SPE 94746
Use of Streamlines and Quality Map in the Optimization of Production Strategy of
Mature Oil Fields
Marcelo S. Guimarães, UNICAMP; Denis J. Schiozer, SPE, UNICAMP; Célio Maschio, SPE, UNICAMP
Copyright 2005, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the 2005 SPE Latin American and Caribbean
Petroleum Engineering Conference held in Rio de Janeiro, Brazil, 20 – 23 June 2005.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300
words; illustrations may not be copied. The proposal must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Proposal
Developed or mature oil fields can be considered as sets of
reservoir that present production strategy already established
and implemented, generally presenting production in the
decline phase. Changes in geologic characterization or
economic or technologic scenarios can demand changes of the
production strategy, requiring the application of an
optimization process. The optimization process of developed
or mature oil fields presents different characteristics from
fields in beginning of production, mainly regarding flexibility
for changes. The existence of production data allows the
improvement of the geological characterization of the
reservoir and the reduction of the uncertainties allowing a
more detailed optimization procedure.
The objective of this work is to develop an automated
optimization methodology combining the utilization of three
important tools: traditional simulation, streamline simulation
and quality map. The traditional simulation is used to give
reliability to the process. The streamline simulation and
quality map are supporting tools employed to give more
reliability and to bring up insights about the optimization
process and speedup to the process, mainly in the control of
water injection and production and in the possible allocation
of news wells.
The proposed methodology had been applied to an
offshore field with water injection. The results show that the
efficient control of fluid flow and the better understanding of
the overall process provided by the supporting tools allow the
determination of production strategy that improved the
economic recovery of the developed or mature field studied.
Additionally, the developed methodology allows the definition
and the organization of the possible changes, making feasible
the automation of part the process.
Introduction
The definition of a production strategy of oil fields comprises
several study phases. In the beginning of the field life the data
are scarce, what reflects in a great uncertainty and in the one
adoption of an approach initial production strategy, without
great refinement of the solution. After the choice of the initial
strategy, appears the necessity of refinement since the
uncertainties diminish in function of the biggest amount of
data, what reflects in the best detailing of the field model to be
simulated. In the third phase occurs the optimization of
production strategy of the developed or mature oil fields. In
this stage, the reservoir knowledge increases, and more
geological are available, reducing the involved uncertainties in
the optimization process.
The strategy optimization is a complex procedure that
demands the analysis of the producer and injector wells
behavior to know which wells can be modified and the
analysis of a great number of variable that influence the
process, as for example, geological and fluids properties,
production and pressure data. Developed and mature oil fields
optimization presents minor flexibility because although the
used models are more realistic, there are more restrictions of
changes after the total or partial implementation of the initial
strategy. Therefore with perforated wells and in operation,
modifications must be made of different form to present
economic viability. This complexity increases with the
difficulty in to apply mathematical optimization methods in
function of the great computational effort required in the
modeling process. This difficulty becomes necessary the use
of support tools as the streamline simulation and the quality
map, that allow to define additional objective functions that
classify the wells and determine which regions of the reservoir
present mobile residual oil to be recovered.
The streamline simulation is used to study the fluid flow
pattern in the field, determining the efficiency of injector wells
and distribution of injection costs for producer wells. The
quality map is used to define regions of the reservoir with
mobile residual oil, defining regions for perforation and
completion of producer wells.
The conventional simulation (based on finite difference) is
used to do the main simulations to the tests that verify the use
of the support tools. All the economic analysis is made based
on conventional simulations output.
2 SPE 94746
Supporting Tools
Mezzomo and Schiozer (2002) work with a methodology to
support managers in the decision making process for water
injection planning optimization for fields in development stage
and under operational and economic restrictions. Nakajima
and Schiozer (2003) [a] developed a methodology to support
reservoir managers to optimize a strategy developed with
horizontal wells. It does not provide unique solutions, since
this optimization problem depends on many factors which may
lead to a set of alternatives. His methodology is based on
economic optimization and the main objective is to maximize
the net present value of the field.
In developed and mature oil fields studies is necessary a
detailed analysis of field behavior, because the geological and
numerical models are more precise, and production history
data permits improve the reservoir chacacterization. The
streamline simulation and quality map assist this analysis,
