sat + Gsat ) / ρ B
3

199
Vs = Gsat / ρ B (15)

Table 1. Parameters used in the modeling

Parameter Value Reference Parameter Value Reference


Quartz bulk modulus 36 GPa 1 Calcite bulk modulus 77 GPa 2
Quartz shear modulus 45 GPa 1 Calcite shear modulus 32 GPa 2
Clay bulk modulus 20.9 GPa 2 Mica bulk modulus 62 GPa 2
Clay shear modulus 6.85 GPa 2 Mica shear modulus 41 GPa 2
Feldspar bulk modulus 76 GPa 2 Organics bulk modulus 2.94 GPa 2
Feldspar shear modulus 26 GPa 2 Organics shear modulus 2.73 GPa 2
1 2
Carmichael, 1982, Mavko et al., 1998

We apply the theory to Ocean Drilling Program (ODP) Leg 146. This is a review of the geophysical surveys, sea floor measurements, and Ocean
Drilling Program (ODP) data that constrain the occurrence, distribution and concentration of gas hydrate and underlying free gas on the northern
Cascadia subduction zone. The Cascadia subduction zone extends from northern California to Vancouver Island, the focus of this summary is the
continental slope off Vancouver Island. In this region, the Juan de Fuca plate is under thrusting the margin approximately orthogonally at about 40
mm/yr. The Juan de Fuca spreading center is located only several 100 km offshore and the oceanic plate is very young. Since the Eocene has formed
a large clastic sedimentary prism. The prism exhibits active folding, thrusting, tectonic consolidation, and pore fluid expulsion. Further landward
where the sediment section grows rapidly in thickness due to tectonic shortening, sediments undergo severe tectonic compaction and distributed
small-scale deformation. Most of the stratigraphic reflectors are disrupted. Slope basin sediments are deposited on the continental slope, ponded
between the anticlinal and faulted structures, and then progressively deformed.

Kumar and Gupta (1998) examine and illustrate the deep-sea benthic foraminifera from Holocene to Late Quaternary sample from North Pacific
Ocean IODP leg 146, Cascadia Margin. According to them the formation of natural gas hydrates require low temperature (12º C), high pressure (2.6
MPa) and high organic carbon content of sediments (2.0% to 3.5%) and water depths between 300 to 1000 meters below sea floor (Claypool and
Kaplan, 1974; Kvenvolden, 1993). Cascadia Margin is a well-established gas hydrate field and provides an ample opportunity to understand methane
formation and eruptions using various proxies during the Quaternary. Apart from other proxies like geophysical seismic surveys and geochemical
studies, benthic foraminifera also help in gas hydrate exploration. Benthic foraminifera are an important component of the marine community and
sensitive to environmental changes. Numerous species of benthic foraminifera have been found in different methane rich marine settings and have
proved to be good indicator of methane releases (Sen Gupta, 1997). Some species prefer to feed on rich bacterial food sources at methane seeps
showing their potential as indicators of methane release in the geological record. Some methane loving taxa include species of Uvigerina, Bolivina,
Bulimina, Chilostomella, Globobulimina and Nonionella, which can withstand such stressful conditions.

Studies indicate that the benthic faunal composition is strongly correlated with productivity of the surface waters and the delivery of organic matter to
the seafloor (Streeter, 1973). Thus benthic foraminifera are good indicators of paleoproductivity especially in areas where carbon flux is high. At
Hole 890B benthic faunal biofacies show a major change near 5 Kyrs BP. From 12-5 Kyrs BP an interval dominated by biofacies Be-Ba, Cm-Bo,
Uc-Gp, Ei-Us, Ba-Bs, Uc-Up, characteristic of intermediate to high rates of organic flux. The dominance of biofacies Be-Ba, Uc-Gp, Ei-Us, Ba- Bs,
Uc-Up during 2 to 3 and 6.5 to10.5 Kyrs suggests warmer conditions marked by sustained and high flux of organic matter with low oxygenation
when the seasonality was low. Taxa like Bolivina, Bulimina, Cassidulina, Chilostomella, Globobulimina, Nonionella and Uvigerina have been
reported from different seep related environments and can be used as proxy for methane rich environments.

Fig. 2. The northern Cascadia margin showing the leg 146, ODP Sites, whose data was used for present study.

200
The studies of deep sea benthic foraminifera from the methane rich environment of Cascadia margin (IODP, Leg 146, Site 890B) for the past 13.49
Kyrs found the major distribution of 29 dominant species. The particular taxa of foraminifera like Uvigerina, Bolivina, Bulimina, Epistominella are
adapted to high organic, low oxygen, reducing environments of modern methane seep environments from Gulf of Mexico. Northern California,
Monterey, CA, USA and Offshore Japan. Uvigerina peregrina species is also found flourishing in hydrocarbon seeps environment of Green Canyon,
Gulf of Mexico and is considered to be facultative anaerobes, Bulimina spathulata association and relatively high abundance are also reported from
methane seep environments of Green Canyon, Gulf of Mexico. The increasing trend of U. peregrina during the Early Holocene from 10.5 to 6 Kyr
BP and decreasing trend of Ei- Us showing the transition period of Mid Holocene to Early Holocene. On the basis of these results and availability of
methane loving taxa, like Uvigerina,Epistominella, Melonis, Bolivina, Bulimina, Chilostomella and Cassidulina, there are strong possibilities of
methane flux in this site. Ocean Drilling Program (ODP) Leg 146 drilled several wells through the gas hydrate on the mid-continental slope. Which
contain the site 889 and 891, in two cases we are finding the elastic wave velocity, gas hydrate (methane) volume and gas hydrate stability zone.

ESTIMATION OF ELASTIC VELOCITIES

For the calculation of the velocity we required the porosity, effective pressure, mineral composition, saturation of pore fluid. Here we are taking all
the data from North Pacific Ocean IODP site. The input porosities for site 889 and 891 are shown in Fig. 3 (a,c) respectively, which are the
decreasing function of depth. The porosity value varies from 45% to 65% in case of site 889 and from 30% to 50% in case of site 891. Fig. 3(b,d)
shows the increasing trend of effective pressure with depth for both sites. The effective pressure is ranging from 1.5MPa to 3.5MPa for site 889 and
from 2MPa to 6MPa for site 891. We are assuming the sediment’s solid phase includes three mineral components these are calcite, clay, quartz. This
is consistent with the description of the sediments as clay large amount of nanofossils. The mineralogy used in this study consists of 35% calcite, 5%
quartz, and 60% clay for site 889 and for site 891 it is 7% calcite, 11% quartz, and 82% clay. The pore fluid may include seawater, methane, and gas
hydrate. The bulkmoduli and densities of methane and sea water were calculated (Batzle and Wang, 1992) at 36MPa pore pressure and 15˚C (12˚C
for site 891) temperature. For the minerals involved, we used standard (Carmichael, 1982) moduli and density values. Pure methane hydrate
properties are from Sloan (1990). In Table 2 we list bulk moduli (K), shear moduli (G), and density (ρ) for the materials used in the study. From these
inputs we calculate the bulk modulus and shear modulus of the mineral phase using Hill`s average formula. Then input the value of the bulk modulus
and shear modulus of the mineral, sediment porosity value, and effective pressure into the Dvorkin and Prasad formula (Dvorkin et al., 1999).

Fig. 3. Neutron porosity variation with depth (a-site 889, c-site 891) and Effective pressure with depth (b-site 889, d-site 891).

Table 2. Mineralogical composition

Mineral Bulk Modulus (GPa) Shear modulus (GPa) Density (g/cm3)


Calcite 76.8 32 2.71
Clay 20.9 6.85 2.58
Quartz 36 45 2.65
Pure Hydrate 5.6 13.8 2.63
Brine 2.5 0 1.032
Methane 0.1 0 0.235

We calculate the bulk modulus and shear modulus of the dry sediment frame (Fig. 4(a-b) for site 889 and Fig. 5(a-b) for site 891) and then calculate
the bulk modulus and shear modulus of saturated sediment (Fig. 4(c-d) for site 889 and Fig. 5(c-d) for site 891) with saturation of pore fluid by
Gassmann equation (Gassmann, 1951). From Figs. 4 and 5 we observe that the shear modulus is not affect by the pore fluid saturation while the bulk
modulus is affected. For site 889, the bulk modulus remains unchanged at depth 60mbsf-120mbsf, 215mbsf onward and suddenly increases with in
120mbsf-215mbsf due to the presence of saturated pore fluid. And for site 891, the bulk modulus is remained unchanged at depth 60mbsf-180mbsf,
230mbsf on ward and suddenly increases within 180mbsf-230mbsf due to the presence of saturated pore fluid.

Then calculate the P-wave (Vp) and S-wave (Vs) of dry and saturated sediment, which is shown in Figs. 6 and 7 for both sites. S-wave remains
unchanged for dry and saturated sediment because the S-wave is independent of bulk modulus. But in case of P-wave the value is suddenly increase
at the depth of 120mbsf-215mbsf (for site 889) because of the presence of gas hydrate and depth of 115mbsf-125mbsf the P-wave velocity is
suddenly fall due to the presence of free gas. For site 891 the sudden increase in velocity happens at depth of 180mbsf-220mbsf

201
Fig. 4. Elastic moduli versus depth of (a-b) dry sediment frame (c-d) saturated sediment (ODP site 889)

Fig. 5. Elastic moduli versus depth of (a-b) dry sediment frame and (c-d) saturated sediment (ODP site 891)

Fig. 6. Elastic velocity versus depth of (a-b) dry sediment and (c-d) saturated sediment (ODP site 889)

202
Fig. 7. Elastic velocity versus depth of dry sediment (a-b) and of saturated sediment (c-d) for ( ODP site 891).

ESTIMATION OF METHANE GAS VOLUME

The physical properties and elastic models for hydrate bearing sediments are taken from Dvorkin et al. (2000). The model for water saturated
sediments is based on the rock physics model of Dvorkin et al. (1999). This model relates the stiffness of the sediment dry frame to porosity,
mineralogy and effective pressure. The effect of water saturation is modeled by Gassman’s (1951) equations. The model is modified by placing gas
hydrate either in the pore fluid or in the sediment frame. Gas hydrate might also act as grain contact cement for sands but this scenario will not be
considered for this study. If gas hydrate is suspended in the pore fluid, its presence affects only the elastic moduli of the fluid and the bulk density of
the sample. If gas hydrate is instead part of the load bearing sediment framework, then the effective solid phase moduli and sediment porosity must
be based on the mineral plus hydrate mixture.

For the leg 146 ODP site 889 the inputs for the models are estimated from Morelock (1967). The mineral phase is assumed to be 60% clay, 35%
calcite and 5% quartz. The bulk and shear modulus of the constituents are taken from Dvorkin et al. (2000). The porosity varies from 65% at the
ocean bottom to 45% at 400 m depth below OB. The effective pressure ranges from 0 to 9 MPa. The rock bulk density varies from 1.61 g/cm3 to 1.97
g/cm3. The density of brine is 1.035 g/cm3. The temperature at the ocean bottom is 39.7 deg F. The temperature gradient below the ocean bottom is
20 deg F/1000 ft (18.3 deg C/1000 m). Finally, two other input parameters are needed in the model, number of contacts per grain and critical
porosity, taken at 8.5 and 0.40, respectively.

In Fig. 8 (a), the inverted p-wave velocity profile using an AVA inversion scheme (Rajesh and Kandpal, 2010) is plotted together with the modeled
velocity profile for 10%, 20% and 30% bulk hydrate in case of site 889. These calculations are made using the hydrate-part-of-the-solid model. The
hydrate concentration estimated by the frame model is in the range of 20 to 30%. Corresponding volume of hydrate is then estimated for 20% and
30% hydrate concentration. The reservoir extends 240 mbsf from the salt tongue from cdp 1000 to 1120. The extent of the salt dome in the direction
perpendicular to the seismic line is estimated to 240 m. The reservoir is 70 m in thickness. The porosity is estimated to 45% at the reservoir depth.
One unit of hydrate, by volume, contains 164 units of methane gas, by volume. For 20% hydrate concentration, the total volume of methane gas is
estimated to 130 billion cubic feet, for 30% concentration, the total volume is 415 billion cubic feet. In case of site 891 the mineral phase is taken to
be82% clay, 7% calcite and 11% quartz. The porosity varies from 30% at the ocean bottom to 50% at 240 m depth below sea bed. All others
parameters are as discussed above. Fig. 8 (b) shows the inverted p-wave profile in case of site 891. Similarly as in case of site 889, the estimated
volume of gas is 210-514 billion cubic feet.

Fig. 8. Estimation of hydrate concentration from inverted p-wave velocity profile. Comparison of inverted Vp profile with Rajesh and Kandpal
(2010) model results assuming that hydrate is part of the pore fluid. Blue, red and green points show 10%, 20%, and 30% saturation respectively. (a)
for site 889, (B) for site 891.

203
PREDICTION OF GAS HYDRATE STABILITY ZONE

We estimate the P-T (Pressure-Temprature) relationship for the boundary by fitting between experimentally data. Dickens and Quinby-Hunt(1994)
were the first investigators to make measurement of methane hydrate stability in seawater, Brown et al. (1996) fit the data of Dickens and Quinby-
Hunt (1994) to a second order polynomial and obtained the following equation:

1
= 3.83 ×10−3 log10 P + 8.64 ×10−5 (log10 P)2 (16)
T

Fig. 9 represents the methane hydrate stability curve calculated from Eq. 16 and the temperature profile in the Cascadia margin as a function of depth
below the ocean bottom. The ocean bottom depth for our sites is 260bsfm. A value of 18.3 deg C/1000m was assumed for the Cascadian margin
temperature gradient below OB and a variant of 0.915 deg /260m from this value is represented on the adjacent curves. For site 889, the methane
hydrate stability zone is found between the ocean bottom and 224mbsf, given by the intersection of the two curves (Fig. 9a). For site 891, the gas
hydrate stability zone is found between the ocean bottom and 239mbsf.

Fig. 9. Empirical prediction of methane hydrate stability zone. The thin curve shows Hydrate stability and thick one shows temperature profile in
Cascadia margin. (a) site 889, (b) site 891.

CONCLUSION

The most information on gas hydrate has come from velocity verses depth curve. The velocity plot at depth 130; the velocity suddenly increase 1.7
km/s to 1.85 km/s , at the depth 220mbsf velocity reduces to 1.6 km/s and at the depth 240mbsf velocity increase to 1.75 km/s. Therefore, from this
velocity curve we can conclude the ODP site 889 contains gas hydrate within the depth 130mbsf -220mbsf and contain free gas within 240mbsf to
240mbsf. The velocity plot of ODP site 891 shows that at depth 180 the velocity suddenly increase 1.8 km/s to 2 km/s , at the depth 220mbsf velocity
reduces to 1.75 km/s and at the depth 240mbsf velocity increase to 1.95 km/s. Therefore, from this velocity curve we can conclude the ODP site 891
contains gas hydrate within the depth 180mbsf to 220mbsf and contain free gas within 220mbsf to 250mbsf. The hydrate stability zone of site 889 is
224mbsf and in case of site 891 is 239mbfs. The hydrate stability zone calculated from empirical equation (Eq. 16) fairly satisfies the hydrate zone
found through the velocity analysis. The estimated volume of methane gas in the reservoir ranges from 130 to 415 billion cubic feet for ODP site
889 and 210 to 514 billion cubic feet for ODP site 891.

REFERENCES

Batzle, M, and Wang, Z (1992). “Seismic properties of pore fluids,” Geophysics, Vol 57, No 11, pp 1396-1408.
Brown, KM, Bangs, NL, Froelich, PN, and Kvenvolden, KA (1996). “The nature, distribution, and origin of gas hydrate in the Chile Triple Junction
region,” Earth and Planetary Science Letters, Vol 137, pp 471-483.
Carmichael, RS (1982). “Practical Handbook of Physical Properties of Rocks,” Vol II, CRC Press, Boca Raton, Fla.
Claypool, G, and Kaplan, I (1974). “The origin and distribution of methane in marine sediments. In: Kaplan, I. (eds) Natural gases in marine
sediments,” Plenum, New York, pp 315-340.
Collett, TS (1993). “Natural gas hydrates of the Prudhoe Bay and Kuparuk River area, North Slope, Alaska,” AAPG Bulletin, Vol 77, No 5, pp 793-
812.
Davis, EE, Hyndman, RD, and Villinger, H (1990). “Rates of fluid expulsion across the northern Cascadia accretionary prism: constraints from new
heat flow and multichannel seismic reflection data,” Journal of Geophysics Research, Vol 95, pp 8869-8889.

204
Dickens, GR, and Quinby-hunt, MS (1994). “Methane hydrate stability in seawater,” Geophysical Research Letters, Vol 21, No 19, pp 2115-2118.
Dvorkin, J, Prasad, M, Sakai, A, and Lavoie, D (1999). “Elasticity of marine sediments,” Geophysical Research Letters, Vol 26, pp 1781-1784.
Dvorkin, J, Helgerud, MB, Waite, WF, Kirby, SH, and Nur, A (2000). “Introduction to Physical Properties and Elasticity Models,” Natural Gas
Hydrate in Oceanic and Permafrost Environments, M.D. Max, ed, Kluwer Acad Pub, Dordrecht, pp 245-261.
Gassmann, F (1951). “Elastic Waves Through a Packing of Spheres,” Geophysics, Vol 16, pp 673-685.
Hashin, Z, and Shtrikman, S (1963). “A variational approach to the theory of elastic behaviour of multiphase materials,” Journal of Mechanics and
Physics of Solids, Vol 11, No 2, pp 127-140.
Hill, R (1952). “The elastic behavior of a crystalline aggregate,” proceding Physical Society London, Ser 65 A, pp 349-354.
Hyndman, RD, and Spence, GD (1992). “A seismic study of methane hydrate marine bottom simulating reflectors,” Journal of Geophysics Research,
Vol 97, pp 6683-6698.
Kumar, A, and Gupta, AK (1998). “Deep sea benthic formation foraminifera as a proxy of methane hydrates from IODP site 890B Cascadia Margin”
Proceedings of the 6th International Conference on Gas Hydrates (ICGH 2008), Vancouver, British Columbia, CANADA, July 6-10, 2008.
Kvenvodlen, KA (1993). “Gas hydrate geological perspective and global change,” Reviews of Geophysics, Vol 31, pp 173-187.
Mathews, MA, and Von Huene, R (1985). “Site 570 methane hydrate zone” in S. Orlofsky, ed. , Deep Sea Drilling Project Initial
Repots, Vol 84, US Government Printing Office, pp 773-790.
Mavko G, Mukerji T, and Dvorkin, J (1998). “The Rock Physics Handbook: Tools for Seismic Analysis in Porous Media,” Cambridge
Press.
Morelock, J (1967). “Sedimentation and mass physical properties of marine sediments, western Gulf of Mexico,” Unpublished Ph.D. Dissertation,
Texas A&M Univ.
Nur, A, Mavko, G, Dvorkin, J, and Galmudi, D (1998). “Critical porosity: a key to relating physical properties to porosity in rocks,” The Leading
Edge, Vol 17, pp 357-362.
Paull, CK, Buelow, WJ, Usslerr III, W, and Borowski, WS (1996). “increased continental-margin slumping frequency during sea-level lowstands
above gas hydrate-bearing sediments,” Geology, Vol 24, No 2, pp 143-146.
Rajesh, RR, Kandpal, SC (2010). “Effect of co-operative fuzzy c-means clustering on estimates of three parameters AVA inversion,” Journal of
Earth System Science, Vol 119, No 2, pp 137-145.
Sen Gupta, BK, Platon, E, Bernhard, JM, and Aharon, P (1997). “Foraminiferal colonization of hydrocarbon-seep bacterial mats and underlying
sediments, Gulf of Mexico Slope,” Journal of Foraminiferal Research, Vol 27, pp 292-300.
Shipley, TH, Houuston, MH, Buffler, RT, Shaub, FJ, Mcmillen, KJ, Load, JW, and Worzel, JL (1979). “Seismic evidence for widespread possible
gas hydrate horizons on continental slopes and rises,” The American Association of Petroleum Geologists Bulletin, Vol 63, No12, pp 2204-2213.
Sloan, ED (1990). “ Clathrate hydrates of natural gases,” Marcel Dikker, Inc., New York.
Streeter, S (1973). “Bottom water and benthonic foraminifera in the north, Atlantic- glacial- interglacial contrasts,” Quat. Res., Vol 3, pp 131-141.
Thakur, NK (2010). “Gas hydrates as alternative energy resource – seismic methods,” Current Science, Vol 99, No 2, pp 181-189.
Tucholke, BE, George, MB, Ewing, JI (1977). “Gas-Hydrate Horizons Detected in Seismic-Profiler Data from the Western North Atlantic,” AAPG
Bulletin, Vol 61, No 5, pp 698 – 707.
USGS, http://www.usgs.gov
Wood, WT, and Ruppel, C (2000). “Seismic and thermal investigations of the Blake Ridge gas hydrate area: A synthesis,” Proc. Ocean Drill.
Program, Sci. Results, Vol 164, pp 253-264.
Xu, W, and Ruppel, C (1999). “Predicting the occurrence, distribution, and evolution of methane gas hydrate in porous marine sediments,” Journal of
Geophysical Research, Vol 104, No B3, pp 5081-5095.

205
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and
National Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

An Investigation into Enhancement of Surface Finish of Pulse Electrochemical drilled


deep hole using rotating electrode
D. S. Bilgi1, P. V. Jadhav2 and A. S.Harel3
1
Principal, B. V. Women’s College of Engineering, Pune, India, Pin: 411043
2,3
Research Scholar, Department of Mechanical Engineering,
B. V. U., College of Engineering, Pune, Maharashtra, India, Pin: 411043
Email:1dsbilgi@yahoo.com,2pvjadhav_19@rediffmail.com,3avinashharel@gmail.com

ABSTRACT
The Electrochemical Machining (ECM) is widely used in machining variety of components used in aerospace, automotive, defense, for deep hole
drilling of blades in steam turbines as ratio of depth to diameter increases it become extremely difficult to produce such holes. Due to low
machining accuracy ECM is yet to be a best alternative process. ECM with pulse current offers an enhanced accuracy control. This paper presents
experimental investigation of PECM parameters such as Voltage, Pulse on time and duty cycle on surface enhancement by rotating electrode
(cathode tool) arrangement. This results shows PECM with rotating tool has enhanced surface finish by 41% as compared to stationary tool in
PECM. (Range1.18 to 0.77 micron). The design of experiments was done by 2k factorial designs. The experimental results were analyzed by
analysis of variance (ANOVA) method and by plotting various graphs.
KEY WORDS: ANOVA; DoE; PECM; rotating tool movement; surface roughness; deep hole.
INTRODUCTION

Electrochemical machining (ECM) process is generally used for machining complex shape and hard materials, ECM generates no burrs, no
internal stress, has a long tool life, higher material removal rate and surface quality. However, due to its relatively low machining accuracy,
difficulties in tool design and electrolyte disposal ECM is not a commonly used technology. Hydrogen gas bubbles and Joule heat generated in the
interelectrode gap (IEG) causes varying local electrolyte conductivity and hence non-uniform distribution of the gap [1].The stray removal in
ECM adversely affects dimensional accuracy and surface quality of machined components [2]. Some flow field disrupting phenomena such as
cavitations and striation in electrolyte flow worsen accuracy and the uniformity of the ECM’d products. Electrochemical machining is an anodic
dissolution process with employ, low D.C voltage across pre-shaped cathode tool and anode workpiece.ECM with pulse current yield higher
accuracy, control [4]. Many attempts have made to improve machining quality with limited success. The progress has been slow because of the
complex nature of the ECM process.

Therefore, this study addresses the improvement of quality of surface finish in ECM by modifying the electrolyte flow distribution. An ECM with
Rotating electrode movement is proposed to enhance the uniformity of electrolyte flow and to reduce or eliminate the flow field disrupting
processes. A significant improvement in surface finish is observed. There are numerous advantages: the ‘cutting tool’ is not subject to wear and
the resulting machined surface is both stress and crack-free. The process works with a DC or pulsed current source and is combined with a
cathode tool that is either fixed or has some linear feed movement. The goal now is to speed up the design process of the cathode tool and also to
improve the design of the final work piece.

ANALYSIS OF EXISTING PROBLEMS

The Machining accuracy in ECM largely depends on the electrolyte flow field distribution. The low field distribution [3]
sometimes results in abnormal dissolution (such as striated dissolution). The machined surface of deep drilled hole wall often
shows evidence of the striation flow due to sharp divergent flow in IEG. And even sparking causing cavitation and striation.

Additionally, these phenomena are often unstable and random, and therefore, further deteriorate the uniformity of ECM products
and the process stability.
Fig.1 ECM Process Fig.2 Velocity distribution and Striation

PROPOSED ROTATING ELECTRODE IN PECM

This study attempts to reduce the effects of the flow field variations to achieve higher process accuracy and uniformity by rotating cathode(tool)
and employing ECM with pulse current(PECM). Rotating electrode (cathode tool) fig.2. Yields shifting of cavitation region due to continuously
varying flow field distribution in IEG and eliminates dead dissolution region and striation [3].
Pulse power generator to supply working voltage across the gap between the cathode and anode. By applying voltage pulses, having short pulse
on time, all the experiments were conducted. It uses small initial electrode gap of 0.1mm for all experiments. Electrochemical dissolution takes
place during on time (ton), [5,6]while no dissolution takes place during off-time (toff) pulse toff allow the electrolyte to carry away the reaction
products of anodic dissolution from the gap. Drilling was carried out with a constant tool feed rate as shown in fig (2).

Fig.3. Rotating tool movement PECM & Gap distribution.

EXPERIMENTAL SET UP.

The experimental set-up used for small hole drilling (2 mm diameter) as shown in the fig.3.It consists of (i) tool feed arrangement (Rotating tool/
stationary tool),
2) Machining chamber, 3) Electrolyte flow system and 4) Pulse power supply. Experiment where conducted using mixed electrolyte (170 gm/lit)
NaNo3 ,10% by Wt HNO3 to avoid wild corrosion and sludge formation during the process [1,3].Current sensing comparator was incorporated
into the tool feed arrangement to avoid short circuiting between tool and work piece, stepper motor controls the tool feed arrangement along z
axis. Servo motor rotates about z axis.

Fig. 4 Experimental Set-up for Rotating tool PECM

Machining chamber consist of table, work holding device, blow off system. Electrolyte flow system used anticorrosive submersible
pump, electrolyte filter, electrolyte tank, pressure gauge and constant discharge flow control valve. Pulse power supply with
constant voltage (CV) made of rectangular pulsed shape was chosen. PECM provides smaller IEG without boiling of electrolyte in
gap, [3], that necessitates limiting the value of IEG across tool and work piece. Pilot experiments with smaller IEG (0.1mm)

208
suffered due to short circuiting between work piece and tool, because of cavitation and sludge formation. The limiting current, to
avoid short circuiting, was compute for controllable IEG, electrolyte conductivity and are of tool, and of operating voltage. [5]
Using stationary and rotating tool feed arrangement. The sufficient electrolyte (Q=12lit/min) was maintained and hole (10 to
15mm) drilled using INCONEL 718 work piece.
Using full factorial design (23) all the experiments were conducted for stationary and rotating tool arrangement. After machining,
the surface roughness value was measured with the help of surf test equipment. Surface roughness produced on the work pieces by
using the rotating electrode & stationary electrode was measured with “Surface Roughness Tester – SJ 201P”.
A) Specification of Machine: PECM power supply , technical specifications; Power rating 3 kva, Working voltage 415 v/ 3 phase,
Vertical travel of electrode: 120 mm, table travel: 120 mm, Pump capacity:12 liters per min Electrolyte tank capacity:
175liters,Machining chamber: 25 liters, Electrode sizes: To be selected depending upon dimensions of work piece. Most of
components are fabricated using stainless steel material or anticorrosive material.
B) Cathode Tool
The tool dimension is slightly i.e. 30% smaller than the size of cavity to allow the overcut (front machining gap) 1.5 times the front
gap.

Fig .5. Tool (All Dimensions are in mm used for drilling)

DESIGN OF EXPERIMENTS

Experiments were planned using full factorial design of experiments for three variables. Experiments where planned using 2k=3
factorial design multiple linear regression to investigate, effect of ‘K’ factor with smallest number of (08) runs for factor screening
experiments [09].
Generalized process model can be expressed as:
Y= f(x1, x2, ------xk) where y= response and f= response function and x1, x2-----xk are controllable variables.
A regression model was fitted to the experimental data and the response surface is given by equation (1)
y = b0+b1x1+b2x2+b3x3+b12x1x2+b13x1x3+b23x2x3+b123x1x2x3 ---------------------------------------------------------------------- (1)
Where y is response, the β’s are parameters / constants whose values are to be determined. The various process variables
represented A, B, C as shown in table no (1).The response surface equation for evaluating surface roughness are obtained
separately by calculating the coefficients of equation 1 and they are given in table (3a, 3b) respectively.

Table no. 1: coded levels and actual values of different parameters

Low High
Parameters Unit Nomenclature
(-1) (+1)

V
Voltage (v) A 12 16

µs
Pulse on time(ton), B 50 500

%
Duty Cycle (%) C 48 80

The plan of experiment is given in the table to all the experiments where conducted stationary tool, rotating tool randomly and surface
roughness were recorded for each experiment.

209
Table2. RESULTS FOR Stationary Electrode and Rotating electrode

Sr. Voltage (v) Pulse on time Duty cycle Surface roughness (µm)
No.
(µs) (%) Stationary Rotating

Electrode Electrode

1 +1 +1 -1 1.15 0.78

2 +1 -1 +1 1.0 0.7

3 -1 +1 -1 1.2 0.82

4 -1 -1 +1 1.1 0.74

5 -1 -1 -1 1.4 0.98

6 -1 +1 +1 1.25 0.84

7 +1 +1 +1 1.18 0.77

8 +1 -1 -1 1.34 0.93

VI ANALYSIS OF EXPRIMENTS:-

To know the significance of the regression equation ANOVA was conducted. It significantly establishes between response surface
and controllable parameter through f test analysis equation (2)

Fo = SSA/VA = MSA ------------------------------------------------------- (3)


SSE/VE MSE

Where SSA stands for sum of square of A, SSE stands for sum of square due to error, VA and VE stands for variance and error respectively.
Anova with 95% confidence interval for stationary tool and rotating tool is given in the table (3ab).

Table 3a: FINAL ANOVA TABLE (stationary Electrode)


Sr. Degrees of Variance or Mean
Factor Sum of Squares Fo
No. Freedom Square

1 A 0.0098 1 0.0098 196.00

2 C 0.039 1 0.039 784.00

3 BC 0.065 1 0.065 1296.00

Pooled error 0.005100 4 0.00012


5

Table 3b: FINAL ANOVA TABLE (Rotary Electrode)

Sr. Degrees of Variance or Mean


Factor Sum of Squares Fo
No. Freedom Square

1 A 0.00281 1 0.00281 225.00

2 C 0.025 1 0.025 2025.00

210
3 BC 0.033 1 0.033 2601.00

5 Pooled error 0.001251 4 0.00031

From f distribution at 95% confidence level we find that F 0.05,1,4 = 7.71i.e. F limit = 7.71 F0 values for stationary and rotating tool are given in the
table () if critical F value (7.71) it implies factor effects of A, C and AC are significant since F0 >F. The regression equation can be return as
below

Y=bo+b1x1+b3x3+b23x2x3--------------------------------------------------------------------------------------------------------------------------- (4)

Therefore final regression equation in terms of coded factors for stationary electrode and rotating electrode is given equation (4, 5) respectively

Surface Roughness (Ra) = +1.20 - 0.035 * A- 0.070 * C + 0.090 * B *C-------------------------------------------------------------------5)


Surface Roughness(Ra) = +0.83-0.019 * A - 0.056*C+0.064*B*C---------------------------------------------------------------------------(6)

RESULTS AND DISCUSSIONS

The surface roughness can be related to many factors such as voltage, Pulse 0n time and duty cycle. A DoE (design of experiment) based study
conducted in the present work that includes above-mentioned three important factors has come up with many important results. A multiple linear
regression model that includes some of the very important factors in predicting surface roughness has been developed. Using the model presented
here, one can select the optimum combination of voltage, duty cycle and interaction between pulse on time and duty cycle parameters for better
performance of surface roughness.
Selecting the Factors for Mathematical Modeling
A Normal Probability plot of the estimates of the effects can be used in selecting the most influential factors. The effects that are negligible are
normally distributed, with mean zero and variance σ2 and tend to lie along a straight line on this plot, whereas significant effects will have non-
zero means and will not lie along the straight line. Thus the mathematical model will be specified to contain those effects that are apparently non-
zero, based on the Normal Probability plot
.

Design-Expert® Software Half-Normal Plot


Surface Roughness

99

p-value = 0.026 Half-Normal % Probability


A: Voltage 95
B: Pulse on time BC
C: Duty Cycle 90

80
C
70
A
50

30
20
10
0

0.00 0.05 0.09 0.14 0.18

|Standardized Effect|

Fig: 6 Normal probability plots for effect estimate.

From the data obtained for this study, the normal probability plot as shown in Fig: 5 can be constructed to pick the significant effects. After
examining the graph 1 the factors selected for the mathematical modeling are main factors A, B & C and interaction between B & C. A multiple
linear regression model developed considering these factors is presented.

Effect of parameters on surface roughness:- The multiple linear regression equation model equation (4, 6) include significant factor that affects
in prediction of surface roughness. Using these models main effects of factors namely A (voltage), C (duty cycle), and interaction between BC
(pulse on time and duty cycle). The given in fig 6 for stationary and rotating tool respectively..
Design-Expert® Software One Factor
Design-Expert® Software One Factor
Surface Roughness 1.42
Surface roughness 0.98

X1 = A: Voltage X1 = A: Voltage
0.905
Actual Factors 1.31 Actual Factors
B: Pulse on time = 0.00
Surface roughness
Surface Roughness

B: Pulse on time = 0.00


C: Duty Cycle = 0.00 C: Duty cycle = 0.00

1.2 0.83

1.09 0.755

0.98 0.68

-1.00 -0.50 0.00 0.50 1.00 -1.00 -0.50 0.00 0.50 1.00

A: Voltage A: Voltage

Fig 7: Voltage vs. Surface Roughness

211
Effect of voltage on surface roughness:

Fig 6: shows the effect of voltage on surface roughness value for both a) stationary electrode and rotating electrode in PECM.As voltage
increased from (12 to 16 volts), surface roughness decreased by about 50% and so (1.18µm to 0.77µm) in ‘a α b’ respectively.
Anodic dissolution takes place in PECM from the work piece, since work piece consist of different consistitents whose electrdes potential are
different lead to preferential dissolution[6].
At low voltage rough surface is produced as current density being low cause eatching effect and highly rough surface is produced (7 ). At higher
voltages difference in electrode potential of the constituents deminish that lead to incresing dissolution resulting decresing surface roughness
value.

Design-Expert® Software One Factor Design-Expert® Software One Factor


Surface Roughness 1.46
Surface roughness 1.01

X1 = C: Duty Cycle
X1 = C: Duty cycle
1.3325
Actual Factors 0.9175
Actual Factors
A: Voltage = 0.00
Surface Roughness

A: Voltage = 0.00

Surface roughness
B: Pulse on time = 0.00
B: Pulse on time = 0.00

1.205
0.825

1.0775 0.7325

0.95 0.64

-1.00 -0.50 0.00 0.50 1.00 -1.00 -0.50 0.00 0.50 1.00

C: Duty Cycle C: Duty cycle

Fig 8: Duty cycle vs surface roughness

Fig 7: shows that at low frequency, ECM removal rates are greater than that of DC and PC , and the surface roughness is also lower due
to the elimination of oxide film reheating during the cathodic cycle. As shown in Figure 9, a minimum in surface roughness is obtained
when the anodic duty cycle, which represents the actual time when material is removed from the surface of the sample, is about 80%.
And when 100% anodic duty cycle refers to DC, which clearly has the poorest surface quality. When the cathodic duty cycle is
changed, the best surface quality is obtained. This parameter (Duty Cycle) still in under investigation for the purpose of surface
roughness.

Interpretation plot for interaction between Pulse on time and Duty cycle:

Design-Expert® Software Interaction


Design-Expert® Software C: Duty cycle
Interaction Surface roughness 0.99
C: Duty Cycle
Surface Roughness 1.42 C- -1.000
C+ 1.000
C- -1.000
C+ 1.000
X1 = B: Pulse on time 0.915
X2 = C: Duty cycle
S u rfa ce ro u g h n e s s

X1 = B: Pulse on time 1.315


X2 = C: Duty Cycle
Surface Roughness

Actual Factor
Actual Factor A: Voltage = 0.00 0.84
A: Voltage = 0.00 1.21

0.765
1.105

0.69
1

-1.00 -0.50 0.00 0.50 1.00 -1.00 -0.50 0.00 0.50 1.00

B: Pulse on time B: Pulse on time

Fig: 9: Interpretation plot for interaction BC


Fig 8: shows the interaction between pulse on time and duty cycle (BC). The two level factorial reveled the powerful interaction between duty
cycle and pulse on time.

Surface Response Plots

Combined effects of various process parameters can be better visualized with the help of three dimensional response surface plots. A response
surface is a plane of y – values (response variable values) generated by various combinations of x1 and x2. These plots for various combinations.
Surface response of interaction BC.

212
Design-Expert® Software
Design-Expert® Software
Surface roughness
Surface Roughness 0.98
1.4
0.7
1 0.97
1.38 X1 = B: Pulse on time
X1 = B: Pulse on time X2 = C: Duty cycle

Surface roughness
X2 = C: Duty Cycle 0.9075

Surface Roughness
1.2975
Actual Factor
Actual Factor A: Voltage = 0.00 0.845
A: Voltage = 0.00 1.215

0.7825
1.1325

0.72
1.05

1.00 1.00
1.00 1.00
0.50 0.50 0.50 0.50

0.00 0.00 0.00 0.00


-0.50 -0.50
C: Duty Cycle -0.50 -0.50
B: Pulse on time C: Duty cycle B: Pulse on time
-1.00 -1.00 -1.00 -1.00

Fig. 10: surface for Electrode diameter vs. duty cycle

The 2- Level factorial approach is very effective as a screening tool. After identifying the vital few factors we move in depth study via response
surface methods (RSM) and central composite rotatable design (CCRD) can be performed [5].

Obtained hole

Comparison of Stationary & Rotating Electrode:-

(Rotating electrode) (Stationary electrode)

Sr. no. 7, voltage= 16v, f=0.8mm/min, Pulse on time= 500µs, duty cycle=80%, speed = 60 rpm, pressure=
0.9kg/cm2
Fig: 11 surface roughness profiles of hole during pecm
Surface Roughness profile obtained for two holes with stationary electrode and rotary electrode arrangement during PECM. For the experimental
Sr. no 7 parameters of above experiments are given in the respective caption (0.8*3=2.4mm). Surface was measured from top face of the hole
with cut off length is 0.25 mm and sampling size is ‘3’. The surface roughness value improved from (1.18 to 0.77µm). Hence Enhancement of
surface finishes ware improved by % value of (%) with PECM using Rotating electrode.

CONCLUSION

Improvement of surface finish of Electrochemical machining using rotating electrode taking input parameters as Voltage, Pulse on time, Duty
Cycle and output parameters as surface roughness following facts can be concluded.
When the Voltage & duty cycle increases the surface roughness value is decreased. Keeping Electrolyte concentration and feedrate constant and
electrode gap (0.1mm).
When stationary electrode is compare to the rotating electrode. The rotating electrode gives better Surface finish than the stationary electrode.
Increase in Duty cycle (C) has effect on surface roughness value (decreased).
The interaction between B (pulse on time) and C (Duty cycle) is important for the response of surface roughness.
From ANOVA analysis, it is found that A, C, and BC are more important factors of surface roughness performance. While the compare between
rotating & stationary electrode the rotating electrode is better surface finish.

213
Design of experiments and analysis of variance- helped in –Identifying the significant factors affecting response factors. Developing regression
models.
Results shows the effect of surface finish on nickel based super alloy 718 using deep hole drilling in PECM with parameters of Voltage, Duty
cycle and pulse on time.

ACKNOWLEDGEMENT

The authors acknowledge the financial support provided by the department of science and Technology, New Delhi, to the project no
DST.SR/S3/MERC-64/2005, entitled Enhancement of Surface finish of electrochemically drilled deep hole”

REFERENCES

O.V. Krishnalal Chetty, Dr. V. Radhakrishanan. (1978) “A study of surface production in electrochemical machining.” Proceedings of the 8th
AIMTDR conformance IIT Bombay, pp.no.562
Ahmad, M.S. and Duffield, (1989) “A Deep hole drilling using ECM, Society of Manufacturing Engineers (SME)” Conf.MS 89-816,
K.P.Rajurkar, D. Zhu, B. Wei. (1998). “Minimization of machining allowance in electrochemical machining” Annals of the CIRP volume 47/1,
pp. no. 165”
K.P.Rajurkar, D. Zhu. (1999). “Improvement of electrochemical machining accuracy by using orbital Electrode movement” Annals of the CIRP
volume 48/1, pp. no. 139”
K.P. Rajurkar,et al (1999) “New development in electrochemical machining” annuls of the CIRP vol.48/2,pp No.567-
D.S.Bilgi, V.K.Jain, Sekhar R., Mehrotra. S. (2004). “Electrochemical deep hole drilling in super alloy for turbine rotor, Journal of material
process technology” vol. 149 .pp 445
D.S.Bilgi, V.K.Jain and Shekhar, (2007) “Predicting radial over cut in deep holes drilled by shaped tube electrochemical machining, journal of
adv Manuf Technology”, pp1185,.
Li Yong, Zheng Yunfel, Yang Gaung, “Localized electrochemical micromachining with gap control”. .
Douglas C. Montgomery, (2001) “Design and Analysis of Experiments”, John Wiley and Sons, Prentice Hall Publication,.
Mishra, P.K., (1997) “Non conventional machining, the institute (India) text book series”, Narosa publishing house, New Delhi,

NOMENCLATURE

A Voltage (V)
B Pulse on time (µs)
C Duty cycle (%)
σ2 population variance,
N total number of observations,
yi ith observation or response data,
i = 1,2,3,……..N
i number of levels for each factor,
Ai sum of observations under ith level,
Ai average of observations under Ai level,
T sum of all observations,
T Average of all observations,
nAi number of observations under Ai level,
kA number of levels or treatments for factor A
kB number of levels or treatments for factor B.
SSA Sum of square for factor A.
ν Degrees of freedom
CF Correction factor
Ra Surface roughness

214
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Marginal gas field development using directional/multilateral drilling


S. R. Joshi
Department of Petroleum Engineering
Maharashtra Institute of Technology Pune, India
E-mail: sanjosh05@hotmail.com

ABSTRACT

Make marginal gas field development and production commercially viable using directional / multilateral drilling technology. The technical objective
of directional / multilateral is to expose hydrocarbon bearings / discrete pockets of H-C reservoir to conventionally drilled mother bore.
The aim of this paper is the planning and drilling two laterals from a vertical well to cover to H-C bearing one at North-East and other is South-West.
A case study of marginal gas field in Rajahmundry (Krishna – Godavari) asset has taken as reference

INTRODUCTION

The marginal field conveys the impression that the field in question may not be economical to produce. However its exact meaning is subject to
specific conditions, which vary from producing regions, operating companies and technical feasibility. Syalkeim (2004) defined the term marginal as
“discovery that for reasons of materiality, risk, economics or capital allocation does not merit allocation of resources to progress’. A marginal field can
be an undeveloped prospect, mature field or an abandoned field. Marginal as applied to the reservoir means that the reserves although present will be
difficult to produce.

