You are on page 1of 16

SPE-170725-MS

Secondary Application of Low Salinity Waterflooding to Forties Sandstone


Reservoirs
S. Law and P. Sutcliffe, Senergy; S. Fellows, CMG

Copyright 2014, Society of Petroleum Engineers


This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Amsterdam, The Netherlands, 2729 October 2014.
This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Low salinity waterflooding (LSWF), versus high salinity waterflooding (HSWF) has been the focus of
significant research at various centres around the world, yet there is still considerable debate over the exact
mechanism that provides incremental oil recovery. The use of the LSWF technique is not widespread in
the United Kingdom continental shelf (UKCS). However, it has been announced that the Clair Ridge
development will deploy low salinity waterflooding (LSWF) in secondary mode from the start of field life,
and a number of companies are currently assessing the applicability of the technique through high level
screening and core flooding. Forecasting the potential oil recovery under LSWF is heavily influenced by
the simulation technique that is used. Presently the most widely discussed approach is the use of a
weighting table with relative permeabilities representing the high and low salinity cases. As the grid block
falls below threshold salinity, the simulator utilises the weighting table to assign an interpolated value of
salinity. This value of salinity is utilised to represent a change in wettability. While this approach
approximates the net effect of LSWF, it does not capture the oil/rock/brine interaction. This study
examines the modelling approach to LSWF utilising an in-house generic Forties Palaeocene model in
CMGs STARS simulator. The conventional approach of modelling LSWF using high and low salinity
relative permeabilities is compared to the latest Multi-component Ion Exchange (MIE) methods by
numerical simulation to assess the impact on incremental oil recovery. A sensitivity analysis is then
carried out on the effects of specific parameters on incremental oil recovery, utilising published data from
fields in the Forties Palaeocene fan system. A discussion is provided. The impact on secondary recovery
was accessed with respect to wettability alteration; injection salinity (LSWF versus HSWF); oil viscosity
and aquifer influx. The application of LSWF in secondary mode to the Forties Palaeocene Sandstones was
found to be favourable for the case of mixed-wet reservoirs.

Introduction
Low salinity waterflooding (LSWF) is an enhanced oil recovery (EOR) technique which is of growing
interest, as it represents a low cost and flexible form of EOR. The technique involves the injection of water
at of a significantly lower salinity, compared to the natural salinity of the reservoir connate water. Until
recently, although it was known that the ionic composition of a fluid flowing in a porous medium does
influence the measured permeability (Schleidegger 1974), the manipulation of this effect to improve oil

SPE-170725-MS

recovery by injecting water of a different salinity and ionic composition to that of the natural formation
water, had not been considered. As compared to the normal method of injecting seawater (HSWF), LSWF
is seen as a viable EOR technique. Further, LSWF offers the potential to increase recoverable oil without
the need for re-engineering of the field, as it can use the existing infrastructure and wells, provided that
facilities space exists topsides for installation of a reverse osmosis plant.
The application of the technique in secondary mode is considered to offer a higher potential gain than
in tertiary mode (Jerauld et al 2008). LSWF in secondary mode is considered to be economic for new
fields, with published field studies in other parts of the world suggesting the technique has been successful
(Vledder et al 2010, Mahani et al 2011). To date in the North Sea, no secondary or tertiary offshore
implementation of low salinity EOR has yet been performed, though plans for the Clair Ridge Phase 2
development, using LSWF in secondary mode, are well advanced.
The merits of applying the LSWF technique in tertiary mode are uncertain. Variable recoveries have
been observed in coreflood results and only one limited field trial in the Endicott field has taken place to
date. Given the uncertainty surrounding incremental oil recovery from tertiary mode, combined with the
cost associated with upgrading older platforms and infrastructure, low salinity EOR in tertiary mode has
not yet been piloted in the UKCS.
The simulation work described in this paper is performed on a generic proprietary Senergy model of
the Forties Sandstone loosely based upon the Arbroath field. The main hydrocarbon bearing reservoir is
the Palaeocene Forties sandstone member which is characterised by laterally extensive submarine
low-stand fan deposits. The facies in the static model have been simplified to consist of channel and
over-bank deposits.
The CMG STARS reservoir simulator was chosed for the modeling work, as in addition to relative
permeability interpolation, it has an Ion Exchange model (MIE) to model the ion exchange between
cations in the aqueous phased and those sorbed on the surface of the reservoir rocks. At this time, it was
chosen to perform the work in isothermal mode.