determining parameters that through the use of conventional
simulation are more difficult to determine.
Streamline Simulation. The streamline simulation currently
is accepted as an effective and complementary tool for
reservoir modeling. This tool allows working models that
possess a raised number of blocks, with more complex and
heterogeneous geology. Thiele (2001) shows that the
streamline simulation functions with six key ideas:
1) Tracing streamlines in three dimensions using time-of-
flight; 2) Recasting the mass conservation equations in terms
of time-of-flight; 3) Periodic updating of streamlines; 4)
Numerical solutions along streamlines; 5) Including gravity
presented a problem; 6) Use in compressible flow.
Batycky et al. (1997) illustrate the application of
streamline simulation method to a real field data set, the
House Mountain waterflood in central Alberta, Canada. The
streamline simulation accurately models overall field historical
data and on a well-by-well basis accurately models
performance in 60% of the wells. The authors had concluded
that streamline method is applicable to three-dimensional field
scale waterflood simulations with only minor approximations
to data input and can either include larger simulation areas
with more wells and/or reduce the need for substantial
upscaling.
Grinestaff (1999) demonstrate the use streamlines
simulation to waterflood management. Your focuses in three
areas: 1) how the streamline models require a different
approach to simulation; 2) using streamlines to find
inefficiencies in the waterflood and set injection targets; 3) the
benefits of a history matched model. The main conclusions of
the author had been: streamline simulation provides a rapid
means to quantify injector to producer relationships and other
fluid displacement in complex waterfloods; streamline models
should not be developed with conventional simulation model
building and history matching techniques; production data
should be the primary source of information used to history
match a developed waterflood area.
Maschio and Schiozer (2002) analyzed the performance of
streamline simulation comparing the results for this tool with
finite differences. It was shown that streamline simulation is
very efficient, in terms of run time and accuracy, for
incompressible or slightly compressible models, with high
degree of heterogeneities and large grid size, which represent a
limitation for conventional finite difference simulators. For
these cases streamline simulation have clear advantages
compared to conventional simulators. Once streamline
simulation must necessarily neglect or simplify some
important physics in order to run quickly, it should not be
intended to match or rigorously predict field recovery when
complex physics related to fluids properties and fluids flow
must be considered. The main limitation is when highly
compressible fluids are present in the model. In these cases,
conventional simulation is required for more rigorous and
reliable evaluations.
This work applies the streamline simulation in a field in
production phase (developed or mature), where the main
objective is to study the flow distribution between producer
and injector wells.
Quality Map. Cruz et al. (1999) introduce the concept of a
quality map, which is a two-dimensional representation of the
reservoir responses and their uncertainties. The quality
concept may be applied to compare reservoirs, to rank
stochastic realizations and to incorporate reservoir
characterization uncertainty into decision making, such as
choosing well locations, with fewer full field simulation runs.
The authors used a very large case study using fifty different
realistic reservoirs to demonstrate the following uses of the
quality maps: 1) well location for a specific realization using
the quality map of that realization; 2) well location in a
manner that is robust with respect to the uncertainty and
reducing the number of scenarios in the “full approach”, using
the lower quartile quality map; 3) reduction of the number of
realizations in the “full approach” to just one, by identifying a
representative realization for each scenario, 4) ranking of
realizations for several purposes using the quality map of all
realizations; 5) reservoir comparisons using the average values
of the mean quality and uncertainty quality maps. The main
conclusions of the author had been: the quality map permits a
simple two-dimensional visualization of how good the area is
for production values and of the uncertainty in those values;
the quality map along with a simple optimization algorithm
may be used to determine good locations for vertical producer
wells; the lower quartile quality map allows the incorporation
of geological uncertainty into reservoir management decision
making; one representative realization can be identified for
each production scenario using the quality maps, allowing
scenario comparisons with similar results to the ones using the
expected value over all realizations but with much less CPU
time expense.
Nakajima and Schiozer (2003) [b] propose the use of a
quality map to guide reservoir managers in horizontal wells
placement. This map represents the regions with production
potential in a reservoir, providing the best place to locate a
well. Three methods of quality map construction are used: (1)
numerical simulation, (2) analytical and (3) fuzzy system.
These three methods were developed to provide a fast
evaluation in order to reduce time-consuming and
computational efforts. The main conclusions the Nakajima and
Schiozer (2003) had been: the use of quality map help the
decision-making process in petroleum field management; the
quality maps can be applied in any stage of reservoir life,
SPE 94746
reducing time-consuming and computational efforts in the