A 1.54 sq. km marginal field on east-coast of India where vertical exploratory well was drilled and field data available as the under compacted gas
bearing below 3025 M in subsurface had flow rate through 16/640 choke was 185,272 m3/day and condensate 65 bbls/day with FTHP (Flowing Tubing
Head Pressure) 4807 psi. Out of there stacks of sand under-compacted clay-stone, the upper layer of sand in the interval of 3053 M to 3058 M has
been selected for the paper. The two pockets of gas bearing, one at the same depth towards north (850 M drift) and other 40 M down the vertical
towards south (650 M drift) as shown in fig. 1.

GEOLOGY OF THE AREA

Presence of a high is interpreted in the area and the well was drilled on the SW margin of the high. A fault is seen truncating the high on the northern
part. The well was drilled up to a depth of 3900 m through Tertiary sequence and part through Upper Cretaceous Sequence. The Cretaceous sequence
is characterized by Claystone with pockets of sandstone and siltstone. The tertiary sequence is represented by alternate sandstone and claystone beds
of varying thickness (figure 2). Three pockets of sand are noticed in the well within the high pressure zone. Log analysis indicated that the sand
horizon located between 3078 – 3081 m is gas bearing with Sw varies from 25 % -30 %. Three more gas bearing zones are noticed between 3000-
3005 m, 3010 – 3017 m, and 3090 – 3098 m. with higher water saturation of 60 % -65 %.
2154 M

2372.68

South-West North

2800 M

3055 M

200 M
200 M 650 M 850 M

3088 M

3096 M

3800 M

Fig 1

216
F i g 2 : Stratigraphy of exploratory well

Detailed study carried out using 2D and 3D seismic has indicated that the gas bearing sand is essentially channel deposits typical of meandering
geometry. Therefore it was essential to plan the well location very carefully to optimize production. Out of there stacks, the upper layer of sand in the
interval of 3053 M to 3058 M has been selected for the paper and two pockets of gas bearing one at the same depth towards north (850 M drift) and
other 40 M down the vertical towards south (650 M drift) as shown in fig. 2. Plan of two multilaterals with double curve design one at KOP1 (Kick Off
Point) and 2nd just above the pay-zone. The objective is to expose two pockets more significantly to mother well bore rather than drilling two separate
vertical wells which is an economic benefit with increase in production.

Fig 3: Sand body in seismic

217
WELL PLANNING

Using local co-ordinate system of a vertical well define target and plan trajectory of the well bore in relation to the target. Accuracy must be important
to achieve the target.

PLANNING WELL TRAJECTORY

To intersect the target, planning should consider the effect of geology on bottom-hole assembly, well bore stability, insitu stresses in horizontal section
and other factors like:

1) From existing vertical well establish target depth, horizontal departure (drift), pay zone thickness.
2) Select kick off point where well is to be deviated from vertical. Depth selection of kick off point is based on geometry of the well path and
geological characteristics of the formation.
3) Build up rate: It is the rate of change of angle from the kick off point up to the maximum inclination. Determine buildup rate by considering
the depth of target, torque and drag, formation characteristics, limitations of drill string, logging tools and completion etc.

In double curve design 1st build up is a KOP and 2nd build up is at the marker zone i.e. thin shale section above the pay zone. End of the build up is the
entry point into the reservoir. If reservoir has a significant dip angle, then the horizontal section must be designed to follow the dip. In this case the
hole is not truly horizontal but may be above or below 900. The drain-hole length and size is based on reservoir area, geological criteria and lease
boundary constrains.

To achieve the object use PDM (Positive Displacement Motor) motors, MWD (Measurement While Drilling) tool, conventional drilling method or
rotary steerable system (RSS). Advantage of RSS over conventional is to reduce the torque and drag, provide smooth well path, avoid stuck up
problems, ease BHA (Bottom-hole Assembly) orientation, save time and cost. Different software’s can be used for designing well path, casing design,
BHA configuration as length of drill collar, stabilizers, weight on bit, motor bend angle, bit side forces etc. Junction stability is affected by fracture
gradient at junction, pore pressure and mechanical properties of the material making up the junction. The main bore is cased and cemented and the
lateral is open or linear dropped. The junction is constructed with either down-hole milling or by installing pre-milled window joint.

CASE STUDY

This planning is for a well in Rajahmundry asset. A vertical well is planned to drill a depth of 3800 M. The casing (csg) policy is as follows:
20” csg – 300 M, 133/8” casing – 200 M 95/8” csg – 2800 M 7” liner from 2700 M to 3800 M. A well is planned to drill horizontal to a depth
3094 M (TVD) toward south west (650 M drift). Depth of KOP1 = 2372.68 M. Build up rate (BUR) φ1 at KOP1 = 50 per 100 ft and build up rate φ2 =
100 per 100 ft, horizontal displacement = 650 M, Tangent angle = 500, Length of drain hole / lateral = 200 M, measure depth including drain hole
length = 3758.97 M shown in fig. 4.

KOP1 R1 BUR1 = 5° per 100 ft (φ1)

V1 L1 R1

TVD
3094M 50°

V2

R2 BUR2 = 10° per 100 ft (φ2)

R2
V3 L1

H1 H2 H3 200M

Fig 4: Lateral for south-west


TVD = 3094 M = 20248.32 ft.
Length of lateral = 200 M

218
BUR1 = 5° per 100 ft.
BUR2 = 10° per 100 ft.
H1 + H2 + H3 = 650 M
Tangent angle = 50°
BUR1 = 5° per 100 ft = Turning radius = 18000 / (3.14 * 5) = 1146.49
V1 = 1146.49 * (cos50 – cos3) = 1146.49 * (0.766 – 0.052) = 818.59 ft.
H1 = 1146.49 * (cos3 – cos50) = 350.82 ft.
L1 = [100 * (50-3)] / 3 = 1566.66 ft.
R3 = 5730 / 10 = 573
V3 = 573 * (sin90 – sin50) = 134ft.
H3 = 573 * (cos50 – cos90) = 573 * (0.642 - 0) = 367.86 ft
L3 = [100 * (90-50)] / 10 = 400 ft.
H1 + H2 + H3 = 650 * 3.28 = 2132
H2 = 2132 – 350.82 – 367.86 = 1413.32
Tangent Length = 1413.32 / sin50 = 1844.95 = L2
V2 = L2 * sin50 = 1413.31 ft
Depth of KOP1 = 10148.32 – V1 – V2 – V3
= 10148.32 – 818.59 – 1413.31 – 134
= 7782.42 ft = 2372.68 M

TVD of KOP2 = 7782.42 + 818.59 + 1413.31


= 10014.32 ft = 3053.14 M

MD2 = (7782.42 / cos3) + 1566.66 + 1844.95 + 400 + (200 * 3.28)


= 7861.824 + 1566.66 + 1844.95 + 400 + (200 * 3.28)
= 12329 ft = 3758.973 M
Set the whipstock below 2372.68 M and mill the window in 9518” casing in and drill 12.25” hole up to 2800 M drop 7 linear and cement if then drill
6” hole up to T. D. Lower 4.5” linear or bore foot depends on well bore stability. Set the bridge plug at 2300 M and isolate mother bore and south west
lateral. Set the whip stock below 2154 M and mill the window in 9518” casing for 2nd lateral to a depth 3054 M (TVD - True Vertical Depth).
Horizontal displacement is 850 M towards north-west. Target depth is 3055 M (TVD). Build up rate 30 per 100 ft φ2 = 100 per 100 ft, drain hole
length 200 M. T.V.D. of KOP2 = 3014.07 M (TVD) fig. 5.

KOP1 R1 BUR1 = 3° per 100 ft (φ1)

L1 R1
V1
H1
B
50°

V2 L2

3012.07M R2 BUR2 = 10° per 100 ft (φ2)

V3 H1 H2 850 M L3 R2
B
H3
200 M Lateral Drain Hole

Fig 5: Lateral towards north-west

219
BUR1 = 3°per 100 ft.
Build up rate 2 BUR2 = 10° per 100ft.
KOP1 = 2154 M (7065.12ft)
Target depth 7VD = 3055 M (10020.4ft)
H1 = H2 + H3 = 803 M
Horizontal displacement towards north = 850M (2788ft)
Drain-hole lateral = 200 M
Top of whip stock = 2110 M
Turing radius R1 (3° per 100ft) = 18000/(3.14 * 3) = 1910.82 ft.
V1 = 1910.82(sin 50 – sin 3) = 1910.82(0.766 – 0.052) = 1364.32 ft.
H1 = 1910.82(cos3 – cos50) = 1910.82(0.998 – 0.642) = 681.45 ft.
L1 = [100(50-3)]/3 = 1566.66 ft.

V2 = L2 * (cos50)
H2 = L2 * (sin50) = 1900.06 * sin50
R3 = 5730/10 = 573

V3 = 573 * (sin90 – sin50) = 573 * 0.234 = 134


H3 = 573 * (cos50 – cos90) = 573 * (0.642 - 0) = 367.86
L3 = [100 * (90-50)]/10 = 400ft.

H1+ H2 +H3 = 2788 M


684.07 +H2 + 367.86 = 2788
Therefore, H2 = 1736.07

Length of tangent section = L2 ft.


H2 = L2 * sin50
Therefore, L2 = 1736.07/sin50 = 2266.27ft.

Hence, length of tangent section = 2266.27ft.


Therefore, V2 = L2 * cos50 = 1456.73ft.
TVD of KOP2 (10° per 100 ft.) = 7065.12 + 1364.32 + 1456.73 = 9886.17 M
= 3014.07ft.

DISCUSSION

KOP of north well should be below 50 /m of 13 3/8” casing shoe to avoid magnetic interference while drilling through sticky formation low and high
viscous LCM (Low Calcium Mud) pill should pump at regular interval and before lowering of linear also. Proper mud weight should be maintained to
avoid stuck up as well as kick problems in abnormal pressure zone. Instead of using SOBM (Synthetic Oil Based Mud), mud formatted brine can be
used to reduce the cost as well as environmental problems.

CONCLUSIONS AND RECOMMENDATIONS

The technology of horizontal and multilateral wells is proven and beneficial over vertical wells. Effective application, software and engineering efforts
can successfully drill the multilaterals. The fundamental variable of well path are target depth, geology, surface co-ordinates etc. Drilling can be done
by slide mode or rotary steerable mode.

Three measurement by MWD tool i.e. inclination, azimuth and tool face orientation are the key factors to know the well path. Dogleg severity at KOP
is also important. Bottom-hole assembly with proper size and position of stabilizers and RSS can produce a gauge hole, smooth well path without dog
leg and key seat. Production from the single completion string can be possible.
Using multilaterals the drilling as well as completion cost reduces effectively.

ACKNOWLEDGEMENTS

Authors are thankful to Dr. L. K. Kshirsagar, Principal, Maharashtra Institute of Technology, Pune, Dr. P. B. Jadhav, Head, Department of Petroleum
Engineering, Maharashtra Institute of Technology, Pune, Dr. S. V. Deshpande, C.E.O., Cambay Petroleum, Pune.

REFERENCES

Recent advancements in multilateral technology SPE International Conference on Horizontal Well Technology, Paper SPE 50390.
New Aspect of multilateral well construction Schlumberger quarterly report by Jose, Ohmer, Tom Pulick Roshron Texas USA, Mike Jordan Caraces
Venezuela, Mirush Kaja Milan Itali, Paez Jakarta, Indonesia, Sotomayor (Petrobras) Brazil.

220
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and
National Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Elastomeric and Metallic Stator Progressive Cavity Pump: A Comparative Study


Ramasamy Karthikeshwaran1 and Abdus Samad2
Department of Ocean Engineering
Indian Institute of Technology – Madras
Chennai - 600036, India
E-mail: 1eskrni@gmail.com and 2samad@iitm.ac.in

ABSTRACT

A progressive cavity pump or a PCP is a primary method of artificial lift in sand laden and heavy oil production. It consists of stator and rotor
member as the primary component in which the rotor is disposed within a cavity in the stator member. The stator is a resilient elastomer member
but the rotor is a rigid metallic member and it fits with stator making interference fit. One of the major drawbacks in this PCP is stator failure due
to the action of chemicals in the crude, wear and tear, temperature, pressure and stress developed within the stator cavity. On the other hand,
metallic stator is used in the metal PCP and a clearance between rotor and stator is allowed so that rotor can rotate inside stator cavity and the
problems related to the elastomer failure are avoided. The negative side of metallic stator PCP is that it can handle only viscous fluid. This paper
discusses the pros and cons of the metallic and elastomeric stator PCPs and reviews the technologies.

KEYWORDS: Artificial lift, elastomeric stator, metallic stator, interference fit, slippage, oil and gas.

NOMENCLATURE

Abbreviations
AL Artificial lift
ED Explosive decompression
CSS Cyclic steam stimulation
ESPCP Electric submersible PCP
NBR Nitrile based rubber
PCP Progressive cavity pump
SAGD Steam assisted gravity drainage

Symbols
b Breadth
C Generic constant
f Friction factor
L Length
Re Reynolds number
S Slippage
w Width
Differential pressure
µ Viscosity

INTRODUCTION

Progressive cavity pump (PCP) used in downhole pumping application is one of the preferred techniques of artificial lift (AL) system in shallow
to long deep wells producing heavy oil with sand. The concept of PCP was proposed by Rene Moineau in 1930 and the first application was in
1970s for the production of oil and gas. The PCP is a rotary positive displacement pump which can deliver fixed volume of fluid irrespective of
head requirement at a given speed. It is a non-pulsating pump and used to lift fluid having viscosity range of 1cSt to 1,000,000 cSt (Nelik and
Brennnan, 2005) and this is high tolerance to contaminant and abrasive particles. It delivers fixed volume of fluid at a given speed and the pump
can lift sand particles; slurry along with wellbore fluid. It is tolerant to entrained air and gases and produce minimal churning and foaming (Nelik
and Brennnan, 2005). The PCP is limited in the extensive use due to high temperature of fluid, chemical incompatibility with gases, liquids and
differential pressure of the pump during operation that affects the run life, performance and volumetric efficiency.

There are PCPs which are metallic stator PCPs or elastomeric stator PCPs and these are called metallic PCPs and elastomeric PCPs respectively.
Several works related to the PCP application were published in the journals and conferences. Bruce (2009) analyzed the challenging issues of
elastomeric stator to improve the performance of PCP. Shepherd (2003) reported the de-bonding of elastomer and its characteristic requirement
for efficient working. Osborne (2001) presented the internal friction exists between stator and rotor that leads to ignition of heat sensitive fluids in
the wellbore environment. Wild et al. (1998) reported that the leakage across seal line is due to the disproportionate pressure distribution in the
cavity. Non-uniform thickness of elastomeric stator causes differential expansion of stator at high temperature (Wood and Spencer, 1998). Zitka
and Murreiy (2005) analyzed frictional force developed by the interference fit and high coefficient of thermal expansion that deleterious the life of
elastomer. Bratu (2005) presented the leak tight contact between the stator and rotor and the tightness leads to appearance of cavitation which
damages the elastomer strength and life. Swelling effect of elastomer creates additional load to the pump (Lane, 2002).

Pessao et al. (2009) worked in the flow model of fluids in different viscosities of metallic stator PCP or simply metal-PCP and they analysed that
the fluid analysis can be made using an unsteady 3D computational model. The computational model predicts the fluid behaviour at different
clearance and fluid properties. Beaquin et al. (2005) presented experimental study in metal-PCP. Field results of metal-PCP in Josyln field and
Cuban field respectively were presented by Beaquin et al. (2007) and Guerrae et al. (2009).

This paper discusses the failure modes of the elastomeric stator PCP and its effects of replacing elastomeric stator by metallic stator to increase
the performance and run life of the PCP. The positive and negative side of these PCPs are presented analyzing applicability, failures and current
technological advancements.

PCP TYPES

A PCP is simply a screw pump in which the flow through the pumping is axial. The screw threads mesh each other and deliver fluid. There is
single and twin screw pumps used as downhole pump and these pumps contain two screws with opposing thread where one screw turns clockwise
and the other turns counter clockwise. The screws are mounted on shafts that run parallel to each other. The single screw pump has a single screw
type rotor and the rotor meshes with stator creating a cavity to deliver fluid. The PCP is a single screw type pump which consists of an
elastomeric or metallic stator and a helical rotor (Fig. 1). The whole rotor-stator system is inserted into the well and a sucker rod rotates this
system. The sucker rod is run by a surface installed motor (Fig. 1c).

Stator
Rotor
assembly
Rotor-

To separator
Motor

Sucker
Rotor
stator

rod
Casing

Tubing
(a) Geometry of stator and rotor

Pump

Perforations

(c) Complete progressive cavity pump


assembly

(b) Eccentric motion of


rotor

Fig 1: Progressive cavity pump

222 
 
Based on the stator material, PCP is classified into elastomeric and metallic stator PCP. The pump is operated by a prime mover placed at the
surface and this type of PCPs is used in all the conventional operations. If the prime mover is inserted into the well bore, the PCP is called electric
submersible PCP (ESPCP). Presently ESPCP are used in the field of high flow rate at high speed. The ESPCP has an arrangement of submersible
motor which drives the PCP placed in the wellbore.

PRINCIPLE OF OPERATION

The stator is the elastomeric or metallic member of certain thickness which is bonded inside a metallic tube. The rotor is a metallic member
having helical shape and inserted in the stator member as shown in Fig.1a. The drive rod or sucker rod string transmits rotational power to the
rotor from the prime mover. The drive rod connects the rotor at the lower end of the sucker rod (Saruwatari, 1998). The rotor is a helical element
that rotates in the stator to lift the fluid. The eccentric motion or orbital rotation of rotor is shown in Fig.1b. The rotor makes a contact with the
stator to form a series of cavities, thereby lifting the fluid progressively along the axis to the pump outlet. Based on the direction of rotor rotation,
the pump displaces fluid from the wellbore to the surface through the progressive action of cavity in which fluid displaces along the pump. The
PCP is a self-priming Pump and it has an inlet to move the fluid into the cavity.

Slippage

The basic difference between the metallic and non-metallic stator PCP comes from the leakage analysis. The metallic stator maintains a clearance
while the elastomeric stator maintains an interference fit with the stator. The pumping capacity depends on the volumetric efficiency of the pump.
The leakage or slippage is the backflow of the fluid between rotor and stator. Leakage is defined as the difference between theoretical flow rate
and the actual flow rate. A general approach for calculating the internal slippage is presented by Moineau (1930) is based on Hagen Poiseuille
flow along the seal lines.

Considering the flow between two parallel plates separated by some distance as the clearance along the pump, slip can be calculated. L, b and w
are the length, breadth and width of the flow area (Fig.2). Pressure drop can be calculated across the sealing line which is defined by the
contact in between the parallel plates:

(1)

where, f and S are the friction factor and slippage respectively. C is generic constant. Slippage (S) is formulated as

(2)

The performance curve (Fig. 3) shows that the slippage increases as the differential pressure increases. The slippage is function of differential
pressure, viscosity and pump geometry. The elastomeric stator fits with rotor in interference fit of zero pressure drops at low pressure. The
cavities are enclosed with seal and the pump flow rate is equal to the discharge flow rate. The deformation of the elastomer increases with
increase in differential pressure. The slippage significantly increases with pressure drop with increase in deformation of interference fit. For the
metallic stator the clearance between the stator and rotor is constant. For the high viscosity fluids, slippage is linear.

ELASTOMERIC STATOR

The elastomeric stator is the heart of the PCP. The elastomer is resilient material made up of rubber, nitrile based rubber (NBR) elastomer, fluro
based rubber elastomer, buna–N, teflon, silicone and plastics. Commonly used resilient material is nitrile based elastomer (Lane, 2002). The stator
fits with rotor in interference fit. The interference fit provides a rubbing contact with stator when the rotor rotates. This rubbing contact provides a
good sealing effect in the pump. Hence there is no leakage along the length of the pump and this is a major advantage of elastomer PCP.

                      

Fig 2: Geometry of plate Fig 3: Differential pressure Vs volume displaced

223 
 
Failures of elastomeric stator

The premature failures of elastomeric stator are due to elastomer swell, distortion in its profile, hardening and brittleness to crack, wear and
pitting (Fig. 4) and these leads to restriction of rotor motion in the stator. Heat generated at the interference causes elastomer to fail and it results
in decrease in fluid intake and running out in dry condition. The rotor maintains an interference fit to the stator. The rotor rotates in eccentric
motion and it compresses the stator to provide a sealing effect during the motion. The leakage flow rate is depends on the geometry and tightness
of the fit between the rotor and stator, vibration and speed of rotor, viscosity of fluid and differential pressure exist between the cavities.

The differential pressure is not uniformly distributed across the length of the stator/rotor. The minimum pressure exists in one cavity permits
leakage or slip (Wild et al., 1998) to the next cavity across the seal. The elastomeric stator starts to deflect when the minimum pressure is
achieved for the leakage. The non uniform pressure distribution along the PCP leads to pressure wash as shown in Fig.5a. The leakage flow rate in
between the stator and rotor is a function of pressure gradient (Fig.3). Leak-tight contact between the rotor and the stator can lead to the
development of cavitation (Bratu, 2005) if the PCP displaces viscous fluid at high flow rate or at low inlet pressure. The cavitation reduces the
strength of the elastomer and reliability of the system. The sealing effect generates high pressure within the cavity and the high pressures
generally develop forces which result in stress along the interference in the elastomeric stator. These forces create friction which generates large
amount of heat during the pumping operation and the heat changes the characteristics of elastomer. The fraction of heat develops and accumulates
in the stator and rotor that can dissipate the heat to the fluid within the cavity. After a period of time, the amount of heat dissipated to the fluid can
ignite the fluid, when the fluid is heat sensitive (Osborne, 2001). The PCP is used to lift abrasive particles along with crude and heat is generated
by friction or rub of the stator by rotor. The rubbing of the elastomer at high temperature causes pitting and tearing out of the stator surface
(Fig.5b). The heat changes the interference fit between the stator and rotor and creates an additional load to the pump (Lane, 2002) as shown in
Fig.5c.

The elastomer has high thermal expansion coefficient and low thermal conductivity (Zitka and Murreiy, 2005). The elastomeric thickness is not
uniform along the length of the stator. The radial thicker sections distort more than the thinner sections of the stator (Wood and Spencer, 1998)
and the distortions creates de-bonding problem (Shepherd and Hobbs, 2003) shown in Fig 5d. Elastomer is mainly NBR and the nitriles are
attacked by chemicals which change the mechanical properties making them softer, harder and weaker. Swelling of elastomer is due to action of
single ring aromatics such as benzene, toluene, xylene and other derivatives. Aromatics impact the physical properties of the elastomer, impacting
the pump performance (Bruce, 2009). Swell is expressed in terms of percentage of thickness of the elastomer (Bruce, 2009). In the cross section
of elastomer, the thickness of the elastomer at the major diameter is significantly different than that of the minor diameter. The higher percentage
of swell occurs at the minor diameter than the major diameter. The elastomer has a characteristic of very good insulation properties. More heat is
stored inside the elastomer in the thicker elastomer and it reduces the run life of the pump as shown in Fig.5e. At high pressure some of the gas 
molecules  are  diffused  into  the  elastomer  matrix  and  a  subsequent  rapid  reduction  of  pressure  can  cause  internal  fracturing  or  explosive 
decompression (ED).  The fracture may be in the form of crack and blisters. All type of gas in well bore result in ED type of failure. Typically CO2 
is  more  damaging  than  other  (Bruce,  2009).  In  the  wellbore  environment,  the  foreign  materials  contaminate  the  elastomer  (Fig.5f)  and  it 
degrades the elastomer by eroding the surface.  

Damages of elastomer

Interference fit Aromatic compounds Gases Foreign materials

• Burn/overheat Blister/swell • Blister/swell


• Blister/swell • Cracks
• De-bonding
• Torn
• Pressure wash

Fig 4: Causes of elastomer damages

(a) Pressure wash (b) Torn (c) Blister

224 
 
(d) Debonding (e) Burn/over heat (f) Contamination of foreign material

Fig 5: Damages of stator (Wagg, 2007)

(a) Worn rotor (b) Crack due to overheat (c) Pitting

Fig 6: Damages of rotor (Wagg, 2007)

Failures of metallic rotor

The elastomer damages the rotor by removing the fine surface of the stator during the rotor is in orbital motion and it results in worn-out shown in
Fig. 6a. The helical edge of the rotor rubs the stator helical edge and cracks are developed (Fig.6b) due to overheat while rubbing. Pitting of the
rotor occurs at the cracked area (Fig.6c). To prevent premature failure, the stator must have desired characteristics for efficient run life for high
performance and reliability of the system.

METALLIC STATOR

The metallic stator is manufactured by Hydro-forging process (Beauquin et al., 2005) and is made up of metal. The metallic stator is otherwise
called as rigid stator. The technology is newly developed to handle the high temperature fluids and not may studies has been taken place except
field results presented by the authors. The metallic stator and rotor are coated with corrosion resistant materials. The stator fits with rotor in
clearance fit; hence there is no rubbing contact between the rotor and stator. This leads to leakage or slip along the length of the pump. A constant
clearance maintains along the length of the pump to permit backflow within the pump.

Failures of metallic stator

Corrosion is one of the major drawbacks in the metal-PCP. The corrosive gases like CO2 and H2S corrode the stator if the wellbore environment is
highly corrosive. The metal-PCPs are set to be use in thermal flooded fields for extracting heavy oil. As the temperature of the field increases, the
percentage of CO2 and H2S gases also increases. Due to clearance fit leakage occurs along the length of the pump. To reduce the leakage, number
of stages of within the pump should be increases. These results in complexity of the pump and power required to drive the rotor for a large
number of stages pump also high.

Failure of metallic rotor

The rotor in the metallic PCP has less failure probability than that of the elastomeric PCP. The rotor mainly faces corrosion failure.

COMPARISON OF ELASTOMERIC AND METALLIC STATOR PCP

The premature failure of the elastomeric stator leads to jeopardize the system. To overcome this elastomeric stator is replaced by non-deformable
material i.e. metallic member. The metallic stator works at a temperature of up to 660ºF and but the elastomeric stator works at a temperature of
less than 350ºF (Nelik and Brennnan, 2005).

The rotor-stator fit is a clearance fit and the leakage flow rate depends on the clearance in metal PCP. The pressure is uniformly distributed over
the length of the stages of stator/rotor. The clearance fit results in uniform temperature distribution. The interference fit in elastomeric PCP
increases temperature.

Generally metal reacts with the chemical lower than the other materials like rubber, polymer, and plastics at high temperature. Alloy and treated
metals are used in the chemical environment not to react with the chemicals. Metallic stators are chrome plated with low sulphur and low
phosphorous for corrosion resistance and to avoid degradation when it is exposed to wellbore environment (Parise et al., 2009). Gases such as
methane and carbon-dioxide are cant able to diffuse into the metallic stator to cause ED.

The metallic stator is very tough and case hardened and it has very high tensile strength. The problem in metallic stator is corrosiveness and it
corrodes in the high corrosive environment of wellbore containing CO2 and H2S (Parise et al., 2009). The metallic stator is more abrasion resistant
for the same pressure per stage as compared to elastomeric stator. Metallic stator has longer life and this stator can pump the abrasive particle

225 
 
sizing up to 200 microns (Nelik and Brennnan, 2005). In metal-PCP, there is no rotor to stator contact and hence the contamination from
elastomer wear particles is eliminated. The surface wear out rate is very low due to clearance fit as compares to that of the elastomeric stator. The
maintenances and inspection of the Metal-PCP is quite easier. Because the clearance fit is maintained in the rotor-stator contact surface (Parise et
al., 2009).

The elastomeric stator has weakness that the rotor rubs the stator surface but it has good fluid delivery capability at any viscosity. The leakage at
lower viscosity fluid is higher for metal-PCP as the clearance fit allows the fluid to slip. The flow loss due to leakage in metal-PCP can be
controlled by increasing the number of stages. Here, less viscous drag force acts on the rotor. For a less viscous fluid, number of stages is to be
more to maintain the differential pressure in the cavity. In Metal-PCP, the overall efficiency is not very sensitive to differential pressure for high
viscous fluids, but it improves with speed. The overall efficiency decreases with increase in differential pressure between the stages.

TECHNOLOGY UPDATE

The viscous oil, tar sands, extra heavy oils and natural bitumen are predicted to be about 5 trillion barrel in the world’s oil reserve (Unleashing the
Potential Heavy Oil, 2006). Most of the world conventional oil is exploited and gradually the world is depending on heavy oil and extra heavy oil.
The mobility of heavy oil and extra heavy oil is very low. Thermal processes are assisting to extract the heavy oil and extra heavy oil by
increasing the mobility. The thermal processes such as cyclic steam stimulation (CSS), steam assisted gravity drainage (SAGD) and steam
flooding are techniques in practise in the heavy oil field. The technology of developing heat resistant elastomeric member at high temperature is
under research to avoid swelling and burning of the elastomer. The technology is being developed to handle high bottomhole temperature and
large amount of corrosive gases and focuses is on the development of the corrosion resistant metals.

CONCLUSION

In recent years, the PCP marked a milestone in artificial lift technique in oil and gas sector. In this article metal and elastomeric PCP technologies
are presented and compared. Elastomeric PCP works satisfactorily at low to very high viscosity. The problem in elastomeric PCP is elastomeric
failure due to rubbing and this stator cannot work at high temperature. Metal-PCP is one of the landmarks in the production of heavy oil in
artificial technique in oil and gas field. This PCP can work at high temperature and can deliver viscous fluid. Metal-PCP needs large number of
stages if it is allowed to deliver thin fluid. Hence both of the PCPs having their own weaknesses and strengths and selection mainly depends on
the fluid property.

REFERENCES

Beauquin, J-L, Boireau, C, Lemay, L and Seince, L (2005), “Development Status of a Metal Progressing Cavity Pump for Heavy-Oil and Hot-
Production Wells”, SPE International Thermal Operations and Heavy Oil Symposium, Calagry.
Beauquin, J-L, Ndinemenu, F, Chalier, G, Lemay, L andSeince, L (2007), “World’s First Metal-PCP SAGD Field Test Shows Promising
Artificial-Lift Technology for Heavy-Oil Hot Production: Josyln Field Case”, SPE Annual Technical Conference and Exhibition, Anaheim,
California.
Brandt, DC and Neil, GR (1998), “Stator Assembly for a Progressive Cavity Pump”, US Patent, Patent No.: US005807087A.
Bratu, C (2005), “Progressive Cavity Pump”, US Patent Application Publication, Publication No.: US2005169779A1.
Bruce, WA (2009), “Improving Progressive Cavity Pump Performance Through New Technology and System Design”, Middle East Artificial
Lift Forum, Bahrain.
Gamboa, J, Olivet, A, and Espin, S (2003), “New Approach for Modelling Progressing-Cavity-Pump Performance”, SPE Annual Technical
Conference and Exhibition, Denver.
Guerra, E, Sanchez, A and Matthews, C (2009), “Field Implementation Experience with Metal-PCP Technology in Cuban Heavy-Oil Field”, SPE
Production and Operations Symposium, Oklahoma City, Oklahoma.
Heavy Oil and Oil sands, volume – 3, (2008), June warren Publishing.
Lane, W (2002), “Adjustable Fit Progressive Cavity Pump Apparatus and Method”, US Patent, Patent No: US0063580271B1.
Nelik, L and Brennan, J (2005), “Progressive Cavity Pumps, Downhole Pumps and Mudmotors”, Gulf Publishing Company, Houston, Texas.
Osborne, AM (2001), “Progressive Cavity Pump with Meltable Stator”, US Patent, Patent No.: US006220838B1.
Paladino, EE, De Lima, JE, Pessoa, PAS and Almeida, RFC (2009), “Computational Three Dimensional Simulation of the Flow within
Progressive Cavity Pumps”, 20th International Congress of Mechanical Engineering, Gramado.
Parise, N, Lehman, M and Amara, AB (2009), “PCM Vulcain TM-Metal-to-Metal PCP Artificial Lift Systems”, Middle East Artificial Lift
Forum, Bahrain.
Pessoa, PAS, Paladino, EE and De Lima, JE (2009), “A Simplified Model for the Flow in a Progressive Cavity Pump”, 20th International
Congress of Mechanical Engineering, Gramado.
Saruwatari, M (1988), “Progressive Cavity Pump”, US Patent, Patent No.: US004773834.
Shepherd, SL and Hobbs, GE (2003), “Polyurethane Stator for a Progressive Cavity Pump”, US Patent Application Publication, Publication No.:
US20030003000A1.
“Unleashing the Potential Heavy Oil” (2006), www.hartenergy.com.
Wagg, BT (2007). “Progressive Cavity Pump Inspection and Reporting”, C-FER Technologies.
Weir, BA (2009), “Improving Progressive Cavity Pump Performance through New Technology and System Design: A Case History of PDVSA
San Tome Operations”, Middle East Artificial Lift Forum, Bahrain.
Wild, AG, Mirza, KZ and Shang, SSL (1998), “Cascaded Progressive Cavity Pump System”, US Patent, Patent No.: US005820534A.
Wood, SM and Spencer, BE (1998), “Progressive Cavity Pump Using Composite Material”, US Patent, Patent No.: US005759019A.
Zitka, MD and Murreiy, WD (2005), “Progressive Cavity Pump/Motor”, US Patent, Patent No.: US006881045B2. 

226 
 
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Current state-of-the art well stimulation technology starting from types of


near-wellbore damages to nano ferro fluid thermal well stimulation
Sudjati Rachmat and Astra.A. Pramana
Petroleum Engineering Department, Bandung Institute of Technology
Bandung, Indonesia
E-mails: sudjati@bdg.centrin.net.id; asnova.nova@gmail.com

ABSTRACT

Using a modified Resource Triangle, the conventional and unconventional reservoirs are mapped and related with their volume of reserve and
technology to produce them. This paper is then focusing on well stimulation which discusses its current state-of-the art technology starting from
types of near-wellbore damages and related to the techniques in hydraulic and foam fracturing, acid stimulation, and electro magnetic thermal heating
using nano ferro fluid thermal well stimulation. The paper also overview the latest developments in stimulating the unconventional reservoirs.
Moreover, many key publications are provided as references for up-dating our knowledge in this discipline.

KEY WORDS: Stimulation, Unconventional Reservoir

INTRODUCTION

Well stimulation is a well intervention performed on an oil or gas well to increase production by improving the flow of hydrocarbons from the
drainage area into the well bore. The process might be cleaning from the near well damages, enlarging conductive channels, or creating new ones1,2
in which the techniques are listed in table 1 3,4.
Table 1. List of current and will be available stimulation techniques
These techniques are initially applied only to unproductive wells after producing the hydrocarbon for a length of period. As the oil prices increased
and the conventional crude oil becomes scarce, the unconventional reservoirs appeared to be economically feasible, though massive stimulation
treatments are needed.

Unconventional reservoirs are defined as those which cannot be produced at economic flow rates without massive stimulation treatment or special
recovery processes. As a consequence, it is more expensive and brings a longer return of investment compared to the easier conventional one. They
are categorized as tight gas sands, oil and gas shale, heavy oil/tar sands/bitumen, coal bed methane, and gas hydrates 5,6,7.

To have a complete view of the use of stimulation techniques to the unconventional reservoirs, a concept of resource triangle is applied. This concept
was introduced by Masters and Gray of Canadian Hunter in 1979 to indicate that low quality, unconventional resources are huge in reserves but may
be developed with more sophisticated technology. This could be realized if the price of high quality resources increased 8. By mapping the
conventional and unconventional reservoirs and related with their volume of reserve taken from the available data 5,6, a modified resource triangle is
produced as in Figure 1. A slight modification with the original concept can be seen which is by putting the higher production cost on the apex of the
triangle while the cheaper conventional resource stayed on the base. Another thing is that the world reserve of gas hydrates which can be produced by
heat/gas generation such as SGN (thermal/chemical) or by electrical heating should be positioned at 13-14 tcf, but was given from the data as 3000
tcf and therefore is excluded from the triangle.

Fig 1. Modified resource triangle related to the stimulation techniques


.
THE DEVELOPMENT OF CURRENT STIMULATION TECHNOLOGY

An overview of the development in stimulation techniques based on the groups given in table 1 will then be discussed as follows.

Mechanical Stimulation

Hydraulic fracturing is the main mechanical stimulation which is a process of pumping water, oil, and/or gases with a proppant at sufficient pressure
to create a fracture in a chosen formation. The goal is to create and prop the openned fracture. The fracture becomes a conductive pathway for fluids
into the wellbore increasing the wells productivity. In the design of the stimulation treatment many factors are considered such as formation
properties and its rock mechanics, fluid dynamics, chemistry and interactions of these elements.

Water-based fracturing fluids are economical and used in most stimulation applications. They are a mix of fresh water, acid or light brine and can be
further sub divided by their viscosity enhancement. By the early 1960’s water based fluids based on guar polymers (such as HPG) were used to
viscosify the water. Guar is a long-chain, high molecular weight polymer composed of mannose and galactose sugars. It has a high affinity for water
which can make it swell and hydrate. By the late 1960’s cross-linker such as Ti and Zr had been introduced for water based frac fluids to produce
higher viscosity at high temperatures. In the 1970’s to 1980’s, alternatives to HPG had been found such as HEC for a cleaner frac fluid and Xanthan
for a better sand suspention. It is then followed by CMHPG and CMHEC in the 1990’s which also can be used for low temperature wells and being
used upto now.

In the late 1990’s to the early 2000’s,polymer-free frac fluids are being used that can be prepared using VES which typically is a quaternary
ammonium salt of long-chain fatty acid. Its molecules are called micelles where the outside hydrophilic head groups are in direct contact with water
phase while the inner hydrophobic tail groups are insulated. When the aqueous environment has an optimum salt concentration, the micelles become
rodlike shaped causes the fluid to be viscous and elastic. Another advantage of these VES fluids is little residue is left after clean-up because of
disrupting mechanism of the micellar structure by a contact with hydrocarbon or formation fluid. As a result, less proppant pack damage is observed 9.

228
As soon as the oil prices started peaking up in the 2004, unconventional reservoirs became economically beneficial. In unconventional tight gas
reservoir, for example, clean-up failures for the gelled liquids reversed back the use of stimulation fluid to water. Moreover, waterfrac is cheaper and
provided, the reservoir is not water sensitive. Due to high friction of water, friction reducer is added which made the water becomes slick 10.
Typically, the friction reducer is polyacrylamide or emulsion polymers which undergo inversion process which then interacts with eddies of
turbulence if migrated to the near-wall boundary layer then promotes the reduction of friction 11. The viscosity is maintained to be about 1 to 2 cp,
compared to that of the gelled which is 15 to 25 cp and will be multiplied as much as 25 times with the cross-linker addition 12. This less viscous
property of the slickwater needs to be pumped in higher rates to achieve wider fracture width and for transporting the proppants (smaller: about
20/40-sized and less dense: about 1.75 ASG). This might increase the cost, though at the end, the preference of using slickwater frac mostly depends
on the availability of water which can exceeds millions of gallons. The slickwater fracturing is estimated to be 30% of the stimulation treatment in
200413.

Another way of avoiding the gelled fluid for low permeability reservoir is by Batch Fracturing which was rebirth in the mid 2000’s. In batch frac,
ultra light weight proppant with specific gravity of 1.05 to 1.75 is used. By having much less dense proppant, the based frac fluid doesn’t have to be
highly viscous and this means no need of viscosifying chemicals such as gelling agents, cross-linkers and breakers. Moreover, the proppant is just
added to the frac fluid in the batch-mixing tanks, without using blenders, and discharges to frac pump then to the well bore. The main issue here is
the proppant selection such as size,crushing, roundness, sphericity, etc. This technique can also be developed for the proppant partial monolayer
which delivers a higher conducting fracture from the void spaces among the mono-layer proppants 14.

For water sensitive formations, oil based fluid fracturing and foam fracturing are used. If it is very sensitive which typically contain large percentages
of clays that react with fresh water or brines, oil based fluid is preferred. Though originally heavy oil was being used for its viscous property, later
aluminum carboxylate then aluminum phospate ester salts were used for gelling the non-heavy hydrocarbon frac fluid. In the 1990’s, on the fly
addition technique for forming the gel while pumping the hydrocarbon to the downhole was found which eliminate the premixing stage 15.

For not very fluid-sensitive or under/low-pressure formations, foam frac is being used with nitrogen or carbon dioxide gas as internal phase and
dispersed in the external phase of water, acid, alcohol-water mixtures, or hydrocarbon liquids. The reason is that because of its easy flow back as
clean-up/displacement and proppant carrying fluid. Besides, it reduces the fluid incompatibility problem with less contact and save about 52% upto
96% fluid 16.

For less water sensitive or less clay content, clay stabilizer can be added to water frac. At a concentration of 0.2 wt%, Al compound stabilizer can
prevent fines migration and clay swelling caused by the contact of formation with low salinity or high pH brines. Initially, KCl and NaCl salts were
used as temporary clay stabilizers. The difficulties in mixing these salts to the water and other drawbacks made hydroxyl aluminum or zirconium
oxychloride as the good clay stabilizer. The most popular clay stabilizers used recently are categorized as cationic organic polimers (COP) 17.