Discussion on the Mechanisms of Low Salinity Waterflooding


The mechanistic processes behind LSWF have been the subject of much debate within academia and
industry since the publication of the original ideas from Morrow and his research colleagues at the
University of Wyoming in the early 1990s (Morrow et al. 1998). Multiple hypotheses have been proposed
during the last two decades, including fines migration, wettability alteration, multi-component ionic
exchange (MIE), pH modification, desorption and double layer effects. Laboratory analysis performed
under varying conditions, has sometimes led to contradictory results. Field trials have been minimal, but
primarily successful. Thus, although it is widely agreed that a LSWF can lead to increased field oil
recovery, the exact mechanism by which this occurs is poorly understood.
Migration of Fines was the mechanism initially proposed by Tang and Morrow in 1999. They
suggested that less saline brine promotes the dispersion of clay and silt in the formation. These fines
become mobile, following high permeability paths, while lodging and blocking-off smaller pore spaces,
thus causing microscopic diversion of the injection water. Subsequent coreflood work from various
studies, showed either an increase in oil recovery without fines production, or significant clay production
with no additional oil recovery (Boussour et al 2009).
Some studies confirmed an increase in pH associated with a LSWF. An increase in brine pH is thought
to have an effect in the reservoir, similar to an alkaline flood, resulting in generation of in-situ surfactants,
changes in wettability, and reduction in the interfacial tension (McGuire et al 2005). Austad et al (2010)
proposed the idea of desorption by pH increase. As with fines production, neither laboratory data nor field
trials show a consistent relationship between the pH of the produced water and the incremental oil
recovery from a LSWF. Alternatively, Lager (2007) suggested that change in pH is the result of cation
exchange between clay minerals and invading water, and proposed the concept of Multi-component Ion

SPE-170725-MS

exchange as the fundamental mechanism resulting from a LSWF. Ionic exchange between the injected
brine and the formation water, leads to the adsorption of divalent ions, promotes mineral dissolution and
changes the ionic composition of the formation water and the state of wettability.
Thus, although it cannot be specifically argued that one single mechanism is responsible for the
increased oil recovery seen subsequent to a LSWF, what is widely agreed is that the result of the
mechanism(s) is an alteration in the wettability of the reservoir towards increased water wetness, leading
to a reduction in Sor, and thus the production of more oil.

Pilot Testing and Field Implementation


Despite extensive laboratory work over the last decade, field implementations, either secondary or tertiary,
remain few. Particularly in the North Sea, any form of EOR, low salinity or otherwise, requires careful
consideration due to the topside constraints of the offshore environment. In this section, the tertiary
Endicott field trial, and the secondary Omar water flood are described, as the reservoirs can be considered
analogous to North Sea Forties reservoirs. In the North Sea, the Clair Ridge secondary LSWF is a flagship
project in the move to the use of LSWF in the offshore environment.
The first comprehensive inter-well field trial of LSWF took place during 2008-9 in BPs offshore
Endicott Field on the North Slope of Alaska, Seccombe (2010). Endicott is the third largest Alaskan North
Slope field with an estimated oil in-place of around a billion barrels. It came on stream in 1987 and has
since been produced by crestal gas re-injection and peripheral seawater injection. Four single well tests
were undertaken in the Prudhoe Bay and Endicott fields (McGuire 2005), with the saturation change
measured using single well reactive chemical tracer tests (SWCTTs). These tests indicated that the
incremental oil recovery from low salinity water injection could be in the range of 6-12% OIIP.
Subsequent laboratory and simulation work, supported by further single well tests, using both log-injectlog and SWCTTs to measure saturation changes, also indicated a positive response to a LSWF. The results
are described by Jerauld (2006), Webb (2008), and Seccombe (2008).
The Omar oil field in Syria is operated by the Shell subsidiary, Al-Furat. Low salinity water injection
was not originally intended as an EOR technique, but was implemented as a consequence of operational
requirements. The light oil bearing Omar Field, (viscosity 0.3 cp), came on stream in 1989, but suffered
from lack of aquifer support, Vledder et al (2010). Water flooding using an adjacent river water source
with salinity 500 mg/L ( 100 mg/L bivalent ions), began in 1991. In contrast, the formation water has
a salinity of 90,000 mg/L. A detailed analysis of the secondary LSWF in the Omar Field was undertaken
by Vledder (2010). Measurements and observations at 21 wells in the Omar Field are considered by Al
Furat and Shell, to provide conclusive evidence of wettability alteration due to low salinity water flooding
at the reservoir scale. Analysis indicated that the decrease in wettability from the original, predominatly
oil-wet state, to the post-flood, substantially water-wet state, would give an expected incremental oil
recovery of 17% OIIP, (compared to high salinity water flooding). However, a subsequent comparison of
high and low salinity water flooding across Al Furats assets indicated that a more conservative estimate
would be an increase in 5-15% OIIP incremental oil recovery from the LSWF in the Omar Field.
The only UK field where LSWF has been evaluated is the new offshore Clair Phase 2 development
known as Clair Ridge, located around 35 miles west of the Shetland Isles. To improve the recovery factor
for the larger Clair Phase 2 development, various EOR techniques were considered with LSWF identified
as offering the best solution. Reservoir condition core flood experiments were undertaken using core from
Clair Phase 1. The incremental recovery from the experiments varied between 7-10% OIIP, (Buikema
2011). Additionally, imbibition experiments demonstrated that LSWF is able to recover incremental oil
under imbibition conditions in Clair, which is important in this fractured system (Mair 2010). Low salinity
recovery at the reservoir scale was obtained by running a series of sector models using the BP Proprietary
VIP code. Reservoir description was determined to exert the greatest influence on recovery from the Clair
Ridge.