production strategy optimization process; the analytical
method and fuzzy system method were considered more
flexible, since it is easier to use.
This paper study the application of the quality map
constructed through numerical simulation to an oil field in
production phase (developed or mature), identifying regions of
the reservoir with residual oil localization.
Methodology
Guimarães (2005) developed a methodology for optimization
of fields in production phase, where the objective-functions
NPV, Np, W, G, Qo_m and Mp are used for to determine the
efficiency of producer well. The methodology used the
streamline simulation and quality map as auxiliary tool in the
optimization process of fields in production phase. This paper
focuses on the feasibility and application of streamline
simulation and quality map in the optimization process,
showing the main procedures to use the techniques
Streamline Simulation Analysis. Two tests are made to study
the application of streamline simulation:
Test one - efficiency of injector well (Ef): Ef is defined as
being the produced oil volume (Vop) divided by the injected
water volume (Viw) responsible for the production of this oil
(Equation 1).
Vop
Ef = (1)
Viw
To verify the application of the Ef, all the injector wells must
have its individually tested shut down. The relation between
the variation of the field NPV and the Ef must be investigated
through graphs that present correlation studies between the
analyzed functions. The NPF for the field is calculated through
conventional simulation.
Test two - distribution of injection costs for producer wells:
Guimarães (2005) uses revenue and cost curves to determine
of producer wells closing date. These curves are constructed
using production data for the behavior forecast period and
revenue (oil sale) and cost (water and oil production) data. The
relation between producer and injector wells is gotten through
of the streamline simulation, where the injected water volume
directed to a producer well has its costs added to the
production costs of the well that receives this water. The
revenue and cost curves for the well are calculated through
conventional simulation.
Quality Map Analysis. The quality map is constructed by
running a flow simulator multiple times and varying the
position of a single well in each run to cover the entire
horizontal grid. Each run evaluates the quality for the
horizontal cell where the well is located. The quality unit is the
NPV a certain time of production. In the simulations, the well
is completed in all oil layers with automatic shut down of the
layer when some water cut limit is reached. No rate limits are
3
imposed, allowing the well to produce the maximum it can.
Only a minimum bottom hole pressure and a minimum oil rate
must be specified in accordance with the real operational
limitations of the well.
Three methods to generate numerical quality map are
tested with the objective to define which of them must be used
in studies of developed or mature fields. These methods are: 1)
single well tested in all the cells of the horizontal grid
(accurate map); 2) single well tested in a mesh of tests with 59
points present in the horizontal grid (intermediate map); 3) 59
points present in the horizontal grid simulated at the same time
(less accurate map). These three maps must be constructed for
the end of the production history period (Initial Quality Map)
and for the end of the forecasting period (Final Quality Map).
The validation of these maps is made through tests where ten
horizontal wells perforated in diverse regions of the reservoir.
Plots of NPV versus quality index indicate the correlation
between them.
In order to test quality maps for injector wells, eight wells
were perforated in several regions of the reservoir.
Application
The study presented in this paper was applied to an offshore
field. The STOIP of the field is approximately 100 MMm3. It
is a developed field, with 1800 days of production history.
This field has useful life of 2922 days (keeping the production
strategy applied during the history period). The field is drained
by 11 oil producer wells and 9 water injector wells. In 1800
days, all the producer and injector wells already are in
operation. The water injection began after 180 days of primary
production.
The simulation model is composed of a grid of 60 blocks
in the x direction, 35 blocks in the y direction and 7 layers
(60x35x7), discretized into a Cartesian grid. Datum depth is
located at 3000 m, water oil contact is located at 3100 m and
gas oil contact is located at 2900 m. In Fig. 1 is presented a
three-dimensional horizontal permeability map of the
reservoir.
Results and Discussion
Initially, each tool was analyzed separately to evaluate its
potential and to define procedures to integrate them in the
optimization process.
Streamline Simulation. In this section are presented the
results of the application of the streamline simulation.
Test one: Table 1 presents the obtained results for the injector
wells shut down in 1800 days. The shut down of injector wells
with low efficiency (Ef) provides high NPV increasing. On the
other hand, when injectors with high efficiency are shut, a
high decrease in NPV is observed. Fig. 2 shows good linear
correlation between NPV and Ef (dashed line in the Fig. 2).
Test two: Revenue and cost curves were constructed for all
producer wells present in the field, but only the results for the
wells PH-02 e PH-14 will be presented. Fig. 3 and the Fig. 4
show, respectively, the revenue (blue) and cost (red) curves
for the PH-02 and PH-14. The green curve is the additional
cost related to the injector costs distributed to the producers.
4 SPE 94746
For the PH-02 (Fig. 3), the red curve crosses the blue curve in