RPM or Relative Permeability Modifiers were introduced in the late 1970’s to reduce the permeability to water and not affecting the permeability of
hydrocarbon. Once the RPM material enters the formation, it attaches to the formation surface and restricting the water ability to flow through
formation. Moreover, it also reduce the amount of water based fluid entering the formation. Therefore, besides being a self diverting fluid to be used
in areas where breaking into nearby waterzone commonly occured, RPM also works as a fluid loss additive which can give longer frac lengths 18

Chemical Stimulation

For carbonate formations which mainly occurred in Europe and Middle East, acid is being used for frac fluid. Acid treatment is an alternative to
hydraulic fracturing in carbonate formations which usually about 95% soluble by acid system, though it also could be applied to sandstone formation
which mainly occurred in North America, for matrix acidizing. Inorganic scales and other acid soluble debris, fines, and solids can also be treated
with acid washing 9.

The objective of matrix acidizing the sandstone is to solve the near-wellbore damages by dissoliving clays and other pore plugging materials
occurred when drilling or completion. The acid formulation is a mixture of HF and HCl and should be optimized based on the sensitivity of the
sandstone which depends on temperature, mineralogy, carbonate content, and its permeability. In carbonates, matrix acidizing is done mostly with
HCl.

About 70% of the stimulation jobs are acid treatments since they are 20%-40% cheaper than that of hydraulic fracturing, though their failure rate is
about 32% compared to the 5% of hydraulic fracturing. These failures mostly attributed to incorrect treatment design or improper identification of the
well damage 19.

Acidizing treatment was first done at Lima, Ohio in 1894. It became a common stimulation for limestone or dolomite formations since1932 after the
invention of arsenic inhibitor which can protect the metal casing and other well equipment from HCl-corrosion. Matrix acidizing of sandstone was
later came into practice in the 1960’s though regular-strength mud acid of 12% HCl-3% HF had been found since 1939.

Acidizing relies on acid etching of the fracture face, rather than placing of proppant to prevent the closing and produce conductive flow channels
often called wormholes 20.

To reduce wellbore friction, the acid is usually viscosified or gelled which also useful to suspend and transport any fines released by the acid, and
retard the acid reaction rate. These optimization techniques are varied such as by emulsifying the acid, the in-situ gelled acid, HCl-formic, or visco-

229
elastic surfactant (VES) 21.

Emulsified acid or Acid internal emulsion (AIE) where diesel is utilized as external phase, can produce higher conductive etch patterns than live-HCl
when closure pressure was applied. This is shown from its retardation factor of 10 which gives deeper acid penetration compared to 1 for HCl and
gelled acids; and 4 for Organic Acid. This AIE technique was widely used since early 1970’s.

VES was used initially as gravel pack carrier fluids in 1980’s, It is then expanded to fracturing fluids for conventional proppant fracturing in 1990’s.
It can also be foamed or added with specific fluids for high temperature application. Self-Diverted Viscoelastic Acid (SDVA) uses VES added to HCl
which reacts with carbonate and increases the pH and the Ca++ ions. This technique was first tested in Kuwait and Gulf of Mexico in 2001 and in
deep/hot water injector wells for acid fracturing in 2003. Currently, SDVA sometimes combined with Acid Internal Emulsion (AIE) to stimulate
multiple (perforated) intervals 22 (Zaeff, 2007)

For higher temperature use, the acid can react more for corroding the tubular casing and the coil tubing which cannot be avoided only by inhibitor.
Therefore, chelating agent is used with several names such as EDTA, HEDTA, DTPA, HACA and GLDA 23 (Mahmoud, 2010).

By combining with high pumping rate, acid fracturing is intended to get “live acid” penetrating the formation as far as possible. At the end of the
treatment, the fracture is allowed to close and the spent acid flowed back soon. The limitation of acid fracturing is that it is applied to harder rock
formation, lower permeability, and tight carbonate formation. The reason is that the soft one will plastically deform into etched width created by the
acid and making it unbeneficial. oreover, the resulting conductivity tends to be smaller than the proppant fracturing.

Thermal Stimulation

Most of the unconventional reservoirs need thermal heating stimulation. The stimulation of Oil and Gas Shale is focused on Mahogany projects
where many in-situ technology are applied. They are: Shell’s ICP, Chevron’s CRUSH, GFC, ExxonMobil’s Electrofrac, EPI’s EEOR, and Global
Resource’s Microwave Heating. Meanwhile, Heavy oil needs CSS or its variants besides Electrical Heating,

For oil shale stimulation, the Shell’s electrical heating project called Insitu Conversion Process (ICP) was started in 1980’s and tested in 1997 in
Mahogany, Colorado, but postponed until 2002. In 2006, Shell got approval for more leases to finalize the ICP from field test to commersialize it.
The oil shale is heated by electric heaters. The oil which was converted from kerogen, produced by this technology is chemically stable, without low-
value residue and can be immediately sent to refineries; unlike the oil produced by exsitu - surface retorting, where near site upgradation is necessary
24,25
(shell, website).

Another technique was developed by ExxonMobil called Electrofrac process where the conductive brine is injected in a hydraulic-fracturing
technique followed by electrical heating (oilshale web). Meanwhile, Raytheon developed Radio Frequency antenna, in which no need of
maintaining the connectivity of the 2 conductors, to heat the oil shale and combined it with carbon-dioxide supercritical fluid. This project was later
sold to Schlumberger in 2008. The abovementioned techniques are still competing with self-fueling Geothermic Fuel Cell (GFC) and Chevron’s
CRUSH 3,26.

GFC utilizes a solid oxide ceramic electrolyte fuel cell (SOFC) put to the solid pay zone and slowly heat the oil. The heat travels through the rock,
lowering the viscosity of the oil in the shale, while creating the ‘off-gases’. Both the oil and gases are moved into collector wells, where they are
extracted by conventional means. A portion of the generated off- gases will refuel the cell 27 (IEPM webs).

Chevron’s CRUSH (Chevron’s for Recovery and Upgrading oil SHale) process by combining explosives and hydraulic fracturing to rubblize the
production zone then heated carbon dioxide as a carrier gas is injected through the wellbore and allows to penetrate to the newly fractured zone then
recycled back to be heated. 28,29(wikipedia chevron website)

Another thermal simulation used for unconventional reservoir is called Self-Generated Nitrogen (SGN). This method was patented in Brasil and
applied to heat and dissociate the crystallized gas hydrate which was clogging the (wet) christmas three cap 30 (Marques, 2004).

The process of creating synthetic light crude oil from oil shale is actually comparable to that of heavy oil or oil sands, although there are some
technical differences. CSS or Cyclic Steam Stimulation is a standard technique to be applied for heavy oil reservoirs with the recovery factor of 20%
to 35%. This method works best with the pay zone thicker than 10m, good horizontal permeability for vertical wells or vertical permeability for
horizontal wells (>1 darcy) and high porosity sands (>30%). CSS process is started with HPHT steam injection to the pay zone for quite a long time
(up to 1 month). It is then followed with soaking in which the well is left idle without any production for about 1 or 2 weeks to diffuse the heat into
pay zone and reducing the viscosity of the resided heavy oil. The last stage is the production time that might be lasted for a year in which the less
viscous oil is electrically pumped out until the production rate is not economical. The cycle is then repeated as many as 15 times and usually followed
by EOR treatment such as Steamflooding. The typical Steam oil Ratio (SOR) for CSS is 3 to 5. 31.
Many other improved variations based on this CSS technique such as 3,32:
• SAGD (Steam Assisted Gravity Drainage) which uses 2 horizontal wells, the upper for steam injection while the lower for the production.
• SAS (Steam Alternating Solvent) where steam injection is done for 6 months then alternated to propane as a solvent for 3 months followed
by the production time.
• LASER (Liquid Addition to Steam Enhanced Recovery (LASER) with the hydrocarbon as the liquid.
• THAI (Toe-to-Heel Air Injection)
• THSF (Toe-to-Heel Steam Flooding) which is like THAI except one uses steam instead of air, and top-down combustion where the flame

230
front forces the heavy oil down to a lower collection (i.e. production) well.
• VAPEX (Vapour Assisted Petroleum Extraction) which uses gas (such as methane, sometimes added with CO2 and solvent).

Duri field in Indonesia has been known as the biggest steam thermal project in the world. Unfortunately, for steam stimulation beside having big heat
loss in the vertical section of the reservoir, it is limited to onshore, shallow formation (with no steam confinement issue), thick pay-zone, non-
swelling clay reservoir content, and no fractures or near-homogeneously permeable reservoir 33,34,35. Meanwhile, stimulation which related to in-situ
combustion technique is still having unresolved safety issue.

Electrical heating (EH) techniques could overcome those steam stimulation’s limitations, and therefore have been introduced as a promising
alternative. A number of EH variants are summarized by Pramana as given in the chart in Fig.2 3 and in details by Callarotti 36 while its Capex
analysis by Koolman et al 37.
Electrical Heating
.

AC DC: EEOR-EPI

Low Frequency(100 Hz): High Frequency(>10 kHz)


ET-DSP
- Microwave (10 MHz-10 GHz)
- Radio Frequency (300 kHz-300 MHz)
- Electromagnetic Induction (<300kHz)

Fig 2. The variants of Electrical Heating techniques 3

The stimulation using Direct Current (DC) Electrical heating technique is named Electro Enhance Oil Recovery (EEOR) which was developed by
EPI. The DC current is applied between cathodes in the pay zone and anodes at the surface 38 Almost similar concept using low frequency AC current
and some electrodes (instead of anodes and cathodes for DC current), ET-DSP (Electro-Thermal Dinamic Stripping Process) was developed 39. Both
of the techniques claimed to have been succeed in the Alberta heavy oil projects though the major drawback of them is that conductive medium
(formation’s brine) between the electrodes should always be maintained. Therefore, if the brine has been heated to vaporized, the electrodes won’t
be connected and the heating process is interrupted 3.

To overcome this drawback, higher frequency AC current electrical heating are developed such as Global Resource’s Microwave heating,
Schlumberger/Raytheon’s Radio Frequency heating, , and Siemen’s Electromagnetic (EM) Induction heating 3,37 (Pramana2, 2010; Koolman, 2009).

In the induction heating, eddy current is generated in the reservoir which produces a contactless way of heating the conductive brine. By increasing
the conductivity of the medium, a much powerful heating can be achieved. This idea has been developed as part of the research at Bandung Institute
of Technology which utilized very conductive stimulant fluids such as graphite or magnetic stimulant fluid such as nano-ferrofluid. By this concept,
heating time of the heavy oil can be much shortened 3, 40.

CONCLUSION

The invention of Coiled Tubing in 1980’s followed by multilateral well, real time directional drilling, Casing while Drilling, sliding sleeves, and
some other technology had pushed the stimulation techniques forward. Some of the these are 41,42,43,44 :
• Pinpoint fracturing to have a precise placement of fracture, especially in the horizontal well. This technique utilizes abrasive jetting which
will minimize the creation of random hydraulic fractured channels occurred in between the packers. Moreover, it leads to hydrajet
perforating or fracturing with water or acid and/or abrasive stimulation fluids.
• Multistage fracturing to treat multiple intervals of thin pay zones of near distances.

These developments can be combined to all the 3 basic types of fracturing for the optimum production.

REFERENCES

http://en.wikipedia.org/wiki/Well_stimulation
http://oilgasglossary.com/stimulation.html
Pramana, A.A., Research Progress Report, Petroleum Eng. Dept., Bandung Institute of Technology, 2010
Heriott Watt University,Production Technology II Lecture Notes, Chapter 5, p.4
http://www.fossil.energy.gov/programs/reserves/npr/NSURM_Documentation.pdf
Holditch, S.A., Tight Gas Sands, Journal of Petroleum Technology, June 2006, p.84-90
Holditch, S.A., Stimulation of Tight Gas Reservoirs Worldwide, SPE Offshore Tehnology Conference, Houston, USA, May 2009.

231
Camperos C., Technology and Economics Affecting Unconventional Reservoir Development, M.Sc. thesis, Texas A&M Univ., Dec. 2008.
Economides, M.J., Nolte, K. (editors), Reservoir Stimulation, 3rd ed., Wiley, 2000.
Hoch, Ottmar, Personal Communication, 2010
Zelenev, A.S. et al. Fast Inverting, Brine and Additive tolerant Friction Reducer for Well Stimulation, SEP International Symposium on Oilfield
Chemistry, Woodlands, Texas, USA, April 2009
Curtice, R., et al., To Gel or Not to Gel, SPE Annal Technical Conference, Lousiana, USA, Oct.2009
Pallisch, T.T. et al., Slickwater Fracturing, Food for Thought, SPE Production & Operations Magazine, Aug. 2010, p. 327-344.
Aboud, R.S., et al., Past Technology Emerge Due to Lightweight Proppant Technology: Case Histories Applied on Mature Fields, SPE Latin
American and Caribbenan Petroleum engineering Conference, Buenos Aires, Argentina, April 2007.
Gulbis, J and Hodge, R., Fracturing Fluid Chemistry and Proppant in Reservoir Stimulation Handbook, Ed. Economides, 2007
Kumar, A. et al., Prospects of Foam Stimulation in Oil and Gas Wells of India, Trinidad and Tobago Energy Resources Conference, Trinidad, June
2010.
El-Monier, I.A., et al., A New Environmentally Friendly Clay Stabilizer, SPE Production and Operation Conference, Tunisia, June 2010
Curtice, R.J., et al., Relative Permeability Modifiers Used in Conjunction With Hydraulic Fract Can Increase Hydrocarbon Production With
Hydraulic Frac.,SPE Eastern Regional Joint meeting in Pittsburgh Pennsylvania, USA , Oct. 2009
Portier, S. et al., Review on chemical stimulation techniques in oil industry and applications to geothermal systems, Centre for Geothermal Research,
Neuchâtel, Switzerland, Technical Report, May 2007
Economides, M.J. and Martin, T. (eds), Modern Fracturing – Enhancing Natural Gas Production, Gulf Publishing Co., 2008
Nasr-El-DinNasr-El-Din, H.A., et al., Fracture Acidizing: What Role Does Formation Softening Play in Production Response ?, SPE Production &
Operation Magazine, May 2008, p. 184-191
Zaeff, G., et al., Recent Acid-Fracturing Practices on Strawn Formation in Terrell County, Texas, SPE Annual Technical Conference, Anaheim,
California, USA, November 2007.
Mahmoud, M.A., et al., Stimulation of Carbonate Reservoirs Using GLDA (Chelating Agent) Solutions, SPE Trinidad/Tobago Energy/Resources
Conference, Trinidad, June 2010.
http://www.shell.com/static/usa/downloads/about_shell/upstream/chronology_fact_sheet.pdf
GeoExpro Magazine, Dec. 2007, p.24-29
http://www.oilshaleassoc.org/documents/Sep-2009-Newsletter-Final.pdf
http://www.iepm.com/pressreleases.htm
http://en.wikipedia.org/wiki/Chevron_CRUSH
http://fossil.energy.gov/programs/reserves/npr/Secure_Fuels_from_Domestic_Resources_-_P.pdf
Marques, LCC., A New Technique to Solve Gas Hydrate Problems in Subsea Christmas Trees, SPE Production & Facilities Journal, Vol.19 No.4,
Nov. 2004, p.253-258)
NPC Global Oil & Gas Study, Working Document #22: Heavy Oil, July 2007
http://www.heavyoilinfo.com/feature_items/interview-dr-ali-suat-bagci
Rangel-German, E.R. et al., Electrical-Heating-Assisted Recovery for Heavy Oil, Journal of Petroleum Science and Engineering 45 (2004) p.213–
231.
Sahni, A. et al., Electromagnetic Heating Methods for Heavy Oil Reservoir, 2000, SPE Regional Meeting. Long Beach, California, USA.
Das, S. Electromagnetic Heating in Viscous Oil Reservoir, 2008, SPE/PS/CHOA 117693, ITOHOSymposium, Calgary, Canada
Callarotti, R., Electrical Heating Method, Petroleum Engineering Handbook 2007.
Koolman, M., Huber, N., Diehl, L, Wacker, B., Electromagnetic Heating Method to Improve SAGD, 2008, SPE International Thermal Operations
and Heavy Oil Symposium.
Wittle, J.K., et al., direct Current Electrical Enhanced Oil Recovery in Heavy Oil Reservoirs to Improve Recover,, Reduce Water Cut, and Reduce
H2S Production While Increasing API Gravity, SPE Western Regional & Pacific Sect. APPG Joint Meeting, Bakersfield, USA, April 2008.
McGee, B.C.W., et al., Electro-Thermal Dinamic Stripping Process, SPE- ITOHOS, Calgary, 2008
Pramana, A.A. et al., Electromagnetic Induction Heat Generation of Nano-ferrofluid and Other Stimulants for Heavy Oil Recovery, American
Institute of Physics Proceedings of Nano-Science & Technology International Conference, Bandung, Indonesia, June 2010.
McDaniel, B.W. et al., Hydrajet (Abrasive) Perforating Can Improve Success of Fracturing Stimulations, IPTC, Kuala Lumpur, Malaysia, December
2008
McDaniel, B.W. et al., Use of Hydrajet Perforating To Improve Fracturing Success sees Global Expansion, CIPC/SPE Gas Technology Symposium,
Calgary, Canada, 2008
Barclay, D., et al., First North Sea Application of Pinpoint Stimulation Technology to Perform a Rig Based Acid Fracture Treatment Through CT,
SPE CT & Well Intervention Conference, Woodlands, Texas USA, April 2009
Brooks, R.T., et al., Improvement in Completing and Testing Multi Zone Open-hole Carbonate Formations, SPE Middle East Oil & Gas Show and
Conference, Bahrain, March 2009.

232
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Deep water drilling platforms - Zenith for efficient operations.


S. Priyadharshini
Petroleum Engineering (b.tech 3rd year)
School of Petroleum Technology, Pandit Deendayal Petroleum University
Gandhinagar, Gujarat, India
E-mails: priya.sbtech08@spt.pdpu.ac.in

ABSTRACT

Drilling for hydrocarbons offshore, in some instances hundreds of miles away from the nearest landmass, poses a number of challenges over drilling
onshore. The actual drilling mechanism used to delve into the sea floor is much the same as can be found on an onshore rig. However, with drilling at
sea, the sea floor can sometimes be thousands of feet below the sea level. Therefore, while with the onshore drilling the ground provides a platform
from which to drill, at sea an artificial drilling platform must be constructed. For offshore drilling, it is crucial to construct a platform to maximize the
drill’s extraction, while at the same time ensuring the safety of the workers by compensating for the water’s natural movement. Advances in drilling
and production technology have increased the possibility that offshore platforms can be controlled from an onshore location. Including a function to
control automatic shut off that will minimize pollution. This paper discusses in detail about the various types of offshore deep water drilling
platforms, their advantages and disadvantages, their maintenance and supply, the risks involved in carrying out operations and their ecological effects
with a few quoted examples and a glimpse of design criteria.

INTRODUCTION

An offshore platform often referred to as an oil platform or an oil rig, is a large structure used in offshore drilling to house workers and machinery
needed to drill wells in the ocean bed, extract oil or natural gas, or both, process the produced fluids, and ship or pipe them to shore. Depending on
the circumstances, the platform may be fixed to the ocean floor, may consist of an artificial island or may float. For offshore drilling, it is crucial to
construct a platform to maximize the drill’s extraction, while at the same time ensuring the safety of the workers by compensating for the water’s
natural movement. Drilling offshore dates back as early as 1869, when one of the first patents was granted to T.F Rowland for his offshore drilling
rig design. This rig was designed to operate on very shallow water. The first offshore well was drilled in Gulf of Mexico in 1947. Since then, offshore
production, particularly in the Gulf of Mexico, has been very successful, with the discovery and delivery of a great number of large oil and gas
deposits.

Remote sub sea wells may be connected to a platform by flow lines and by umbilical connections; these sub sea solutions may consist of single wells
of a manifold centre for multiple wells. Since the land that is going to be drilled through cannot provide a base for offshore drilling as it does for
onshore drilling, an artificial platform must be created. This artificial platform may take many forms, depending on the characteristics of the well to
be drilled, including how far under water the drilling target is. One of the most important pieces of equipment for offshore drilling is the sub sea
drilling template. Essentially, this piece of equipment connects the under water drill site to the drilling platform on the surface of the water. The
drilling template, secured to the sea floor and attached to the drilling platform above with cables, allows for accurate drilling to take place, but allows
for the movement of the platform above, which will inevitably be affected by shifting wind and water currents. In addition to the drilling template, a
BOP is installed on the sea floor which is same as that used in onshore drilling; it prevents any oil or gas from seeping out into the water. Above the
BOP, a specialized system known as ‘marine riser’ extends from the sea floor to the drilling platform above. It is designed to house the drill bit and
drill string, and yet be flexible enough to deal with the movement of the drilling platform. Strategically placed slip and ball joints, ij the marine riser
allow the sub sea well to be affected by the pitching and rolling of the drilling platform. The cost of deep water platforms has risen, principally as a
function of the environmental conditions and as a consequence larger funds are now required to make production profitable in these severe
environments. Fundamentally the need is for a load bearing system, which is sufficiently high, with respect to the water surface, so that various
mechanical operations can take place in the dry. When production capability is required in addition to the drilling capability in shallow waters, an
individual production platform may be constructed which would have the necessary facilities for treating the gas condensate, oil prior to shipment.
However a combined drilling/production platform is encountered in deeper water where higher capital investment increases the desire to be able to
drill and produce from the same platform. This type of self-contained platform can normally provide both functions.

Offshore platforms can broadly categorized in two types:


Fixed structure that extend to the sea bed
• Steel jacket
• Concrete gravity structure
• Complaint towers

Structures that float near the water surface-Recent development

• Tension leg platforms


• Semisubmersible platforms
• Spar platforms
• Drill Ships

DESCRIPTION OF VARIOUS TYPES OF PLATFORMS


Fixed Platform

These platforms are built on concrete or steel legs, or, both anchored directly onto the sea bed, supporting a deck with space for drilling rigs,
production facilities and crew quarters. Such platforms are, by virtue of their immobility, designed for very long term use. Fixed platforms are
economically feasible for installation in water depths up to about 1700 ft or 520m.

Jacketed Platform

ƒ Space framed structure with tubular members supported on piled foundations.


ƒ Used for moderate water depths up to 400 m.
ƒ Jackets provides protective layer around the pipes.
ƒ Typical offshore structure will have a deck structure containing a Main Deck, a Cellar Deck, and a Helideck.

234
ƒ The deck structure is supported by deck legs connected to the top of the piles. The piles extend from above the Mean Low Water
through the seabed and into the soil
• Underwater, the piles are contained inside the legs of a “jacket” structure which serves a bracing for the piles against lateral loads.
• The jacket also serves as a template for the initial driving of the piles.( The piles are driven through the inside of the legs of the jacket
structure).
• Natural periodint (usually 2.5 seconds) is kept below wave period (14 to 20 seconds) to avoid amplification of wave loads.
• 95% of offshore platforms around the world are jacket supported.

Many parameters influence the design of the jacket, such as required strength, fatigue, load and lifecycle. The Pile design results in a balanced
combination of diameter, penetration, pile wall thickness and spacing.The design of the pile is very important in the design of the jacket structure
itself and the cost of the pile foundation and installation could be as much as 40% of total cost of petroleum structure.

These structures can withstand immense vertical loading and overturning moments as they are designed to be resistant to toppling from very large
wave fronts. It can be assumed that many fixed offshore installations can withstand vertical loadings and the overturning moments imposed by
renewable energy devices. It is not advisable to impose any further lateral loading on the offshore installation as this may affect the overall strength
of the platform which could create a potential hazard. An individual structural analysis will however have to be carried out to determine the
suitability of reusing each fixed installation.

Complaint tower

• Narrow, flexible framed structures supported by piled foundations.


• Has no oil storage capacity. Production is through tensioned rigid risers and export by flexible or catenary steel pipe.
• Undergo large lateral deflections (up to 10 ft) under wave loading.
• Used for moderate water depths of 1500-3000ft.
• Natural period (usually 30 second) is kept above wave period (14 to 20 seconds) to avoid amplification of wave loads.

With the use of flex elements such as flex legs or axial tubes, resonance is reduced and wave forces are de-amplified. This type of rig structure
can be configured to adapt to existing fabrication and installation equipment. Compared with floating systems, such as Tension-leg platforms
and SPARs, the production risers are conventional and are subjected to less structural demands and flexing. This flexibility allows it to operate
in much deeper water, as it can 'absorb' much of the pressure exerted on it by the wind and sea. Despite its flexibility, the compliant tower
system is strong enough to withstand hurricane conditions which is the greatest advantage.

Concrete gravity structure

• Fixed-bottom structures made from concrete


• Heavy and remain in place on the seabed without the need for piles
• Used for moderate water depths up to 300 m.
• Part construction is made in a dry dock adjacent to the sea. The structure is built from bottom up, like onshore structure.

235
• At a certain point, dock is flooded and the partially built structure floats. It is towed to deeper sheltered water where remaining
construction is completed.
• After towing to field, base is filled with water to sink it on the seabed.
• Advantage- Less maintenance

The design of offshore gravity platforms necessitates a dynamic response analysis to account for dynamic load effects on the structure.
The environmental loads acting on the platform are dynamic, and this affects the design analysis w.r.t:
1. Extreme loads (waves, wind, earthquakes) .
2. A fatigue analysis for the anticipated platform life.
3. Local vibration levels (i.e. at top deck) to satisfy facilities and equipment design.

FLOATAR TYPE PLATFORM


Tension Leg Platform (TLP)

ƒ TLP has excess buoyancy which keeps tethers in tension. Topside facilities, no. of risers etc. have to fixed at pre-design stage.
ƒ Used for deep water up to 1200 m.
ƒ It has no integral storage.
ƒ It is sensitive to topside load/draught variations as tether tensions are affected.

A Tension Leg Platform (TLP) is a buoyant platform held in place by a mooring system. The TLP’s are similar to conventional fixed platforms
except that the platform is maintained on location through the use of moorings held in tension by the buoyancy of the hull. The mooring system is a
set of tension legs or tendons attached to the platform and connected to a template or foundation on the seafloor. The template is held in place by
piles driven into the seafloor. This method dampens the vertical motions of the platform, but allows for horizontal movements. The topside facilities
(processing facilities, pipelines, and surface trees) of the TLP and most of the daily operations are the same as for a conventional platform. The
foundation is the link between the seafloor and the TLP. Most foundations are templates laid on the seafloor, then secured by concrete or steel piles
driven into the seafloor by use of a hydraulic hammer, but other designs can be used such as a gravity foundation. The foundations are built onshore
and towed to the site. As many as 16 concrete piles with dimensions of 100 ft in diameter and 400 ft long are used (one for each tendon). The hull is a
buoyant structure that supports the deck section of the platform and its drilling and production equipment. A typical hull has four air-filled columns
supported by pontoons, similar to a semisubmersible drilling vessel. The deck for the surface facilities rests on the hull. The buoyancy of the hull
exceeds the weight of the platform, requiring taut moorings or “tension legs” to secure the structure to the seafloor. The columns in the hull range up
to 100 ft in diameter and up to 360 ft in height; the overall hull measurements will depend on the size of the columns and the size of the platform

A template provides a frame on the seafloor in which to insert either conductors or piles. Not all TLP’s use templates; if used, they are typically the
first equipment installed at the site. There are several types of templates that may be used in conjunction with a TLP to support drilling, foundation
integrity, or the integration of the two. Drilling templates provide a guide for locating and drilling wells; they may also be a base for the tie-in of flow
lines from satellite wells or for export pipelines and their risers. Foundation templates may be one single piece or separate pieces for each corner. The
foundation piles are driven through the foundation template. An integrated template is a single piece that contains all drilling support, anchors the
tendons, and locates and guides the foundation piles. Separate templates allow each part to be installed individually. They also use smaller pieces that
weigh less and are easier to install. The drilling template can be installed and drilling can begin while the foundation template is being designed and
built. Tension Legs (tendons) are tubular that secure the hull to the foundation; this is the mooring system for the TLP. Tendons are typically steel
tubes with dimensions of 2-3 ft in diameter with up to 3 inches of wall thickness, the length depending on water depth. A typical TLP would be
installed with as many as 16 tendons. A production riser conveys produced fluids from the well to the TLP surface production facilities. Riser system
for a TLP could be either a single-bore or dual-bore (concentric pipe) arrangement. The dual-bore riser is rated for low pressure and marine riser that
serves as an environmental barrier is rated for high pressures.

236
Installation of a TLP is done in stages; often the design work on one section of the TLP is being done while another part is being installed. For
example, the wells will often be predrilled while the TLP is being designed and constructed. The wells may be drilled to their total depth, or partially
drilled and the conductor casing set. Additional well drilling and completion operations can be done from the TLP. The template is built onshore and
towed to location for installation. A drilling rig (mobile offshore drilling unit [MODU]) is preferred for installation because it eliminates the need for
additional vessels. However, drilling rigs cannot lift large payloads and have limited lowering capacity. Large templates may need a crane for
installation; they will also require costly handling systems and rigging. The upper section of a TLP consists of the hull, the deck, and the surface
facilities. The surface facility modules are built onshore and typically assembled at a shallow-water location near shore, then towed to the site. The
modules may be attached to the hull either inshore or at the site. Well drilling for the TLP often begins after well template installation. Predrilling
involves using a mobile offshore drilling unit (drillship or semisubmersible) to batch drill and case the wells to a convenient depth, normally through
the shallow water flow zone or other potential hazard. Predrilling may also be suspended just above the production zone. Some wells may be drilled
to total depth and completed. The Sonat George Richardson semisubmersible drilling vessel is an example of the type of vessel used to predrill.

Sea star drilling platforms:

They are like miniature tension leg platforms. A lower hull is fitted with water when drilling, which increases the stability of the platform against
wind and water movement. Sea star platforms also incorporate the tension leg system employed in larger platforms. They can operate in water
depths up to 3500 ft.

Spar Platform

A spar is a deep-draft floating caisson, which is a hollow cylindrical structure similar to a very large buoy. Its four major systems are hull, moorings,
topsides, and risers. The spar relies on a traditional mooring system (that is, anchor-spread mooring) to maintain its position. About 90 percent of the
structure is underwater. Historically, spars were used as marker buoys, for gathering oceanographic data, and for oil storage. The spar design is now
being used for drilling, production, or both. The distinguishing feature of a spar is its deep-draft hull, which produces very favorable motion
characteristics compared to other floating concepts. Low motions and a protected center well also provide an excellent configuration for deepwater
operations. Water depth capability has been stated by industry as ranging up to 10,000 ft (Ultra deep water depth).

The number of wells, surface wellhead spacing, and facilities weight determine the size of the center well and the diameter of the hull. In the classic
or full cylinder hull forms, the upper section is compartmentalized around a flooded center well containing the different type of risers. This section
provides the buoyancy for the spar. The bottom section (keel) is compartmentalized to provide buoyancy during transport and to contain any field-
installed, fixed ballast.The mooring lines are a combination of spiral strand wire and chain. Each mooring line is anchored to the seafloor with a
driven or suction pile. The hull end of the line passes through a fairlead located on the hull below the water surface, then extends up the outside of the
hull to chain jacks at the top, usually 50 ft or more in elevation. Excess chain is stored in the hull. Design- Depending on hull size and water depth,
the moorings can vary in number up to 20 lines and contain 3,700 ft of chain and wire. The topsides configurations follow typical fixed platform
design practices. The decks can accommodate a full drilling rig (3,000 hp) or a work over rig (600-1,000 hp) plus full production equipment.
Production capacities range up to 100,000 BOPD and 325 MMcfgpd. The type and scale of operation directly influence deck size. The larger topsides
would be consistent with drilling, production, processing. Total operating deck load, which includes facilities, contained fluids, deck structural and
support steel, drilling/work over rig, and work over variable loads, can be 6,600 tons or more. There are three basic types of risers: production,
drilling, and export/import. Production – Each vertical access production riser is top tensioned with a buoyant cylinder assembly through which one
or two strings of well casing are tied back and the well completed. This arrangement allows for surface trees and a surface BOP for work over. The
drilling riser is also a top-tensioned casing with a surface drilling BOP, which allows a platform-type rig to be used. Due to space restrictions in the
core, number of risers has to be pre-determined.

237
Semi-Submersible Platforms

The semi-submersible has been adapted to perform numerous functions offshore unconnected with drilling but still operating as a mobile unit. When
used as an early production unit however, it is required to remain on a particular site for several years. A semi-submersible is a complex structural
unit, it is made up of several components, each performing a Particular function necessary for overall structural integrity and possessing flexibilities
and stiffnesses, which vary by several orders of magnitude. There is a large deck area, of 50m wide and 70m long with plate thicknesses0 of 7-10mm
for equipment, stores, housing, and work space. There may be several deck levels with an overall depth of 8m, with intermediate transverse and
longitudinal bulkheads, forming an extremely stiff box type structure. There may be several cut-outs, For example, for drilling or pipe laying and
there are usually heavier bulkheads with plate thicknesses varying between 8-12mm, where additional support is required, for example, the derrick or
crane tub support. The deck plating and support structure are designed to a large extent by local loads, pipes, mud, forklift trucks, mobile cranes, etc
.
A semi-submersible obtains its buoyancy from ballasted, watertight pontoons located below the ocean surface and wave action. The operating deck
can be located high above the sea level due to the good stability of the concept, and therefore the operating deck is kept well away from the waves.
Structural columns connect the pontoons and operating deck. With its hull structure submerged at a deep draft, the semi-submersible is less affected
by wave loadings than a normal ship. This is its main advantage. With a small water-plane area, however, the semi-submersible is sensitive to load
changes, and therefore must be carefully trimmed to maintain stability. Unlike a submarine or submersible, during normal operations, a semi-
submersible vessel is never entirely underwater. A semi-submersible vessel is able to transform from a deep to a shallow draft by deballasting
(removing ballast water from the hull), and thereby become a surface vessel. The heavy lift vessels use this capability to submerge the majority of
their structure, locate beneath another floating vessel, and then deballast to pick up the other vessel as a cargo.
Used for ultra deep water depths even of up to 10,000 ft.

Drill Ships

A drill ship is a vessel designed to drill subsea wells. Its drilling tower is located midship, where an opening in the hull allows the drill string to pass
through. The drill ship positioning system, consisting of acoustic sensors, propellers and computers, cancels out effects of the wind, waves and
currents that tend to move the ship from its position. While drill ships look similar to a tanker or cargo vessel, there are a couple of major differences.
Drill ships are equipped with a drilling derrick and moon pool. Additionally, drill ships have extensive mooring or positioning equipment, as well as
a helipad to receive supplies and transport staff. Drilling equipment is passed through the vessel's moon pool and connected to the well equipment
below via riser pipe, a somewhat flexible pipe that extends from the top of the subsea well to the bottom of the drillship.

Drill ships are differentiated from other offshore drilling units by their easy mobility. While semisubmersible rigs can also drill in deep waters, drill
ships are able to propel themselves from well to well and location to location, unlike semisubmersible, which must rely on an outside transport vessel
to transfer them from place to place. The greatest advantages these modern drill ships have is their ability to drill in water depths of more than 2500
meters and the time saved sailing between oilfields worldwide. Drill ships are completely independent, in contrast to semi-submersibles and jack up
barges Drill ships were first developed in the late 1940s by marine architects. Created to overcome water depth challenges offshore California in the
Pacific Ocean, the first drillship was a surplus U.S. Navy patrol craft that was equipped with cantilevered drilling equipment. Seeing some success,
the next drill ship incorporated a moon pool and an onboard drilling derrick. After its success, companies soon ordered the construction of subsequent
drill ships. While capable of drilling in deep and ultra-deep waters, a disadvantage to using a drillship is its susceptibility to being agitated by waves,
wind and currents. This is especially troublesome when the vessel is actually drilling, because the drillship is connected to equipment thousands of
feet under the sea. A proper mooring system on a drillship is integral to drilling successful wells. In shallower waters, drill ships are moored to the
seafloor with anywhere from 6 to twelve anchors. Once the water depth becomes too deep, drill ships depend on dynamic positioning systems (DPS)
to keep the vessel in place while drilling. DPS relies on several thrusters located on the fore, aft and mid sections of the ship, which are activated by
an onboard computer that constantly monitors winds and waves to adjust the thrusters to compensate for these changes. Sometimes, both positioning
systems are used.

238
Typically employed in deep and ultra-deep waters, drill ships work in water depths ranging from 2,000 to more than 10,000 feet (610 to 3,048
meters).

REAL WORLD EXAMPLE OF OIL RIG EXPLOSION

The Deepwater Horizon drilling rig explosion refers to the April 20, 2010 explosion and subsequent fire on the Deepwater Horizon semi-
submersible Mobile Offshore Drilling Unit (MODU), situated about 40 miles (60 km) southeast of the Louisiana coast in the Macondo Prospect oil
field, owned and operated by Transocean. The explosion killed 11 workers and injured 17 others; another 98 people survived without serious
physical injury. It caused the Deepwater Horizon to burn and sink, and started a massive offshore oil spill in the Gulf of Mexico; this environmental
disaster is now considered the largest in U.S. history.

The Deepwater Horizon was a floating semi-submersible Drilling Unit — a fifth-generation, ultra-deepwater, dynamically positioned, column-
stabilized drilling rig owned by Transocean and built in Korea. The platform was 396 feet (121 m) long and 256 feet (78 m) wide and could operate
in waters up to 8,000 feet (2,400 m) deep, to a maximum drill depth of 30,000 feet (9,100 m). Press releases from Transocean state the platform had
historically been used for deeper wells, including the deepest underwater gas and oil well in history at 35,055 feet (10,685 m) in 2009. The platform
commenced drilling in February 2010 at a water depth of approximately 5,000 feet (1,500 m). At the time of the explosion the rig was drilling an
exploratory well. The planned well was to be drilled to 18,360 feet (5,600 m) below sea level, and was to be plugged and suspended for subsequent
completion as a subsea producer. Production casing was being run and cemented at the time of the accident. Once the cementing was complete, it
was due to be tested for integrity and a cement plug set to temporarily abandon the well for later completion as a subsea producer.

There were few indications of any trouble with the Deepwater Horizon before the explosion

According to Beyond Petroleum (formerly British Petroleum, or BP), the rig was drilling 18,000 feet down to get to pockets of gas and oil under
pressure when it caught fire. The rig reportedly lacked a last-ditch safety valve, an “acoustic switch,” that could have potentially averted the massive
oil spill. Such safety mechanisms are common in many oil rich countries around the world, but are not mandated in the U.S. because of their high
cost.
Some of the causes of explosion and fire of rig were reported as:
• Abnormal pressure accumulation inside the marine riser which expanded and ignited rapidly as it came up
• A bubble of methane gas escaped from the well and shot up the drill column, expanding quickly as it burst through several seals and
barriers before exploding. Basically, it was a blowout.

Explosion was followed by a fire that engulfed the platform.

Though the actual or the root cause of the explosion is not yet fully known, the investigations came out with different predictions such as:-
• Sudden catastrophic failure of the cement, the casing or both.
• Misinterpretation pressure test and ignorance of ominous signs such as a pipe called a riser losing fluid.
• Lack of centralizers.
• Non-reduction of flow of flammable gases.

ENVIRONMENTAL REQUIREMENTS

Various parameters like water depth, wind loading and soil conditions have to be looked into while performing the required operations. Each of these
parameters individually may have significant impact on the cost and configuration of the structure and collectivity may have devastating impact.

239
Increasing water depths, of course, involve additional materials, which result in greater cost, and increasing wave size, with its larger loading, has a
similar effect. Wind loads are relatively small; however, for high winds and larger projected areas, they form a significant part of the overall loads
imposed on the structure.

Foundation conditions are one of the most important environmental criteria because a platform must transmit all of the afore-mentioned loads to the
sea floor through the foundation system. Traditionally this has been achieved by utilizing piling. However, hard soils create difficulty because it is
difficult and expensive to obtain the necessary pile penetration in order to connect the platform to the sea floor; soft soil on the other hand often
yields a condition whereby no strength may be obtained during the initial pile/soil connection. In some cases the piling can be unsupported for
distances upto 50 feet. Thus, weak soils from a structure point of view may create a deeper water condition than actually exists.

CONCLUSION

This paper has discussed about the various types of offshore deep water drilling platforms available, their construction, advantages and disadvantages
of each type of platform, operating depth, their maintenance, supply, design aspects to some extent and also about the environmental considerations
with a few quoted examples. It would appear that the development of deep water reserves is feasible not only from an operational point of view but
from economics as well. This is dependent upon a careful evaluation of all factors-environmental, reservoir characteristics and use of a fully
integrated development system that provides a reasonable maintenance capability for the life of reservoir. These concepts for the production of oil
and gas from deep water locations are being presented for review and evaluation from great number of sources. Advances in drilling and production
technology have increased the possibility that offshore platforms can be controlled from an onshore location. Including a function to control
automatic shut off that will minimize pollution. The operating depth criteria has to be very carefully looked into before choosing a type of
platform for particular operation and also the various parameters mentioned in the ecological considerations are very important in order to ensure an
economical and expeditious platform solution.

ACKNOWLEDGEMENT

Iam very thankful to my University and College for supporting me financially and Helping me to attend this International conference on Drilling
Technology which I hope to be a great platform for upcoming Petroleum Engineers like us. I would also like to thank my parents for encouraging me
through out in preparing the paper. Iam also thankful to my professors for encouraging and boosting confidence in me.

REFERENCES

Books

“Oil and gas technologies for Arctic and Deepwater” by Dr.John H. Steele (Washington DC; U.S Congress, office of Technology assessment, OTA-
o-270, May1985)
Internet sources
www.wikepedia.org
www.naturalgas.org/.../extraction-offshore.asp
www.marinetalk.com
www.offshore-technology.com
www.subsea.org
www.globalsecurity.org
www.nola.com/news/gulf-oil-spill
www.freepatentonline.com
www.csmonitor.com/.../gulf-of-Mexico-oil-spill
Related Research papers from www.onepetro.com

240
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Downhole Pump Selection Methods


Abdus Samad
Department of Ocean Engineering, Indian Institute of Technology Madras
Chennai-600036, India
E-mail: samad@iitm.ac.in

ABSTRACT

Downhole pumps are the subsurface part of the artificial lifts or ALs used to pumping fluid from the oil and gas wells. As the reservoir pressure
declines, the well fluid ceases to flow to the surface and the pumps add energy to the fluid to serve the purpose. There are several ALs; such as Sucker
rod pump, electric submersible pump, progressive cavity pump, hydraulic pump etc and the ALs selection depends on the well condition and fluid
properties. A typical well bore might have four phases fluid and the phases are water, gas, oil and sand. Selecting a right type of AL needs the
expertise of pump as well as well fluid condition. Hence proper selection of downhole pump is necessary to reduce the workover cost as well as to
keep good production rate. Present article discusses about the selection methods of downhole pumps.