SPE-170725-MS

Elsewhere in the North Sea, based on laboratory work and SWCTTs carried out over the last few years,
other fields are now being proposed for LSWF. Statoil have suggested that the Heidrun, Snorre and
Gullfaks fields are all being considered for low salinity pilots (Spangenberg 2008). More recently, the
Varg Field, operated by Talisman (RezaeiDoust, 2010), and the Fry Field, operated by DetNorskOljeselskap (Hadia 2011), have been evaluated for their low salinity potential.

Numerical Simulation of LSWF


The modelling of LSWF has been undertaken by a number of different researchers with most of the early
studies modelling the LSWF response based on established modelling approaches for chemical EOR.
More recent techniques have concentrated on the implementation and use of ion exchange models which
aim to replicate the geochemical reactions which are believed to occur when the low salinity water
interacts with the reservoir environment. Early LSWF numerical simulation studies by Jerauld et al
(2006), Idowu et al (2011) and Kristensen (2011) used net wettability alteration as the simulation
mechanism, but did not address the chemical reactions which are taking place between the brine, the oil
and the reservoir rock e.g. the effects of ion exchange and desorption.
BP modelled LSWF using their customised VIP simulator code, Jerauld et al (2006). They modelled
salt (or total dissolved solids) as a single component in the aqueous phase which could be initialised,
injected and tracked. The salinity determined the viscosity and density of the aqueous phase. Upper and
lower salinity thresholds were set, and relative permeability and capillary pressure were made a function
of salinity between the threshold values. Above the upper salinity threshold, the high salinity curves were
used; below the low salinity threshold, low salinity curves were used. In between, the curves were
interpolated as a function of the salinity. Additionally, residual oil saturation as a function of salinity was
utilised to provide greater definition of the wettability shift. Hysteresis between imbibition and secondary
drainage was addressed, and an allowance made for isolating a proportion of the connate water to
represent the impact of connate water banking. Jerauld et al (2006) also provided an extensive discussion
on the importance of mixing during LSWF, and as a consequence, the impact of dispersion on numerical
modelling. They proposed that numerical dispersion may be used to represent physical dispersion.
Fine-grid resolution is necessary to accurately represent the appropriate physical dispersion, but is
impractical at the field scale. In these circumstances, the authors proposed pseudoizing the relative
permeability curves and modifying the salinity dependency to better match finer-grid results to coarse grid
simulations.
Mahani (2011) described the addition of LSWF capability to the Shell in-house simulator. Their
modeling process was slightly different to that described by Jerauld (2006). A slug of a low salinity phase
was defined to which was assigned a set of low salinity oil-water relative permeability curves. As long
as the proportion of low salinity water in a grid block remained beneath a specified threshold, the original
high salinity relative permeability curves were used, but once the threshold was exceeded (typically
around 90% in the case presented), the low salinity relative permeability curves were used. There was
apparently no interpolation between the two sets of relative permeability curves.
Verma et al (2009) utilised the EMpower simulator developed by Exxon-Mobil to model LSWF in
clastic reservoirs. A salt tracer associated with the aqueous phase was introduced to enable the tracking
of salinity variations. Relative permeability, capillary pressure, residual saturations, viscosity, density and
formation volume factor were modelled as a function of salinity. Water properties at a minimum and
maximum salinity were interpolated to derive the properties at intermediate salinities. Additionally, an
allowance for salt adsorption/desorption was calculated. The approach is comparable to that of Jerauld
(2006). Subsequently, Gupta (2011) extended the modelling work to include Advanced Ion Management.
These early modelling studies demonstrated the key behaviours associated with LSWF. However, the
approach of treating salt as a single-lumped component in the aqueous phase which was then used to

SPE-170725-MS

Table 1High level screening criteria utilised for Forties Palaeocene Sandstone based on Prudhoe Bay/Endicott field
Suitability Indicator

Endicott/Prudhoe Bay

Reference

Forties Formation

Reference

Initial wettability state


mixed-wet

Mixed-wet

Jerauld et al (1997)

Mixed-wet to water-wet

Hughes et al (1990)
Turran et al (2002)
Hughes et al (1990)
Davis et al (2009)

Clay content of the


formation between
7%-30%
Reservoir temperature

Kaolinite content between


9 to 45%, Average
20%
210F

Almond et al (1990)

Clay content 12 to 30%,


Average 19%

McGuire et al (2005)

Presence of polar
compounds in the
oil
pH of formation water
less than 7
Presence of soluble
minerals in
formation
Moderate Formation
Salinity

Acid number 0.12mg


KOH/g

Buckley (1998)