2895 days and the green curve in 2779 days. The difference
between the crossing points is about 100 days. For PH-14 this
difference is approximately 300 days. This can influences the
estimation of the well shut down date.
Quality Map. In this section are presented the results from the
quality map analysis, perforating new producer and injector
horizontal wells.
Quality map analysis. Fig. 5 shows the accurate map
generated for 1800 days. For the construction of this map were
necessary 853 simulations. Fig. 6 and Fig. 7 present,
respectively, the intermediate map and the less accurate map.
The intermediate map required 59 simulations, and the less
accurate map required 1 simulation. The accurate map (Fig. 5)
shows larger concentration of regions with high unity quality.
Figs. 8, 9 and 10 show, respectively, the accurate map, the
intermediate map and the less accurate map for 2922 days.
This map also presents larger concentration of regions with
high unity quality. The difference between the maps
constructed for 1800 days and 2922 days is the size of the
regions that present high unit quality, what can be explained
by the smaller oil volume in the field in 2922 days.
Perforation of producer wells. For testing and validation of all
constructed quality maps, 10 horizontal wells were perforated
one at time in several regions of the reservoir. Each new well
corresponded to an individual model, in order to evaluate the
individual contribution of each well. Figs. 11, 12 and 13 show
field NPV increase versus quality index, for the accurate map,
the intermediate map and the less accurate map for 1800 days,
respectively. Figs. 14, 15 and 16 show the same for 2922 days.
Linear correlation is indicated in all plots by dashed lines.
The accurate maps generated for 1800 and 2922 days
present the best linear correlations, but its construction needed
853 simulations. The less accurate maps were generated with 1
simulation, but they show worse correlations. The choice of
the construction method of quality maps depends on the
objective. In the case of the user to need more precision, the
accurate map can be desirable. Intermediate and less accurate
maps can be used when computational and time resources are
limited.
Perforation of injector wells. Eight injector wells were also
perforated one at a time. Table 2 shows the NPV generated
with the perforation of these wells. Fig. 17 and Fig. 18
present, respectively, its locations in the poor quality region
for map constructed for 1800 and 2922 days. The wells
represented by yellow symbols provided better results. They
are predominately located in the neighborhood of the high
quality area in the map generated for 2922 days.
Optimization process. In order to evaluate de benefits of the
support tools, two optimization processes are presented: the
first one without using the tools (optimization 1) and the
second using them (optimization 2). In Fig. 19 is presented the
evolution of the optimization process without the use of the
support tools. In the y-axis is the percent variation (with
respect to the initial model) of the NPV, Np, Wp, Wi and Gp
and in the x-axis is the number of runs in the optimization
process. Fig. 20 shows similar plot for optimization 2. The
comparison of the results from the two optimization processes
shows following benefits related to optimization 2: (1)
decreasing the effort in terms of time consuming for the
manage the process and in terms of number of iterations
(optimization runs), for optimization 1 was necessary 50
iterations, while optimization 2 required 39; (2) increasing 79
% in NPV against 73 % from optimization 1; (3) small
increasing in water injection (Wi), approximately 2 %, while
in the optimization 1 the water injection increase 17 %; (4)
decreasing 55 % in water production (Wp) against 37 % for
optimization 1. Although increasing Np was smaller in
optimization 2, a better water management allowed improving
NPV.
Conclusions
This paper presented the application of streamline simulation
and quality maps for optimization of production strategy that
can be used for mature and developed field. These support
tools permitted to reduce the total computational and time
effort, allowing the improvement of the optimization process.
Some particular points can be emphasized:
1. The application of the Ef function determined through
streamline simulation presented good results for analysis of
injector wells.
2. Distribution of injection costs can be generated using
streamline simulation and applied to producer wells analysis.
3. The accurate map presents the best correlation between
quality index and NPV increase. However, more computation
effort was necessary. The use of more or less accurate maps
depends on the compromise between time and computational
resources and the desirable precision.
4. Final quality maps showed better results for producer and
injector wells allocation.
Nomenclature
Ef = efficiency of injector well
G = function related to gas production
IHT = injector well for test
IHW = injector well
Np = cumulative oil production
NPV = net present value
PH = producer well
Qo_m = average oil flow rate
Vop = produced oil volume
Viw = injected water volume
W = function related to water production
Wi = cumulative water injection
Wp = cumulative oil production
References
1. Batycky, R. P., Thiele, M. R., Blunt, M. J. “A Streamline-Based
Reservoir Simulation of the House Mountain Waterflood.”
Stanford Center for Reservoir Forecasting (SCRF). Annual
Report, Stanford U., 1997.
2. Cruz, P. S., Horne, R. N., Deutsch, C. V. “The Quality Map: A
Tool for Reservoir Uncertainty Quantification and Decision
Making.” In: SPE Annual Technical Conference and Exhibition,
Houston, Texas, EUA, SPE 56578, 3-6 October 1999.
SPE 94746 5