KEY WORDS: Downhole pump, artificial lift, SRP, ESP, PCP, gas lift, hydraulic pumping.

INTRODUCTION

Downhole pumps (DPs) are the subsurface part of the artificial lifts (ALs) used to pump fluid from the oil and gas wells. Initially, the oil or gas
reservoirs maintain high pressure and the well fluid flows from the well to the surface naturally. After a period of time, the well pressure gets reduced
and the well is unable to deliver the fluid economically. The DPs which add energy to the fluid so that fluid can flow from a lower pressure pay zone
are used to pump fluid from the wells. In gas wells, if well pressure gets reduced and the liquid column builds up, the gas cannot enter into the well
through the casing perforations. The ALs remove liquid from the perforation zone and gas starts entering into the well. The gas needs low pressure to
move to the surface. Sucker rod pump (SRP), electric submersible pump (ESP), progressive cavity pump (PCP), hydraulic pump (HP) etc. are the
common type of pumps which are used for liquid pumping from the well. The other lifting methods which are included in the AL category are gas
lifts.

The pump selection depends on the well condition and fluid properties. The typical well bore may have four phase fluid and the phases are water, oil,
gas and sand. If a pump works good for water may not work for gassy situation. Similarly, sand is problem for certain type of pumps. Inflow
performance relationship (IPR) is an important criterion to select an AL. High inflow rate needs a high volume delivery pump. Dewatering pumps of
gas wells are typically low delivery pumps. Heavy oil may contain sand and the sand causes wear and tear to the submersible pumps. High viscosity
fluid is handled by positive displacement pumps and kinetic pumps are good for low viscosity fluids. Depth, deviation (horizontal or vertical or
inclined well) etc of the wells are also the factor for selection of pumps. If the pump ceases to deliver fluids economically, it is inevitable to do
expensive ‘work over’ operation. Hence, proper selection of pump is necessary to reduce the work over cost as well as to keep good production rate.

Several authors have reported the selection criteria of ALs on the basis of different criteria. Reservoir, well considerations, environmental impact etc
were considered to select an AL (Neely et al., 1981). About 87 % of the producing wells in the world are operated using different AL methods.
Among those SRP, ESP, gas lift and other AL methods have shares of 71, 14, 8 and 7% respectively (Ghareeb et al., 2007). A software has been
developed to select an AL for different wells (Alemi et al., 2010 and Jiang et al., 2007). Cost analysis basis AL selection method was presented by
Heinze et al. (1995) and Clegg (1988). Espine et al. (1994) presented and expert system development to select an AL for a well or group of wells.
Clegg and Bucaram (1993) compared the main selection attributes for the major AL methods and on the performance and operating capabilities of the
methods. Khasanov et al. (2010) discussed main AL selection steps for Vankor oil field considering flow rate, water cut, gas-oil-ratio (GOR) and
productivity index. Case study of AL selection from Golfo Sand gorge basin was presented by Hirchfeldt and Ruiz (2009) and Hirchfeldt et al. (2007).
The article highlights several methods commonly used for selection and also includes some examples of costs and profits over time calculated to the
present time as a tool of selection (Lea and Nickens, 1999). An electric submersible PCP and jet pump was used for high viscosity fluid pumping and
multi zone testing and study was presented by Lamki et al. (2007).

Present article reviews and discusses about the selection method of ALs on the basis of multiphase flow along with the other criteria and presents of a
matrix for quick selection of a downhole pump.

TYPES OF ARTIFICIAL LIFT

The artificial lifts (ALs) works mainly on the principle of addition of energy by pumps, injecting liquid or injecting gas into the well so that the liquid
can be lifted to the surface. The pumps add energy to the fluid. The energy is supplied through the reciprocating or rotary motion produced by a motor
or internal combustion engine. Gas is injected from the surface to the downhole to facilitate the fluid to be lifted to the surface. In hydraulic pumping
principle, liquid is injected into the well to lift the well fluid to the surface. The ALs are broadly classified and shown in Fig. 1. The following sections
discusses about the main types of DPs or ALs.

PUMPS

A pump displaces a volume of fluid by physical or mechanical action. Major types of pumps which are used as AL are positive displacement or kinetic
pumps (Fig. 1). The positive displacement pumps (PDPs) delivers fluid by the reciprocating action of a plunger or rotary motion of rotor. Typically an
SRP is a reciprocating pump having a plunger-barrel assembly while a PCP is a rotary pump having a rotor which rotates inside a stator. The kinetic
pump or centrifugal pump adds energy to the fluid by the rotation of impellers. Conventional ESP is a centrifugal pump. The positive displacement
pumps are generally low speed pump and have high head producing capability. On the other hand, the centrifugal pumps deliver fluid having lower
head and higher flow rate. Fig. 2 shows the difference between these pumps. Flow rate is fixed irrespective of head developed for a specific speed of
the pump in the case of positive displacement pump. Pumps delivering any fluid irrespective of area of application must be selected on the basis of
fluid properties such as acidity or alkalinity, operating temperature, solid concentration and particle size, specific gravity, viscosity and vapour
pressure of fluid.

Positive displacement pumps: The positive displacement pumps (PDPs) are divided into two main classes: reciprocating and rotary. The PDPs
delivers the same flow at a given speed (RPM) irrespective of discharge pressure. That’s why these are called constant flow machines. A PDP operates
against a closed valve on the discharge side of the pump because it has no shut-off head like the centrifugal pumps. SRPs are in the category of the
PDP and a SRP consists of a cylinder with a reciprocating plunger in it. A rotary pump contains a rotor which rotated inside a stator. Details will be
discussed in the SRP/PCP section.

Artificial lifts

Gas lift Hydraulic Pumps

Reciprocating pump Jet pump


Positive Kinetic
(or a positive
displacement
displacement pump)

Reciprocatin Rotary Centrifugal


(ESP)
Continuous Intermittent
flow SRP
flow
Radial Mixed
flow flow
Conventiona Plunger lift Chamber lift Insert Tubin pump pump
l intermittent pump g
gas lift
Multiple Single screw
screw pump pump or

Elastomeri Metallic
c stator stator PCP

Fig 1: Major artificial lift systems


 

242 
 
Sucker rod pump: An SRP is a reciprocating PDP and widely used in the oil and gas wells. The pump has a subsurface plunger barrel assembly and a
pumping unit (Fig. 3). The pumping unit gives reciprocating motion to the plunger and fluid is delivered only in upstroke. Fig. 4 shows a reciprocating
pump or SRP and this pump operates vertically. It has a plunger which travels inside a barrel. The plunger is connected to a valve rod and the rod
supplies tensile or compressive force to the plunger giving a reciprocating motion. ‘Up’ and ‘down’ strokes are defined by the plunger movement
opposite to or to the gravitational pull. The plunger has a travelling valve (TV) which opens only during the down stroke of the plunger. A standing
valve (SV) which seats on a valve seat threaded to the plunger barrel opens only during the upstroke of the plunger. The valve seats and valves (TV
and SV) are made up of hardened metal because during pumping operation the valves constantly will hits the seat. This creates high rate of wear. The
valves are simply ball valves. Valve cages are fitted at the top of the balls so that valves cannot move too long distance from valve seat during opening
period. Cavity C1 is the volumetric space between the TV and SV. Similarly, cavity C2 is the volumetric space above the TV. The valves are gravity
controlled; hence the pump does not work horizontally (Samad, 2010). There are several types of SRPs such as insert type or tubing type and these are
selected on the basis on pump depth, sand and wax content etc. Rod pumping is not normally considered to be particularly applicable for high-rate AL
systems (Clegg, 1988).

Screw pumps: Screw pumps are the positive displacement pumps in which the flow through the pumping is axial. The screw threads mesh each other
and deliver fluid. There is single and twin screw pumps used as downhole pump. These pumps contain two screws with opposing thread where one
screw turns clockwise and the other counter clockwise. The screws are mounted on shafts that run parallel to each other. The single screw pump has a
single screw type rotor and the rotor meshes with stator creating a cavity to deliver fluid. The progressive cavity pump (PCP) is a single screw type
pump which consists of an elastomeric stator and a helical rotor (Fig. 5 and 6). The rotor rotates inside the stator. The whole rotor stator system is
inserted into the well and a sucker rod rotated by a surface installed motor transfers the torque. The rotor speed varied from 10 to 500rpm and viscosity
of fluid which these pumps can handle is up to 100000cSt. The PCPs have good priming capability. Entrained air can be allowed for a short while but
the air gets compressed due to polytropic compression and temperature gets increased if gas enters into the cavities for a longer period of time. Higher
temperature of fluid can be handled if metallic stator is used instead of polymeric stator (Nelik and Brennan, 2005).

Kinetic pump: A kinetic pump adds kinetic energy to the fluid by increasing the flow velocity by impellers. A centrifugal pump is a kinetic pump that
uses rotating impellers (Fig. 7) to increase the fluid velocity and the velocity is converted into static head by a diffuser or volute. The centrifugal
pumps can be classified as radial, axial and mixed flow type. The radial flow pumps delivers fluid with high head and lower flow rates than the axial
and mixed flow pumps. The axial pumps are high flow and low head pumps. ESPs are typically multistage centrifugal pump having radial or mixed
flow impellers.

ESP: The ESPs are single- or multi-stage centrifugal pumps and the pumps are connected in series. A stage is constituted by the assembly of an
impeller and a diffuser. An ESP consists of stages along with protector, motor and electric power cable (Fig. 8). The ESP can handle a very wide range
of flow rates from 200 to 90,000 barrels per day and bottomhole assembly depth can be up to 10,000 ft. An ESP can be placed in vertical, deviated, or
horizontal wells.

GAS LIFTS

Gas lift is a secondary recovery method that uses injected gas to supplement the pressure in an oil reservoir or field so that produced fluid can be lifted
to the surface. In most cases, a field will incorporate a planned distribution of gas-injection wells to maintain reservoir pressure and affect an efficient
sweep of recoverable liquids (Oil field glossary, Schlumberger). Gas lift is an AL method (API gas lift manual) and used to lift the fluid by reducing
the density of fluid. It closely resembles the natural flow of fluid from the well bore. During upward movement of the fluid, the gas from the liquid
comes out and gas bubbles gradually increases its volume which reduces the density and weight of the fluid column (Fig. 9). The reduction of the
column weight produces the pressure differential between the well bore and the reservoir that causes the well to flow. If formation water is higher,
dissolve gas amount is lower and hence a gas injection is required. This gas will act in the same way as if there is dissolve gas in the oil (API gas lift
manual). This system is free from any mechanical moving part inside well, hence this can handle solid and gas at a high temperature. There are
intermittent and continuous gas lift methods and gas is injected on the basis of well property by these methods. In continuous gas lift method, the gas
is injected continuously while gas is injected intermittently in the intermittent method.

- - - Centrifugal
Positive
Head, H

Flow
rate, Q
Fig 2: Positive displacement Vs Centrifugal pump

243 
 
Horse
Walking beam head Polished
Pitman rod
arm

Stuffing
box

Sampson post  Plunger

Tubing

Casing

(a) Down stroke (b) Up stroke

Fig 3 Complete sucker rod pump assembly Fig 4 Plunger barrel assembly 

To separator
Motor
Stator
Rotor

Sucker
rod
Rotor-stator
assembly

Casing

Tubing
Rotor

Pump
M

Perforations

Fig 6 Progressive cavity pump rotor and rotor-stator


Fig 5 Complete progressive cavity pump assembly assembly.

244 
 
To separator Power cable

VR Vr

Vt
Impeller
Impeller Casing
rotational
direction

Tubing

Pump
d

D Motor

Perforations

Fig 7 Centrifugal impeller (Vr, Vt and VR are the radial,


tangential and resultant velocities and d and D are the inner Fig 8 Complete electric submersible pump
and outer diameters of the impeller)

To separator
High Pressure power
fluid injection
Gas injection Spent power fluid
and production up

Casing
Casing

Tubing Gas lift valve


Tubing Hydraulic
pump

Packer
Packer

Perforations
Perforations

Fig 9: Gas lift Fig 10: Hydraulic pumping

245 
 
HYDRAULIC PUMPING
 
Hydraulic pumping systems transmit energy to the bottom of the well by means of pressurized power fluid that flows down in the wellbore tubular to a
subsurface pump (Fig.10). There are mainly two types of hydraulic subsurface pumps: reciprocating piston pump and jet pump. jet pump works on the
principle of venturi. These systems are versatile and have been used in well depths of 1000 ft to 18,000 ft, low rate wells with production in the tens of
barrels per day to wells producing in excess of 10,000 barrels per day (Lea et al., 2008). Hydraulic pumping systems are suitable for the deviated wells
where conventional pumps are not feasible. These pumps are sensitive to solids.

EFFECT OF FLUID PROPERTIES ON DOWNHOLE PUMPS

Solid or sand effect: The well producing sand along with fluid faces problem in selecting pump if reservoir pressure is not sufficient for natural flow.
Pumping can have problem related to the choking of valves or wear in centrifugal ESP, SRP, hydraulic pumping and plunger lift. During well
completion, sand screens are deployed so that the well can have less sand production. The sand screen adds pressure drop and reduces production rate.
If the well still produces sand, proper pumping method should be chosen to handle sand related problems. For high volume flow, gas lift will work.
PCP can handle solid although the life of the pump gets reduced. A limited amount of sand can be handled by ESP if proper material of impeller and
diffuser is selected.

Gas interference: Gas interference has big issue when pumping using SRP, ESP or PCP. SRP faces ‘gas lock’ or ‘fluid pound’, production gets
reduced and pump failure occurs due to gas interference (Samad, 2010). Gas lift has a positive effect as this requires less gas injection due to extra gas
present in the well. Jet pump can work under gassy environment. Centrifugal pump will not deliver any fluid as this causes cavitation and this will lead
to an uneven vibration and failure of the pumping system. PCP allows gas for a short while but not for a longer time.

Temperature: The high temperature has adverse effect if proper material has not been selected for a pump. Seal failure is the biggest problem in ESPs.
Increasing temperature will increase the chances of seal failure. The SRP does not have much effect because normally metal to metal sealing is used
for deep and long wells and a clearance is maintained. Elastomeric stator strength becomes weak and life gets reduced in PCPs. Metal to metal stator
can work for high temperature operation. Jet pump, gas lift and plunger have good pumping capability in high temperature.

Flow rate: Gas lift and ESPs are the highest volume pumps (Fleshman and Lekic, 1999). SRP is a low volume pump and volume delivery reduces as
pump setting depth becomes higher. Plunger lift, jet pump and PCP are the low volume pumps.

Net positive suction head, priming and dry running: Net positive suction head (NPSH) is the inlet pressure required for a pump. The well should
have pressure more than NPSH required. ESP needs highest NPSH while PCP and SRP requires less because these pumps delivers less volume of
fluid and these are low speed pump as compared to ESPs. Higher speed ESP needs higher NPSH and an inducer is placed at inlet to reduce the NPSH
requirement. Jet pump needs minimum NPSH while gas lift and plunger lift virtually does not need any NPSH.

ESP needs priming or filling the whole impeller and diffuser system with liquid else the gas at inlet will not allow the pump to deliver any fluid. Dry
running is not allowed for centrifugal pump. The SRP faces problem because of friction, heating and gas interference if the pump is not filled with
liquid or wells pressure gets reduced below the NPSH required by the pump. PCP can run dry for a short time because of friction and the elastomer
can get over heated and life will be reduced. The PCPs and the jet pumps are considered as self priming pumps. Plunger lift and gas lift do not have
such requirement.

Using the above discussion, a pump matrix has been prepared (Table 1). The table will help selecting a proper pump for a well having specific well or
fluid property. The table contains approximate data which can be found in the different literatures cited in this article. The digits placed in the cells of
matrix cannot be considered as absolute.

CONCLUSION
The present article presents the selection of different artificial lift or downhole pumping methods. The SRP, ESP, PCP, Gas Lift and Hydraulic
pumping solutions are analyzed and used for comparison. The pumps have their inherent characteristics of handling fluids. Any pump cannot be
selected randomly for any well. The SRP has good capability in the well has high temperature and low volume flow but has weakness of handling
solids. ESP has capability of lifting high volume flow from verticle or deviated wells and fluids should have low viscosity. PCP can handle solid with
high viscosity at lower flow rate. Gas lift has the good volume delivery capability but has limitation of unable to handle low volume economically.
Using the above discussion, a downhole pump matrix has been prepared for quick selection of a downhole pump.

REFERENCES

Alemi, M, Jalalifar, H, Kamali, G and Kalbasi, M, (2010), “A Prediction to the Best Artificial Lift Method Selection on the Basis of TOPSIS Model”,
Journal of Petroleum and Gas Engineering, Vol. 1(1), pp 009-015.
“API Gas lift Lift Manual, Book 6 of the Vocational Training Series”, Third edition, 1994, American petroleum Institute, Dallas, TX.
Aversante, OL and Morrow, S, (1995), “Plunger-Lift: Gas Assisted”, 1995, Ferguson Beauregard, Inc.

246 
 
Clegg, JD, (1988), “High-rate Artificial Lift”, Journal of Petroleum Technology, Vol. 40, No. 3, SPE 17638.
Clegg, JD, Bucaram, SM and Heln, J, (1993), “Recommendations and Comparisons for Selection of Artificial Lift Methods”, Journal of Petroleum
Technology, 45(12): 1128 -1167.
Espin, DA, Gasbarri, S, Chacin, JE and Intevep, SA, (1994), “Expert System for Selection of Optimum Artificial Lift Method”, SPE Latin
America/Caribbean Petroleum Engineering Conference, Argentina, SPE 26967.
Fleshman, R and Lekic, HO, (1999) “Artificial Lift for High Volume Production”, Oilfield Review, Vol. 11, Issue 1, March 1999.
Ghareeb, MM, Shedid SA and Ibrahim, M, (2007) “Simulation Investigations for Enhanced Performance of Beam Pumping System for Deep High-
Volume Wells”, Society of Petroleum Engineers, Paper Number 108284-MS.
Heinze, LR, Herald, W and Lea, JF, (1995), “Decision Tree for Selection of Artificial Lift Method”, SPE Production Operations Symposium,
Oklahoma, SPE 26510.
Hirschfeldt, M and Ruiz, R, (2009), "Selection Criteria for Artificial Lift System Based on the Mechanical Limits: case Study of Golfo Jorge Basin",
Society of Petroleum Engineers, Paper number: SPE 124737.
Hirchfeldt, M, Martinez, P and Distel, F, (2007), "Artificila Lift Systems Overview and Evolution in a Mature basin: Case Study of Golfo San Jorge",
Society of Petroleum Engineers, Paper number: SPE 108054.
Jiang W, Zhong X, Chen K and Zhang S (2007), “TOPSIS Model as a Data Classifier”, Fourth International Conference on Fuzzy Systems and
Knowledge Discovery, Hainan, China.
Khasanov, MM, Khabibullin, RA, Pashali AA and Semenov AA, (2010), “Approach to Selection of the Optimal Lift Technique in the Vankor Field”,
SPE Annual Technical Conference and Exhibition, Florence, Italy, September 2010, SPE 134774.
Lamki, A, Hajri, SA, Salmi, MA and Riyami, MA, (2007) “Artificial-Lift Selection & Design for Down-Hole Sampling and Heavy Oil Multi-Zone
Testing,” Middle East Artificial Lift Forum, February 2007, Muscat.
Lea, JF and Nickens, HV, (1999), “Selection of Artificial Lift”, Society of Petroleum Engineers, SPE 52157.
Lea, JF, Nickens, HV and Wells, MR, (2008), “Gas Well Deliquification”, Elsevier Inc., Oxford, UK.
Neely, B, Gipson, F, Capps, B, Clegg, J and Wilson, P, (1981), “Selection of Artificial Lift method”, SPE Annual Technical Conference and
Exhibition, Dallas, Texas, SPE 10337.
Nelik, L and Brennan, J, (2005) “Progressing Cavity Pumps, Downhole Pumps and Mud Motors”, Gulf Publishing Company, Houston, TX.
Oilfield Glossary, http://www.glossary.oilfield.slb.com/ accessed on September 30, 2010.
Samad, A, (2010), “Gas Interference in Sucker Rod Pump”, International Conference on Modelling, Optimization and Computing (ICMOC 2010),
National Institute of Technology Durgapur, India.

247 
 
Table 1: Downhole pump matrix 

Pumps Æ SRP ESP PCP Jet pump Plunger lift Gas lift

Maximum pump
16000 15000 6000 15000 20000 15000
setting depth, ft

Capacity, Q
10-500 70-1000 4-600 6-10 < 100 Wider flow rates
(bbl/day)
NPSHr, ft 75-100 150-2000 60-100 3ft maximum - -

High viscosity, Cp Good Bad Upto 100 000 Good - -


Sand/ solid particle Limited Bad Good Good Bad Good
Elastomeric stator:
Temperature limited. Good at high
High temperature Good Bad Good Good
Metallic stator: Higher temperature
temperature
Speed, N Some travels/
~12 3700 10-500 - -
(RPM or SPM) day
Cost High High Moderate Cheaper Cheaper Relatively inefficient
Dry running Bad Bad Bad Bad - -

Gas/ oil well Any well Any well Oil well/ heavy crude oil - Good Gas/Well well
Gas lock, rotor wear,
Gas lock, Rod Cavitation, Cavitation, Plug
Failure method elastomer swelling/rod - Leakage, valve stuck
failure Seal failure off
failure
Reservoir Pressure,
Good Good Good Good - Good
500 1500psi
150-500 psi Pump off required Bad Good Good - Good
Plunger lift in some
<150psi Good Bad Good Good -
cases
Priming Needed Needed Self priming Self priming - -
Bad/Pump off
High GOR Bad Bad Good Good Good
needed
Small or no
Well deviation Large deviation Good Good Bad Good
deviation

248 
 
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Transfer of experience for improved oil well drilling


Pal Skalle, Agnar Aamodt and Odd Erik
Norwegian University of Science and technology, NTNU, Norway

ABSTRACT

The drilling process is getting increasingly more complex as oil fields mature and technology evolves. At the same time, the amount of information is
increasing in volume and frequency. Although technology is advancing, failures do occur, leading to loss of valuable time. Whenever the process is
running smoothly or is failing, valuable experience is gained. However, experience attrition is a well-documented problem. To take advantage of
established and continually growing new experience a formalized methodology, case-based reasoning, was applied in a decision support system for
capture and reuse of drilling operators’ experience. A case describes an episode during drilling, its circumstances and method to repair the problem. A
general domain knowledge model, an ontology, supports the case-based process. In this paper we focus on how experience can be captured and
transferred from different information sources and then stored in cases. It was demonstrated how the system was able to recommend how problems
can be solved when they arise, while at the same time bridging the gap between new and experienced personnel. The quality of selecting the most
relevant experience was proven through performance testing of 62 field cases. The system was used to identify root causes of problems that had
happened as well as to suggest repair actions.

KEYWORDS: Oil well drilling, experience transfer, ontology, drilling failure, downtime, case-based reasoning, symptoms

INTRODUCTION
The motivation behind the work presented here is to advance computerized methods for helping the petroleum industry in reducing unwanted
downtime. We have studied a novel combination of two technologies that have proved to work well in other industries: Case-based reasoning
(Cheetham and Watson, 2005) and Ontologies (e.g. (Obrst, Liu and Wray, 2003), and have recently been applied to the oil&gas domain (Kravis and
Irrgang, 2005).

Case-based reasoning (CBR) is a method and a technology for solving new problems in complex processes by comparing them to previous situations,
and reusing the experience from the most similar situations in solving the new problems (Aamodt and Plaza, 1994). CBR can be described as adoption
of common human problem-solving behavior for computer use (Popa et al., 2008). In practical systems, the technology incorporates different types of
information, which empowers the system to learn and adapt from an ever-growing case base of new experiences. The ultimate goal of our research is
to improve process quality and efficiency by systematically capturing useful human experience during a drilling process, and make relevant past
experience available on-line when needed.

In CBR, a specific situation that is occurring or has occurred before is referred to as a case. The parameter data and other information that constitutes a
case need to be represented in an appropriate case structure. The set of cases contained in a system is called its case base. In some CBR systems all
the knowledge possessed by a system is contained in the case base. In the method presented here additional knowledge in the form of general domain
relationships between significant oil drilling concepts constitutes an additional type of knowledge in the system. This part of the knowledge base may
be viewed as an ontology, and principles of ‘ontology engineering’ (Staab and Studer, 2009) have been adopted, through which method the knowledge
has been structured and interrelated in a logical manner. The knowledge gap created by “The Great Crew Change” that exists in most companies has
been well documented and discussed (e.g. (Mc Cormack, 2010)). The problem is not one of just filling the gaps. There are sufficient numbers of
people entering the workforce to do that. The problem is one of “experience attrition”. The immediate challenge is therefore how to transmitting the
soft and hard skills necessary to quickly bridge the gaps between new and existing personnel. Productivity is an ongoing training concern. Today,
training is moving closer to sites of operations – a trend that will only increase as the number of new entrants to the industry increases. At the end of
the day, companies want a measurable return on investment. They want to achieve a reduction in accidents, an improvement in oil and gas
measurement yield, and fewer lost days of production. That means skills that are transportable between companies and among different industries.
CBR is a methodology that enable computer systems to assist in achieving all these tasks.
The goal of the work described in this paper is to investigate how knowledge transfer in drilling can effectively be realized by combining the re-use of
situation-specific experiences (cases) with justifications and explanations generated by a general domain model (ontology). The combined approach is
referred to a “knowledge-intensive CBR”. We have developed an experimental system that is able to read data from a drilling process and capture
interesting parts of it. After having captured and stored human experience accompanied by technical information, we demonstrate different ways of re-
using the experience, and point out especially two applications in larger detail:

1. Determine the root cause of a problem


2. Determine the optimal repair of a problem

The next chapter discusses some approaches to experience transfer, in order to set our method in a wider perspective. This is followed in chapter 3 by
a description of how the case knowledge and general knowledge are represented, while chapter 4 describes the various sources of information and
knowledge. The quality of the CBR systems, in terms of finding relevant past cases, has been tested on a small scale, and the results are presented in
chapter 5. In chapter 6 the above two application foci (identifying root cause and suggesting repair action) are outlined, including the way the case-
base and the ontology are combined to achieve the results. A discussion of the results and future plans concludes the paper.

EXPERIENCE TRANSFER

The transfer of knowledge in the CBR paradigm involves three main processes: observing and interpreting a situation, identifying and capturing
interesting information in the situation, and searching for and re-using it as knowledge in new situations. Episodic experience from drilling operations
is either stored in writing, e.g. as different types of drilling reports, or as human experience in the heads of the people involved in the operations.
Experiences typically contain the drilling operators’ understanding of the process; how they handled the situation; their ability to point at root causes,
etc. In depth studies to identify and share successful drilling practices across companies (Brett et al., 1998) resulted in benefits on the order of 10 %.
Thorsvoll and Grotmol, (1999) reported a similar approach; joint business development between operator and service companies added value to both
parties after a systematic appoarach to improve quality and communication (experience transfer). Integrated operations represent a resent method of
experience transfer (Milter et.al. 2005). Real time data transferred from offshore to land enables support of drilling operations in an efficient manner.
Support from onshore results in a much better utilization of engineering resources and experience.

Few safety-related incidents over the past twenty years have been noticed by so many people worldwide as the loss of the Space Shuttle Challenger in
1986, the crash of Air France's Concorde in 2000, and the loss of the Space Shuttle Columbia, 2003 (Gordon, 2006). Any piece of information,
observation, referred to as “deviance of normalization” was used to find the root cause. A particular challenge in computerized experience transfer
systems is the knowledge representation issue, i.e. how to represent the symbol structures that stand for human experience in formal data structures in
a computer. CBR applications are still at its beginning in the oil industry, although CBR itself is a well-established technology. Previous work within
the oil industry related to how experience is organized and represented to make it fit for reuse can be divided into 1) situation-specific approaches, and
2) generalized approaches.In situation-specific approaches case-based reasoning is used, as already mentioned. Experience is represented as a
collection of parameters and other information that describes an episode, i.e. an interesting situation. Typically, cases are clustered into classes with
the same solution. The CBR task is to match new cases with historical cases by retrieving them from the correct class. Previous work include Mendes
et al (2003), Abdollahi et al (2008), Popa et al (2008), in addition to earlier work by the authors.

SCIRO documented one of the first applications of CBR in the petroleum engineering domain (Karvis and Irrgang, 2005). Alternate drilling plans
were derived from comparison with previously drilled wells. Each well represented one case. Bhushan at al. (2002) applied CBR to globally search for
reservoir analogues as an important step in planning and development of oil fields. Such information would be useful when appraisal information is
limited. The key lies in characterizing each reservoir by a set of attributes which describe the reservoir and can be used to differentiate it from other
reservoirs. Khajotia and colleagues (Khajotia et. al., 2007) took a non-typical approach in applying CBR within a predictive mathematical model. The
method was designed to mimic the approach to the problem taken by experienced field personnel, by taking knowledge of corrosion rates from
existing cases and apply them in new fields having somewhat similar parameters.In generalized approaches to experience transfer, rule-based
reasoning is the typical method. A rule, whether a final decision rule or an intermediate inference step, can be viewed as a generalization over a set of
cases that share many of the same properties. whether general or specific. Rules are empirical and may be based on sound, if not well understood,
physical, economic, social, or other principles (Brown et. al. 2005). They allow us to shortcut some thinking processes and, in doing so, can also cause
us to make costly mistakes. As we grow more and more experienced, we personally develop and adopt new rules of thumb. But rules often have
exceptions. In a rule-based setting, cases can be viewed as exceptions to the rules, opening up for combined systems. In model-based reasoning
knowledge is represented as multi-relational dependencies between parameters and other concepts, as opposed to the single relation (i.e. the “if-then”
implication) of rule-based methods. In the knowledge-intensive CBR approach taken in our research, model-based reasoning – based on an ontology
model – constitutes the generalization-based method.

EXPERIENCE AND GENERALIZED KNOWLEDGE

In our approach to knowledge-intensive CBR the case base and the general domain knowledge play together in helping the user identify possible
problem of an ongoing drilling operation. The structure and contents of these knowledge types are as follows.

250
Case structure and experience content

The problem type we study is restriction in the wellbore. It is experienced during tripping operations and may escalate to failures. Persistent
reoccurrence of such problems calls for reaming activity to hinder the restriction to escalate into stuck-pipe and lost circulation. Since wellbore
restrictions represent a complex problem it was selected as our first test application.

We learned in the previous chapter that there is a large variety of ways to organize and transfer experience; through mathematical models, through
well plans etc. Our approach has been to build structured cases in order to describe the problem in a purposeful manner. A case contains several types
of experience, and it is structured to help facilitate the knowledge and the experience that stems from users. Three common requirements dictate case
structure:

1. The need of case indexes for retrieval and similarity assessment :


All information embedded in the cases before triggering of the case (BToC ) characterizes the episodes, and is used to discriminate
between cases (similarity assessment)
2. The selected repair method:
The actions taken after the case was triggered (AToC) encompasses repair
3. The result after repair:
Outcome of selected repair, and gained experience. Can repair method be recommended?

Figure 1 illustrates the structure of a case, divided into three sections as described above.

Case retrieval info BToC is composed of items shown in Figure 1, upper section. The data contains indicators and statistics of the real time drilling
parameters, including interpreted events and type of activity on the rig. These are needed as symptoms of the state of the drilling process. Finally it
contains description of the problem and how the case was triggered. The middle case section contains direct experience or case re-use information
AToC. It describes how the personnel involved handled the episode and what was learned. It contains statistics on time spent on well cleaning; repair
actions outcome of the repair; groups of classes (classes are used to evaluate matching ability); lesson learned; Outcome is a measure of what effect
the recommended repair action had. This is useful information the next time the drill string and its Bottom Hole Assembly is passing by the
problematic part of the well where the case episode took place. For example: After completing the specific well section, a steel casing is routinely
installed and cemented in place. Loss of time during casing installment and cementing operations caused by wellbore cleaning related problems is a
poor outcome. In this case repair was not optimal one could state. A very useful property of a CBR system is its ability to capture and store a problem
just solved as a new case. However, an initial case base is typically constructed by manual means, utilizing principles of knowledge engineering.
Figure 2 exemplifies the process of filling the case structure with contents through the identification of relevant info that describes an episode.

Ontology structure and content

The purpose of the ontology is to serve as a knowledge model for model-based reasoning to assist the CBR process. The ontology can be viewed as a
semantic network, where each node in the network corresponds to a concept in the knowledge model, and each link corresponds to a relation between
concepts. A concept may be a general definitional or prototypical concept, and may describe knowledge of domain objects as well as problem solving
methods and strategies. A network view to concept definition is taken, in which each concept is defined by its relations to other concepts. Figure 3
illustrates the three main types of knowledge in the model: a top-level ontology of generic, domain-independent concepts, middle-level ontology of the
domain-specific knowledge, and a bottom-level that constitutes instances, i.e. cases. Data from the real world environment enters into cases – possibly
after some transformation or abstraction. Cases are linked into the ontology model by the fact that each parameter that describes a case is represented
as a concept in the ontology.

Ontology is a term used in philosophy, encompassing the study of what is. The application of Ontology within Information Technology and
Engineering is more resent, and has replaced and enhanced terms like knowledge model, data model, term-catalogue etc. All ontologies make some
assumptions about the world that it represents. In our ontology, the top-level concept Thing stands for anything in the world worth naming or
characterizing. Everything we want to talk about is a subclass or instance of Thing. Thing has three subclasses; Entities (objects in the real world),
Descriptive Things (description of Entities) and Relations (bi-directional relations between subclasses and instances). Figure 4 illustrates a part of the
ontology. Only class-subclass relations are shown in the figure. The leftmost concept, State, is a subclass of Entity. Over the last years an extensive
ontology of oil and gas terms has been developed, and also made into a standard, the ISO 15926 (Batres et. Al, 2007). Our model sets out to define
drilling engineering terms in accordance with the international standard ISO 15926 whenever practically feasible. Unfortunately that turned out to be
quite difficult, with the result that our ontology differs substantially from that standard. In our ontology’s middle layer all petroleum related
information used in the reasoning process is formally defined entities, e.g. the activity of pulling the drill string up the well is given the entity name
Tripping Out.

To enable reasoning, symbols are connected through relations that facilitate default inference as well as different types of property inheritance.
Examples of relationships are;

Swelling may lead to wellbore enlargement after long exposure time


Enlarged wellbore sometimes cases tight spots in horizontal wellbores

Such relationships will assist in pointing out the root cause and / or in explaining the episodes. More on that in chapter 6.

251
SOURCES OF EXPERIENCE

A case is the basis for reasoning and all relevant information must be gathered. Figure 5 indicates two basic sources of experience:

On-line interpretation of drilling data. We have developed a system for on-line interpretation of drilling data (Verdande, 2009). Surface and
downhole data are logged typically every 5th second. These data are continuously evaluated and interpreted. Anomalic behavior is transferred into
either Interpreted Events, Inferred Parameters or Indicators, data types that are important for indexing and similarity matching of cases. Events show
up through signs such as tight spot, packing-off, hard stringers etc. Indicators are mathematical functions for estimating, e.g. formation chemical
stability. Inferred parameters are logical groups of parameters, e.g. TVD/MD, MW-Pore Pressure. Interpreted drilling data are necessary in order to
improve the understanding of downhole processes and enable us to find the causes of hole restrictions.

Experience stored in documents. An expert performs an initial investigation of drilling operation logs and documents and retrieves information
which may characterize a case such as lithology, history of the operation both in a time and depth view, etc. Drilling engineers rarely have time to
analyze the entire dataset before making recommendations. And since access to oil industry experts for the purpose of knowledge acquisition is
limited, knowledge is being extracted from discrete information sources in client databases.

IDENTIFYING THE CORRECT SOLUTION CLASS - TEST RESULT

A test of the implemented system was undertaken in order to evaluate its ability to select cases from the right solution class. A standard testing
algorithm (cross-validation) was used. Seven different wellbore sections drilled in an offshore oil field during 2004 – 2006 were
studied, resulting in 62 cases of wellbore cleaning episodes. Three groups of solution classes were defined as potential target classes:

Root Cause
Repair Time
Consequence

Root Cause represents the ultimate test. However, we had insufficient knowledge (real time data interpretation) of the downhole processes to be able
to reveal downhole root cause with high enough probability. Of the two remaining classes, Repair Time represents a defined and measurable quantity
and was therefore selected. Consequence refer to the additional time it took to repair problems encountered during subsequent casing running and
cementing operations in the same well. Repair Time has four sub-classes:

Class 1: Insignificant Repair Time < 1 hour


Class 2: Short Repair Time 1-3 hour
Class 3: Long Repair Time 3-15 hour
Class 4: Gave Up Well > 15 hour

The test was performed in a stepwise fashion; first, all 62 cases were tested against manually analyzed and tagged events, then against automatically
generated triggering events, and finally against completely new logs. Manual tests implied that cases were defined and tagged (marked) manually in
the real-time log. Testing of the 62 cases against manually tagged triggering events resulted in acceptable if not excellent test results. However, many
of the 62 triggering events were not “seen” by the automatically performed test. We introduced several modifications in preparation of the next test
round:

• We allowed only cases which could be automatically detected through on-line interpretation
• Due to low number of cases, they were grouped into only two classes,; Class A contained Insignificant and Short Repair Time, Class B
Long Repair Time and Gave Up Well.
• Cases which had repair time shorter than 0.1 h (6 minutes) were excluded

These requirements reduced the number of cases to 22. The matching-test gave as result that class A cases were retrieved correctly in 82 % and class B
in 73 % of the time. The baseline percentage was approximately 50 % for each class. All together the test results strongly indicates the feasibility of
the method. Access to a sufficient amount of data of sufficiently good quality and sampling rate has been problematic. More data will be needed in
order to perform a more elaborate evaluation.

TWO APPLICATIONS OF THE GAINED EXPERIENCE

As previously mentioned, wellbore restriction is our initial target problem. The problem is experienced during tripping operations and may escalate to
failures if the wellbore is not properly cleaned. The causes of wellbore restriction failures can be grouped into three classes related to either equipment,
wellbore or wellbore wall. Examples of failures within each of the three main classes are; Drill pipe twist off, Poor Hole Cleaning (cuttings, cavings)
and Swelling Wellbore. Insufficient smoothness of the wellbore path can become evident in the form of pack-offs, took weight, high torque or high
drag. Reoccurrence of such problems calls for reaming activity to hinder the restriction to escalate into stuck-pipe and lost circulation.

Determine the root cause of a failure


The objective is to determine the root cause starting out from three types of parameters: Direct observations – i.e. measurements, inferred parameters –
i.e. values derived from observations, and interpreted events – i.e. particular concepts describing important states which require particular awareness

252
or action. The parameters and causes are related in the ontology through intermediate concepts, such as intermediate causes or other states. Figure 6
illustrates the anomalous state of the wellbore and the formation. This state may return to normal by itself, and sometimes turn into a failure state as
previously illustrated in Figure 4. In this paper we have specified root causes through the name of the failure. By specifying the subclasses of a failure
as exemplified through Poor Hole Cleaning, the root cause is determined through pinpointing the failure to e.g. ‘Cleaning Of Enlarged Hole’ – the root
cause is Hole Enlargement.

The task of the system is to determine which root causes or intermediate states are entailed or likely, given the parameters. Only some paths provide
support for such a conclusion. In Figure 7 we have exemplified a case with the following reported observations (the observations are circled in red)

High Mud Flow Rate (here MFI Erosion)


Increasing Formation Exposure (here Increasing-I-Fm-Chem)
Pack Off
Took Weight
Low ROP
Narrow Event Distribution (along the axial length)
Increasing Drag
Low Wellbore Inclination

There is a path from all the observations either directly or indirectly to two different root causes; Enlarged Hole and Cuttings Accumulation. The
strength of the paths is the product of the strength of each relation leading from the observation to the target entity:

where n is the number of serial relations. We observe that there are several explanatory paths from an observation to each target entity (the root cause
entity).
The total explanation strength for each target entity is determined with the equation below.

where m is the number of paths. The strength of a relation and which relation to apply between entities was decided by drilling experts, including
relations like ‘indicates’ (with a strength of 0.4), ‘leads to’ (0.6) ‘causes’ (0.8). Calculated explanation strength will be a good indicator for identifying
the possible root cause. In this example the explanation strength of Enlarged Hole was 0.088 while only 0.025 for Cuttings Accumulation.

Determine optimal repair activity through evaluation of outcome

A matching case can be used to inform the user about;

Time needed to clean (the most) similar wells


How was it done? Can it be performed alternatively?
Potential problems during later casing operations?

Correct diagnosis of the problem will lead to treatment that is appropriate and will result in efficient repair actions, leading to significant cost
reductions. The assumption is that treatment of historical cases of class A also is suitable for new cases of this class. We claim this is a fair assumption
since 82 % of the cases in this class were similar enough to be selected. However, since the downhole cause of the problem is normally not 100 %
known, the repair action is not necessarily directly related to the cause. It is for this reason difficult to improve the repair efficiency. The operator will
naturally select repair methods which are commonly in use to treat hole-cleaning problems, but again, repair methods which are not necessarily well
suited for that specific type of root cause since stated cause may be wrong.

The solution and / or the repair applied in historical cases must be checked if it corresponds with recommended actions for the specific situation. After
an incident has taken place, the user tries out a repair strategy. When the outcome of the selected repair action becomes known, the repair strategy can
be evaluated. Did it work? Experts are consulted through interviews or through investigation and evaluation reports.

As it is now, we just recommend the Repair Activity applied in the best matching class. One could claim that Repair Activity in the selected case
represents the accumulated experience in the field, However, the experience could be wrong, e.g. based on wrong interpretation of downhole behavior.
If outcome-evaluation is not considered, the system will become static.

CONCLUSION

An on-line software tool has been successfully developed. It has the capability of recognizing episodes and relating them to situations that have
occurred before. The episodes are stored as cases consisting of three separate parts; circumstantial information and gained experience, explanation of
why the situation arose and how it was handled and outcome of repair. Experience embedded in a case is retrieved from data streams, and from
documents, but also from the user and other experts during case generation and ontology building; during evaluation, etc. Cases representing complex
domains like oil well drilling operations need to be related to a rich ontology since the case space is large. Experience is translated into a symbolic
language and then stored in the ontology as interrelated entities.