169F
205F
0.09mg KOH/g

Brand et al (1996)
Hughes et al (1990)
Chevron (2013)
BP (2013b)

7 to 8

McGuire et al (2005)

5.6

Mitchell et al (1980)

97 ppm to 247 ppm of Ca


25 ppm to 156 ppm of
Mg
28,000ppm (32 kg/m3)

McGuire et al (2005)

3110 ppm of Ca 480


ppm of Mg

Mitchell et al (1980)

McGuire et al (2005)

55,000ppm (55kg/m3)

Carter et al (2003)

control the wettability shift, neglected the more complex oil/brine/rock reactions. Subsequent studies by
Sorbie and Collins (2010), Wu et al (2009), Omekeh et al (2012) and, most recently, Dang et al (2013),
sought to address the complex geochemical processes in the reservoir.
Dang et al (2013) modelled LSWF behaviour by developing a comprehensive ion exchange model with
geochemical processes including intra-aqueous and mineral reactions, which they coupled to the multiphase, multi-component flow equations in the equation-of-state compositional simulator GEM. They
demonstrated that the new model was able to capture most of the important physical and chemical
phenomena that ensue during a LSWF, in particular, intra-aqueous reactions, mineral dissolution /
precipitation, and wettability alteration. The use of the Ca equivalent fraction as the interpolant was
found to be sufficient to match the experimental coreflood data. The GEM model was validated against
the ion-exchange model of the geochemistry software PHREEQC, and showed strong agreement in ion
evolution. Additional validation was undertaken using the results of low salinity coreflood studies on a
North Sea reservoir, Fjelde et al (2012), and a Texas reservoir (Rivet 2009). In both cases, the new model
was able to generate excellent matches to the effluent ion concentrations, and effluent pH and oil recovery
as measured in the laboratory. From these results, it was deduced that the new model was accurately
capturing the main mechanisms of LSW, providing a robust tool for LSW design and process performance
interpretation in the field.

LSWF Screening of the Forties Sandstone Reservoirs


The potential for LSWF is dependent upon the initial reservoir wettability, formation pH, salinity,
temperature, clay content, presence of polar compounds in the oil, and soluble minerals in the formation.
Where research and analysis indicates, there are good opportunities for the use of the low salinity EOR
technique to enhance incremental oil recovery, but there may also be additional holistic benefits with the
potential to reduce reservoir souring and realise additional cost savings in fluid handling (Collins et al
2011).
When considering possible low salinity targets in the North Sea, the Forties Palaeocene sandstone
formation has similar characteristics to the Prudhoe Bay/Endicott field which may make it suitable for the
application of LSWF (Table 1). The Forties Palaeocene sandstone formation includes the Arbroath,
Arkwright, Forties, Montrose, Nelson, and Pierce oilfields. The fields were deposited within the same
depositional Forties fan system and have similar properties in terms of reservoir depth, pressure,

SPE-170725-MS

Table 2Depth, Initial Pressure, Temperature and Static Properties of the Forties Sandstone Fields (Kunka et al 2003, Hogg 2003, Carter et al
2003)

Field Name

Depth
to
Crest
(feet)

Initial
Pressure
(psia)

Arbroath
Arkwright
Forties
Gannet South
Montrose
Nelson

8030
8500
6660
5,550
8040
7192

3700
3700
3215
2,640
3744
3322

Average Static Properties


Temperature
F
245
250
96
169
257
224

Formation
Age

Porosity
(%)

Palaeocene
Palaeocene
Palaeocene
Palaeocene
Palaeocene
Palaeocene

Permeability
(mD)

Net to
Gross

1 to 4,000
20 to 50
700 to 4,000
200 to 2,400
80 to 2,000
7 to 1,610

0.50
0.78
0.65
0.70
0.50
0.7

24
19
0.27
28
24
23

Table 3Reservoir fluid properties of the Forties Sandstone Fields (from Kunka et al 2003, Hogg 2003, Carter et al 2003)
Field
Name
Arbroath
Arkwright
Forties
Gannet
South
Montrose
Nelson

Petroleum
Saturation
(%)

Oil
Gravity
( API)

Oil Viscosity
(cP)

55
51
85
unknown

38-42
38-42
37
19.6

0.4
0.43
0.76
15.3

55
unknown

40
40.6

0.32
unknown

Gas to
Oil Ratio
(scf/stb)

Oil FVF
(rb/stb)

Formation Water
Salinity (ppm)

1991
2670
1142-1390
1,142

490
700
300
114

1.327
1.456
1.22
1.085

135,000
55,000
55,000
unknown

2348
1550-1699

600
555

1.467
1.357

111,000
84,000

Bubble
Point
(psia)

temperature and static properties for the fields (Table 2). The reservoir fluid properties are similar in terms
of API gravity (Table 3), but vary in terms of other parameters, in particular oil density, oil viscosity, and
formation water salinity.