3. Grinestaff, G. H. “Waterflood Pattern Allocations: Quantifying 20 2

the Injector to Producer Relationship with Streamline R = 0.8686
Simulation.” In: SPE Western Regional Meeting, Anchorage, 15

Addition NPV (millions US$)

Alaska, EUA, SPE 54616, 26-28 May 1999. 10
4. Guimarães, M. S. “Methodology for Optimization of Drainage
Strategy for Oil Fields in Production”, 129p. M.Sc. Dissertation, 5
Faculdade de Engenharia Mecânica e Instituto de Geociências,
0
Universidade Estadual de Campinas – UNICAMP, Campinas,
2005. [In Portuguese]. -5 0.2 0.4 0.6 0.8 1
5. Maschio C., Schiozer, D. J.: “Analysis of the Performance of
-10
Streamline Simulation”. In: Second Meeting on Reservoir
Simulation, Buenos Aires, Argentina, 5-6 November 2002. -15
6. Mezzomo, C. C., Schiozer, D. J.: “Methodology for Water
-20
Injection Strategies Planning Optimization Using Reservoir Injector Efficiency (Ef)
Simulation”. In: Canadian International Petroleum Conference,
Calgary, Alberta, Canada, 11-13 June 2002. Figure 2 – Linear correlation between NPV and Ef.
7. Nakajima, L., Schiozer, D. J. “Automated Methodology for Field
Performance Optimization Developed with Horizontal Wells.” In:
SPE Latin American and Caribbean Petroleum Engineering Income
200
Conference, Trinidad, West Indies, SPE 81031, 27–30 April 2003 Production cost
[a]. Production + Injection Costs
8. Nakajima, L. e Schiozer, D. J.: “Horizontal Well Placement 150
Optimization Using Quality Map Definition”, Canadian

US$ * 1000
International Petroleum Conference, Calgary, Alberta, Canada,
10-12 June 2003 [b]. 100
9. Thiele, M. R. “Streamline Simulation.” In: 6th International
Forum on Reservoir Simulation, Schloss, Austria, 3-7 September
50
2001.

Acknowledgments 0
The authors would like to thank the Fundação de Amparo a 1800 2300 2800 3300 3800
Pesquisa do Estado de São Paulo – FAPESP, PETROBRAS, Time (days)
CNPq and FINEP (CTPETRO) for their financial support. Figure 3 – Revenue and cost curves for the PH-02.