253
Initial real-time runs have proven the tool’s functionality and pointed out potential user support. Tests have showed that the CBR system matched and
retrieved cases of the correct class to an acceptable degree, even with the sparse data available for the experiments. We also claim that cases occurring
in wellbores that exhibit escalating restrictions have built-in prediction capability. These wells are typically deep or long wells associated with long
exposure time. Improvements are pending; new Indicators; new Events; higher quality of existing Events. All improvements will allow similar
episodes to become more similar. They will open for formal and reliable prediction of root cause. Up to now, good engineering judgment dictates root
cause. The correct root cause will improve quality of advises. Research is continuing on refining the system in response to industry feedback. We have
exemplified the utilization of a CBR system directed towards reduction of downtime. It is at this stage merely a claimed reduction since the system has
yet to be field-tested. When run in the field, the user will start testing recommendation from the system in a real setting. After evaluating the outcome
of recommended repair actions and eventually adjusting the cases, the user should gradually be witnessing improved outcome. In fact, after the results
reported here, Verdande Technology has developed and extensively tested a commercial and customer-targeted CBR system that utilizes many of the
design principles of the reported system. That system has shown excellent accuracy and performance properties in real drilling settings
(www.verdandetechnology.com).

REFERENCES

Aamodt, A. and Plaza, E. (1994) ‘Case-based reasoning: Foundation issuers, methodological variations and system approaches’, Artificial Intelligence
Communications, Vol. 7, No 1, pp. 39-52.
Abdollahi, J., Carlsen, I.M., Randhol, P., Tenold, E., Haga, H.B. and Jakobsen, T. (2008) ‘A Case-Based Approach to Understand the Complexity of
Causal Connections Related to Well Integrity Problems’, presented at the 2008 IADC/SPE Drilling Conference, Orlando, 4–6 March.
Brett, J., Ford, G.M.K. and Gahan, B.C. (1998) ‘Successful Drilling Practices - An Experiment in Cooperation’, IADC/SPE Drilling Conference, 3-6
March, Dallas.
Bhushan, V. and Hopkinson, S.C. (2002) ‘A Novel Approach to Identify Reservoir Analogues’, European Petroleum Conference, Aberdeen, 29-31
October.
Brown, L.P., Brown, L. and Hawkes, R.V. (2005) ‘Rules of Thumb in Well Testing: What Works and Doesn't Work-and Why’ Journal of Canadian
Petroleum Technology, Vol. 44, No. 5.
Cheetham, W. and Watson, I. (2005) Fielded applications of case-based reasoning. The Knowledge Engineering Review , Cambridge University Press,
Vol. 20, pp 321-323.
Datta, B.K., Ratnayake, C., Saasen, A. and Omland, T.H. (2005) ’Hole Cleaning and Pressure-Loss Prediction From a Bulk Transport Perspective’,
Paper SPE 96315 presented at Offshore Europe, 6-9 September, Aberdeen.
Gordon, H.J. (2006) ‘Integrating Learning into Safety’, ASSE Professional Development Conference and Exposition, Seattle, June 11 – 14.
Khajotia, B., Sormaz, D. and Nesic, D. (2007) ‘Case-based Reasoning Model of CO2 Corrosion Based on Field Data’, Corrosion 2007, NACE
International, Nashville, March 11 – 15.
Kravis, S and Irrgang, R. (2005) ‘A case bsed sy7stem for oil and gas well design with resk assessment’, Applied Intelligence, Vol. 23, pp 39-53.
Obrst, L., Liu, H. and Wray, R. (2003) ‘Ontologies for Corporate Web Applications’ AI Magazine, Vol. 24 No. 3.
Mendes, J.R. P., Morooka, C. K. and Guilherme, I. R. (2003) ‘Case-based reasoning in offshore well design’, Journal of Petroleum Science and
Engineering, Vol. 40, pp 47– 60.
McCormack, G. (2010) ‘An Industry upturn with a downside: Talent needs still loom large’, JPT, May, 44 – 46.
Milter, J., Bergjord, O.G. Høyland, K. and Rugland, B. (2006) ‘Use of Real-Time Data at the Statfjord Field Anno 2005’, Intelligent Energy
Conference and Exhibition, Amsterdam, 11-13 April.
Popa, A, et al. (2008 ) ‘Case-based reasoning approach for well failure diagnostics and planning’, SPE paper 114 229, presented at the 2008 SPE
Western Regiounal and Pacific Section AAPG Joint Mtg, Bakersfield, March 31 - April 2.
Staab, S. and Studer, R. (2009) ‘Handbook on Ontologies’, 2nd. ed. Springer, 811 pgs.
Valipour, S.S. and Skalle, P. (2009) ‘Enhancing Decision Making in Critical Drilling Operations, SPE paper 12290, presented at SPE Middle East Oil
& Gas Show and Conference, Bahrain, March 15–18.
Torsvoll, A. and Grotmol, Ø. (1999) ‘Optimizing Contract Management and Cooperation Between Operator and Service Company’, Journal of
Petroleum Technology, Vol. 51, No. 2.
Valipour, S.S., Skalle, P. (2009) ‘Enhancing Decision Making in Critical Drilling Operations, SPE paper 12290, presented at SPE Middle East Oil &
Gas Show and Conference, Bahrain, March 15–18.
Valipour Shokouhi, S.S., Aamodt, A., Skalle, P. and Sørmo, F. (2009) ‘Determining root causes of drilling problems by combining cases and general
knowledge’, Proceedings of the 8th International Conference on Case-Based Reasoning, ICCBR 2009, Seattle, Lecture Notes in Artificial
Intelligence 5650, Springer Verlag, ISSN-0302-9743, pp. 509-523.
Verdande Technology (2009), ' Harvesting knowledge to improve drilling performance in real time; A perspective on lowering NPT in difficult and
critical wells’. White paper. (http://www.verdandetechnology.com/images/stories/verdande/publications/vt_whitepaper.pdf).

254
Fig 1. Selected format of case structure

Fig 2. Evaluation and initiation of a case building process, here on basis of Daily Drilling Report.

255
thing
generic concepts

general
domain concepts

cases
case case case
039 76 112

data

Figure 3. Example of Ontology. It consists of general knowledge in the upper levels and of specific knowledge in the lower levels. Most specific
knowledge are stored in cases as defined by the process data.

Figure 4. Failure State and its sub classes are indicated by an arrow. Symptoms activated during an episode are point to a specific Failure Subclass.

Fig 5. Sources and transfer of experience during case building: Real-time data (left) and Documents (to the right). An expert is needed in the process
of creating cases, either they are created from historical logs/reports or from real time data.

256
Fig 6. Abnormal State. Its sub classes are indicated by arrows.

Fig 7. Failure State. Symptoms activated during the episode, pointing to a specific Failure.

257
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Computer Aided Design of Air/Gas Drilling System for Tight Gas Reservoirs
S. M. Mushhoor1 and R. Sharma2
1 and 2
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
E-mails: 1muhammadmashhoor@gmail.com,2rajivatri@iitm.ac.in

ABSTRACT

In this paper we present, a computer aided design model of air/gas drilling system for tight gas reservoirs. Our aim is to develop a field penetration
model based upon depth of a well, weight on bit, rotary speed and bit type for KG (Krishna-Godavari) basin’s soft and hard formations. We plan to
incorporate the following input parameters: rotary speed; formation type; bit weight; and formation depth. The mathematical equations for the wells
will be generated to study the effect of bit size on the rate of penetration. And, we plan to generate separate models for soft and hard formations.

KEY WORDS: air/gas drilling, rate of penetration, formation depth, CAD model.

INTRODUCTION

Air and gas drilling technology utilizes compressed air or other gases as drilling circulating fluid. The compressed air or other gas (e.g., nitrogen or
natural gas) can be used also or can be injected into the well with incompressible fluids such as fresh water, formation water, formation oil, or drilling
mud. There are some distinct operational applications for this technology: air or gas drilling operations (using only compressed air or other gas as the
circulating fluid), aerated drilling operations (using compressed air or other gas mixed with an incompressible fluid), and stable foam drilling
operations (using compressed air or other gas with an incompressible fluid to create a continuous foam circulating fluid). Some of the advantages of
this method are:
• Improved penetration rate.
• Formation damage avoidance.
• Loss of circulation avoidance.
• Borehole cleaning with low fluid volume.
• No differential sticking.

The advantages mentioned above gives economic benefit also; increased penetration rate cuts down the drilling time and avoidance of loss of
circulation utilizes less amount of drilling fluid both in turn cuts down the cost. There are some disadvantages for this method like inability to control
high pore pressure and inflow of formation water; and combustion due to oxygen content. In general, the use of air and gas drilling methods is confined
to mature sedimentary basins within mature geologic provinces. When the basin-forming geologic time is long, these basin sediments become well
cemented (usually with CaCO3). These basins form competent rock formations at depth and are referred to as mature sedimentary basins. These basins
usually have dispersed hydrostatic columns of retained ancient sea waters or fresh lake waters and often have no retained water in their formations at
depth. In young or immature sedimentary basins, the cementing processes have not yet fully cured. Such immature rock structures cannot fully support
themselves in the buoyant environment of retained ancient sea and fresh water. In general, under-balanced drilling and completions operations are
basically restricted to mature sedimentary basins. Air and gas drilling is a broad term that not stick just use of compressed air as the circulating drilling
fluid for rotary drilling operations. But the majority of these drilling operations use compressed air as the circulating drilling fluid. In some oil and
natural gas recovery drilling operations it is necessary to drill with a gas that will not support down-hole combustion. This is particularly the case when
drilling in or near the hydrocarbon producing formations. The technology for air and gas drilling accounts for approximately 30% of the world’s land
oil and gas drilling operations. Although modern air and gas drilling technology began in the United States in the 1930s, it is presently being used in
drilling operations throughout the world’s many oil and gas producing geologic provinces. The basic layout of the air/gas drilling is shown in Fig. 1.

In the past, air and gas drilling methods have been a small segment of the petroleum recovery industry. Currently, air and gas drilling methods has a
comparable role in industry. This in turn gives various types of advantageous additions to the research area. Basically industry subdivides the whole
operation of air/gas drilling Dry air (dust), Gas, Mist Foam, and Gasified fluids (aerated mud). The lighter the fluid in the annulus lesser the confining
pressure this in turn gives the advantages for this method over conventional drilling fluid method. The advantages of different fluids that are used in
this method is compared in Fig. 2.The fluid densities of different chemicals used in this industry air/gas drilling are summarized in the Fig. 3.

Fig 1: The basic layout of the air/gas drilling.

Fig 2: A comparison on advantages of different fluids that are used in air/gas drilling, with conventional mud
(Direction of arrow indicates Increasing order).

It is in this environment that the master’s thesis of the first author of the present paper has been planned. The present reported work is part of his
planned thesis, and in our work (including master’s thesis), broadly we have the following objectives:
• Our aim is to develop a field penetration model based upon depth of a well, weight on bit, rotary speed and bit type for KG (Krishna-Godavari)
basin’s soft and hard formations.
• We plan to incorporate the following input parameters: rotary speed; formation type; bit weight; and formation depth.
• The mathematical equations for the wells will be generated to study the effect of bit size on the rate of penetration.
• We plan to generate separate models for soft and hard formations.

However, our larger aim is too broad to be addressed in one paper. Therefore, in this paper we restrict ourselves to the conceptual development of a
computer model for computer aided design model of air/gas drilling system for tight gas reservoirs. This paper will discuss the considerations in

260
developing such programs with particular reference to design of air/gas drilling system for KG basin. Also, we shall show typical output from such a
program. The remaining of this paper is organized as follows: Section 2 describes the brief review of literature; Section 3 presents the basic CAD
model; Section 4 describes the computer model briefly, and Section 5 concludes the paper with identifications of future areas of research and
application.

Fig 3: The fluid densities and chemicals that are used in air/gas drilling.

BRIEF REVIEW OF LITERATURE

Adewumi and Tian (1992) presented a study on a model that is demonstrably capable of predicting the pressure-drop profile in the annulus and the
optimal lifting velocity, an essential ingredient in the optimal design of air drilling. To effectively utilize air drilling, one must be able to understand
and predict such design parameters as optimal transport velocity, power requirements, choking velocity, etc. The authors identified variables which
characterize the transport of solid particles in the well-bore (independent variables: physical properties of the drill cuttings, including size and density;
thermo physical properties of the transporting fluid such as density and viscosity; penetration rate; geometrical configuration of the well-bore, fluid-
transport velocity) and design parameters (i.e. optimal transport velocity, and associated pressure loss). This problem is basically that of gas-solid flow
through concentric circular pipes with vertical orientation. A generalized two-dimensional multiphase hydrodynamic model has been formulated based
on the following assumptions: (1) solid particles are of uniform size and density; (2) system is isothermal and there is no inter phase mass transfer; (3)
particulate phase is incompressible while the gas-phase behaviour is describable by the ideal gas equation of state; (4) pressure gradient is assumed to
exist in the gas phase only; and (5) the radial and tangential components of the velocity vector are assumed to be negligible. The authors used ‘method
of lines’ to solve the set of partial differential equations. The model presented in their work is used to investigate the effect of all the parameters that
have been mentioned on pressure drop and the optimal air velocity. There exists an optimal gas velocity which would yield the minimum pressure loss
during the transportation of drill cuttings from the bottom of the hole to the top. Gas flow rate, rate of penetration, and solid density significantly affect
the pressure loss in solid cuttings removal during air drilling.

Bizanti (1989) presented a field penetration model based upon depth of a well, weight on bit, rotary speed and bit type for Louisiana's soft and hard
formations. The effect of depth, weight on bit, rotational speed, and bit characteristics on the rate of penetration is investigated. He suggests that an
accurate drilling rate model which will increase the penetration rate and reduce the number of trips and thus reducing in rotary-drilling operations cost
could lead to possible economic recovery of smaller reservoirs. The primary tasks of this paper are: (1) to investigate the important factors that
influence penetration rate. (2) To generate and develop an equation or group of equations which will predict with reasonable accuracy the drilling rates
in the state of Louisiana. The author pointed out the following factors that are incorporated in the drilling rate model (input parameters): rotary speed;
formation type; bit weight; and formation depth. The empirical equations generated from wells drilled during the period 1955- 1983 in the state of
Louisiana. These functions indicate that as the bit size decreased (as depth increased), the rate of penetration tends to decrease. The empirical relations
generated strongly dependent on rotary speed in shallow formation, and less dependent in deeper formations. Separate models are generated for soft
and hard formations. The author has suggested drilling rate model for soft and hard formations. In his work, regression analysis is used for this purpose.
The model reached in following conclusions: (1) in the predicted drilling rate model, it was concluded that as bit weight increases the coefficient of
rotary speed exponent decreases. (2) In the predicted drilling rate model, it was concluded that as rotary speed increases the coefficient of bit weight
exponent decreases. (3) For the wells investigated in this study, the rate of penetration was strongly dependent on rotary speed at shallow formations
and less dependent at deeper formations. (4) In soft Louisiana formations, insert bits performed better than steel bits did. (5) Drilling rates were much
higher in soft formations than in hard formations. (6) In soft Louisiana formations, penetration rate was highest for low weight-on-bit and high rotary
speed. (7) In hard Louisiana formations, penetration rate was highest for high weight-on-bit and high rotary speed. (8) Penetration rate appears to be
strongly dependent on depth, weight-on-bit, and rotary speed in soft formations and weakly dependent in hard formations

In Weatherford (2010) the benefits of air/gas drilling have been discussed, especially the increase in ROP and reduce drilling costs through proven
engineering and service delivery excellence. It also gives brief idea about where this type of drilling is being done round the globe, and tells about
down hole and surface air drilling equipments .It summarizes the benefits of air/gas drilling as: 1) Decreased costs - In hard-rock formations, air
drilling provides an ROP that is up to five times faster than drilling with conventional fluids. This increased ROP not only reduces costs but brings

261
wells on line faster. 2) Minimized deviation tendency in faulted formations - Air drilling produces straighter holes with lower bit weights and results in
fewer deviation issues while maintaining a good ROP. 3) Minimized lost circulation - Drilling with air, mist and foam reduces hydrostatic pressures
and helps manage lost circulation without resorting to costly and time-consuming techniques. The article suggests various air drilling equipments as
following: Down hole: Percussion Hammers and percussion hammer bits; Surface: Air-compression equipment (primary and booster), and Rotating
control device (RCD), Two-phase separation. Table 1 summarizes the area at which each fluid system can be used.

Table 1: The applicability of various fluids of air/gas drilling.

Type of Air Drilling (Fluid System)


Application
Hard-rock and dry formations
Dry air (dust) Zero water production
Formation cannot withstand hydrostatic pressure
Mist Hard-rock formation, Small amounts of water influx
Foam Large hole diameters, Medium amount of water influx
Gasified fluids Loss-circulation zones ,Unconsolidated formations

Fig 4: The rate of penetration comparison between conventional drilling and air/gas drilling.

Adewumi and Tian (1993) presented a study in which a viable well-bore hydraulics model is utilized as the basis. Several important parameters that
influence air drilling operations are analyzed, such as drill cutting size and size distribution, hole size changes, attrition, and particle shape. This work
defines several areas that need special investigation and sheds light on several previously unanswered questions. It is important to note that there are
two fundamental differences between well-bore hydraulics associated with air drilling and conventional drilling. The first is the high compressibility of
air compared with drilling mud, and the second is the large density difference between air and drill cuttings. These differences preclude the wide
experience already acquired for conventional drilling from being directly applied to air drilling. The authors identified two basic variables which have
always been emphasized in designing air drilling programs are air volumetric requirement and compression power requirement. In hydrodynamic
terms, these translate to air velocity and pressure loss respectively. According to the authors independent variables affect these basic design parameters
include:
(1) The geometrical configuration and size of the well-bore/drill string annulus.
(2) The geometrical and physical properties of drill cuttings including size and size distribution, density and shape.
(3) Penetration rate as it determines the solid generation rate and hence solid loading.
(4) Particle attrition as it traverses the vertical height of the well-bore.

The authors formulated the model with following assumptions:


(1) Solid particles of uniform size and density constitute a phase thereby enabling us to assume a single set of hydrodynamic parameters (such as solid
velocity distribution) for each particulate phase.
(2) The system is isothermal and there is no inter-phase mass transfer.
(3) The particulate phase is assumed to be incompressible while the gas phase behaviour is described by the ideal gas equation of state.

262
(4) Pressure gradient is assumed to exist in the gas phase only.
(5) The radial and tangential components of the velocity vector are assumed to be negligible. For this reason, pressure is assumed to vary only in the
axial direction but uniform in the radial direction. In their work, the ‘method of lines’ was used in solving the set of partial differential equations.
Numerical method of lines is implemented in FORSIM VI package. This package was used in integrating the set of quasi-linear partial differential
equations formulated by above assumptions. By solving the model authors arrive in following conclusions:
(1) In a multi-sized particulate system, the coarsest size has the dominant effect of the well-bore hydraulics.
(2) The changing hole size, which is a rule rather than an exception in practice, has a very significant effect on air drilling well-bore hydraulics and
hence cannot be ignored in drilling program design. It affects the pressure drop profile in particular.
(3) The phenomenon of particulate boundary-layer was predicted, and it helps to explain important phenomenon of choking.
(4) Particle attrition cannot be ignored when analyzing air drilling well-bore hydraulics. All predictions demonstrate a significant effect on well-bore
hydraulics. Furthermore, the functional form of the attrition kinetics shows significant influence.
(5) Particle angularity was found to aid the lifting capacity of air. It drastically reduces the pressure loss incurred as a result of transporting the cuttings
from the bottom of the hole to the surface.

Fig 5: The structure of standard BHA for most hammer operations for surface and intermediate hole sections.

Akgun (2007) introduced a different approach to assess drilling performance, and to alleviate the problems of learning curve analysis. The new
approach suggests that the drilling rate is compared with a newly introduced parameter, called as drilling rate technical limit. It will be defined as the
maximum achievable drilling rate without risking drilling safety. This method is superior to learning curve analysis because; one, it does not depend
on the previous drilling records, two, it aims to drill a well at the fastest rate possible without jeopardizing the safety of drilling operation. It has a
couple of disadvantages; one, the proposed method can only compare the drilling rates, two, it is laborious. The author identified the rate of drilling
can be improved for a given field until it reaches its technical limit. The drilling rate technical limit (DRTL) can only be achieved by carefully
selecting all critical drilling parameters, which influences DR like Mud weight (MW), weight on bit (WOB), rotary speed or rotation per minute
(RPM), bit type and hydraulic parameters, such as; flow rate (Q) and impact force (Fj) are among the controllable drilling variables. He found some
variables as uncontrollable like Formation properties such as; pore pressure, compaction, in-situ stresses and mineral content. The paper investigates
the maximum WOB, RPM, and minimum MW and recommended Q; and found what will be the drilling rate or DRTL, if drilling parameters are set
as above values. Analyzing each and every parameter the author arrives in following conclusions: MW is selected as the bigger of following two
values; a) minimum MW needed to prevent formation fluid kick, and b) minimum MW needed to prevent borehole collapse. WOB is selected as the
smaller of following two values; a) optimum WOB for minimum cost per footage, and b) minimum critical buckling load of drill string portion that
will be put in compression. RPM is selected as the smallest of following three values; a) maximum allowable RPM for rotational drag, b) maximum
allowable RPM for bit hang up c) optimum RPM for minimum cost per footage. Q is selected as the bigger of following two values; a) minimum
flow rate needed to prevent cuttings bed formation in inclined holes and vertical holes, b) the flow rate need for an effective cleaning beneath the bit
tooth.

BASIC CAD MODEL

In order to make reasonable predictions of the characteristics air and gas drilling operations, (whatever the fluids ; air ,gas ,form etc) it is necessary to
derive a consistent theory that can be used, with certain simplifying limitations, to develop specific equations to model each of the aforementioned
operations. All three basic drilling fluid circulation models, air and gas, aerated, and stable foam, must utilize a combination of mathematical theory
and empirical correlations to develop a complete calculation model for each. There are number of parameters which influence design of air drilling
system. E.g. physical properties of the drill cuttings, thermo physical properties of the transporting fluid, penetration rate, geometrical configuration
of the well-bore, fluid-transport velocity and associated pressure loss. In computer aided design of air /gas drilling system we like to optimize these
factors keeping some of them constant. The optimization gives us an idea which system should be used (air, gas, form, etc.) in a given reservoir

263
conditions. Also we can predict the input parameters (flow rate etc) that should be provided at the surface for the system so as to take out cuttings and
achieve maximum rate of penetration, which in turn cuts down the cost. The basic structure of the CAD model that minimizes flow rate is shown in
Fig. 6.

Fig 6: The basic structure of the CAD model that minimizes flow rate.

In order select different parameters for the design we should have an objective for the use of this method. If we are looking for high rate of
penetration the parameters of interest will be depth of a well, mud weight, weight on bit, rotary speed and bit type. But in case of a well-bore
hydraulics model our interest will be upon cutting size and size distribution, hole size changes, attrition, and particle shape.This data come apart from
the fluid properties in both cases. In design, we plan to explore the flow in the following three types: three-phase flow in the annulus, two-phase flow
through the bit, and two-phase flow in the drill string. In design of model this system in well-bore along with the fluid, we assume Direct circulation
which is defined as the injection of drilling fluid into the inside of the top of the drill string, the flow of the fluid down the inside of the drill string,
through the bit orifices or nozzles, the entraining of the rock cuttings into the drilling fluid at the bottom of the borehole, and then the flow of the
drilling fluid with the entrained cuttings up the annulus between the outside of the drill string and the inside of the borehole. If the trajectory is
opposite it is called reverse circulation by larger diameter shallow and intermediate depth wells are usually drilled.

CONCLUSIONS
In this paper we have presented a basic CAD model of air/gas drilling system for tight gas reservoirs. Our model is primarily a field penetration
model based upon depth of a well, weight on bit, rotary speed and bit type for KG (Krishna-Godavari) basin’s soft and hard formations. Our
presented model incorporates the following input parameters: rotary speed; formation type; bit weight; and formation depth. In our CAD model, the
mathematical equations for the wells are generated to study the effect of bit size on the rate of penetration. And, we have explored the generation of
separate models for soft and hard formations. However, our presented CAD model is only conceptual at the present state. It has not been fully
developed and hence is not operational. We need to develop it fully, and then test it against the available data from soft and hard formations of India
and abroad. Our future work shall go in this direction and currently this is under investigation.

ACKNOWLEDGEMENTS

Some of the data used in the present work has been supplied by our industrial research collaborators (e.g. Weatherford Inc., USA).
The authors gratefully acknowledge their support and encouragement for the present research.

REFERENCES
Adewumi, M. A. and Tian, S, (1992), “Determination of optimal air flow rate in air drilling”, Journal of Petroleum Science and Engineering, Vol.
8, pp 1-11.
Adewumi, M. A. and Tian, S, (1993), “Multiphase hydrodynamic analysis of pneumatic transportation of drill cuttings in air drilling”, Powder
Technology, Vol. 75, pp.133-144.
Akgun, F., (2007), “Drilling Rate at the Technical Limit”, International Journal of Petroleum Science and Technology, Vol 1, No. 1,
pp. 99–118.
Bizanti, M., (1989), “Drilling Rate Model for Louisiana Formations”, Journal of Petroleum Science and Engineering, Vol 2, pp. 261-267.
Lyons, W.C., (2009), “Air and Gas Drilling Manual Applications for Oil and Gas Recovery Wells and Geothermal Fluids Recovery Wells”, Third
Edition 2009, Gulf Professional Publishing, USA.
Lyons, W.C., (2010), “Working Guide to Drilling Equipment and Operations”, First edition 2010, Gulf Professional Publishing
Weatherford, (2010), Article on “Air drilling services”.

264
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Design of Integrated Casing Drilling System for Deepwater Applications


D. VamshiKrishna1 and R. Sharma2
1 and 2
Department of Ocean Engineering,
Indian Institute of Technology Madras, Chennai, India
E-mails: 1vamshikrishna9@gmail.com,2rajivatri@iitm.ac.in

ABSTRACT

Though the concept of drilling-with-casing system is deceptively simple to state, its actual implementation from experimental
procedure to on-field applications remains challenging. Our focus in the proposed thesis will be to design an integrated drilling-with-
casing system for deep water offshore applications. We shall focus end-to-end connected system where standard oilfield casing will be
used to drill the well first and then it will be left in the proper place to case the well. Our integrated process will make it possible to
speed up drilling (e.g. anticipated gains at around 15 – 25 %) by eliminating drill-string tripping and the problems associated with it.
In addition we plan to address the drawbacks of drilling-with-casing system, e.g. difficulty in assembling, stuck casing, and limited
drilling distance because of fatigue life of linear connections.

KEY WORDS: Casing Drilling; Casing Drive System; Bottom Hole Assembly (BHA).

Nomenclature:

CWD: Casing While Drilling


BHA: Bottom Hole Assembly
DLA: Drill Lock Assembly
ROP: Rate of penetration
RSS: Rotary Steerable Systems
CDS: Casing While Drilling
DCwD: Directional Casing While Drilling

INTRODUCTION

Though conventional drilling is performed all around the world and it is in use for many years since oil and gas exploration has
started, it has its limitations. Such as tripping, loss of circulation of mud, time taken for casing and cementing, bore hole stability
especially in depleted reservoirs and in fracture zones where the pore pressures will be more and it is time consuming especially in off
shore industry where economics play a major role. A new technology used for simultaneously drilling and casing a well where the
casing is used to transmit mechanical and hydraulic energy to the bit, instead of using a conventional drill-string. The fundamental
value of Drilling with Casing lies in improving drilling efficiency by eliminating flat spots in the drilling curve. It will eliminate the
use of drill pipe, significantly reduce tripping times and reduce time lost due to unscheduled events such as reaming, fishing and
taking kicks while tripping. This technology utilizes a top drive to rotate the casing and uses a combination of traditional components
such as standard oil field casing, bits, mud motors, MWD/LWD tools, as well as some newly developed components that allow
drilling of most hole sizes from any drilling unit, with the ability to recover the drilling assembly using wire-line. Since these tools are
run and retrieved inside casing, they are protected from the harsh down-hole environment while in transit. In casing drilling, drill
collars can also be used to add stability and to provide better fatigue resistance to the drill-string. Since there are a variety of possible
configurations for Drilling with Casing systems, the casing may or may not be rotated during the drilling process; and the drilling
tools may be integrated into the casing string or be part of an assembly that extends below the casing shoe. There are two basic
technologies used for Drilling with Casing:

• The first configuration employs a bottom hole assembly comprised of a positive displacement motor (PDM), drill bit and hole
opener. The assembly is latched to the first joint of casing. The hole is drilled by this assembly while the casing is lowered into
the hole, either in a static or rotated mode. Upon reaching total depth (TD), the latch-on bottom hole assembly is recovered with a
special retrieval tool. A valve system is run and installed before cementing commences.

• The second configuration utilizes only casing to transmit rotary torque and weight to the drill bit. A drillable drill bit and valve
assembly is made up and run with the first joint of casing. The casing string is rotated during drilling, typically with a spear
assembly that provides rotation to the casing. Upon reaching TD, there is no need for an additional trip and the casing can be
cemented immediately.

Casing drilling system has been designed primarily for multi-well offshore platforms, multi well operations on land, deep-water
operations, and for situations requiring operators to drill through and place casing across problem formations quickly. This technology
was applied successfully to drill through depleted reservoir (problems: wellbore instability, mud losses into the depleted zones) as an
alternative to the underbalanced drilling, which requires special equipment.

The major systems for Casing Drilling are as follows:

• Casing Drilling Rig: Rig required for Drilling with casing operations is similar to the conventional drilling rig. The only optional
equipment required for this operation is a top drive system. Conventional drilling rigs can be converted to drill with casing. The
conversion to Drilling with Casing rig requires addition or replacement of several key components. Conversion costs can vary
depending on the rig design and layout. There are four basic changes from a conventional drilling rig.
1. A wire-line BOP and double pack-off above the top drive in order to seal off around the wire-line.
2. Addition of pipe-handling tools to handle casing instead of a conventional drill string.
3. A wire-line winch which is capable of running and retrieving the BHA. It included an electric line for future tool
actuation.
4. A travelling block ,crown and a top drive in order to facilitate running the wire-line down through the casing
• Casing Drive System (CDS): Casing Drive System is required to rotate the casing string similarly as the drill string in case of
conventional drilling. The casing string is rotated for all operations, except slide drilling with a motor and bent housing assembly
for oriented directional work. The hoisting equipment must hold the weight, apply rotational torque and contain pressure. The
Casing Drilling system utilizes a top drive to rotate the casing. Single joints of casing are picked up off the pipe rack and set in the
mouse hole. The top drive, with extend feature, is connected to the top of the joint which is then stabbed into the top of the casing
string in the rotary table and drilled down in a conventional manner. The casing string is attached to the top drive via a casing
drive system without screwing into the top casing connection. The CDS includes a slip assembly to grip the pipe exterior and an
internal spear assembly to provide a fluid seal to the pipe. It is operated with the hydraulic top drive control system. Use of the
CDS speeds up the casing handling operation and prevents damage to the threads. There are two types of CDS: internal – for
greater casing radius, and external – for smaller casing radius. It is controlled automatically from the driller’s cabin with
Programmable Logic Control. A typical casing drive system grips and supports the casing during drilling acting like a
combination Kelly and elevator. This is shown in Fig. 1.

Fig 1: The grip and support of a typical casing drive system adapted from Avery. et al. (2009).

• Bore Hole Assembly (BHA): It consists of: External BHA - Pilot bit, Stabilizer, Pony Collar, Under-reamer, Rotary steerable
system, Internal BHA - Spacer Collar, Stabilizer, Drill Lock Assembly, and Casing Stabilization - Two Crimp-on stabilizers,
Casing, Crimp on stabilizer, and Profile nipple.

266
• Drill Lock Assembly (DLA). Connecting the BHA to the bottom of the casing is achieved by the DLA, which works equally
well whether a directional mud motor or RSS is used. The DLA locks into the casing profile nipple (CPN) and allows both torque
and weight-on-bit (WOB) to be applied by the casing during drilling. The DLA/CPN lock into a special lower casing shoe that
shrouds and protects the mud motor and internal stabilizers during shipping and rig-up. During drilling, using the chosen RSS
BHA has several advantages. Firstly, the 8½-in. PDC bit is employed to drill and steer a pilot hole, with steering applied by the
RSS immediately above the bit. At-the-bit MWD and Gamma Ray measurements can be transmitted to the LWD string by means
of a short hop electromagnetic telemetry link. The LWD system comprises all necessary geo-steering and formation evaluation
measurements along with mud pulse telemetry, and these can be conducted in the freshly drilled, optimally sized pilot hole. The
next step is borehole smoothing and enlargement using a series of roller reamers, followed by the 12 ¼-in. collapsible under-
reamer tool. Rotation is achieved via a combination of mud motor and casing rotation. The correct ratio is determined to keep the
under-reamer within its design speed. Critical factors that must be balanced include mud flow rate, which governs motor speed
and optimizes cuttings transport, and casing rotation to reduce friction and support hole cleaning. Using flow rate to maintain
equivalent circulating density (ECD) of 1 lb/gal above static mud weight is another critical factor to help control borehole
stability. In addition, external casing centralizers were crimped onto the casing as it was run to help reduce sliding friction and
maintain annular uniformity so correct cement distribution could be achieved after the casing point was reached and casing was
set.

• Drillable Drill Shoe: Reduced costs when drilling with casing are realised by eliminating conventional drilling operations prior to
running casing, while the ability to drill down the casing and cement in a single operation provides a means of dealing with
problematic well issues such as lost circulation zones & borehole instability. This is currently achieved using “drillable” tools
such as the existing and field proven Drill Shoe I; Drill Shoe II & Drill shoe III.

ƒ Drill Shoe I: The DS 1 product (Figure 1) comprises a central Aluminium core with integral cutting blades. The blades are
coated with a hard material to provide resistance to the abrasion imparted by drilling. Drillable nozzles located between the
blades direct the drilling fluid for efficient cooling and cuttings removal. This central “drillable” core is housed inside a steel
body where the profiles of the central blades are continued around the leading edge of the outer body and along the outer
diameter. The steel blades on the body extend the blades out to the cutting OD and feature a Carbide cutting structure to drill
out to full hole diameter. A raised section is continued along the body behind the cutters in spiral geometry to provide the
gauge area. The gauge pads consist of Carbide Bricketts held in a metal matrix, which gives the necessary resistance to
abrasion.

Fig 2: The wire-line retrievable BHA adapted from Međimurec et al. (2005).

ƒ Drill Shoe II: The DS 2 product (Figure 2) has the same general construction as the DS 1. It features an upgraded cutting
structure on the blades, which consists of a series of individual TSP cutters mounted in the face of the blades. This provides
the ability to drill harder formations over longer intervals. The steel portion of the blades has PDC cutters out to gauge
diameter.
ƒ Drill Shoe III: Inevitably the nature of the DS 1 & DS 2 cutting structure will not be suitable for applications where more
competent formations and longer intervals need to be drilled. In order to avoid excluding such applications from the benefits

267
of drilling with casing a third generation DrillShoe 3 (DS 3) product has been developed, which combines the benefits of a
PDC cutting structure as featured on standard PDC drill bits, with an ability to displace this “non drillable” cutting structure
into the annulus, so leaving only drillable materials in the path of the following drill string. The steel blades are
manufactured separately, with the cutter pockets in place, and are then fabricated to the Blade Sub. The PDC cutters are then
brazed in place utilizing established procedures. The Main Housing provides the length and guidance required for the Inner
Sleeve Assembly. An internal shoulder controls the stroke of the tool and provides a positive stop. Sufficient stroke is
required to ensure that the blades are fully displaced into the annulus, with the stop preventing the Inner Assembly from
being ejected from the end of the tool. A Latch Ring is contained between the opposite face of the internal shoulder and the
end the face of the Blade Sub. Inner Sleeve Assembly consists of the following components: Aluminum Nose (with blade
holding profiles), Steel Inner Sleeve, Drillable Nozzles, and Aluminum Ball Seat Funnel. The various types of drill shoes are
shown in Fig. 4.

Fig 3: The location of stabilizers in the pilot hole on the straight hole drilling assembly adapted from Warren, Lesso et al. (2005).

268
a. Drill shoe 1 b. Drill shoe 2

c. Drill shoe 3 d. Displaced drill shoe


Fig 4: The various types of drill shoes, adapted from McKay et al. (2003).
Drilling operations are moving into deeper water and well depths, and there has been a lack of consistent and sustained performance
improvement in exploration and appraisal drilling. A paradigm shift in well design philosophy is critical to the future success and
economic viability of deepwater drilling. It involves managing drilling risks in the shallow hole sections, where the well costs are
minimum, rather than the current practice of incurring risks after significant investment has been made. With these wells’ daily
operating cost often approaching USD$1 million and requiring 100 days or more to drill, it is critical to the economic success of
deepwater field development to reduce well costs and improve attainment of well objectives. The proposed well design model uses the
shallow and rapid growth of the pore pressure/fracture gradient (PP/ FG) environment to optimize casing seats. Drilling with casing
can be an enabling technology that mitigates shallow hazards, safer and more effectively than the existing model. The fundamental
premise is to use this technology to set the first – and possibly the second casing strings, across the shallow hazards, significantly
deeper than the current practice.

269
Fig 5: The various types of nozzles adapted from Galloway et al. (2003).

Fig 6: Stabilizer End Sub and Main Housing McKay et al. (2003).

The proven ability of drilling with casing to mitigate many similar drilling hazards as those encountered in deepwater drilling would
allow the casing seats to be based on the prevailing PP/FG environments, rather than being influenced by shallow hazards. This could
allow for the following improvements: Fewer casing strings required to meet the well objectives. This reduction of hole sections
would reduce the time required to construct the well. Flat time activities in deepwater typically are very time-consuming, making a
reduction of hole sections significant. In general, the fracture pressure at a given depth increases as the cumulative weight of
overburden above it increases. As water depth increases, the hydrostatic pressure exerted by the seawater column, in effect, replaces
the pressure that would be exerted by overburden in shallower waters. This reduction in overburden pressure dramatically decreases
the fracture gradient, especially in ultra deepwater locations. As water depths increase and more casing strings are required, it becomes
more critical to use effective design and drilling practices to maintain wellbore pressures within the operating window, while
achieving well objectives. Casing drilling is more effective in drilling the depleted reservoirs, in off shore industry it is a time
consuming factor with conventional drilling. Casing drilling is also more effective in lost circulation areas because of the reduced bore
hole size. It improves well site safety and well control which is more important in offshore industry.

BRIEF REVIEW OF LITERATURE

Directional casing while drilling (DCwD) wells require the use of a retrievable and re-runable BHA to recover the expensive
directional drilling and guidance tools, to replace failed equipment before reaching casing point, and to provide quick and cost
effective access to the formations below the casing shoe. The Tesco Casing Drilling system provides the versatile tools required for
directional drilling operations. A wire-line retrievable directional drilling assembly, positioned at the lower end of the casing, replaces
the directional tools used in a conventional BHA. These tools can be retrieved and re-run at inclinations exceeding 90o and the casing
can be reciprocated and circulated while running or retrieving the tools to assure that the casing does not become stuck. The casing
directional drilling system is composed of down hole and surface components that provide the ability to use normal oil field casing as
the drill string so that the well is simultaneously drilled and cased. A description of the system is briefly recapped below. A wire-line
retrievable drilling assembly is suspended from a profile nipple located near the bottom of the casing. The drilling fluid is circulated
down the casing ID and returns up the annulus between the casing and well bore. The casing is rotated from the surface with a top
drive for all operations except for oriented directional work when slide drilling with a steerable motor and bent housing assembly.
Individual joints of casing are picked up from the V door with hydraulically activated single joint elevators. Each joint is attached to
the top drive with a quick connect assembly that grips the casing without screwing into the top casing coupling. An internal spear
assembly provides a fluid seal to the pipe. The top drive is used to make up the connections to the casing string. The quick connect
prevents damage to the threads, allows casing connections to be made as fast as drill pipe connections, minimizes floor activity while
making a connection, and increases rig floor safety.

270
The casing directional BHA includes a pilot bit with an under-reamer located above it to open the hole to the final wellbore diameter.
The pilot bit is sized to pass through the casing and the under-reamer opens the hole to the size that is normally drilled to run casing.
Other down-hole tools in addition to the bit and under-reamer are used as appropriate. For vertical drilling, stabilization is included on
the assembly between the pilot bit and under-reamer. For directional drilling, a steerable motor or rotary steerable system, MWD, and
non-magnet collars are included in the BHA. The drilling assembly is attached to the bottom of the casing with a Drill-Lock-Assembly
(DLA). The DLA provides the ability to connect conventional drilling tools with rotary-shouldered connections to the casing and
facilitates running the tools in and out of the casing. It provides the capability to axially and torsionally lock and un-lock the drilling
BHA to the casing, seals in the casing to direct the drilling fluid through the bit, locates itself in the profile nipple without relying on
precise wire-line measurements, and bypasses fluid around the tools for running and retrieving. A releasing and pulling tool is run on
wire-line to release the DLA and pull the BHA out of the casing in a single trip for vertical and low angle wells. In the unlikely event
that the BHA cannot be pulled on the first attempt, the releasing tool can be disconnected from the DLA so that remedial measures can
be taken.

In some directional wells, particularly those with high inclination, it is desirable to release the DLA with a pump down dart before
running the wire-line to pull the BHA. This has the advantage of allowing the casing to be slid down (about 80 to 120 ft) over the
directional BHA to protect it while the wire-line is being run. For high angle wells, the retrieving tool can be simultaneously pumped
down and used to pull the wire-line into the well. Pump pressure is used to generate the axial force required to release the DLA for
both the direct pulling method and the dart release method.

o Rotary Steerable Systems (RSS): Drilling with steerable motor assemblies in both test wells and commercial wells has identified
three considerations that are somewhat unique for casing directional drilling. The assembly geometry required to achieve a specific
curvature is different than for drilling a conventional well. Selection of the particular motor for use with casing requires some
additional considerations. Field operational practices when drilling with casing may be different from those used when drilling with
a steerable motor with a conventional drill string.

o Assembly Geometry: The typical casing directional drilling steerable motor assembly places the under reamer below the motor,
immediately above the pilot bit .This arrangement requires the motor to power both the pilot bit and under reamer to allow drilling
without pipe rotation when sliding. A full gauge (pilot hole) stabilizer is incorporated into the under reamer body immediately
below the cutters to assist in drilling a smooth curve and to provide directional control.