Modelling approach in STARS


Both the CMG STARS and GEM simulators provide the capabilities to model a range of effects
considered to be involved in Low Salinity Waterflooding. It was decided to perform the current work
using the STARS simulator. Accordingly, the description below focuses on the STARS implementation.
GEM provides enhanced Low Salinity modelling capabilities in an equation of state compositional
simulator, thereby extending the range of phenomena which can be simulated.

Multi-Component Ionic Exchange


STARS has an Ion Exchange (MIE) model to model the ion exchange between cations in the aqueous
phase and those sorbed on the surface of rocks. The ion exchange capacity reflects the varying level of
clays, (kaolinite, montmorillonite, etc.), in the porous media and is represented by a charge balance
equation of the adsorbed ion pairs. As clay content varies, the ion exchange capacity varies according to
rock type, and this dependency can be built into the simulation. The ion exchange reaction itself is
represented as an exchange equilibrium reaction. STARS uses a format to represent the exchange
equilibria which allows for a generalization to multiple rock types by means of an ion exchange capacity
input parameter. With respect to multiple rock types, only one pair of cations may be specified in a data
set, but different exchange parameters can be specified for each rock type within the individually defined
adsorption sets.

Wettability
Ion exchange facilitates adsorption/desorption of surface active components, encourages mineral dissolution and leads to an increase in the pH of the formation water, all of which promote a shift in the

SPE-170725-MS

wettability of the rock/water interface. In STARS, the resulting change in the shape and end points of the
relative permeability curves can be modelled using ion exchange as the mechanistic effect. This is
accomplished by combining the relative permeability interpolation keyword, with a sub-keyword which
links the rock-fluid interpolating parameter to variation in a specific component. In this context, the effect
of wettability alteration is modelled by shifting the relative permeability curves from a less water wet to
a more water wet scenario. Multiple relative permeability tables can be defined for a reservoir rock type,
where each table corresponds to one value of a specified interpolant. Typically, two sets of relative
permeability curves representing low salinity and high salinity conditions are considered and an interpolation between these two curves is performed. The interpolation parameter can be the aqueous ion
concentration, the adsorbed ion concentration, the surface active acid component concentration in the oil
phase, or the adsorbed species concentration. Further adjustment to the shape of the curves may be
implemented by use of curvature interpolation parameters which can be defined to correspond to
laboratory results as required by the user.

Rock Water Interactions


Rock matrix and clay reactivity due to both ion exchange and the effects of changing water pH, vary
considerably both within a given rock type, and between rock types. Porosity and permeability may be
significantly altered by these processes which may be chemical (dissolution, precipitation), or mechanical
(eg. blockage, dispersion), or some combination of these mechanisms. With respect to mechanical
phenomena, permeability alteration often accompanies sorption. The simulator accounts for this via region
dependent resistance factors RRFT which allow correlation of local permeability with local sorption
levels. In this manner, STARS allows a phenomenological description of these phenomena, which the user
can apply in specific contexts, according to the relevant coreflood results.

Combined EOR Techniques


Low Salinity Water-flooding is increasingly being considered in conjunction with proven EOR techniques
such as Polymer and Surfactant flooding. The STARS simulator allows for two levels of interpolation of
relative permeabilities, namely multiple relative permeability sets within a rock type, and relative
permeability interpolation across rock types. Thus different wettability states can be defined as different
rock types, and interpolation modelled between rock types eg. by a suitable interpolation parameter such
as aqueous Ca. Within a rock type, the IFT reduction effect of a surfactant can be represented by
interpolation of different log Nc values, where Nc refers to Capillary Number. This can be further
combined with modification of the aqueous phase properties resulting from a Polymer flood. (STARS
User Guide 2013). The flexibility provided by a three-phase multi-component thermal and steam
additive simulator such as STARS, facilitates the modelling of the multiple phenomena involved in
combining a LSWF with Polymer and Surfactant injection and provides the ability to address a range of
rock brine interaction scenarios.

Additional Low Salinity Functionality


Additional Low Salinity functionality associated with geochemical processes such as intra-aqueous and
mineral reactions, (dissolution and precipitation), is provided within GEM via a full Geochemical reaction
model. This allows the modeling of carbonate equilibrium, including calcite mineral dissolution, and also
the hydrogen sodium (H -Na) ion exchange, both of which reactions are implicated in the increase of
pH observed in some low salinity floods. Further, the ion exchange model in GEM allows for multiple ion
exchange equilibria to be defined according to the users requirement, and as determined by laboratory
coreflood results. GEM has the additional ability to model enhanced low salinity processes such as the low
salinity CO2 WAG process. (Dang et al 2014).