30 Income Production Costs

Production+Injection Costs

20
US$ * 1000

10

Figure 1 – Three-dimensional horizontal permeability map.

0
1800 2300 2800 3300 3800
Table 1 – Injector well shut down in 1800 days. Time (days)
Wells NPV (millions US$) Ef Figure 4 – Revenue and cost curves for the PH-14.
Initial strategy 105,4 -
IHW-01 107,1 0,610
IHW-02 112,6 0,602
IHW-03 103,6 0,626
IHW-04 113,2 0,639
IHW-05 89,1 0,951
IHW-06 119,2 0,313
IHW-07 94,3 0,920
IHW-10 119,8 0,389
IHW-12 106,7 0,753
6 SPE 94746

Figure 5 – Accurate map generated for 1800 days. Figure 8 – Accurate map generated for 2922 days.

Figure 6 – Intermediate map generated for 1800 days. Figure 9 – Intermediate map generated for 2922 days.

Figure 7 – Less accurate map generated for 1800 days. Figure 10 – Less accurate map generated for 2922 days.
SPE 94746 7

500 500

-500 0.0 0.2 0.4 0.6 0.8 1.0 -500 0.0 0.2 0.4 0.6 0.8 1.0
Addition NPV (US$ *1000)

Addition NPV (US$ *1000)

-1500 2 -1500
R = 0.4486 2
R = 0.5598
-2500 -2500

-3500 -3500

-4500 -4500
-5500 -5500
-6500 -6500
Unit Quality Unit Quality
Figure 11 – NPV x Unit Quality for accurate map generated in 1800 Figure 14 – NPV x Unit Quality for accurate map generated in 2922
days. days.

500 500

-500 0.0 0.2 0.4 0.6 0.8 1.0 -500 0.0

Addition NPV (US$ *1000)

0.2 0.4 0.6 0.8 1.0

Addition NPV (US$ *1000)

-1500 -1500
2
R = 0.4377
-2500 2 -2500
R = 0.2443
-3500 -3500
-4500 -4500
-5500 -5500
-6500 -6500
Unit Quality Unit Quality
Figure 12 – NPV x Unit Quality for intermediate map generated in Figure 15 – NPV x Unit Quality for intermediate map generated in
1800 days. 2922 days.

500 500

-500 0.0 0.2 0.4 0.6 0.8 1.0 -500 0.0

Addition NPV (US$ *1000)

0.2 0.4 0.6 0.8 1.0

Addition NPV (US$ *1000)

-1500 -1500

-2500 2
R = 0.2694 -2500
2
R = 0.3654
-3500 -3500
-4500 -4500
-5500 -5500
-6500
-6500
Unit Quality Unit Quality
Figure 13 – NPV x Unit Quality for less accurate map generated Figure 16 – NPV x Unit Quality for less accurate map generated in
for 1800 days. 2922 days.
8 SPE 94746

100%
Table 2 – NPV resulting of perforation of injector wells. NPV Np Wp
80%
Wells NPV (millions US$) Wi Gp
60%
Initial strategy 105,4

Percent variation
40%
IHT-01 110,0
20%
IHT-02 101,8
0%
IHT-03 101,7

IN

R5

R10

R15

R20

R25

R30

R35

R40

R45

R50
-20%
IHT-04 102,6
-40%
IHT-05 115,3
-60%
IHT-06 117,6
-80%
IHT-07 125,1
-100%
IHT-08 126,6 Optimization runs
Figure 19 – Evolution of the optimization process without the use
of the support tools.

100%
NPV Np Wp
80% Wi Gp
60%

Percent variation
40%
20%
0%

IN
R3
R6
R9
R12
R15
R18
R21
R24
R27
R30
R33
R36
R39
-20%
-40%
-60%
-80%
-100%
Optimization runs
Figure 20 – Evolution of the optimization process with the use of
the support tools.

Figure 17 – Injector well locations in less accurate map generated

in1800 days.

Figure 18 – Injector well locations in less accurate map generated

in 2922 days.