The directional response of a conventional steerable system is primarily determined by the motor bend angle, motor size compared to
hole size, bit characteristics, and stabilization placed on and/or above the motor. The same factors determine the directional response
while directional drilling with casing, but the options for modifying the factors are somewhat more limited. Rotary drilling with casing
is more efficient than drilling with a motor in many situations, even for straight hole CwD applications. The ROP generally decreases
when using motors and the wear and tear on the down hole tools is often significantly increased. This observation leads to suggesting
that the benefits of rotary steerable directional systems that have been demonstrated for conventional drilling should also be attractive
for CwD applications. A rotary steerable system run in the pilot hole below the under-reamer should provide an ideal assembly for
directional control. It would eliminate all the potential problems associated with motor performance. It would also provide a smoother
well-bore that should reduce torque and would eliminate problems associated with sliding. Rotary steerable tools are available that can
be used with casing sizes of 7-in. and larger, but they present an economic hurdle for less expensive rig operations.

The BHA design for the 7-in. Casing directional drilling has several innovative features. It is about 110 ft long with 85-ft stick-out
below the shoe of the 7-in. casing. Starting from the top inside the casing is the DLA. Below the DLA is a tandem casing stabilizer.
This has two stabilizer blade sections that are gauged to the internal drift diameter of the 7-in., 23-lbm/ft of 6.26 in. This stabilizer is
designed to take most of the vibrations of drilling away from the DLA, reducing wear on it. A spacer drill collar is usually run below
the stabilizer to extend the BHA out below the casing shoe. A straight PDM motor may be run instead of a drill collar. It would be a
relatively large motor with a 6-in. diameter. The motor’s sole purpose is to allow surface drill string rotation to be lowered when
dealing with vibrations issues. The motor adds the rpm's back into the BHA and bit to maintain a good ROP. Casing directional
drilling is similar to drill pipe drilling, because similar bits are used and therefore similar rates are desired, typically 120 to 180 rpm.
Whirl problems could limit surface rotation to 50 rpm. The motor would add 100 turns to the drill string, bringing it to the optimum
range. The under-reamer is placed directly below the mud motor or spacer drill collar. This opens the pilot hole to 8 1/2 in. or greater.
When the pumps are off, the arms of the under-reamer collapsed so that the maximum diameter of the tool is slightly less than 6 1/8
in. To be retrievable, all components of the BHA have to be smaller than the internal diameter of the casing. A jet sub is added to the
BHA below the under-reamer to divert 20% of the flow from the bit. The pilot hole below the under-reamer is 6 1/8 in.

Below the jet sub is the external tandem stabilizer. This configuration is designed to reduce vibrations and wear from the under-
reamer. The gauge of these two stabilizer sections is 6 1/8 in. minus 1/16 in. Below these sections is a slim MWD system. The RSS, a
push-the-bit type, is installed below the MWD with a PDC bit. Rotary steerable directional drilling is controlled by making a series of

271
“downlinks” to the tool. These downlinks usually consist of a tool face direction and a power setting. The tool face is usually based on
a compass heading or magnetic tool face when kicking off the directional well or until the inclination greater than 5o.A well that is
building angle in a due east direction would use a 90o magnetic tool face. Once the well is kicked-off, tool face is switched over to
gravity mode. Here 0o is straight up while 90o is 100% to the right. The RSS power setting is usually expressed in terms of a
percentage of time that the tool is thrusting referred to as the duty cycle. Thus a downlink of 135om @ 60% means that a RSS is
building angle in the direction of 135o (south-east) 60% of the time (in the other 40% the tool is functioning in a neutral mode).

For more details about the above mentioned please see Warren and Lesso (2005).

Casing drilling at south Texas lobo: Five wells are drilled. The phase two program proved that casing drilling could eliminate the
formation related trouble time experienced with conventional rigs. This allowed additional wells to be drilled that would otherwise be
uneconomical. The wells were not drilled trouble-free, but the trouble was associated with the mechanical equipment (about 25%). In
the phase two drilling casing related problems were encountered, i.e,
1) The intermediate casing was tripped out several times without wear or damage on pipe body. Some couplings in lower parts of the
casing were worn on one side. However, this problem was solved by providing wear bands on the lower half of the 7 inch casing.
Lower end of wear band includes 1inch tungsten carbide hard facing material to that used for wear protection on drill pipe.
2) Stabilization: Stabilizers are used on the lower end of all three strings of casing to function as cementing centralizers and on 7inch
casing key seat wipers spaced at 229m along the length.
3) More challenging one is lateral drill string vibrations, these vibrations had resulted in two casing connection Fatigue.
4) Buckling: A significant difference between drilling with a conventional drill-string and casing drilling is that drill collars are not
used to provide weight-on-bit. The lower portion of the drill-casing will support only a limited compressive load before it buckles.
Buckling occurs when the compressive load and casing/hole geometry create a sufficient bending moment so that the casing becomes
unstable.

Effects of Buckling: A) Lateral contact force between casing and bore hole wall cause wear on the casing and extra torque is needed to
rotate the casing.
B) Buckling causes the casing to assume a curved geometry within the borehole that increases the stress in the pipe and may increase
the tendency toward lateral vibrations.

For more details about the above mentioned please see Međimurec, (2005).

Qatar Casing Drilling: Formation: Shale formation directly above carbonates formation. The formation is extremely unstable
formation and has plagued well construction Efforts for years.

Conventional drilling: Occidental petroleum Qatar limited (OPQL) is developing a site using 12.25inch horizontal, production and
injection wells. Problems encountered are as follows:

• Unstable shale has played havoc with conventional drilling and casing operations, due to highly conductive faults, which cause
sudden loss of circulation leading to severe caving of the shale with subsequent collapse on BHA. So severe has been the
borehole instability issue that often the drill-pipe and BHAs are LIH (lost in hole).
Solution: Changes has been done using directional casing drilling (DCwD). Advantage is that casing is protected all the length
except for few feet of the hole. If a sudden loss of circulation occurs at the dangerous shale/carbonate interface, less than 100 ft
(30.5 m) of drilling would be required to get the casing to its landing point.
• Even if the unstable shale caves in, the BHA can still be retrieved through the casing, allowing landing of the casing at that point,
cementing it in place, and drilling out. However, narrow casing-to-formation annulus, coupled with continuous rotation of the
casing and application of hydrostatic pressure on the shale using NaCl brine drilling fluid, reduces the chance of hole problems
commonly encountered with conventional drill-pipe drilling. This capability allows bits and reamers to be changed out as needed.
It also ensures that the MWD/LWD string can be retrieved and changed in the event of a failure. Casing can be rotated and/or
reciprocated during these trips and circulation can be maintained to resist the possibility of sticking.
• Well construction in mature fields with standard drill-pipe sometimes requires extra casing strings to avoid well stability problems
caused by depleted formation pressures. Because they improve wellbore stability, the DCwD techniques developed may reduce
the number of casing strings needed.
Solution: When circulation is lost or the wellbore collapses and excessive cuttings come to the surface during conventional
drilling, the usual and often ineffective remedy is to condition the mud and ream the hole. If an hour is spent conditioning and
reaming for every 2 hours spent drilling conventionally, DCwD techniques can very likely save rig time and reduce well
construction cost. Because directional casing while drilling minimizes annulus size, casing rotation causes a plastering effect also
called smear or stress cage that mechanically strengthens the borehole wall.

272
For more details about the above mentioned please see Avery, Stephens, (2009).

BASIC DESIGN OF AN INTEGRATED DRILLING-WITH-CASING SYSTEM FOR DEEP WATER OFFSHORE


APPLICATIONS

Casing Weight, Grades And Connections

Casing weight determination: Casing weight is determined by density of steel and the dimensions of the casing body API Spec. 5CT
or ISO 11960 for calculating the nominal casing weight of the plain pipe without threads or couplings:
2 2
W = C ( Do − t )t or W = C ( Do − Di )
4
Where
W=Plain end casing weight
C=conversion Factor 10.69 for oil field units and 0.0246615 for metric units
Do= Outside diameter, Di= inside diameter, t=wall thickness

Casing Grades and connections

Table 1: Basic PI casing grades, yield strength and tensile strength.

Yield Strength (kilo pounds per


square inch(ksi)) Minimum Tensile
Sl. No Grade
Strength
Minimum Maximum
1 H-40 40 80 60
2 J-55 55 80 75
3 K-55 55 80 95
4 N-80 80 110 100
5 M-65 65 85 85
6 L-80 80 95 95
7 C-90 90 105 100
8 C-95 95 110 105
9 T-95 95 110 105
10 P-110 110 140 125
11 Q-125 125 150 135

Connections: Many types of connections are used for casing. But there are three basic types 1) Coupling, 2) integral and 3) Weld on.
The most common type is threaded pipe with Coupling. A plain joint of pipe is threaded externally on both sides and an internally
threaded couplings or collar.

API 8-rd Connections: The most common type of connection in use is API 8-rd connection, where 8-rd or eight threads per inch and
slightly rounded profile. The profile is a V or wedge shape but slightly rounded at the crest and valleys of the threads. The API 8-rd
connection is made in either ST&C(short thread and coupling) or LT&C(long thread and coupling).
CASING DEPTH, SIZE AND LOAD DETERMINATION

Factors affecting Casing Depth and size


Depth Parameters
1. Pore pressures(PP)
2. Fracture pressure
3. Experience in an area
4. Bore hole stability problems
5. corrosive zones

273
6. Environmental considerations
7. Regulations
8. Company Policy

Pore pressure: Is the pressure of the fluid that fills the pore spaces in the rock. This pressure normally determines the lower limit of
drilling fluid. The pressure of the fluid depends on the depth of the rock and the density of the fluid.

Eaton’s Pore pressure Prediction from resistivity data:


a
⎡ R ⎤
G pp = Gob − (Gob − G N ) ⎢ o ⎥
⎣⎢ R N ⎥⎦
Where Gpp= Pore pressure gradient of a given depth (psi/ft)
Gob= Over burden pressure gradient at given depth (psi/ft)
GN== Normal pore pressure gradient (0.465psi/ft)
Ro= Shale resistivity at the given depth (ohm-meters)
a= exponent coefficient, it is normally 1.2, but can have range from 0.9 to 2.00.

Eaton’s pore pressure prediction from sonic or seismic travel time data:
a
⎡ ∆t ⎤
G pp = Gob − (Gob − G N ) ⎢ n ⎥ Where
⎣ ∆t o ⎦
where
Gpp= Pore pressure gradient of a given depth (psi/ft)
Gob= Over burden pressure gradient at given depth (psi/ft)
GN== Normal pore pressure gradient (0.465psi/ft)
∆tn= Normal shale travel time (micro-seconds/ft)
∆to= observed shale travel time ( micro seconds/ft)
a = exponent coefficient, it is normally 3, but can have range from 2.00 to 4.00

Fractures Pressure: One true definition is that the fracture pressure is the pressure at which a formation matrix opens to admit whole
liquids through an actual crack in the matrix of the rock as opposed to invasion through natural porosity of the rock or the pressure at
which the matrix of the rock physically fractures and admits the entrance of whole liquids.

Eaton’s Fracture gradient prediction:


ν
FG = (Gob − G pp ) + G pp
1 −ν
where
FG= Formation fracture gradient (psi/ft)
Gob= overburden pressure gradient (psi/ft)
Gpp= pore pressure gradient (psi/ft)
ν =Poisson’s ratio
Factors affecting conductor, surface, intermediate and production casing depth

Conductor Casing depth: The conductor casing is the largest diameter casing run in the well. It often supports the weight of the
subsequent tubes run in the well bore and maintains some minimal amount of bore-hole integrity while drilling the surface hole for
surface casing. Main purpose of conductor casing is
1) To prevent washing out under the rig while drilling the surface hole.
2) Support structure for the well and must support a small platform attached to the well head and production equipment.
Setting depths must be determined by soil bearing tests mainly in offshore industry.
Surface Casing Depth: Factors affecting are
1) Pore pressure (PP)
2) Fracture pressure (FP)
3) Depth of fresh water bearing zones
4) Legal regulations and requirement
From design point first two (PP, FP) are important

274
Intermediate and production Casing depth: Main reason for Intermediate casing is
1) Bore hole below surface casing require a mud density too high
2) Mud density may fracture exposed weak zones or a mud density too low allows high pressure zones in to the well bore
3) Presence of unstable zones and corrosive zones
4) Instability, usually shale may make it impossible to drill to total depth without isolating these zones

Setting depths using pore pressures and fracture pressures (PP and FP): Setting depths typically are selected using PP and FP,
plotting PP and FP versus Depth is the best way. Drillers use to plot in this way to determine mud densities at various depths.
1) Mud density slightly greater than formation pressure
2) Mud density less than fracture pressure
Drillers likely to keep the maximum density slightly less than fracture pressure because running the drill string back in to the hole
causes a positive surge pressure but the maximum also is considered a “kick margin”, so that during a well control event, the
formation is not fractured in the process of killing the well.

Casing Depth and Mud Specific gravity Determination for the given formation data

Design Data: The following is the well data for XXX location in offshore in which water depth is 2000ft and the reservoir depth is
11000 ft from sea bed, Surface temperature is 45oF and bottom hole temperature is 326oF and the pore pressure and fracture gradient is
given in the following figure. For this data the design of casing and casing while drilling system is based on.

Fig 7: Pore pressure and Fracture Pressure versus depth curve.

After getting Pore pressure gradient and Fracture pressure gradient curve, casing depth and mud weight is determined as follows
• First Estimate pore pressure curve
• Estimate mud weight curve. The mud weight should balance highest pore pressure in the open hole with a 0.5ppg or 0.06Specific
gravity trip margin.
• Plot estimated actual fracture gradient curve and designed fracture gradient curve (Kick Margin which is 0.5ppg or 0.06 specific
gravity less than fracture gradient), which is estimated actual fracture gradient, less allowance for well control.
• Start from bottom on the mud weight curve at point A draw vertical line to the designed fracture gradient curve at point B from B
extend it horizontally to Mud gradient curve to point C. By getting all points across the curves by the same procedure
• From the graph casing depth and mud density for corresponding casing is decided as follows
a) Point A to Point B is production casing 2500 ft and mud specific gravity is1.74.
b) From point C to D it is Intermediate casing or protective liner 2000 ft, Mud specific gravity 1.64.
c) From E to F is intermediate casing 2500 ft( assuming from 6000ft to 8500ft from water level even though it is from 5900 to
make calculation part easy), Mud specific gravity 1.48
d) From G to H it is Surface casing 3000 ft (assuming it up to 6000ft from surface to make calculation part easy), Mud specific
gravity 1.27
e) Up to H from sea bed to 1000 ft depth is conductor casing, mud specific gravity 1.05

275
Casing Size selection: Two important things to know about the selection of casing size
1) Hole size determines casing size
2) Hole size at any point in the well except the surface is determined by the previous string of casing
The size of the last String of the casing run in a well generally is determined by the type of completion that will be employed (it
depends on Reservoir, Production and drilling persons). Once we know the diameter
1) Determine the hole size for final string of casing.
2) Determine what diameter casing allows that size bit to pass through it. That is the size of next casing string.

For the Below Figure 8 casing size is as follows


1. Surface casing 4000 ft from sea bed = 16inch
2. Intermediate casing 6500 ft from sea bed = 13.375inch
3. Protection liner 8500 ft from sea bed = 9.625inch
4. Production casing 11000 ft from sea bed = 7inch

Conductor casing depends on soil properties which we are not considering in the design.

G F
C

Fig 8: Graph showing depth of casing and its corresponding mud


B specific gravity.
CASING LOADS

Types of primary casing loads

1) Collapse loads: These are differential pressure loads where the external pressures exceed the internal pressure tending to cause
the casing to collapse.
2) Burst Loads: These are differential loads in which the internal pressure exceeds the external pressure tending to cause to rupture
or burst.
3) Axial Loads: These are tension or compression loads caused by gravitational and frictional forces on the pipe.

First two are dictated by well conditions and anticipated operations in the well (PP FP and Drilling fluid). The third type of load axial
is a function of casing selection process itself.

Magnitude of loading (worst case of loading): The process consists of determining the following

Collapse Loading
1) Minimum Internal pressures
2) Maximum external pressures
Burst Loading
1) Maximum internal pressures
2) Minimum external pressures

276
Some typical sources of these parameters are
1) Formation Fluids: water, oil and gas
2) Drilling fluids: mud, mud filtrate
3) Unset cement: whole cement, cement filtrate
4) Stimulation fluids
5) Ocean or Surface water
6) Atmosphere

Collapse Loading:
Internal loads
1) Empty casing,
2) gas, oil,
3) fresh water,
4) field salt water or stimulation fluids, and
5) drilling or work over fluids
6) Combinations and partial columns of above
Most serious collapse loading is atmospheric pressure (an empty well bore mostly in surface casing).
External Loads
1) Fresh water
2) Salt water
3) Formation pressure
4) Drilling fluid
5) Cement

Burst Loading
Internal loads
1) Gas
2) Oil
3) Water
4) Combination of gas and liquids
5) Stimulation fluids in combination with pumping pressures
External Loads
1) Atmospheric pressure
2) Seawater pressure
3) Fresh water
4) Salt water
5) Formation pressure
6) Drilling fluid
It is never assumed that hardened cement will give support in burst, even though it will.

Formulas to determine loads:


M ( h2 − h1 )

= Pshoe e
ZRTavg
Gas pressure at Surface: Pn , where PShoe is the pressure at shoe or fracture pressure, Pn is the pressure at any point
lb. ft
from shoe to the surface, Z=1 for methane, R=1544 (ideal gas constant at standard gravity, oil field units), h1= vertical height
mole.o R
from casing shoe, h2 = vertical height from shoe surface = zero(0), M= molecular mass of gas(16(lb/mole) for methane, Tavg=
average absolute temperature at point shoe and point "n".

Hydrostatic pressure due to mud: Pmud= γmud x h, γmud= specific gravity of mud, h= length of casing

Fracture pressure: Pfrac= γfrac x h, γfrac= mud specific gravity at which formation fractures, h= length of casing.

Loads on Surface Casing

Collapse Load:
1) Internal pressure is atmospheric pressure or zero.

277
2) External pressure is mud pressure.

Burst Load:
1) Internal pressure is equivalent to gas kick that fractures and flows into formation below the casing shoe.
2) External pressure equal to freshwater gradient.

Loads on Intermediate Casing

Collapse Load: Mud gradient outside and no pressure inside or pressure of freshwater is taken.
Burst Load: Fresh water or mud outside and gas pressure equal to fracture at the shoe or alternate approach sometimes is used when
the surface BOP and well head selection limits the burst rating at the surface.

Maximum Burst Load method: This method of burst design assumes that the minimum rating of the wellhead equipment is one
limiting factor and the injection pressure of gas or liquids into the formation just below the casing shoe is the other. By knowing the
working pressure of the BOP, the fracture pressure of the formation below the shoe, and the density of the mud, we can calculate the
length of the column of mud and gas from the following equations of prentice (1970).
P frac − Psurf − γ h csg
h mud =
gas

γ mud − γ gas

P frac − Psurf − γ gas h csg


h gas =
γ gas − γ mud
where,
hcsg = vertical length of the casing, hmud = vertical length of the mud column
hgas = vertical length of gas column Pfrac = fracture pressure at shoe
Psurf = surface equipment pressure rating, γmud = specific weight of mud, γgas = Specific weight of gas
Loads on Production Casing

Collapse Loads: The most common approach in the production is to assume that the worst collapse scenario is one in which the
casing is empty and open to atmosphere. Another possibility is to assume that the well always has some amount of liquid or pressure
inside it, equal to the formation pressure at the time.
Burst Loads: Most common procedure is to assume that the casing must withstand the maximum shut-in formation pressure in the
form of gas column for a gas well (or oil for oil well)
Determination of loads on Casing for the given formation data
Given Data Temperature at surface 45oF, Bottom hole temperature is 326oF, all the values of Pore pressure and fracture for
determination of loads is taken from figure 8.

SURFACE CASING

Collapse load determination


Bottom Hole temperature=326Fo at 11000ft from sea bed
Average Surface Temperature = 45 oF
Assume Internal casing pressure=0 psi, Mud Specific Gravity = 1.27
External pressure at 4000ft is due to mud pressure
∆Pshoe= γmud x h – 0= 0.052x1.27x8.33x4000
∆Pshoe=2200psi
Calculating external pressure at each depth by the same procedure
3000 ft depth ∆Pshoe = 1650
2000 ft depth ∆Pshoe = 1100
1000 ft depth ∆Pshoe = 550
Collapse load at the surface is zero because external and internal pressure is = 0.

278
Fig 9: Surface casing collapse load curve.

Burst Load determination: at shoe the burst load is fracture pressure of the formation below the casing less the external pressure at
the casing shoe, which is equivalent to the freshwater gradient.
From Figure 8 γfrac= 1.55
∆Pshoe=Pfrac - Pwater= (γfrac- γwater) x h
∆Pshoe= 0.052x(1.55-1)x8.33 x4000= 950psi
Now Calculate the burst load at the surface, worst case would be gas up to the surface that is maximum pressure at surface
The net burst presseure at the surface is due to the column of gas whose pressure at the casing shoe is equal to the fracture pressure at
the shoe, so we calculate the fracture pressure first.
Pfrac= 0.052x1.55x8.33x3000= 2690psi

Assume temperature gradient is linear.


Surface temp= 45oF, Bottom Hole Temp= 326oF
By interpretation T= 147oF at 4000ft
The average temperature = 45 + 147 + 460 = 556 o R
2
Now surface gas pressure is:
(16 ( 0− 4000 )
(1544 X 556 )
Psurf = 2690
e =2500psi.

Now by the same procedure the net burst pressure at each interval is. The values are as follows: Psurf= 2500 psi, P1000 = 2110 psi, P2000
= 1340 psi, P3000 = 1340 psi, ∆Pshoe = 950 psi.

The Burst load curve for surface casing is shown in Fig. 10.

INTERMEDIATE CASING:

Collapse load Determination


Depth=6500 from sea bed
Specific gravity of mud γmud= 1.48, γfrac= 1.7
Bottom Hole temp= 326oF and surface temp= 45oF
Consider worst possible case for collapse in which there is fresh water on the inside and the mud it was run in on the outside.

Net collapse load at bottom


∆Pshoe = (γmud-γwater) x h

279
=0.052(1.48-1.0)(8.33)(6500) =1350psi, now by Calculating net pressure for every 1000 feet depth , at surface collapse is zero
P6000 = 1240, P5000 = 1040,
P4000 = 830, P3000 = 620,
P2000 = 420, P1000 = 210
At surface collapse load is zero, the collapse load curve is as shown in Fig. 11.

Fig 10: Surface casing burst load curve.

Fig 11: Intermediate casing collapse load.

Burst load Determination for Gas


Assumptions:
Internal pressures: 1) Fracture pressure at the Shoe, 2) Gas inside with gas pressure at the surface.
External Pressure: Freshwater gradient behind the casing.
γfrac = 1.7

Net burst pressure at the bottom is: ∆Pshoe=(γfrac - γwater) x h= 0.052(1.7-1.0)(8.33)(6500) = 1970psi.

280
The net burst presseure at the surface is due to the column of gas whose pressure at the casing shoe is equal to the fracture pressure at
the shoe, so we calculate the fracture pressure first:
Fracture Pressure Pfrac= γfrac x h= 0.052(1.7)(8.33)(6500) = 4790psi

By interpolation temperature at 6500 = 211oF


Now average temperature = (45+211)/2 +460oR= 588oR
[16 ( 0 − 6500 )]
[1544 ( 588 )]
Psurf = 4970 e = 4260psi

Now by the same procedure the net burst pressure at each depth interval values are as follows
∆P6000 = 1970, P6000 = 2150, P5000 = 2500, P4000 = 2850, P3000 = 3200, P2000 = 3550,
P1000 = 3910

Burst load curve is as follows:

Fig 12: Intermediate casing Burst load curve for the well with full gas column.

PROTECTIVE LINER

Collapse load Determination


Depth=8500 from sea bed
Specific gravity of mud γmud= 1.64, γfrac= 1.8
Bottom Hole temp= 326oF and surface temp= 45oF
Consider worst possible case for collapse in which there is fresh water on the inside and the mud it was run in on the outside.

Net collapse load at bottom


∆Pshoe = (γmud-γwater) x h
=0.052(1.64-1.0)(8.33)(8500) =2950psi,
Now by Calculating net pressure for every 1000 feet depth , at surface collapse is zero
∆P8000 = 2220, P7000 = 1940,
P6000 = 1662, P5000 = 1390,
P4000 = 1110, P3000 = 830,
P2000 = 550
P1000 = 280The collapse load curve is as shown in Fig. 13.

281
Fig 13: Protective Liner Collapse load curve.

Burst load Determination for Gas


Assumptions
Internal pressures: 1) Fracture pressure at the Shoe, 2) Gas inside with gas pressure at the surface.
External Pressure: Freshwater gradient behind the casing, e.g. γfrac = 1.8

Net burst pressure at the bottom is: ∆Pshoe= (γfrac - γwater) x h= 0.052(1.8-1.0)(8.33)(8500) = 2950psi.

The net burst presseure at the surface is due to the column of gas whose pressure at the casing shoe is equal to the fracture pressure at
the shoe, so we calculate the fracture pressure first.

Fracture Pressure Pfrac= γfrac x h= 0.052(1.8)(8.33)(6500) = 6630psi


By interpolation temperature at 6500 = 262oF
Now average temperature = (45 + 262) +460oR= 614oR
2
Next gas pressure at surface
[16 ( 0 −8500 )]
[1544 ( 614 )]
Psurf = 6630 e = 5740psi
Now by the same procedure the net burst pressure at each depth interval values are as follows
∆P8000 = 3120, P7000 = 3450,
P6000 = 3770, P5000 = 4100,
P4000 = 4430, P3000 = 4760,
P2000 = 5080, P1000 = 5412The burst load curve is as shown in Fig. 14.

MAXIMUM BURST LOAD METHOD

Fracture pressure= 6630 psi


Gas pressure at surface =5740 psi
Surface pressure of BOP= 5000 psi
Bottom hole temperature = 326oF, Surface Temperature = 45oF
Maximum Specific gravity of mud = 1.74
Burst load for Gas
6630 − 5740 lbf
Now γ gas = = 0.105 2
8500 in . ft

282
lbf
And γmud = 0.052(1.74)(8.33)=0.754
in 2 . ft
P frac − P surf − γ h csg 6630 − 5000 − (0.105)(8500)
h mud =
gas
= = 1136 ft
γ mud − γ gas 0.754 − 0.105

Pint= Psurf + γmud x hmud =5000 + 0.754(1136) = 5860 psi


∆Psurf = Psurf – 0 = 5000 psi
∆Pint = Pint – γwater x hmud = 5860 - 0.052(1.00) (8.33) (1136) = 5370 psi
∆Pshoe = Pfrac – γwater x hcsg = 6630 – 0.052(1.00) (8.33) (8500) = 2950 psi
Maximum burst load curve for protective liner is as shown in Fig. 15.

Fig 14: Protective Liner Burst load curve.

Fig 15: Protective Liner maximum burst load curve.

PRODUCTION CASING

283
Collapse load determination
Depth=11000 from sea bed
Specific gravity of mud γmud= 1.74, Pore pressure = 1.68
Bottom Hole temp= 326oF and surface temp= 45oF
Consider worst possible case for collapse in which there is fresh water on the inside and the mud it was run in on the outside.

Fig 16: Production Casing Collapse Load curve.

Fig 17: Production Casing burst Load curve.

Collapse load at bottom

Assume no water inside the casing


∆Pshoe = (γmud-γwater)h
=0.052(1.64-1.0)(8.33)(8500) - 0 =8290 psi,
Now by Calculating net pressure for every 1000 feet depth, at surface collapse is zero
P10000 = 7540, P9000 = 6790,
P8000 = 6030, P7000 = 5280,

284
P6000 = 4520, P5000 = 3770,
P4000 = 3020, P3000 = 2260,
P2000 = 1510, P1000 = 760.
Collapse Load curve for production casing is shown in Fig. 16.

Fig 18: Chart showing complete methodology of the design.

Burst Load Determination: Assume fresh water on outside of casing and packer fluid might fail during production, so that packer
fluid is produced with the gas resulting in a full column of gas in the annulus between tubing and production casing.

285
Formation pressure is equivalent = 1.68 specific gravity.

The net burst pressure at the surface is due to the column of gas whose pressure at the casing shoe is equal to the fracture pressure at
the shoe, so we calculate the fracture pressure first.

Pshoe = 0.052x1.68x8.33x11000 = 8010psi

Gas pressure at surface


Bottom hole temperature= 326oF, surface temperature = 45oF
16 ( −11000 )
45 + 326
Tavg = + 460 = 645 , now Psurf = 8010e 1544 x 645
= 6712
2

Now net burst at casing shoe due to water:

∆Pshoe = 8010 – 0.052x8.33x1x11000 = 3250,


By finding the values for every 1000ft depth interval.
P10000 = 3570, P9000 = 3880, P8000 = 4200, P7000 = 4510, P6000 = 4820, P5000 = 5140,
P4000 = 5450, P3000 = 5770, P2000 = 6080, P1000 = 6400

Burst load curve for production casing is shown in Fig. 17. The complete methodology of the design is shown in Fig. 18.

CONCLUSIONS

This paper has presented an integrated design approach for drilling-with-casing system for deep water offshore applications. In this paper, we have
focused on the end-to-end connected system where standard oilfield casing is used to drill the well first and then it is left in the proper place to case
the well. Our integrated process makes it possible to speed up drilling (e.g. anticipated gains at around 15 – 25 %) by eliminating drill-string tripping
and the problems associated with it. In addition we have addressed the drawbacks of drilling-with-casing system, e.g. difficulty in assembling, stuck
casing, and limited drilling distance because of fatigue life of linear connections.

However, our presented design approach is confined to the domain where the data is available to us. A fully generic model has not been developed
and hence it is not operational across the board. We need to develop it fully, and then test it against various formations of India and abroad. Our
future work shall go in this direction and currently this is under investigation.

ACKNOWLEDGEMENTS

Some of the data used in the present work has been supplied by the Weatherford International Limited, USA. The authors gratefully acknowledge
their support and encouragement for the present research.

REFERENCES

McKay.D, Galloway.G. and Dalrymple.K, (2003), “New Developments in the Technology of Drilling with Casing: Utilizing a Displaceable Drill-
Shoe Tool”. Presented in World Oil Casing Drilling Technical Conference held at the Westchase Hilton, Houston, Texas, March 6 - 7, 2003.
N.G.Međimurec, (2005), “Casing Drilling Technology”, Technical Report, University of Zagreb, Faculty of Mining, Geology and Petroleum
Engineering, Croatia.
Gupta A.K., (2006), “Drilling With Casing: Prospects and Limitations” Paper SPE 99536 presented at western regional/AAPG pacific section/GSA
cordilleran section held in Alaska 8-10 may 2006.
Tommy Warren, Bill Lesso, “Casing Directional Drilling” presented at the AADE 2005 National Technical Conference and Exhibition, held at the
Wyndam Greens point in Houston, Texas, April 5-7, 2005.
Byromted G. “Casing and liners for drilling and completions”, by gulf publishing company, Houston, Texas 2007.
Avery.M., Stephens.T, Al-Hadad A.K., Moellendick.E., Turki.M., Abed.M and Lesso.B., (2009). “High Angle Directional Drilling with 9-5/8-in.
Casing in Offshore Qatar” presented at the SPE/IADC Drilling Conference and Exhibition in Amsterdam, March 17-19 2009.

286
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Drilling to exploit groundwater in “Vedanattam” Model village in Malattar Basin, Tamilnadu

R. Ramasamy, R. Sundaravadivelu, SP. Subramanian and P. Shanmugam


Department of Ocean Engineering,
Indian Institute of Technology, Madras, Chennai 600036, India

ABSTRACT

Huge quantity of meteoric water was infiltrated into the subsurface over 20 pluvial periods of continuous rainfall for several centuries during the
beginning and retreat of global Quaternary Ice Ages. The water is still moving towards seabed of Bay of Bengal at varying depths between 0 and
4000m. A universal piezometric surface develops interconnecting adjoining groundwater basins through horizontal and deep fractures. Potential
aquifers are found relatively at shallow levels in Government lands along the low lying streamlets. A rural development scheme is proposed for
drought prone Vedanattam Village located in Malattar Basin near Veppalodai coast, Tuticorin District, Tamil Nadu. Drilling process of deep tube
wells are site specific. Application of cutting edge geological geophysical methods and logging while drilling bore-hole data, a cluster of 30 deep tube
wells of 100-200m bgl commanding 10 ha each successively are to be sunk along streamlets to tap groundwater without saline water contamination.
The village can support 50000 individuals with self sufficient income.

KEY WORDS: Drilling techniques, Logging While Drilling, Groundwater Exploration, GPR Imaging, Water Jet Drilling, Tube Well, Streamlets

INTRODUCTION

Water is more essential than hydrocarbons. Hence more emphasis is to be given for exploration and exploitation of water than any
other mineral resources. Tamilnadu is a monsoon rain fed region. India is agriculture based country. Both SW and NE monsoon
prevail in this country for a period of 3 months each. Many rivers, streamlets and nallahs fill tanks, lakes and dams feeding water to
agriculture lands. Besides dug wells / tube wells improve irrigation. However, monsoon failure, drought and water scarcity in shallow
dug wells induce crop failure. Hence, most villages in Tamil Nadu are deserted due to migration of villagers to urban centre in search
of alternate jobs. Timely water supply for irrigation may improve crop production. Only 30 per cent of deep tube wells have sunk in
hard rock terrain yield good water supply. The percentage of tube wells with high specific yield decreases as one goes deeper into the
ground. Failure of deep tube wells without adequate water supply involves loss of huge amount of money and energy.

GROUNDWATER POTENTIAL

The groundwater infiltrated / recharged in aquifers below nallahs, tributaries and major rivers which lie on the lowest topographical elevation of the
terrain of Tamil Nadu is ultimately discharged through the seabed. The streamlets are generally guided by deep fractures with varying attitudes. Along
the courses of such streamlets and nallahs, piezometric surface is opened and spring seepages out or flowing water suddenly disappear owing to
descending water. These fractures often interconnect adjoining fracture systems. Geology, structure and rock types of Tamil Nadu are a complex
system evolved more than 3000 million years. The age of groundwater recharge and slow migrant flowage at depth for a few tens of meters below the
ground level (bgl) also varies from a few days to several thousands of years. Therefore, statistics of rainfall data for a few decades can not express any
valid data for groundwater potential beneath the surface. Again specific-yield of a dug well or tube well does not indicate the groundwater potential of
the site. Land Sat-TM/IRS IC/DLSS-3 imageries show a huge quantity of seepage of groundwater discharged in the seabed. Compared the quantity of
such water discharge on the seabed, the exploitation of groundwater in the major cities like Chennai, Pondicherry, Cuddalore, Nagapatinum,
Ramanathapuram and Tuticorin is negligible. In coastal regions, where multiple aquifer horizons are encountered water quality of one aquifer differs
from other due to their flow paths and solubility of minerals in their host-rocks Major part of Tamil Nadu is covered by Precambrian charnockites,
amphibolites, granites and gneisses. The groundwater infiltered from highly rain-fed Western Ghats moves towards east in Tamil Nadu and into the
seabed of Bay of Bengal. Rock structures, fault planes and shear zones are tilted, refolded and bent because of super impositions of structural features
of one tectonic episode with others during the course of several millions of years since their formation. The geology one site varies from the other.
Multiple aquifers take shape as that of fracture patterns present in rock units. In geological history, several decades of pluvial periods of continuous
rainfall took place just before and after the period of glacial and interglacial periods counting over 20 times are found in the Quaternary Ice Ages. The
intermittant pluvial periods were extended over 50000years between130000 and 20000ybp-years before present- (Wilson et.al. 2000). The rainwater
had infiltrated into considerable depths and they still slowly migrating towards sea. Recent studies indicate that huge quantities of groundwater are
discharged into the sea (Burnett et.al. 2006). The slow migration of groundwater into aquifers at depth, rainfall infiltrated into interconnected shallow
level semi-confined and deep level confined aquifers, indicate the zone of meteoric water extends and spreads over to a depth of 4000 m below ground
level of seabed at the Bay of Bengal. A universal piezometric-surface develops interconnecting adjoining groundwater basins through sheet and deep
fractures.

GROUNDWATER EXPLORATION IN MALATTAR BASIN

Malattar is a very dry river and flash flows are in one or two days during heavy shower of north-east monsoon rainfall during the months of October to
December. It originates from Kalugumalai high land and runs along sheared fault lineaments trending NNW- SSE with a number of acute angle
conjugative tributaries. The basin is covered with 3 to 5 m thick black cotton soil over granite gneisses. The rocks trend ENE-WSW directions with
dipping towards NNE or rarely SSW. Field studies indicate that the region is deeply faulted and sheared and extensive development of quartzite,
granite intrusions and carbonatite diatremes dyke-lets with gypsum mineralization.. Malattar basin is not isolated; it is interconnected with adjoining
groundwater basins of Vaippar, Tambraparani, Kundar, Vagai, Cauvery basins and even to distant Ganga basin through deep fractures and lineaments
crossing across adjoining water divides. Hence a universal piezometric groundwater level exists. Groundwater migrates from one basin to other
making equilibrium level with atmospheric pressure (Ramasamy, 1988). Interconnecting confined aquifers exist at shallower levels in beds of nallahs
and tributaries are potential resources to exploit groundwater with minimum expenditure. Vedanattam village (8o58’08”N; 78o08’49”E) is situated in
Malattar Basin, located 5 km west of Veppalodai coastal tract near Tuticorin. The width of the Malattar Basin in this place exceeds 2km. The
groundwater table fluctuates between 5 and 10m bgl (Survey of India, 1997). Groundwater exploration is carried out by State Groundwater Division
Govt. of Tamil Nadu, Central Groundwater Board (Balachandran 2009) and several research Institutes and Universities. A review of their studies
reveals that there exist multiple layers of aquifers at varying depths within a depth of 70m bgl. Piezometric surface (Mondal, et.al. 2009) in deeper dug
wells and tube wells always exists at relatively 1 or 2 m lower levels than shallow dug wells. The water levels in deep wells show little seasonal
variations (Fig. 1a & 1b). In these diagrams a dot represents parameters of a specific well located in this area. The groundwater quality in deep wells is
good with low EC values (Fig. 1c, 1d, 1e & 1f). Pumping tests and specific yield capacities of wells indicate that storativity decreases with increasing
transmissivity (Fig. 1g.) Increasing static volume of water in a well slightly increases discharge volume (Fig.1h). Draw down volume / (per m3)
negatively correlated with static volume of water (Fig.1i) in these wells (Mondal et.al. 2009). Natural recharge volume decreases as pumping
discharge volume increases (Fig.1j). These factors indicate only the nature of variation of specific yield in a particular well. The specific yield does not
directly represent groundwater potential of the particular site. It indicates only natural recharge of water into a well. It depends only on nature of rock
units permitting the flow of groundwater into the well. The groundwater table fluctuates in shallow dug wells between 5 and 10m bgl even at extreme
dry seasonal condition. Based on geological, remote-sensing imageries and field studies, it is proposed to drill 30 bore holes within and adjoining the
village boundary, along small nallahs and tributaries of the Malattar River Basin in this village (Plate 1)

REVITALIZATION

Most dug wells of poor specific yield are revitalized by sinking vertical boreholes, directional and horizontal drill holes for a distance of 5 to 10 m to
crosscut adjoining potential aquifers. By this technique the recharging capacity of dug or tube wells may be enhanced. For this field investigations are
necessary to explore wet tree fracture system in the side walls and floors of the well bottom. Bore-hole GPR imaging and resistivity logging will give
valid data for such exploration. Extensions of wet- tree fracture system tapping adjoining confined aquifers can be detected at depths revealing at low

288 
 
resistivity zones. In most cases the fracture planes often refolded and curvilinear in nature. Ground Penetrating Radar (GPR) imaging at the floors and
side walls of dug well, will give important clues on recharging capacity of the well and attitude of wet fracture planes.

Variation of water table against depth of wells


Depth versus static ground water table in post
during pre-monsoon period
monsoon period

Static groundwater table


20 15
Static water table m

10

in m
10
5

0 0
1a  0 20 40in m 60 80 1b  0 20 40in m
Depth 60 80
Depth

Variation of Ec against static ground water table Variation of Ec against static ground water table in
during pre-monsoon 20000 postmonsoon period
20000

10000 10000

Ec
Ec

0 0
1c  1d 
0 5 10 15 20 0 5 10 15
Static groundwater table in m Depth of static water table in m

Ec versus depth of wells Ec against depth of wells


20000 20000
15000 15000
10000 10000
Ec

EC

5000 5000
0 0
1e  0 20 40 60 80
1f  0 20 40 60 80
Depth in m Depth in m

Discharge of water m3 perminute against Static


Storativity against Transmissivity
water volume in wells below water table
Discharge of water m3 per

1 15
0.1 0 20 40 60 80 100
10
Storativity

minute

0.01
0.001 5
0.0001
1h  0
0.00001
1g  Transmissivity
0 100 200 300
Static water volume m3 in well

Static water in well against draw down m3/s Pumping discharge m3 per minute against recharge
volume in m3 per minute
Recharge m3 per minute
draw down volume m3 per

10 0.1
0 2 4 6 8 10 12
second

1 0.01
0 50 100 150 200 250 300

0 0.001
1i  1j 
Static water volume in well Pumping discharge volume m3 per minute

289 
 
Fig. 1a to1j indicate that water levels in deeper dug wells and tube wells show little seasonal variations.
Water quality and specific-yield improve in deeper wells. Specific–yield negatively correlates with discharge volume.
 
Plate1. Location of tube-well sites selected, however actual sites will be successively proposed during starting of initial wells by GPR imaging and
well logging methods Malattar River bed shows a series of fracture lineaments. Side boreholes by water jet drilling for a distance of 1 to 5 m may
increase recharge potentiality of the dug well. An experienced geologist can suggest the direction and distance to be drilled by examining the well
sections, rock types and structural features of the terrain. With the help of seismic timer and acoustic meter, it is possible to measure the strength and
structural defects of the well sections for some distance or depths.