SPE-170725-MS

Figure 1Location of the fields within the Forties Sandstone of the UKCS (From Hogg et al 2003)

Sector Model Description of the Forties Sandstone Reservoir


Static Model
The static model is an unfaulted structural model populated with facies and properties based on published
data for the Arbroath/Montrose cluster. The Montrose and Arbroath fields, (including Arkwright), are
located in the Central Graben of the UK North Sea (Figure 1). The fields comprise three separate four-way
dip anticlinal structures that lie in the centre of the main Forties sandstone member depositional fairway
to the south-east of the Forties and Nelson fields (Hogg, 2003). The main hydrocarbon bearing reservoir
in the cluster is the Palaeocene Forties sandstone member. The reservoir is characterised by laterallyextensive submarine low-stand fan deposits. Localised thick channels trend NW-SE are widespread and
in places amalgamation has led to sheet like morphology. The individual 5 - 10 ft thick sedimentary
packages are stacked forming 50 - 100 ft thick sand bodies. Each thick sand body is separated by
laterally-extensive mudstone which is sometimes thick enough to form local seals and baffles. The facies
in Montrose and Arbroath are interpreted as turbidite channel or sandlobe deposits.
The facies in the static model have been simplified to consist of channel and over-bank deposits. An
assumption of 50 % channel and 50 % overbanks has been applied for this model. The channel thickness
to width ratio was obtained from a regional analogue database for the Forties. The channels in the Forties
sandstone are known to exhibit relatively good reservoir quality. The levees or overbanks are known to
contribute to flow in places. Net to gross ranging from zero has been assumed to introduce some form of
baffle to flow which is characteristic of the Forties. Mudstones which are laterally extensive in the system
are not modelled explicitly but are set as two no flow boundaries in STARS (Figure 2). The reservoir static
properties (net to gross, porosity and permeability) are based upon published data sourced primarily from
Hogg (2003). The 3D view of the sector model is shown in Figure 3 with the well locations. It should be
noted that the static model is not a schematic of the Arbroath or other Forties sandstone fields rather it is
constructed to be representative of the structure, geology and petrophysical properties found in Arbroath
based upon public domain information.

SPE-170725-MS

Figure 2Laterally extensive mudstones (red) represented in the Forties generic reservoir model

Figure 3Oil water contact in the model with well locations

Dynamic Model Properties


The model contains simplified compositional PVT to represent fluid behaviour in the reservoirs. The oil
and water formation volume factors are 1.1 rb/stb and 1.0 rb/stb respectively. The reservoir operates above
the bubble point during field life and as such the field is modelled as dead oil with only oil and water
phases in the reservoir. The initial wettability of the reservoir is modelled as mixed-wet with a wettability
shift during low salinity causing the reservoir to become water-wet. Three active producers Prod_D1,
Prod_D2 and Prod_D3 produce from the sector model and are supported by two wells Inj_D4_New and
Inj_D5_New. The injectors are sited close to the oil water contact to allow most of the pressure support
to enter the oil leg during production. The producers are operating on a minimum BHP of 1500 psia with

10

SPE-170725-MS

Figure 4 Comparison of oil recovery observed for LSWF modeling approach and injection salinity

no maximum flowrate constraint. The Injection wells are set to voidage to balance the production well
withdrawal rates. Although ion exchange properties for the Forties sandstone vary across the fan system
published, constant values for sandstone are utilized in the absence of measured coreflood data.

Discussion and Results


Sensitivity analyses were run to demonstrate the potential for low salinity in the Forties sandstone
reservoirs focusing on parameters which were most likely to impact recovery:

Modelling approach: comparison of the effect of incorporating ion exchange between the rock
surface and aqueous phase, versus standard relative permeability interpolation on salinity.
Salinity Injection response: benchmark the recovery under low salinity waterflood compared to the
high salinity case.
Oil Viscosity: the oil viscosity is known to effect mobility and therefore recovery under both high
and low salinity waterflood. A comparison for the range of oil viscosities in the Forties sandstone
reservoirs was run.
Aquifer input: the forties sandstone reservoirs are pressure supported by both basal and edge
aquifers. Aquifer influx effects were modelled for cases of both basal and edge scenarios to assess
the impact on LSWF.

Modelling approach to LSWF


In the literature there is an emerging body of research which suggests that Multi-component ion exchange
is one of the main mechanisms associated with LSWF. Until recently the approach has been to model
LSWF utilising two sets of relative permeabilities representing the high and low salinity case. While this
method is guided by corefloods it tends to consider the net effect i.e. the net wettability shift and does not
account for the complex oil/brine/rock reactions. In STARS, ion exchange capacities for multiple rock
types, can be set to characterize the interaction between the oil/rock/brine. Generic constants for ionic

SPE-170725-MS

11

Figure 5Comparison of Oil recovery for HSWF and LSWF in the sector model for different oil viscosities

exchange in sandstones were used to run sensitivities to assess the influence on oil recovery (Figure 4).
The difference between the traditional relative permeability interpolation and the use of an ion exchange
model for this case appears relatively small. The difference of 4.8% in recovery factor for the LSWF case
with and without MIE does however confirm that when conducting this type of analysis, the oil/rock/brine
reactions should be considered. The presence of clay will influence the ion exchange capacity of the rock
and should be obtained from laboratory testing. It should be noted that core taken for use in testing may
not be fully representative of the reservoir as a whole.