DRILLING TECHNIQUES

Groundwater exploration and exploitation through deep bore-wells are complex problems. Even in a terrain of homogeneous rock exposure, texture
and structure widely vary in depth from one drilling site to the neighbouring site. Hence, it is difficult to predict or apply the same technique to the
other. The drilling technologist should be aware of geological conditions. He should be able to adopt suitable techniques and measures required then
and there. Drill hole may encounter rock units of different structural trends and attitudes, textures, mineral compositions, strengths and fissilities at
varying depths. He should control the speed of the drilling process according to the rock type encountered. Using several types of bore well logging
geophysical instruments, it is necessary to study groundwater potential of multiple aquifers encountered at wall rocks of the drill hole. Horizontal
drilling, gunshot drilling is essential to trap groundwater aquifers at various depths lying adjoining the drill hole. Tube well cuts shallow unconfined
aquifers or deep confined aquifers yielding water with different hydrostatic heads and qualities. Periodic testing of water quality and specific yield is
essential to tap adequate water supply within the shallow depths. During drilling, the speed of drilling rig movement indicates the nature of the rock.
At normal speed the time taken for drilling through quartzite veins takes prolonged period of time, since the hardness of the rock is very high. Along
shear zones, siliken sides, gauge materials, cavities closely jointed formation the speed of the drilling rig may decrease significantly. Reaching at the
contact zones of different rock formations the speed may suddenly vary depending upon their strength, hardness and density of the rocks. While
crossing large aquifers, the speed will suddenly decrease. Drilling in fissile rocks or rocks with close-spaced jointing and rock units with different
hardness and strength borehole deviation may be encountered. Logging While Drilling (LWD) will give valuable information about the nature of the
rock. Automatic LWD will guide the driller to enhance driller’s drilling capacity. Interpretation of LWD gives information about multilayer aquifer
hydrostatic pressure heads. At very high piezometric pressure, drilling may be stopped at an optimum supply of specific yield. By increasing borehole
dimension from 10 cm to 25 cm, the specific yield of recharge may be increased. Within minimum depth of sinking bore wells an optimum
requirement of water supply may be reached. This will also reduces the cost of sinking boreholes and minimize the number of boreholes for a specific
area. Handling of LWD units, borehole camera, piezometer, GPR imaging units in the borehole is much easier in the boreholes of larger dimensions.
However, sinking borehole with larger dimension is highly expensive and depending upon the budget allotment the dimension of bore wells may be
selected.

290 
 
WET / DRY TREE FRACTURES AND LEAKAGE ZONES

Understanding the nature of dry tree fracture and wet tree fracture systems, it is possible to measure the specific yield of aquifers at the opening site of
the aquifer in a tube well during LWD measurements. The leakage through dry tree fractures can also be measured by LWD. Proper sealants can be
applied wherever leakage to be encountered. Locations of slotted pipes will be provided are to be detected by LWD. Water quality is measured by
water quality measurement meters attached with LWD units. Wherever contaminated saline water encounters, there adequate sealants would be
applied after identifying the sites by LWD. By LWD investigations, it is possible to demarcate potential aquifer cuttings to insert slotted pipes to
recharge the tube wells. During LWD accompanied with borehole GPR imaging units, it is possible to demarcate nearest potential aquifer escaped
from drill hole cutting. By demarcating such aquifers, directional drill holes of 2-5 cm dimensions may be made from borehole walls using portable
units of water-jet drilling nozzle through which passing of garnet power slurry and compressed air. The pressurized slurry pot may be kept on the
ground surface and compressed air, water and garnet mixed slurry may be passed through a cable pipe to remotely operated water-jet drilling nozzle.
The directional drilling may be made to a distance from 5 to 10m distance from the borehole wall. By LWD with borehole camera, it is possible to
measure the attitude of (strike, dip direction, plunge direction, dip and plunge direction) of joint patterns, fault planes and bedding planes of rock
formations. Studying the borehole GPR 2D, 3D tomographic images, it is possible to identify the nearest potential aquiferous fractures from the
borehole walls. Such aquifers located far from the distances above 5 to 10 m to the borehole walls a blast hole of 10 m may be made and the end
portion of the blast-hole will be charged with gelatin dynamite and exploded to induce hydro-fracturing to tap water from the aquifer.

WELL DESIGN AND ASSEMBLY

Well design is made after careful geological investigations and locating of borehole site locations. LWD data, it is possible to get sufficient
information about the nature of casing and requirement of double casing to be inserted. Information about directional drilling and hydro-fracturing is
also obtained. Within the borehole, sites encountered with dry fractures leading to leakage and escape of groundwater moving up with high hydro-
static head may be ascertained and escape of water from the dry fractures may be prevented. Further, wet fracture planes can be opened to increase the
recharge capacity of the tube-well. Based on these factors potential groundwater yielding tube well may be designed. Well assembly is made by
determining the specific yield of the tube well measured with piezometer during LWD at various depth intervals from the encountered aquifers. The
cumulative specific-yield may be determined by overall studying specific-yield of aquifers. From these facts, the dimension of tube well pipes to be
installed, and the horse power of the submersible motor is to be fitted may be calculated. From these calculations appropriate tube-well casings and
submersible motors are to be installed for maximum continuous recovery of groundwater for uninteruped water supply by improving the efficiency of
the tube well. During well assembly cleaning of boreholes and directional drill holes are very important.

CASE STUDIES

Similar types of projects were already implemented by Tamil Nadu State Government in Muthulapuram village of Ettaiyapuram Taluka, Tuticorin
District for Social Rural Agriculture development programme in early 1970. Two dug wells of radius 6m each and 7 and 8 m depths respectively were
sunk about 2 km apart each other in the Vaippar River bed for each 5 H.P. motor was fitted. Pipe lines were connected to supply water for 10 ha
command area. Initially two groups of 7 landowners used the water for their agriculture. During the course of over 30 years of duration 7 member
groups are reduced only 2 or 3 landowners who are monopolizing the wells. Failure is also due to some of the members leaving their land as
uncultivated. However the success of the project is rated to 40% and it is an achievement and it is still continuing its application. Similarly in 1988 the
Indo-Dutch Tubewell Project (IDTP) carried out in Uttar Pradesh, India (Alberts, 1998) with funding of project cost of Rs. 54.18 Crores in
collaboration with Netherlands and Government of India. About 547 tube wells up to a depth of 250 m bgl were sunk in 776 km2 in the villages of
Sultanpur, Faizabad, Basti, Gorakpur, Deoria, Ballia, Bahraiach and Gada. The tube wells of 25 cm diameters yield water supply of 150m3/hr with a
command area of 100ha. The project was constrained with inadequate power supply, poor maintenance and lack of public awareness. Yet the project
was successful and achieved 70% of the planned construction targets. About 130000 persons had potential access to public tube wells water making
significant achievement in green revolution in Uttar Pradesh State. The failure may be rectified by the formation of Village Farmers Forum. The
landowners migrated from villages may be treated as sleeping partners and their lands may be cultivated by giving due share at the rate of Rs 25000 /
ha /year to be given to the landowners by the Village Farmers’ Society. Collective forming reduces agriculture expenditures considerably.

CONCLUSION
In order to suggest rural agricultural development, in Malattar Basin, a model village of Vedanattam near the coast is chosen as a pilot scheme study
proposal. Preliminary field work and geological studies were carried out and site specific 30 tube wells are proposed successively to be sunk for a
depth ranging between 100 m and 200 m for a scheme for agriculture based economic development for 300 hectares in this village. According to this
scheme, 30 tube wells to be sunk can supply necessary groundwater for dry crops like cotton, sweet potato, chilli, bajira, maize, black and green grams
with water supply of 40 m3 /hr / well of10 ha command. Monsoon rainfalls add increase in crop production. This would raise annual net profit per
hectare. If two members of a family of 5 individuals with a hectare land work, the income would be sufficiently enhanced 150 thousand per annum to
run a family without any financial problem. A land owner having 1 or 2 ha of land rarely encounters high yield aquifer between 100 and 200m bgl in
his land. Increase of dimension to 25 cm of tube wells, it is possible

291 
 
Table 1 Project Investment Cost and Repayment of Loan

S.No Particulars Details Amount ( Rs)


1. Project Investment 30 tube wells 100-200m 12 l/s 100,00,000
2 30 tube wells 20 hrs pumping @ 300 days @Rs10/hr 40 c.m /hr 18,00,000
3 Maintenance 20,000
4 5 part-time labours @ Rs 3000 1,80,000
5 Interest on investment @10% + annual payment of loan 16,00,000
Payment of loan in 12 years duration

Table 2 Annual Profits and Expenditures Statement

S.No Particulars Details Amount ( Rs)


1 1 Tube well command area 10 ha investment 40 c.m /hr 3,00,000
2 Water supply 40 c.m 20 hr/day 2.5ha/day @ Rs 10/hr Shift irrigation
3 3.2 cm water column/sq.m 5 hrs/day once in 3 days Two crop/year 6000
4 75 rice bags/ha 100kg/bag Rs 15/kg Twice/year Wet crops 2,25,000
5 Net profit Expt Rs100000 1,25,000
6 2 cm water column/sq.m 5 hrs/day once in 5 days Dry crops -3/y 2,700
7 30 black gram bags/ha 50kg/bag Rs 40/kg 3 crops 1,80,000
8 Net profit Expt Rs 60000 1,20,000
A family owning 1 ha works with 2 persons income 120000+60000 1,80,000

Table 3 Economic Uplift of Village Community

S.No Particulars Individuals


1 Farmers owing 300 ha land 300 families (A family 5 individuals) 1500
2 Agriculture labours family- individuals (1000) 5000
3 Supporting families –infrastructure involving other professions 1500 7500
4 Peripheral supporting families 5000 Schools, shops 5000 25000
5 Floating population and development 11000
Total beneficiaries 50,000

to locate good aquifers within 100m bgl yielding over 150m3 water to command 30 ha and the numbers of tube wells may be reduced to 10 for the
same investment. However, it depends on wet fracture pattern of the terrain which can be predicted by interpretation of LWD measurement data.
Hence, groundwater exploration along the course of streamlets is stressed. By this scheme, the Vedanattam village will economically support
population of 50,000 individuals (Table 1, 2, and 3). Formation of Village Farmer’s Society is essential to explore and exploit groundwater along
streamlets which normally form along fault lineaments with potential piezometric surface at shallower level in Government lands. Since, the
groundwater potential is very high in the subsurface column; no environmental degradation will be raised. Government may help the village-
community for exploring and sinking public tube wells in such Government lands. Since the tube well occupies less than 1 m2 of land in the bed of
streamlet, no adverse environmental impact will be encountered. The Groundwater Division of Government of Tamil Nadu must equip with modern
drilling and LWD units to explore groundwater particularly along streamlets lying along deep fractures. Every village has such type of nallahs and dry
streamlets with high potential shallow aquifers. Such projects may be extended along the course of streamlets and geological structural fractures for
green revolution and economic development in rural areas of the country.

REFERENCES

Alberts, J.H. (1998) “Public Tube well irrigation in Uttar Pradesh India”, International Institute for Land Reclamation and Improvement, Wageninen,
The Netherlands, pp. 1-74
Burnett, W.C., Agarwal, P.K., Aureli, A., Bokuniwicz, H., Cable, J.E., Charette, M.A., Kontar, E. Krupa, S., Kulkarni, K.M., Loveless, A., Moore,
W.S., Oberdorfer, J.A., Oliveira, J., Ozyurt, N., Povinec, P., Privitera, A.M.G., Rajar, R., Ramessur, R.T., Scholten, J., Stieglitz, T.,
Taniguchi, M. and Turner, J.V. (2006) “Quantifying submarine groundwater discharge in the coastal zone via multiple methods”, Science of the
Total Environment, vol. 367, 2-3, pp. 498-543
Balachandran, A. (2009) “Technical Report Series Groundwater Brochure Thootukudi District”, Tamil Nadu, CGWB, Chennai, pp. 1-21
Wilson, R.C.L, Drury, S.A. and Chapman, J.L. (2000), The Great Ice Age, Climate Change and Life, London, Routledge.
Survey of India (1997)”District Resource Planning Map” Tuticorin District, 1997, Dehra Dun.
Mondal, N.C. , Singh, K.S. and Rangarajan, R. (2009) “Aquifer characteristics and its modeling around an industrial complex modeling around an
Industrial complex, Tuticorin, Tamil Nadu”: A Case study, J. Earth Syst.Sci. 118, No. 3, pp.231-244
Ramasamy, R. (1988) “Concept of modeling for a universal piezometric surface of Regional aquifer system of Indian Sub-Continent”, Nat. Seminar
on Hydrological Teaching and Development in the National Water Resources, Andhra Univ, Waltair, Abst. vol. pp. 33-36

292 
     
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Drilling Problems in Lakadong Sandstone – Practical Experience at the Captive Limestone


Mine of M/s ACL, Meghalaya
P. Pal Roy1 and C. Sawmliana2
Central Institute of Mining & Fuel Research (CIMFR)
Dhanbad, Jharkhand, India
E-mail: 1ppalroy@yahoo.com, 2sawmliana@yahoo.co.uk

ABSTRACT

A severe drilling problem was faced with Lakadong sandstone at the captive limestone mine of M/s Adhunik Cement Limited, Meghalaya. The mine
encountered a thick bed of 5-6 m, fine to medium grained Lakadong sandstone, overlying the Lakadong limestone. Removal of overburden was
carried out by drilling and blasting method to expose the limestone deposits. Pneumatic crawler drill (ROC 203 PC Model) of Atlas Copco, down-
the-hole hammer with 115 mm drill diameter, button-bit type was used for drilling in both sandstone and limestone deposits. During the drilling
operation in Lakadong sandstone, it was observed that the life of drill bit was only 100 - 130 m whereas its life in other sandstone and limestone
strata was more than 600 m. The Lakadong sandstone had the density, compressive strength and Young’s modulus in the range of (2.4-2.6) g/cc,
(45.48-75.43) MPa and (9.32-12.50 GPa) respectively. The Schmidt Rebound hardness values ranged between 54 and 68 whereas Mohr Scale
Hardness varied between 7.5 and 8.5. The average primary and secondary wave velocities were 5332 m/s and 2497 m/s respectively. The evaluated
values of the Lakadong sandstone varied between 0.65 and 0.78. The main reason for premature damage of drill bits in Lakadong sandstone was
found to be higher strength properties with higher silica content causing more abrasiveness of the rock. Modification in the drill bit design along
with its feed force and rotation speed was made to increase the life of drill bits for Lakadong sandstone. The paper briefs the scientific approach to
get the job done in order to overcome the problem in an amicable manner. A generalised approach is also suggested to encounter such practical
problems being faced by the mining industry.

KEY WORDS: Bit wear; abrasiveness, hardness, compressive strength, feed force, rotation speed, flushing, bit design

INTRODUCTION

Adhunik Cement Limestone Mine is located in Thangskai village, Lumshnong of Jaintia Hills District in Meghalaya, India. The total leasehold area
of the mine is 49000 sq. mts. with the total reserve of 4.58 million tonne of limestone. The area is situated on the western side of the National
Highway No. 44 running from Shillong to Silchar and south-west to Lumshnong village. It is well connected through rail and road. The nearest
railway station is 80 km from the mine at Badarpur on Guwahati - Silchar section of NEF Railway. The nearest airport is Silchar, 125 km and
Guwahati, about 225 km away from the site. The mining leasehold area is surrounded by rugged and mountainous topography. The area is
surrounded by bushes. The overview of the mine is given in Figure 1. In the mining lease area, Lakadong sandstone beds were present between
Umlah Limestone and Lakadong Limestone deposits. The thickness of Lakadong sandstone beds varied between 5 and 6 m. Removal of these
sandstone beds were required to expose and extract the Lakadong limestone. Removal of the overburden as well as extraction of limestone was
carried out using drilling and blasting operations. Pneumatic crawler-mounted drill machines (ROC 203 PC Model) of Atlas Copco, down-the-hole
hammer with 115 mm drill diameter, button-bit was used for drilling in sandstone as well as in limestone. Blasthole depth varied between 5.0 to 10.0
m and spacing between holes varied between 3.5 to 4.0 m. An explosive charge factor of about 0.60 – 0.65 kg/m3 was generally used to obtain
desired fragmentation for Lakadong sandstone formation.

A practical drilling problem was encountered in the Lakadong sandstone deposits where frequent changing of drill-bits was required due to frequent
wearing of the drill-bit. Life of the drill-bit in Lakadong sandstone was found to be about 100 -130 m only, whereas the expected life was more than
600 m. The possible causes of premature wear of the drill-bit in the Lakadong sandstone were thoroughly investigated and suitable recommendations
were made to overcome the problem. This paper briefly describes the scientific approach adopted during field investigations to overcome the said
problem.

Fig. 1: View of Adhunik Cement Limestone Mine, Jaintia Hills, Meghalaya

FACTORS AFFECTING WEARING OF DRILL BITS

There are many factors that affect the wear and tear of a drill-bit. Such factors can be broadly divided into three categories viz. (i) geological; (ii)
machine-related and (iii) operational. Rock and rock mass properties like hardness, strength, elasticity, plasticity, abrasiveness, texture, schistosity,
bedding planes, joints diabase etc. are geological factors that affect the wear of the bit as well as rate of penetration or drillability of rock. Ralf et al.
(2003) classified different wear types for button-bit being used for hydraulic rotary percussive drill on the basis of rock properties as below:

(1) Normal Wear – Observed when tool-body and hard metal insertions are more or less evenly abrasive wear. This wear type is typical for
abrasive rocks with high compressive strength e.g. quartzite gneiss, granite or quartzitic sandstone.

(2) Wear of Tool-Body with possible Breaking out of Buttons – This phenomenon occurs while drilling in weak rock as generally observed in
poorly-cemented or weathered sandstone, sandy-marlstone or weather granite and gneiss. In this rock, the bit penetrates deep into the rock mass and
produces lot of debris, so that both steel body and buttons are more or less evenly exposed to hard-rock and debris.

(3) Wear of the diameter – A typical phenomenon for unstable or highly stressed abrasive rock mass when the drillhole deformed during drilling.
Abrasive rock material forces the bit from the drillhole walls. First, the peripheral buttons begin to show wear on the outer side, later the tool-body
itself gets affected, the diameter of the bit starts reducing and ultimately the peripheral buttons break.

(4) Macroscopic failure of button - Mostly influence by the rock’s strength and fabric, properties of the rock mass, machinery, tools and support
method.

Abrasivity or abrasiveness of the rock is considered to be a major impediment and main factor which controls the rate of wear that ultimately affects
the penetration rate. It is one of the important properties of rock that is measured in order to assess their suitability for mechanical and economical
excavation. Abrasiveness is the capacity of the rock to wear away the contact surface of another body that is harder, in rubbing or abrasive process
during movement. The abrasiveness of rock increased with hardness of grain (quartz content), angular grain and its porosity. Various researchers
such as, Singh (1968 & 1973), Shiromura and Takata (1958), Farmer et al. (1979), Miranda and Mendes (1983) etc. reported that bit wear increased
with abrasivity and consequently penetration of the drag-bit decreased. There are different methods available for determination of abrasiveness of
rock. Rocks with more silica content are usually having higher abrasiveness. The most common methods are Burbank Test, Microbit Drilling Test
and Cerchar Test.

The machine operational factors include tool characteristics, flushing, feed, rotating velocity, temperature, tool handling etc. Failure of a bit-shaft is
mainly as a result of manufacturing problems or bad handling. Wearing of the bit may occur due to insufficient flushing system. If no or insufficient
flushing is available, thermal wear may occur and the wear types equal those of abrasive wear and wear due to material failure (Ralf et al., 2003).
The velocity of air-flow for efficient cleaning with air generally goes from 15 to 30 m/s. The required air velocity depends on rock density and
diameter of the rock particle (i.e. drill chip). The minimum air velocity may be calculated as:

294
va = 9.55 (1)

Where,
va = velocity of air flow (m/s),
ρr = rock density (g/cm3) and
Dp = diameter of particle (mm).

From equation (1), the quantity of air-flow that should be supplied by the compressor can be calculated as:

(2)

Where,
Qa = quantity of air flow (m3/min),
D = blasthole diameter (m) and
d = diameter of the rods (m).

Feed (thrust load), rotation speed and percussion are the important functions affecting the penetration rate, bit wear and the overall drilling efficiency.
Thrust load is necessary to maintain the drill bit in continuous contact with the rock at the hole bottom. Underthrusting causes excessive bit wear
because the load on the bit teeth does not exceed the compressive strength of rock and drilling progress results from wear and minor crushing on the
surface (Singh, 1973). Overthrusting load causes clogging which reduces the penetration rate and also tends to cause rods to bend and hole to deviate.
As far as wear of the bit is concerned, a low rotation speed was preferable. When drilling with insert bits, the most common rotation speed oscillates
between 80 and 150 rpm with angles between indentations of 10 to 20o. For button-bits from 51 to 89 mm, the speed should be lower, between 40
and 60 rpm, that brings turning angles between 5 and 7o. Bits of larger diameters require even lower speed (Jimeno et al. 1997).

DETERMINATION OF ROCK PROPERTIES

Various tests were conducted during field studies as well as in the laboratory to know the important properties of Lakadong sandstone. These include
rebound hardness values using Schmidt Hammer Instrument, Mohs’ Scale Hardness of the rock, compressive strength, tensile strength, Young’s
Modulus, primary (P) and secondary (S) wave velocities and indirect test of abrasiveness etc.

In order to determine the rebound hardness using Schmidt Hammer in the field, the surface of the rock was cleaned and dressed properly. The surface
was then divided into a grid (15 - 20 mm size) of at least 25 points as shown in Figure 2. Impact values were recorded using N-Type Hammer in each
grid. Single impact in each grid was considered and any test that causes cracking or any other visible failure on the rock mass was rejected. The
average of the 10 highest readings was taken. The Schmidt Rebound hardness values tested on the Lakadong sandstone varied between 54 and 68
with the average value of 64.

Fig. 2: Testing of rebound hardness of Lakadong Sandstone using Schmidt Hammer Instrument

The rock specimen was also tested for Moh’s Scale of hardness as well as primary and secondary wave velocities. Table 1 shows the results of Moh’s
Scale of hardness, P-wave and S-wave velocities of Lakadong Sandstone. Mohs’ Scale of hardness varied between 7.5 and 8.5. The average primary

295
and secondary wave velocities were 5332 m/s and 2497 m/s respectively. The strength properties of Lakadong sandstone such as compressive
strength, tensile strength, Young’s Modulus and Poison’s ratio were determined in the laboratory. Core samples were collected from different
boreholes. The different properties of Lakadong sandstone determined in the laboratory for different boreholes are given in Table 2. Compressive
strength varied between 45.48 and 75.43 MPa and tensile strength varied from 2.29 to 4.74 MPa. Poisson’s ratio and specific gravity varied from
0.12 to 0.32 and 2.39 to 2.81 respectively.

Table 1: P and S wave velocities and Moh’s Scale of hardness values of Lakadong sandstone
Specimen No. Borehole Rock Type P-wave S-wave Velocity Hardness on
No. Velocity (m/s) Mohr’s scale
(m/s)
12/1 BH 51 Lakadong Limestone 5332.49 2497.27 8.5
8/1 - Lakadong Limestone 6532.77 2755.26 8.5
11/1 BH 50 Lakadong Limestone 6623.80 2515.30 8.5
10/1 BH 50 Lakadong Sandstone 2632.06 1671.34 7.5
1/4 - Lakadong Sandstone 2331.58 1479.23 7.5
7/1 BH 46 Lakadong Sandstone 4340.31 2300.18 7.5

Table 2: Properties of Lakadong sandstone Tested in the laboratory


BH No. Run Rock Type Compressive Density (dry) Specific Tensile Young’s Poisson’s
Strength gravity strength Modulus Ratio
(m) (MPa) Avg. (kg/cm3) Avg. Avg. (MPa) (GPa)
7.00 Lakadong Sst. 50.59 2.45 2.46 10.51 0.14
to - 45.48 2.42 2.43 2.43 2.44 - - -
9.00 40.36 2.45 2.45 8.68 0.12
- 2.43 2.44 - -
9.00 Lakadong Sst. 63.52 2.45 2.46 12.50 0.32
BH#10 To - 54.21 2.45 2.44 2.46 2.46 - - -
10.00 44.52 2.41 2.42 9.50 0.23
10.00 - 10.20 Lakadong Sst. 53.05 53.05 2.43 2.44 2.45 2.45 2.29 11.65 -
11.80 -12.00 Lakadong Sst. 51.38 51.38 2.38 2.48 2.39 2.39 3.49 10.90 -
BH#49 41.00 - 42.50 Lakadong Sst. - - 2.66 2.66 2.67 2.67 4.74 - -
13.70 - 14.00 Lakadong - - 2.44 2.44 2.45 2.45 3.90 - -
BH#49 Sst.(Block 2)
18.30 – 18.50 Lakadong - - 2.79 2.79 2.81 2.81 - - -
Sst.(Block 2)

The abrasiveness of the Lakadong sandstone was determined indirectly using the relationships between different types of rock abrasiveness and their
strength properties derived by Giri et al. (1997). The authors correlated uniaxial compressive strength and tensile strength with the abrasiveness of
rocks determined through Cerchar Test. The relationships between compressive strength and tensile strength for different kinds of rocks with
abrasiveness are given by:

A= (3)
A= (4)

Where,
A = abrasivity of rock determined from Carcher Test,
σc = uniaxial compressive strength of rock (MPa) and,
σt = tensile strength of rock (MPa).

From the Eqs. 3 and 4, the abrasiveness of Lakadong sandstone comes to 0.65 to 0.78. The Lakadong sandstone present in the mine also showed
higher percentage of silica as well as angular grain formation.

OBSERVATIONS AND DISCUSSIONS

The nature of bit wear observed in the field was similar to the ‘Normal Wear’ described by Ralf et al. (2002). This type of wear was observed for
rocks of high abrasiveness and high compressive strength. Figure 3 shows the carbide button-bit where uniform wear was observed at the periphery
of the bit. More or less uniform wear was observed for those button-bits next to the periphery of the drill bit. Lesser wear was observed in the central
button-bit. Maximum worn was observed in those button-bits located at the periphery of the diameter where more bit-to-rock contact occurred during
drilling.

296
Fig. 3: View of button-bit wearing after drilling in Lakadong sandstone

The size of the drill-chips coming out from the blasthole during drilling showed a mixture of finer-grained as well as coarser-grained mixtures
(Figure 4). The visual observation of the drill-cutting chips could also reveal more silica and quartz contents in the Lakadong sandstone. Due to
higher specific gravity of the rock, high air velocity and more quantity of air would be required for flushing the hole effectively. The thermal wear
due to insufficient flushing could contribute wearing of the bit to a large extent. Taking 0.25 mm as the average particle size of the drill-cuttings in
the Lakadong sandstone, the approximate air velocity for flushing as estimated from Eq.1 was 29.50 m/s. This indicated that higher air velocity and
more quantity of air would be required for effective flushing of the hole.

Fig 4: View of drill cuttings obtained in Lakadong sandstone

The feed force of ROC 203 PC-DTH (Atlas Copco) was 8 kN with working pressure of 6-25 bar. The rotation speed varied between 25 and 80 rpm.
Various rotation speeds and working pressures were tested in the field in order to examine their effects on the bit wear while drilling in Lakadong
sandstone. However, it was observed that change in rotation speed and feed force could not provide significant improvement in reducing the bit wear.
No significant change was observed in the bit life. Modification of the drill bit design, use of stronger tungsten carbide/cobalt materials as well as
careful assessment of rotation speed and feed force could increase the bit life in Lakadong sandstone. Apart from these, followings are a few points
for achieving economy with button drill bits as suggested by Chugh (1985):

297
ƒ Ensure that compressed air with required pressure is supplied to the hammer for proper hammering and flushing.
ƒ It is essential on the part of operator to ensure proper feed and rotation depending on the formation of rock encounter.
ƒ While drilling in abrasive rock formation and at the sign of moisture in hard rock, use water or foam injection to prevent collaring and to
keep the borehole clean.
ƒ The most economical rotating speed is that which gives the highest rate of penetration without excessive bit wear and therefore depends on
specific field conditions.
ƒ The bit, after its use, should be regraded at proper intervals.

CONCLUSIONS

The main reason for increased rate of bit wear and short life in Lakadong sandstone was observed to be higher strength properties as well as high
abrasiveness of the rock. Angular grain formation and higher content of quartz and silica of the Lakadong sandstone increased the abrasiveness of the
rock. Use of higher strength properties of tungsten carbide/cobalt bits as well as careful assessment of feed force, rotation speed and flushing
efficiency could increase the bit life in the Lakadong sandstone rock formation.

REFERENCES

Chugh, C. P., 1985. Manual of Drilling Technology, Pub. Oxonian Press Pvt. Ltd., New Delhi, 567 p.
Farmer, I. W., Hignett, H. I. and Hudson, J. A., 1979. The role of geotechnical factors in the cutting performance of tunnelling mechanics in rocks,
Proc. 5th Congress of International Society Rock Mechanics, Montreux, 1:371-377.
Giri, A. K., Sawmliana, C., Singh, T. N. and Singh, D. P., 1997. Strength properties and their relations with abrasiveness of some Indian rocks, 1st
Asian Rock Mechanics Symposium-A Regional Conference of ISRM, Seol, Korea, 13-15 Oct., pp. 153-157.
Jimeno, C. L., E.L. Jimeno, and F. J. A. Carcedo. 1995. Drilling and Blasting of Rock. Geo-mining Technological Institute of Spain, ISBN 90 5410
1997, Pub. A. A. Balkema/Rotterdam/1995.
Miranda, M. A. and Mendes, F., 1983. Drillability and drilling methods, Procs. International Society of Rock Mechanics, 5:E195-200.
Ralf J. Plinninger, Spaun, G. and Thuro, Kurosch, 2002. Predicting tool wear in drill and blast, Tunnels & Tunnelling International Magazine, April.
Shiromura, Y. and Takata, A., 1958. Research on relation of physical properties of rocks and drilling rate of diamond bit, Mining & Metallurgical
Institute, Japan, 74 (844): 854-860.
Singh, D. P., 1968. Drillability and physical properties of rocks, M. Sc. Thesis, University of Melbourne, Australia.
Singh, D. P., 1973. A study of drill bit wear in rotary drag drilling, Australia Mining: 17-20.
Copyright ©2010 Indian Institute of Technology Madras, Chennai, India. All rights reserved.

298
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Cementing Practices in In-Situ Combustion wells to Enhance Heavy-Oil Recovery.


Pradip Chakravarty*, K.M.Suyan and S. Dutta
Institute of Drilling Technology, ONGC, Dehradun, India
*e-mail: pradipchak2006@rediffmail.com

ABSTRACT

In the heavy oil fields the primary recovery are very low due to high viscosity of oil. The secondary recovery with the application of in-situ
combustion (ISC) process, that is based on the principle of improving oil-mobility by reducing the viscosity of oil by heating it within the reservoir.
The heavy oil belt of Mehsana Asset of ONGC comprising of Santhal, Balol, Lanwa and Becharji fields discovered three decade earlier covers an
area of about 70 Sq.Km and it has an in-place oil of around 150 MMT having API gravity of about 13-17 0. Due to high viscosity of oil and presence
of active aquifer, the primary recovery from all these fields were low in the range of 6—15% of Oil in place (OIIP). However, the application of in-
situ combustion (ISC) process has enhanced the recovery more than double. Completion of Fireflood wells present great challenges of wellbore
integrity to maintain during the ISC operation. Since, cement sheath behind casing has to provide stability and durability at high temperature thermal
cycling when exposed to fire zone. Portland cement (API Class G) with 35% silica flour used earlier to provide thermal stability upto 3000C. A higher
temperatures(>3000C) the cement sheath of Portland cement disintegrated and increases the gas permeability and decreases the compressive strength
which allowed the flue gas to charge the shallow water table. This paper presents a new process for development of a cementing composition to
provide ductility and thermal stability to the cement sheath in thermal recovery wells subjected to high temperature (4500C) thermal cycling during
ISC process. This formulation has been used successfully in many wells of Balol & Santhal and Lanwa fields of ONGC in Mehsana asset and the
problems of shallow water table charging and blow-out at shallow depth during drilling has been minimized by application of thermally stable
cement.. This paper describes the details of formulation, application and success story of field validation in ONGC oil fields.

KEY WORDS: Heavy Oil; In-Situ Combustion; Thermal Cement Slurry; Mehsana ONGC; Portland Cement.

INTRODUCTION

More than 6 trillion barrels of oil in place have been attributed to the heaviest hydrocarbons and has been recently become an important resource of
energy as conventional oil reservoirs are in declined. Heavy oil is any type of crude oil / hydrocarbons which does not flow easily. It is referred to as
"heavy" because its density or specific gravity is higher than of light crude oil. Heavy crude oil has been defined as any liquid petroleum with an API
gravity less than 20°, meaning that its specific gravity is greater than 0.933.

Mehsana asset, located in Northern part of Gujarat State, is having oil fields producing both heavy and light crude o i l with API gravity
0
ranging from 13-42 . Heavy oil fields at Mehsana were d i s c o v e r e d during 1970-71 and covers an area of about 70 sq km contains
around 150 MMT OIIP. Heavy oils are characterized by high viscosity that increases as API gravity decreases, l o w h y d r o g e n /carbon
ratios, l o w gas/oil ratio, as well as significant asphaltenes, sulfur and heavy metals. In short, heavy oil reservoirs generally present low-energy
and low productivity wells. These characteristics make recovery challenging. The poor primary recovery and less effective -artificial
lift method necessitated search for suitable enhanced oil recovery method for improving ultimate recovery from these fields-In-situ
combustion technique.  

IN-SITU COMBUSTION PROCESS

In-situ combustion process is an enhanced oil recovery process to improve the recovery of heavy crude oil. The process is based on the principle
of heating the crude within the reservoir using part of the reservoir oil as fuel for combustion. The processes of in-situ combustion are initiated
through injector wells and allow reservoir oil to move to-wards oil producing well ( f i g - 1 ) . The reservoir temperature at the fire-front end may
reach around 650-7500C for some hours and as front move away from the well mouth, temperature may normally fluctuate upto 450-5000C for long
time (1-2). It is achieved by injecting heat in the form of hot air for some pre-calculated period of time, with the help of gas burners or
electric heaters lowered in the well. Air is injected into the formation to provide oxygen for continuous combustion of the oil in the reservoir. This
process is very much effective as a small amount of heavier fraction of in-place oil is consumed while the rest is d i sp l a ce d towards p r o d u c e r s .
In order to utilize enthalpy of t h e burnt s a n d behind the advancing combustion front and to improve the sweep efficiency, wet
combustion phase is also carried out. It involves injection of a calculated volume of air in a cycle of six days followed by one day water.
During in situ process periodical quenching combustion front by water causes temperature falls as a result of which the cement behind the casing are
subjected to thermal cycling.

The success of in-situ combustion process depends on the reservoir and geological considerations. Since in situ combustion is an inter-well drive
process of good horizontal continuity and is critical for the success of the project. Lack of good sand continuity, fractures and joint trends may create
preferential flow channels causes of failure of in-situ process. Gaps in formation overburden or leaky inter-zonal seals in stratified reservoirs can allow
fluid to leak into overlying strata and reduce the effectiveness of the ISC process.

CEMENTATION PRACTICES ADOPTED –WORLDWIDE FOR IN-SITU COMBUSTION WELLS

Thermal recovery wells are always cemented to surface to prevent casing failure and to prevent the loss of gas (combustion gases) into overlying
permeable layers through the uncemented casing open hole annulus. Moreover because of thermal expansion, high levels of stress are built up in the pipe
and cement sheath and therefore the strongest possible pipe/cement and cement/formation bonds are necessary. Failure of bonds could allow inter-zonal
communication and pipe expansion. The ultimate results would be casing failure by buckling or telescopic. Special well completions are required in ISC
wells as the cement behind the casing is exposed to extreme stresses due to thermal cycling as shown in figure-2. Under the temperature cycling
conditions conventional portland slurries with 35% silica flour fails to retain its strength and permeability for long exposure at temperature above 320 Deg.
The requirement is obviously a thermally stable slurry to withstand the rigors of temperature cycling in the temperature range of 400-5000C experienced
for long time during ISC process.

World wide, for cementation of thermal recovery wells, alumina cement slurry are being used as tail slurry for placement against the combustion zone
and cove above 150 ft for withstanding the high temperature cycling during the in situ combustion process. Since formations are fragile and un-
consolidate so hydrostatic head will be low (3-4). Therefore hydrostatic head of the cement column has to keep low to avoid the slurry loss in to
formation during placement. Portland cement slurries of low sp gr are pumped ahead of the alumina slurry as lead slurry to cover the upper section of
the well upto surface. But the major problems of using alumina cement are difficulty in retardation as well as compatibility with portland cement
which may result in flash set without permitting minimum operational time to complete the cementation job if sufficient precaution is not taken during
the job. To avoid flash setting due to intermixing during placement of two incompatible materials such as Alumina Cement and Portland cement, even
if various special precautions are taken, the serious risks involved were unavoidable (5-7).

WELL COMPLETION

Presently the ISC process in ONGC is mainly confined to heavy oil belt of western onshore fields only where the depth of the wells varies from 1100-
1200mt. The wells are generally completed by lowering 9-5/8” surface casing and then 7” final production casing upto the drilled depth. Formations
encountered in these wells are usually problematic having weak and unconsolidated zones with low fracture pressures. Earlier cementation pracitce of 7”
casing in these fields were generally designed to take cement to surface with 40% silica flour slurry. However, in this type of completion of ISC wells, two
major constraints were being faced. Firstly, due to low fracture gradient, losses were often encountered during cementing with conventional normal density
slurry and desired cement rise upto surface could not be obtained in most of the wells. Secondly, the set cement sheath of the silica flour slurries failed to
contain the strength and permeability for exposure at high temperature near the combustion zone especially when associated with thermal cycling in course
of the ISC process. This two combined adverse effect has resulted in a major problem of charging of the subsurface shallow layers during the in-situ
combustion process. (Fig-2)

The fallout of this adverse effect is that many incidents of near blowout situations at shallow depths were observed during drilling and work-over
operations in and around the EOR affected areas. Due to this problem of emission of gases from the charged shallower zones, the air injection rate has
to be reduced. Though this had lessen the gas migration activity but the ISC process and its impact on oil production had to be partially sacrificed for
operational safety.

APPROACH TO WELL COMPLETIONS

Earlier the production casing cementation of EOR wells were generally designed to take cement to surface with 35% silica flour but as a result of low
fracture gradient, lost circulation problems were often encountered during cementing and desired cement rise upto surface could not be obtained in
most of the wells. Cement slurry used in earlier completed wells with silica flour were evaluated under in-situ combustion conditions. It was observed
that wells completed with Portland cement plus 35%silica flour extends thermal stability upto250-3000C keeping compressive strength and
permeability intact. Laboratory studies of silica flour slurry cured at 700C and thermal cycled to about 3500C confirmed the visible large cracks or in
some cases were broken into pieces (Fig-3).

300 

 
This study reiterated that the conventional slurries with silica flour used
earlier for cementation were not suitable for ISC process and this is the main
reason of charging of the subsurface shallow layers during the in-situ
combustion process.

Since high alumina cement does not provide sufficient time for placement
with Portland cement causes flash setting or hardening in minutes. Therefore,
Moulds No-I, 2, 3 thermal cycled upto 300 deg C the object of the present study was to find out an unique cementing solution to
Moulds No-4, 5, 6 thermal cycled upto 400 deg C this problem, the innovative approach adopted was to provide a cement slurry
with API Class G cement capable of retaining sufficient compressive strength
and low liquid permeability and thermally stable upto 4500C thermal cycling
for in-situ combustion wells(8).

LABORATORY STUDIES- PROCEDURES ADOPTED

Four different compositions of cement slurries were prepared with API Class
G cement blended with graphite and metakaolin as per API specification-10B
Fig – 3: Evaluation of Portland cement Slurry with Silica at procedure, by using the branded commercially available dispersant, retarder
High Temperature Cycling
and fluid loss additives to get the desired parameters.

The cement molds (2”x2”) were cured at 700C and 3000psi pressure for 24 and 96 hrs and compressive strength recorded. These molds were removed
from curing chamber and kept in muffle furnace at 1000C for 24hrs. After drying 24 hrs in muffle furnace, temperature was raised to 3500C (dry heat)
and was maintained for 08 hrs, thereafter started cooling to attend temperature 70-800C in 12hrs. This process of thermal cycling was carried out for
three consecutive days. After three days, compressive strength and permeability were recorded. Another set of molds were subjected further for higher
temperature (4500C) thermal cycling in the similar fashion as carried out for temperature (3500C) for three consecutive days. All molds were removed
from furnace after completion of high temperature (4500C) thermal cycling and determined permeability and compressive strength (Table-1).

RESULTS AND DISCUSSIONS

Cement slurries of different compositions were designed to evaluate the effect of high temperature thermal cycling on compressive strength and liquid
permeability. Only two such compositions has been taken here for discussion. All additives were taken percent by weight of cement and were dry
mixed in cement except dispersant and defoamer in water. All cement slurries were evaluated as;

Cement Slurry Composition -1

The cement slurry composition as mentioned below was prepared and cured in curing chamber at 700C for 96 hrs and determined the water
permeability which is less than 0.1md.

Cement Water Metakaolin Graphite CD-32 n-TBP

100 65 12.5 15 0.6 0.1

The study of high temperature thermal cycling was carried out from 70↔350↔4500C. It is observed that no appreciable change in compressive
strength was found (Table-1). However the liquid permeability was increased from 0.1md to 2.0md. It may be due to dehydration of calcium
hydroxide gel to lime at high temperature. Consequent upon it may develop connecting porosity to enhance permeability. Therefore this formulation
has revealed that dehydration of calcium hydroxide gel at higher temperature has not affect the set cement matrix permeability during thermal cycling.
Metakolin provides high temperature stability and graphite provides high temperature durability and ductility. Both aggregates were further optimizing
to get the good rheological slurry. During hydration process, cement releases the calcium hydroxide which reacts with silica to form additional CSH
gel resulting reducing permeability and increasing compressive strength. Another cement slurry composition-2 was designed with high solids to check
the effect of thermal cycling.

Cement Slurry Composition -2

Furthermore, cement slurry composition as stated below was formulated with reduced water content / optimized by keeping graphite and metakaolin
again constant.

Cement Water Metakaolin Graphite CD-32 N-TBP

100 65 35 15 1.6 0.1

The cured cement molds were processed for thermal cycling upto 4500C. It is observed that the set cement sheath has retained the liquid permeability
and compressive strength upto 4500C temperature cycling (table-1).

301 

 
Metakaolin proportion (10-15%) in cement system must be effective to impart resistance to retrogression of the strength and permeability due to the
effects of high temperature thermal cycling. The cement composition containing metakolin 30-40% and graphite 10-15% by weight of cement
provides excellent formulation for in-situ combustion wells subject to water contain is low. The slurry thus has pumpable consistency till pumped in
conventional fashion into the portion of the well to be cemented. Accordingly, the greater or lesser quantity of water and dispersant may be used to get
the desired consistency and density of the slurry.

Above studies revealed that cementing composition no 2 is most suited for EOR wells. This cement slurry composition of sp gr 1.78 can be used as
tail slurry for placement against the fire-zone (interval) and will provide durable sealing in EOR wells. Cement slurry with this composition is
formulated and used in the EOR fields of Mehsana asset.

Lead Cement slurry

Cement slurry of sp gr 1.60 was designed as lead slurry with 50% silica fume and 90% water. The cured molds were subjected to high temperature
thermal cycling upto 2000C and thereafter determined the compressive strength and water permeability. Study revealed that no significant changes in
compressive strength and water permeability, after five thermal cycling 70-2000C, indicates designed cement slurry will provides stability and
durability below 2000C. Temperature of upper section of the well upto 400-450 mt from surface will not exceed to 200-2500C. Hence, designed light
weight lead-cement slurry of sp gr 1.60 meet the required permeability and high compressive strength.

EFFECT OF CO2

During in-situ combustion process heavy hydrocarbons generated flue gases. It is observed that flue gases contained around 14-18% carbon-dioxide
which may have the carbonation of set cement resulting increases the permeability of set cement and may allow the flue gases to charge the
subsurface. . Therefore, besides high temperature thermal cycling, the cement sheath behind the casing also comes in contact with carbon dioxide,
along with flue gas produced during the in-situ combustion process. The flue gases contained carbon dioxide which may have the long term
detrimental impact on API Class G cement as it contained lime. It was recognized that the Portland cement would convert to calcium carbonate in the
presence of carbon dioxide which may impact the zonal isolation(9).

API Class G cement contains 60 % lime and on hydration it forms CSH gel. Some free limes will also available as CaO to react carbon-di-oxide to
form calcium carbonate which is fragile and disintegrated to damage the cement sheath. To minimize the un-reacted lime in the slurry, metakolin
increases to reduce the lime in solids from 60% to 34%. Cured molds were kept for two months at 700C under carbon-di-oxide environment.
Promising results were obtained with the approach of reducing the relative concentration of Portland cement (free lime) with respect to other
aggregates in the designed slurries. With this approach, the designed slurry systems have improved long term résistance to carbonation which is also a
major factor in long-term impairment of zonal isolation. Compressive strength and permeability studies established that this cement slurry composition
would provide the high temperature thermally stable, less permeable, and more carbonation resistance cement sheath as required for in- situ
combustion wells.

FIELD IMPLEMENTATIONS
 Case – 1. (Well# X)
In this well of Balol field, 7” production casing was lowered upto 1127m. The recorded bottom hole temperature was around 700C. The high
temperature resistance tail slurry was planned to be placed against the zone to cover the annular space upto 900m from surface and the remaining
interval near to surface was planned to be covered with thermally stable slurries.

Post job computer simulation studies were performed for the well. The simulation results have established the merits of the efforts and the success of
the field implementations. A representative simulation is presented in figure-4. Excellent cement bondage was observed throughout the cemented
interval as evident from the CBL/ VDL logs (Fig - 5, 6).
Case Study – 2 (Well # Y)

In this well of Santhal field, 7” production casing was lowered upto 1131m. The bottom hole temperature recorded in this well was around 700C. The
desired rise of the high temperature resistance tail slurry to cover sufficient above the zone of interest was 800m from surface and hence the remaining
interval near to surface was planned to be covered with thermally stable lightweight lead slurry.

Post job analysis showed the top of the cement column to be at upto the surface as per the logging record. Very good cement bondage was observed
throughout the cemented interval as shown in the CBL/ VDL logs (Fig – 7 & 8).

Evaluation at the field of the developed formulation has been extremely encouraging as not only cement could be raised near to surface in these wells
but excellent cement bondage was observed throughout the cemented interval as evident from the CBL/ VDLs. In all these wells no annular
communication problem was observed till date.
CONCLUSIONS

• The thermal cement formulations developed for ISC wells are not based on alumina cement. API Class G oil well cement was used in
conjunction with indigenously available non-conventional additives. This development is not just an alternate cement formulation for ISC well

302 

 
completion but a cementing solution which has eliminated the bottleneck endangering the success of the ISC process itself in the heavy oil belt of
ONGC.
• The field implementation of the recommended formulation has been extremely encouraging as not only cement could be raised near to surface in
these well but also excellent CBL/ VDL was observed throughout the cemented interval.
• The developed formulations have been successful in addressing all the problems pertaining to cementation of ISC wells.
• The thermally stable tail slurry (1.78 sg) has no significant effect on compressive strength and liquid permeability at high temperature thermal
cycling upto 6500C, thereby making it suitable for placement against combustion zone.
• The blend of thermally stable cement slurry contained low lime (34%) in respect to cement (60%) resulting availability of free lime in set cement
mass is low. Therefore, carbonation of set mass will minimize and permeability of cement will not effect due to presence of CO2.
• The lightweight lead slurry (1.60 sg) developed sufficient compressive strength and remains thermally stable to temperature thermal cycling upto
2000C experienced at the upper section of the well during the ISC process, thereby making it suitable for placement in the upper interval of the
annulus.
• Both the developed slurries are lighter compared to industry used formulations, which would not induce losses during cementation and cement
could be raised to desired height.
• The technology is user friendly, environment friendly, cost effective and does not require additional resources than what is presently being used
in the field. The technical advantages of the slurries are definitely most attractive for their use as better alternative to present industry practices
and these advantages relates to better well completion where long term durability of cement is obtained.

ACKNOWLEDGEMENTS

The authors are grateful to management of Oil and Natural Gas Corporation for permission to present this paper. The authors would also like to thank
Cementing Services, Mehsana Asset for rendering valuable services in this implementation programme. The authors acknowledge the fine work and
contributions of the colleagues of Cementing R&D and Technology Group of IDT, especially of slurry design laboratories for providing their help and
co-operation.

REFERENCES

William M. Schulte and Arnold s. de Varies “In-situ combustion in naturally fractured Heavy Oil reservoirs, SPE-10723.
4 Ludwiq, N. C. and Pense, S. A., “Properties of Portland cement pastes cured at elevated temperatures and pressures”. Jour. of Am. Concrete Inst.
V27, No. 6, 678(1956).
Thermal Recovery Process:-SPE reprint series-7, page-427
Kalousek,G. I.,and Chow,S. Y. Research on cements for geothermal and deep oil wells, Paper SPE-5940 presented in SPE-AIME 51ST annual
conference New Orleans-1976
Portland cement”, this paper was presented in SPE of AIME International Symposium Louis H. Eilers and B. “Nelson, Effect of silica particles size on
degradation of silica stabilized on oilfield and geothermal chemistry held in Houston, Texas, Jan 22-24,1979
D.J. Daulton etc. “Successful cementing of the Horizontal In-situ combustion” GRI/DOE, SPE-21684
Erik B. Nelson, Improved cement slurry designed for thermal EOR wells, Oil and Gas Jr. Dec 1, 1986
Well cementing edited. By B. Nelson, chapter-9
K.M. Suyan, D. Dasgupta and S. P. Garg,; Novel Cement Composition for Completion of Thermal Recovery(ISC) Wellbores, SPE / IADC-
101848, Conference held at Mumbai,2006
J. C., Shen,and D. S. Pye; Effect of CO2 attack on Cement in High Temperature Aapplications, SPE / IADC-18618, Conference held in New
Orieans, Louisiana, 1989

Table-1: STUDY OF HIGH TEMPERATURE THERMAL CYCLING ON CEMENTING COMPOSITIONS

Cement slurry composition-1:-

Sp.Gravity:-1.78, Rheology at 500C:- PV=108, YP=39

Temperature Duration of Curing Cumulative Compressive Strength Liquid- Remarks, if


( oC) ( hrs) Duration of Curing (psi) Permeability any
( hrs) ( md)

70 24 24 1750 - -
70 72 96 2310 0.07 -

350 24 120 2200 0.08 Thermal


Cycling
450 24 144 1950 1.71 Thermal
Cycling

303 

 
            

Cement slurry composition -2

Sp.Gravity:-1.78, Rheology at 500C:- PV=177, YP=45

Temperature Duration of curing Cumulative Compressive Strength Liquid- Remarks, if


( oC) (hrs) Duration of Curing (psi) permeability any
(hrs) ( md)

70 24 24 2350 - -
70 72 96 2625 0.08 -
350 24 120 3150 0.11 Thermal
Cycling
450 24 144 2975 0.08 Thermal
Cycling

Fig-1: In-situ Combustion Process

C e m e n t in g – IS C W e l ls
C o n v e n t io n a l
C o m p le t io n

C h a rg e d
  A q ui f er

I n je c t o r W e l l P r o d u c e r W e ll

Fig 2 Cementation of In-situ Combustion wells.

304 

 
g/cm3
m
0.8 1.2 1.6 2.0 2.4

0.9 1.2
Well Head Pressure Hydrostatic
Acquired WHP Min. Hydrostatic

WHP (psi) (x 1000)


Max. Dynamic
Min. Dynamic
Frac
Pore

0.3 0.6
400.0

0
0 20 40 60
Time (min)

Depth = 1000 m

2
800.0

Annular Velocity

Ann. Velocity (m/s)


1
1200.0

0 20 40 60
Fluid Sequence Time (min) Dynamic Well Security  

Fig 4 Post placement simulation results shows attainment of high annular velocity for good mud removal
& proper cement bondage and also cement rise near to surface without losses with the developed
thermal slurry systems. (Well#X)
 
CBL/ VDL Recording - Well # X

Top of
cement
Zones of interest

     

 
Fig 5 Cement rise near to surface. Fig 6 Excellent cement bondage throughout the entire cemented interval

305 

 
               

CBL/ VDL Recording - Well # Y

Top of
cement

 
 

Fig 7 Cement rise near to surface Fig 8 Excellent cement bondage throughout the entire cemented interval

306 

 
Advances in Drilling Technology - E-proceedings of the First International Conference on Drilling Technology (ICDT - 2010) and National
Workshop on Manpower Development in Petroleum Engineering (NWMDPE - 2010), November 18-21, 2010.
ISBN: 978-93-80689-03-6, Copyright © 2010 by the Indian Institute of Technology Madras, Chennai (TN) - 600 036, India.

Experimental Investigation on Water Alternate Gas Injection Methodologies for Enhanced Oil Recovery
for an Indian Brownfield Application
Jigar Bhatia1, J. P. Srivastava2, Jitendra Sangwai1,3 *and Abhay Sharma1,4
1
School of Petroleum Technology, Gandhinagar, India
2
Institute of Reservoir Studies, Oil and Natural Gas Corporation, Ahmedabad, India
3
Department of Ocean Engineering, Indian Institute of Technology Madras, Chennai, India
4
Department of Mechanical Engineering, Indian Institute of Technology Hyderabad, Medak, India
E-mails: jigar_bhatia_123@yahoo.co.in, jp4200@hotmail.com, jitendrasangwai@iitm.ac.in, abhay@iith.ac.in

ABSTRACT

The present work reports laboratory investigation on water alternate gas injection methods for the enhancement of the production from a Brownfield in
India. Effect of different parameters, such as, water alternate gas (WAG) injection cycles, mainly single cycle WAG and five cycle WAG; tapered
WAG (with increasing and decreasing WAG ratio) and the change in the type of gas are studied and presented. The experiments are carried out on the
recombined fluid (oil and gas) obtained from the separator at separator conditions and recombined at the reservoir conditions of 120 oC and 230
kg/cm2 at reservoir gas-oil ratio (GOR) to become representative of the reservoir fluid. The actual core sample is obtained from the field and is used
for the study. Experiments are carried out using the core flooding apparatus for different WAG processes. Two types of gases are used, namely,
hydrocarbon gas (collected from the field and having about 90 % methane by composition in mole %) and CO2 gas. The results are compared on the
basis of incremental displacement efficiency after water flooding, effect of tapering and effect of different gases (hydrocarbon gas and CO2), etc.
During the study, it is observed that increasing the number of cycles for 1 pore volume (PV ≈ 60 cc) of fluid injection does help to increases the
production of the oil from the core sample. The number of cycles in the WAG injection process affects the recovery of oil from the core sample. An
incremental recovery of about 17 % of HCPV (Hydrocarbon Pore Volume) is observed in five cycle WAG injection process as against to about 13 %
of HCPV in single cycle WAG injection process. It is observed that the tapering in WAG injection process helps to improve the recovery of HCPV.
The gas tapering with increasing and decreasing WAG ratio gives incremental recovery of 23.92 % and 16.91 % of HCPV in the core pack,
respectively. The experiments revealed that the CO2 gas with five cycles WAG process gives an incremental recovery of 40.20 % of HCPV, which is
much higher than the five cycle WAG process using hydrocarbon gas of around 17.16% of HCPV. The studies provide an insight into the suitability of
the WAG mechanism for the particular Brownfield and indeed require better understanding and optimization of these processes on economical
parameters. The results of this work can be used for the further simulation study and optimization of operating parameters, like, flow rate, slug size,
WAG ratio, etc. for the actual field applications.

KEYWORDS: Enhanced Oil Recovery, Hydrocarbon Pore Volume, Incremental Oil Recovery, Water-alternate-Gas (WAG)
INTRODUCTION
Oil and natural gas are an important source of energy for the mankind. The total energy requirement of the world has been increased drastically in the
past decade due to the increased demand from the developing countries like India, China and Brazil. India currently imports about 70% of the total
requirement of crude oil and natural gas, thus accounting for 40 % of the total import bills. It is, therefore, important to gear up to invent new oil and
gas discoveries and to increase the production from the existing but mature reservoirs (brownfields) in order to keep the economic growth rate on the
track. It is an established fact that substantial amount of oil usually remains in a reservoir after primary and secondary processes. In some cases it may
be as high as 70% of Oil Initially in Place (OIIP). There is an enormous incentive for development of a field through Enhanced Oil Recovery (EOR)
methods aimed at recovering some portion of the remaining oil. Several methods have been proposed as a part of EOR techniques to be applied on the
brownfields. These methods generally incorporate an external agent in the form of fluids (gas or liquid) to be injected in to the reservoir for pressure
maintenance and to increase the sweep efficiency of the reservoir. Enhanced oil recovery methods can generally be classified as shown in Figure 1.
These are, mainly, thermal techniques, chemical injection methods, miscible or near miscible gas injection methods, and microbial injection methods
(Green and Willhit, 2003). Gas injection based EOR methods are one of the most preferred methods due to their simplicity and economical
advantages.
Fig 1: Different methods of Enhanced Oil Recovery (EOR), (Green and Willhit, 2003)

Gas Injection Method

Gas injection based EOR methods are the second largest methods used for enhanced oil recovery (Kulkarni, 2005) purposes. In this method
hydrocarbon or inert gas is injected in to the reservoir containing residual oil. The components of the gas get dissolved with the lighter components of
the oil which helps to reduce the viscosity and increase the sweep efficiency in the presence of a chasing fluid such as water. The component exchange
processes between the injected gas and reservoir oil causes heavy and light compositions in the reservoir which separately moves towards the
production side. Different gases are used in the gas injection methods, such as, Nitrogen, Hydrocarbon Gas (HC), Flue Gas and CO2 Gas. Some of the
injectants such as, CO2, help to increase oil production by means of oil viscosity reduction, oil swelling and solution gas drive (Green and Willhit,
2003). The use of specific gas depends on the availability of gas at the field. Previously Liquefied Petroleum Gas (LPG) and hydrocarbon gas were
used for injection. But gradually as price of natural gas increases, their priority got reduced. Gas injection method can broadly be classified as
immiscible and miscible gas injection, depending upon their miscibility with the oil at reservoir condition. In immiscible gas injection process the gas
is injected at lower pressure into the reservoir. It is further classified as dispersed gas injection and crestal gas injection according to the injection
region. In dispersed gas injection, gas is directly injected in to the oil bearing zone of the reservoir. This method is used in the thin production zone. In
crestal gas injection method, gas is injected in to the gas cap above the oil bearing zone. For this process, vertical permeability of the reservoir should
be high in order to push the oil towards the production end. Miscible gas injection method can be classified broadly as high pressure dry gas miscible
displacement, enriched gas miscible displacement and miscible slug flooding.

A large change in the mobility of gas and oil is observed in case of the gas injection methods due to difference in the viscosity of gas to the oil and
water at the reservoir conditions. This results in early breakthrough of the gas to the production side due to its high sweep velocity. In order to control
the sweep velocity of the gas, water and gas are injected intermittently. This method is called as Water Alternate Gas (WAG) injection method. Oil
recovery by WAG injection has been attributed to contact of upswept zones, especially recovery of attic or cellar oil by exploiting the segregation of
gas to the top or the accumulating of water toward the bottom. Because the residual oil after gas flooding is normally lower than the residual oil after
water flooding, and three-phase zones may obtain lower remaining oil saturation, WAG injection has the potential for increased microscopic
displacement efficiency. Thus, WAG injection can lead to improved oil recovery by combining better mobility control and contacting upswept zones,
and by leading to improved microscopic displacement. Reservoir specific parameters such as wettability, interfacial tension, connate water saturation
and gravity segregation add complexity to the design of a successful WAG flood. The WAG injection methods can be classified as miscible WAG,
immiscible WAG, hybrid WAG and simultaneous water alternate gas (SWAG) methods (Christensen et al., 2001). Several screening criterion are to
be considered before the application of WAG technique for any particular field operation. These are mainly, reservoir pay thickness, vertical
permeability of the reservoir, availability of the gas, type of formation, mobility ration, etc.

EXPERIMENTAL DETAILS

The aim of the current work is to evaluate the performance of the different gas injection methodologies for a given brownfield in India. It includes
comparative studies on different water-alternating-gas (WAG) injection methods and to verify their effects on the production enhancement from the
given field. Core flooding experiments are performed at close to the reservoir conditions of the pressure and temperature to identify:
a. The effect of WAG injection method for various WAG cycles
b. To analyze recovery efficiency for different methods using different gases like hydrocarbon gas and CO2 gas at reservoir condition
c. The effect of tapering on the WAG performance

WAG processes which have been studied and discussed in this work (on the basis of WAG cycles) are,
1. Single cycle WAG using HC gas
2. Five cycles WAG using HC gas
3. Tapered WAG (with increasing and decreasing WAG ratio) using HC gas
4. Five cycles WAG using CO2 gas

The experiments were performed using the in-situ core sample obtained from the reservoir and is fitted in the core pack, which is then kept
horizontally during all the experiments. The gas and oil samples are collected from the separator and recombined in the laboratory with given gas-oil
ratio (GOR) so as to become representative of the in-situ reservoir fluid. The recombination process is discussed in detail elsewhere (Bhatia, 2010).
The experiments were performed using the recombined separator fluid as a reservoir fluid in the core sample and the hydrocarbon or CO2 gas with
water as a mean for injection in the core sample during WAG process. The water has been injected at 20 cc/hr and gas has been injected at 10 cc/hr,
which remains same for all the experiments as mentioned above. The water and gas ratio remain same except for the experiments where the effect of
tapering was studies. The details on the ratio of water and gas used are described in later section.

308
Properties of the Experimental Fluids and the Reservoir

The composition of the hydrocarbon gas used for injection is given in Table 1, which is obtained by using Gas Chromatographic technique. The major
component of the injection gas is methane (about 90%) of the total concentration. The gas contains around 2% CO2. The gas gravity is observed to be
0.8351 gm/cc. Another gas used for the WAG process was pure CO2. The basic reservoir data and rock properties are given in Table 2. The given
reservoir is a sandstone reservoir and is under depletion. API gravity of the oil is about 42o which indicates the light oil reservoir.

Experimental Set-up

High pressure apparatus has been selected for the core flooding experiments. All the flooding experiments are performed at the reservoir pressure of
230 Kg/cm2 and temperature of 128 ˚C. The schematic of the core flooding experiment is as shown in Figure 2. The heart of the set-up is the core pack
which holds the actual core sample at reservoir conditions. The core pack is placed in the oven which is maintained at reservoir conditions of
temperature. Pressure gauges are used to indicate the pressures at inlet and outlet of the core pack. The pressure in the core pack is maintained at
reservoir conditions by using positive displacement pump (Ruska®) which injects the fluid (gas, oil, water) at different flow rates in the core pack. The
inlet pressure is regulated by same positive displacement pump through which kerosene has been used as a displacing fluid to displace any gas or
liquid from the gas cell/buffer cell/rocking cell into the core pack. The gas cell contains gas (HC or CO2) to be injected during the WAG process. The
buffer cell contains water (2 % KCl) which is used as a buffer to displace oil or water. The rocking cell is used to prepare the live oil (recombined
fluid) from the oil and gas samples collected from the separator. A backpressure regulator regulates the flow from the core outlet by maintaining
constant pressure difference at the input and the output side. The produced fluid (water, gas and oil) collected in the separator flask at the outlet of the
core pack will indicate the quantity of the produced fluid and one end of the flask connected to the gas meter will indicate the quantity of the produced
gas during the WAG process. Steel pipe of 1/8' diameter are used for the fluid supply between the experimental components. The experimental setup
described here was same for all the experiments carried in this work. Before carrying our actual experiments on the WAG process, initial preparation
is done on the recombination of reservoir fluids using the separator sample of the oil and gas and the determination of gas-oil ratio and formation
volume factor (FVF) of the recombined reservoir fluid. The GOR and FVF of recombined fluid is then matched within the accepted limit with the
reservoir GOR and FVF in order to check the reliability of the recombination process.

Experimental Procedure

The experimental procedure for all the WAG case studies mainly consists of the preparation of the core pack, cleaning and drying of the core pack,
evacuation of the core pack, determination of the pore volume (PV) with saturation of the brine solution, determination of the hydrocarbon pore
volume (HCPV) by displacing the brine solution with heavy and light paraffin oil. The obtained value of the PV (≈ 60 CC) and HCPV will give the
connate water saturation inside the core. The core pack will then be cleaned using kerosene for studies with recombined fluid. The cleaned core pack
will then be saturated with the recombined oil prepared in the laboratory. This will be followed by the secondary water flood until oil saturation in the
core reaches to the residual oil saturation. After completion of the water flooding, to produce the residual oil from the core, WAG injection will be
started. The overall process resembles to the actual recovery process a reservoir may undergo during its production phase. The details on other
experimental procedures related to core pack preparation and cleaning, absolute permeability determination, PV and HCPV determination, oil
saturation and water flooding procedure can be found elsewhere (Bhatia, 2010) and not discussed here. We, mainly, but briefly discuss the main
process of tertiary gas injection of WAG method as below.

Tertiary Gas Injection

The tertiary gas injection is carried out mainly by WAG process using hydrocarbon gas and CO2 gas. Different WAG methods that have been applied
for enhanced oil recovery are single cycle WAG, five cycle WAG (with HC gas and CO2 separately), tapered WAG (with increasing and decreasing
WAG ratio). For single cycle WAG and five cycle WAG OF total 1 pore volume (1 PV = 60 cc with ± 0.5 cc) of gas and water has been injected
intermittently with the WAG ratio of 1:1 at the end of water flooding experiment. For tapered WAG injection method a total of 1.5 PV gas and water
has been injected intermittently at the end of water flooding experiment. In tapered WAG (with increasing and decreasing WAG ratio) WAG ratio as
given in Table 3 was selected and used for the experimental study.

The hydrocarbon gas collected from the adjacent field and pure CO2 obtained from other sources has been used as injection gas. The pressure and
temperature conditions of the core pack are kept at reservoir condition and the injection rate for water is maintained at 20 cc/hr and for gas is
maintained at 10 cc/hr to avoid the early breakthrough of the gas. The brine, oil and the gas volumes produced at the end of the experiment are
measured from the separator (flask) and gas meter readings and tabulated as a function of time. Material balance procedure is used to calculate the
saturations of oil, gas and water components.

Chasing Water post WAG Process

The chasing water is injected to get the additional recovery of HCPV after the WAG injection process. Chase water helps to push the trapped gas and
water in the core pack, with that combined oil also gets produced at the production side. In this experimental study maximum of 0.5PV (around 30cc)
chasing water has been injected after the completion of the WAG injection process. The results are tabulated during process and discussed in the
following section.

309
Table 1: Composition of the injected gas in mole fraction obtained by Gas Chromatography

Component Mole Fraction


N2 0.00000
CO2 0.02400
C1 0.90739
C2 0.05237
C3 0.01310
i-C4 0.00089
n-C4 0.00094
i-C5 0.00040
n-C5 0.00048
C6 0.00020
C7 0.00014
C8 0.00005
C9 0.00001
C10 0.00000
Total 1.00000

Table 2: Basic data for the reservoir and the core sample experiments

Details on the reservoir and the core sample

Sr. No. Parameters


1 Reservoir Rock Type Sandstone
2
2 Initial Reservoir Pressure (kg/cm ) 292.7
2
3 Current Reservoir Pressure (kg/cm ) 230
2
4 Bubble Point Pressure (kg/cm ) 269.6
5 Reservoir Temperature (ºC) 128
6 Density of oil (gm/cc) at 128 ºC 0.5142
7 Stock Tank Oil Density at 15.5 ºC 0.8161
8 ºAPI Gravity of Oil 41.5
9 Oil FVF (v/v) 1.84
10 Specific Gravity of Gas 0.8364
11 Solution GOR (v/v) 222
12 Core Length (cm) 20
13 Core Diameter (cm) 3.8
15 Avg. Permeability (mD) 323.23

Table 3: Injection WAG ratio for different cycle of Tapered WAG methods

Cycles WAG Ratio for Tapered WAG (Water : Gas)


Increasing WAG ratio Decreasing WAG ratio
1 3:1 3:5
2 3:2 3:4
3 1:1 1:1
4 3:4 3:2
5 3:5 3:1

310
Fig 2: Schematic diagram of the experimental set-up

RESULTS AND DISCUSSION

The core-flooding experiments are carried out to verify the effect of different parameters of the WAG injection methods. The main objectives of this
work are to study the efficiency of different WAG processes and the parameters affecting the production enhancement, viz., tapering of gas, WAG
cycles, type of the injected gas. The results have been discussed with respect to the incremental oil recovery over water flooding.

Total of five experiments, namely, single cycle WAG (with hydrocarbon gas), five cycle WAG (with hydrocarbon gas), five cycle WAG (with CO2
gas), and tapered WAG suing HC gas (with decreasing and increasing gas tapering) have been investigated and discussed. In core-flooding
experiments, total pore volume (about 60 cc) of the core has been divided according to the number of cycles. In a single cycle WAG process, pore
volume (about 60 cc) is divided as 0.5 PV (about 30 cc) gas and 0.5 PV (about 30 cc) water and are injected accordingly. Similarly, for five cycles
WAG process, 0.1 PV of gas and 0.1 PV of water are injected in five cycles sequentially so as to make total injection equal to 1 PV (about 60 cc).
Attanucci et al. (1993) observed that in case of tapered WAG process an injection of total 1.5 PV gives better results. In this study we carried out the
experiments for tapering WAG using 1.5 PV of gas and water as total injections. In case of tapered WAG (with decreasing WAG ratio) more amount
of gas has been injected in the first cycle and is gradually decreases in the subsequent cycles. Amount of water to be injected during each cycle
remains constant. In case of tapered WAG (with increasing WAG ratio) similar procedure is followed but in reverse direction. The details on the
quantity of gas and water used for each of the above process are given in the Table 4.

The oil recovery from the above experimental data can be expressed as the displacement efficiency or recovery in percentage of the total HCPV and
can be calculated as,

Displacement Efficiency (% HCPV) = HCPV − (Qo × FVF ) − V L ∗100 (1)


HCPV

Where, Q0 is the flow rate of oil and VL is the total line volume. VL is actually a kind of dead volume of oil remained in the production tubing of the
core flood apparatus and which need not to be accounted as oil recovered. The results on the displacement efficiency (in % of total HCPV) with
respect to the total pore volume of fluid (water/gas-water) injected for different WAG processes have been shown in Figures 3 to 7. It is to be noted
that for all the experiments, a water flooding is carried out prior to each WAG process to represent the secondary oil recovery using water flooding.
The water flooding process required a total of about 1.25 - 1.5 PV of the water to be injected in the core sample (see Figure 3-7 and Table 5).
Percentage of recovery obtained using the water flooding prior to WAG process is also given in Table 5 and observed to be in the range of 47 to 57 %.
Actual WAG process starts after the end of water flooding. The WAG processes consume the fluids (gas and water) in different ranges of PV and have
been shown in Table 4 and are cumulative on the x-axis of Figures 3 to 7 after the pre-water flooding section. Recoveries in percentage of
hydrocarbon pore volume (HCPV) have been shown for each of these processes in the following figures. At the conclusion of each WAG process,
chasing water is flown through the core pack to see any incremental recovery. The recovery obtained for different phases of each process are visible in
figures 3 to 7 and are tabulated in Table 5.

The maximum recovery has been noticed in CO2 five cycle injections (about 97.86 % of HCPV), and next maximum recovery is in tapered WAG
injection (decreasing WAG ratio) (about 72.48 % of HCPV). The maximum incremental recovery over the waterflood has been seen with CO2 gas
with five cycle WAG injection (about 40.2 % of HCPV), and the next has been noticed with tapered WAG HC gas injection with increasing WAG
ratio (about 23.92 % of HCPV). The maximum recovery with CO2 is obtained probably due to its better miscibility with the crude oil (in the core
pack) at reservoir conditions as compared with the HC gas used in WAG processes. Better recovery is obtained in case of tapered WAG injection
(decreasing WAG ratio) as against all WAG processes using HC gas is due to an increased sweep efficiency governed by an initial dissolution of
maximum amount of gas with the crude oil in the first cycle, thus helping better mobility in the pore of the core sample. This results in an increased
relative permeability of oil in the core sample which is enhanced by the subsequent water cycle in the WAG process.

The effects of different parameters like WAG cycles, type of the injected gas, tapering, etc., has been discussed below.

311
Effect of Number of WAG Cycle

Zhang et al. (2010) observed that by increasing the number of the WAG cycles in gas injection methods helps to get more recovery of the oil from the
reservoir. The effects of WAG cycle are also studied in this work to see the applicability for the given reservoir. The results obtained are shown in
Figures 3 and 4 for single cycle WAG and five cycle WAG process using HC gas. During single cycle WAG and five cycles WAG injection, a total of
1 PV volume (about 60 CC) of the fluid (water and gas) has been injected after the initial water flooding process. In this work, the single cycle WAG
process using HC gas shows 12.74 % incremental recovery (recovery obtained after the initial water flooding) and five cycle WAG process using HC
gas (no tapering) shows about 17.16 % of HCPV incremental recovery over the water flooding. This indicates that the number of cycle affects the
recovery of HCPV. Increasing in the number of WAG cycle improves the recovery for the same amount of gas utilization. However, in some of the
studies it is observed that the recovery does not improve significantly even increasing the number of WAG cycles, probably due to increased water
saturation and reduced discontinuity of the oil phase (Dong et al., 2005).

Table 4: Details on the amount of water and HC gas used for tapered WAG process

Type of Process/ Tapered WAG Tapered WAG


Number of (increasing WAG ratio) (decreasing WAG ratio)
Cycles
Amount of (cc) Fraction of total PV WAG ratio for each Amount of (cc) Fraction of total WAG ratio for
Water Gas per cycle cycle Gas Water PV per cycle each cycle
1 9 15 0.40 3:5 3 9 0.20 3:1
2 9 12 0.35 3:4 6 9 0.25 3:2
3 9 9 0.30 1:1 9 9 0.30 1:1
4 9 6 0.25 3:2 12 9 0.35 3:4
5 9 3 0.20 3:1 15 9 0.40 3:5
Total 45 45 1.5 --- 45 45 1.5 ---

Table 5: Summary of results for WAG injection methods


Amount of Water injected pre-WAG

Total PV injected including that for


Chasing water injected after WAG

Recovery (%HCPV)

Residual oil saturation (%HCPV)


Recovery with water flooding

Incremental recovery during


Incremental recovery over
WAG process
process (PV)

process (PV)
injection gas
Type of

Total recovery

chase water
water flood
WAG injection pattern

Single cycle WAG HC gas 1.246 0.25 2.378 51.57 12.75 64.32 0 30

Five cycle WAG HC gas 1.028 0.5 2.478 52.05 19.30 71.30 2.1 30.06

Tapered WAG (increasing WAG ratio) HC gas 1.255 0.35 3.127 48.44 23.92 72.36 0 29.9

Tapered WAG (decreasing WAG ratio) HC gas 1.232 0.33 3.063 51.33 16.91 72.48 4.23 16

Five cycle WAG CO2 gas 1.452 0.28 2.723 57.67 40.2 97.86 0 29

312
70
GAS WATER chase water

Displacement Efficiency (%HCPV)


WATER INJECTION
60
FLOODING

50

40 WAG cycle

30

20

10

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Pore volume injected

Fig 3: Displacement efficiency vs. PV injected for single cycle WAG injection using HC gas as injectant

90
Water

Water
Water
Water

Water
Displacement Efficiency (%HCPV)

80
Gas

Chase Water
Gas

Gas
Gas

Gas

Water Flooding
70

60

50

40

30 WAG
Cycle
20

10

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Pore Volume Injected

Fig 4: Displacement efficiency vs. PV injected for five cycle WAG injection using HC gas as injectant

313
90

water

water

water
water

Ga s
CHASE

water

Gas
Gas
WATER

Gas
Displacement Efficiency (%HCPV)

Gas
80 WATER
FLOODING
70

60

50

40

30 WAG
Cycle
20

10

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Pore Volume injected (cc)

Fig 5: Displacement efficiency vs. PV injected for tapered WAG injection (with increasing WAG ratio)
using HC gas as injectant

80
Water

Water

Gas
CHASE WATER
Water

Gas
Gas

70
Water

Water

WATER 
Gas

Gas

FLOODING
60
Displacement Efficiency (%)

50

40
WAG
30 Cycle

20

10

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Pore volume injected (cc)

Fig 6: Displacement efficiency vs. PV injected for five cycle WAG injection (with decreasing WAG ratio)
using HC gas as injectant

314
120

Water
Water
Water
CHASE WATER

Gas
Gas
Water
Gas
WATER

Water
Displacement Efficiency (%HCPV)
100

Gas
Gas
FLOODING

80

60

40
WAG
Cycle
20

0
0 0.5 1 1.5 2 2.5 3
Pore volume injected

Fig 7: Displacement efficiency vs. PV injected for five cycle WAG injection using CO2 gas as injectant

Effect of Tapering

The increase or decrease in the water to gas ratio during the WAG cycles is known as tapering phenomena in the WAG process. It is also known as the
hybrid WAG method (Christensen et al., 2001). In the tapered WAG process, gas injection after the initial water flood pushes the oil to the production
well in case the oil saturation is high, and if it is low then it will displace the oil to the higher water saturated channels and some of the gas stays in the
small channels and resist the water mobility (Dong et al., 2005). Increasing the tapering form a WAG ratio (water:gas) of 1:1 to 1:2 to 1:3 followed by
chase water increases the efficiency of the oil recovery (Attanucci et al., 1993). In this experimental study tapering with both increasing and
decreasing WAG ratio has been studied and are shown in Figs 5 and 6. The results show that increasing WAG ratio shows more recovery than
decreasing WAG ratio during each cycle. The large quantity of gas injected during the first (refer Table 5) cycle helps to get connected with more
residual oil and makes it move towards the high permeable area. Subsequently, the water cycle flowing through this highly permeable area (channels)
helps to push the oil-gas system toward the production well. The experimental results for tapered WAG with increasing WAG ratio show that there is
no further oil production after three cycles of WAG injection, which indicate that tapered WAG injection method can be efficiently used for lower
WAG cycles for the equivalent oil recovery. It is also observed that the chase water also plays an important role during tapering of gas. The oil
recovery obtained in case of tapered WAG with decreasing WAG ratio is about 68.25 % of the HCPV before the chase water flooding. An additional
recovery of 4.2 % is achieved after the chase water flooding at the end of WAG process resulting in a total 72.48 % recovery from the given
experiment. The chasing water is a concluding process and acts similar to water cycles in a WAG process only except that sufficient quantity of water
is injected in the core in order to get the maximum possible recovery from the reservoir.

Effect of Injecting Gas

In this experimental study, hydrocarbon gas and CO2 gas has been used in the five cycle WAG injection process at the reservoir conditions. The
results for five cycle WAG using HC gas and CO2 gas are shown in figures 4 and 7, respectively, and tabulated in Table 5. The results shows that CO2
injection in five cycles WAG gives recovery of about 97.86% of HCPV, which is very high compared to the recovery of five cycle injection of
hydrocarbon gas (about 71.3% of HCPV). This is due to the fact that the CO2 gas is having better miscibility as against the hydrocarbon gas with the
crude oil at the reservoir condition by helping to increase the solution GOR of the oil and also help to reduce the viscosity of the oil. This probably
results in the solution gas driven production and increasing the relative permeability of the oil phase. The reservoir condition of pressure and
temperature of 230 kg/cm2 and 120 oC shows that the CO2 gas may be at in near supercritical state at the reservoir condition thus helping better
miscibility with the reservoir fluid.

We would like to emphasize that though we have studied the effect of different WAG methods for the application for a given brownfield in India, we
have studied the process experimentally in the laboratory based environment using the actual core sample obtained from the field. The actual recovery
which may be obtained from the field using the best WAG method may differ from the laboratory results due to, mainly, the effect of the type and the
pattern of injection. However, the laboratory information will certainly give us knowledge of best recovery method for the given field application. The

315
results of this study are to be extended for actual field study and hence the effect of the type of injection techniques on ultimate recovery needs to be
investigated for better ultimate recovery.

CONCLUSION

The implementation of any Enhanced Oil Recovery (EOR) process, experimental verification of process parameters and cost effectiveness of the
process should be intended. The experimental study, followed by simulation and a pilot project implementation provides the better estimation of the
process parameters at the field scale. This study consist of comparative study of different WAG injection process for the core sample collected from
the brown field and the live oil prepared in the laboratory, from sample of oil and gas collected from the field separator. The experimental work is
done at the reservoir temperature at 120 ˚C and pressure at 230 kg/cm2. The single cycle WAG (with HC gas), five cycle WAG (with HC gas and CO2
gas) and tapered WAG using HC gas (with increasing/decreasing WAG ratio) have been investigated experimentally. As the results of experiments on
WAG injection processes, there are some important conclusions are derived for implementation of WAG process. It is observed that the WAG
injection process gives the better sweep control, mobility control of water and gas phases and improves the total recovery of the HCPV. The number
of cycles in the WAG injection process affects the recovery of oil from the reservoir. The results shows that for the same volume of injecting fluid in
single cycle and five cycle process the incremental recovery of HCPV has been noticed as 12.74% and 17.16% correspondingly. The tapering during
the WAG injection process helps to improve the recovery of the HCPV. Increasing WAG ratio during the process gives better incremental recovery
than decreasing WAG ratio. The tapered WAG process with decreasing WAG ratio doesn’t have the further recovery after three cycle of WAG
process, which indicate that further cycles of WAG process is not economically feasible. Chasing water injection after WAG process helps to get the
incremental recovery of oil. The tapered WAG injection method with increasing WAG ratio gives around 4.2% incremental recovery during chasing
water flooding after WAG process. The five cycle WAG process using CO2 gas shows higher oil recovery compared to five cycle WAG using HC gas
possible due to its presence at near-critical conditions in the reservoir help to have better miscibility with the reservoir oil. The exact choice and
applicability of the above process also depend on the economic feasibility of the given process and thus need to be investigated.

REFERENCES

Attanucci, V, Aslesen, KS, Heji, KA, and Wright, CA (1993) “WAG process optimization in Rangely CO2 miscible flood”. SPE 26622. Presented at
68th annual technical conference and exhibition held in Huston, Texas, October 1993.
Bhatia, J (2010) “Comparative studies of gas injection methodologies for enhanced oil recovery”, M. Tech. thesis, PD Petroleum University, Gujrat,
India.
Christensen, JR, Stenby, EH, and Skauge, A (2001) “Review of WAG field experience”, SPE Reservoir Evaluation and Engineering, Vol. 4, (2), pp
97-106.
Dong, M, Foraie, J, Huang, S, and Chatzis, I (2005) “Analysis of immiscible Water-Alternate-Gas (WAG) injection using micromodel tests”, Journal
of Canadian Petroleum Technology, Vol. 44, (2).
Green, WD, and Willhite, GP (2003) “Enhanced Oil Recovery”. 2nd Edition. SPE text book series.
Kulkarni, MM, and Rao, DN (2005) “Experimental investigation of miscible and immiscible Water-Alternating-Gas (WAG) process performance”,
Journal of Petroleum Science and Engineering, Vol. 48, pp 1 – 20.
Zhang, YP, Sayegh, SG, Luo P, and Huang, S (2010) “Experimental investigation of immiscible gas process performance for medium oil”, Journal of
Canadian Petroleum Technology.\, Vol. 49, (2), pp 32-39.

316
GOLD SPONSORS

Oil and Natural Gas Corporation Limited (ONGCL) (incorporated on 23 June 1993) is a state-owned oil and gas
company in India. It is a Fortune Global 500 company ranked 152nd, and contributes 77% of India's crude oil
production and 81% of India's natural gas production. It is the highest profit making corporation in India. ONGC
is one of Asia's largest and most active companies involved in exploration and production of oil. It is involved in
exploring for and exploiting hydrocarbons in 26 sedimentary basins of India. It produces about 30% of India's
crude oil requirement. It owns and operates more than 11,000 kilometres of pipelines in India. The vision and
mission of ONGCL is to be a world class Oil and Gas Company integrated in energy business with dominant
Indian leadership and global presence. For more information, visit the ONGCL website at
http://www.ongcindia.com.

Transocean is the world’s largest offshore drilling contractor. It provides the most versatile fleet of mobile
offshore drilling units to help clients find and develop oil and natural gas reserves. Building on more than 50
years of experience with the highest specification rigs, its employees are focused on safety and premier offshore
drilling performance. At present, Transocean holds 19 of the past 23 world records for drilling in the deepest
waters. Transocean’s ultra-deepwater drillship ‘Discoverer Deep Seas’ has set the current world water-depth
drilling record in 3,051 meters of water in the U.S. Gulf of Mexico working for Chevron. For more information,
visit the Transocean website at www.deepwater.com.

SILVER SPONSOR

Aban Offshore Limited (AOL) is India's largest offshore drilling contractor in the private sector, offering world-
class drilling and oil field services for offshore exploration and production of hydrocarbons to the oil industry in
India and abroad. AOL's innovative and cost effective solutions have made them one of the most efficient
international drilling contractors. AOL consists of a highly qualified professional team backed by years of
experience and expertise. In modern environment, AOL continues to foray into newer waters. For more
information, visit the AOL website at http://www.abanoffshore.com.

You might also like