Modelling the injection salinity response


The simulation requires accurate coreflood data which can be used to predict the response of the reservoir.
In the absence of such measured reservoir data two sets of relative permeability curves were utilised with
the thresholds described by McGuire et al (2004). These thresholds define the salinity at which a reaction
will be observed. The Ca ion concentration in the injected fluid guides the interpolation between the
high and low salinity relative permeability curves. The ion exchange model controls the sorption of Ca
between mineral surfaces and the aqueous phase, thus further impacting the salinity of the LSWF. The
injected fluid enters the reservoir and mixes with the connate brine. As the salinity in the grid block falls
below threshold salinity the simulator switches from high salinity to low salinity relative permeability.
The Ca ion concentration varies according to the salinity of the injected fluid. The influence of the
injected low salinity water was assessed for high and low salinity (Figure 4). The selection of a threshold
for the reaction in the simulation is crucial to approximate LSWF behaviour. If the threshold is too low
and the injection fluid salinity is too low there is a risk of clay swelling and formation damage. Therefore,
selection of an optimal low salinity is crucial to predict recovery and avoid unwanted effects in the
reservoir. The relative permeability values for high salinity water flood HSWF and LSWF are conservative so as not to over-estimate the response. Further, the default ion exchange properties used for the

12

SPE-170725-MS

Figure 6 Comparison of Watercut development for HSWF and LSWF in the sector model for different oil viscosities

facies will limit the potential for oil/brine/rock reactions and specific properties should be utilised when
assessing the actual Forties type field specific potential for LSWF. In this context, the incremental
recovery due to LSWF of 8.1 % in this study is considered low.

Oil Viscosity
Oil viscosity is a key parameter in assessing the potential recovery of a reservoir. The literature of the
Forties sandstone fields (Table 3) identifies the viscosity in the fields as varying in the range of 0.32 cP
to 15.3 cP. The influence of viscosity on the oil recovery is assessed for values of viscosity of 0.32cP,
0.76cP and 15.3cP (Figure 5). As would be expected, it is observed that under HSWF by increasing the
viscosity of oil the oil recovery is reduced. This is to be expected since higher viscosity oil tends to have
a less favourable mobility ratio. The introduction of low salinity water into the reservoir is observed to
increase recovery for all values of viscosity. The change in relative permeability endpoints between
HSWF and LSWF combined with the different oil viscosities leads to different mobility ratios. The
recovery is a function of the relative permeability end points since the viscosity of water is held constant
regardless of salinity. The incremental oil due to LSWF was found to be in the range of 4.2 %, 3.5 % and
2.3 % OOIP for oil viscosities of 0.32 cP, 0.76 cP and 15 cP. The influence of LSWF on watercut behavior
(Figure 6) suggests that utilsing LSWF can reduce the degree of water breakthrough in secondary mode.
The stepped nature of the plot is influenced by the facies in the Forties sandstone model. Initially, the
channel facies water out for the HSWF cases with later contributions to flow coming from the overbank
deposits. When LSWF is utilised in secondary mode the injected fluid enters the reservoir displacing the
oil and connate water bank ahead of the LSWF. The effect of the LSWF is to create a net wettability shift
which reduces residual oil saturation which leads to a slower development of watercut as more oil is
mobilised towards the producer.

SPE-170725-MS

13

Figure 7Influence of aquifer type and influx on oil recovery

Aquifer Influx
The Forties sandstone reservoirs are pressure supported from a large regional aquifer for which the influx
terms will vary according to the specific field, as the fields within the Forties fan system are supported
by different aquifer configurations. The Forties (Wills et al. 1991) and Nelson fields (Kunka et al. 2003)
are supported by basal aquifers. The Arbroath field contains oil bearing sandstones which have a
mudstone layer at their base (Crawford et al 1991) and can be considered to be edge aquifer supported.
Sensitivities were run to assess the effect of edge and basal water drives on oil recovery. Scenarios
representing a basal aquifer plus LSWF and an edge aquifer plus LSWF, were compared to the base case
LSWF.
There is a positive response in oil recovery, ( 4% incremental recovery), to the presence of the
aquifer, but the results are inconclusive, with both edge and basal aquifer support resulting in similar
recoveries. Potential dilution effects on the LSWF due to the higher salinity aquifer influx are masked.
Possible reasons for this include the small sector scale of the model, in which the injectors are located
close to the OWC, and both producers and injectors are preferentially located in the channelized facies,
which limits the ability of the aquifer to dilute the LSWF significantly and thus adversely impact oil
recovery. Further work is required in this area.

Conclusions
It has been demonstrated that the incorporation of ionic exchange into the modeling process, versus
relative permeability interpolation on salinity alone, impacts the modeled oil recovery. It is concluded that
this is a more physical representation of the reservoir processes during a LSWF.
Based upon this modeling technique, the following conclusions could be drawn from this high level
screening and simulation study of the Forties Generic Sandstone Reservoir Model:

The properties of the Forties sandstone reservoirs are analogous to the Endicott/Prudhoe bay
reservoirs in terms of initial wettability, Kaolinite clay content, reservoir temperature, acid number

14

SPE-170725-MS

and the presence of soluble minerals in the formation. As such, the Forties reservoirs could be well
suited for low salinity waterflooding.
A generic model of the Forties sandstone based on the geology of the fan system consisting of
channels and overbanks was utilised. Conservative relative permeability data was assigned to both
the channels and overbanks to represent the differences in facies response. The relative permeability curves influenced the recoverable oil from both the HSWF and LSWF cases. In the absence
of measured ion exchange data for the Forties sandstone reservoirs, generic values were assigned
in STARS. The ion exchange capacity influences the oil/rock/brine reactions in the reservoir and
as such the low increase in incremental recovery is attributed to the parameters selected., The ion
exchange reactions will be specific to the sandstones of specific reservoirs within the Forties
Sandstone and should be sought before simulating the effects.
The oil recovery was found to be dependent upon viscosity which is consistent with expected
behaviour under high salinity waterflooding (HSWF). The LSWF application was found to be
beneficial in all cases with an incremental increase in recovery of 4.2 to 2.3 % for oil viscosities
of 0.32 cP to 15.3 cP.
The effect of aquifer influx was assessed to compare basal and edge aquifer effects. For the
specific case of the Drumoak model in STARS the type of aquifer was found to have a small effect.
The placement of the injection wells at the OWC, and the location of the injection/production wells
in the channel facies is thought to limit the potential aquifer influx.

Recommendations for further work


The effect of ion exchange constants should be examined for the range of values in

the Forties sandstone reservoirs to assess the differences in recovery when modeling
utilising the ion exchange and relative permeability options.
The effect of aquifer influx should be to reduce the dilution effect of the low salinity
waterflooding. The degree of dilution will be affected by model size, aquifer influx
and the salinity of the reservoir initially. Further study should be undertaken to
consider these parameters at a larger scale.
Additional low salinity functionality associated with geochemical processes such as
intra-aqueous and mineral reactions, should be accessed using a simulator with a
full Geochemical reaction model as well as Ion Exchange Model, such as the CMG
GEM simulator.
The existing work was performed under isothermal conditions. It is recommended
that further work should look at the impact of temperature dependency on the LSWF
process.
Acknowledgements
The authors wish to acknowledge the assistance given by both LR Senergy and CMG in assisting in the
preparation of this paper.

References
Austad, T., Rezaeidoust, A., Puntervold, T. (2010), Chemical Mechanism of Low Salinity Water
Flooding in Sandstone Reservoirs, SPE 129767
Boussour, S., Cissokho, M., Cordier, P., Bertin, H., Hamon, G. (2009), Oil Recovery by Low Salinity
Brine Injection: Laboratory Results on Outcrop and Reservoir Cores, SPE 129767

SPE-170725-MS

15

Buikema, T.A., Mair, C., Williams, D., Mercer, D., Webb, K.J., Hewson, A., Reddick, C.E.A.. and
Robbana, E. (2011), Low Salinity Enhanced Oil Recovery Laboratory to Day One Field Implementation-LoSal EOR into the Clair Ridge Project, IOR 2011
Carter, A. and Heale, J. (2003), The Forties and Brimmond Fields, Blocks 21/10, 22/6a, UK North
Sea, Geological Society, London, Memoirs 2003, v.20, p557561
Collins, I.R. (2011), Holistic Benefits of Low Salinity Waterflooding, B31, 16th European Symposium on Improved Oil Recovery, Cambridge, UK, 12th to 14th April 2011.
Crawford, R., Littlefair, R.W. & Affleck, L.G. (1991), The Arbroath and Montrose Fields, Blocks
22/17, 18, UK North Sea, Geological Society, London, Memoirs 1991, v.14, p211217
Dang, C.T.Q., Nghiem, L.X., Chen, Z.J., Nguyen, Q.P. (2013), Modeling Low Salinity Waterflooding: Ion Exchange, Geochemistry and Wettability Alteration, SPE 166447
Dang, C.T.Q., Nghiem, L.X., Chen, Z.J., Nguyen, N.T.B., Nguyen, Q.P. (2014), CO2 Low Salinity
Water Alternating Gas: A New Promising Approach for Enhanced Oil Recovery, SPE Improved Oil
Recovery Symposium, 12th 16th April, Tulsa, Oklahoma USA. SPE 169071.
Fjelde, I., Asem, S.M., Omekeh, A.V. (2012), Low Salinity Water Flooding Experiments and
Interpretation by Simulations, SPE Improved Recovery Symposium, 14th 18th April, Tulsa, Oklahoma
USA. SPE 154142.
Gupta, R., Smith, G.G., Hu, L., Willingham, T., Cascio, M.L., Shyeh, J.J., Harris C.R.., (2011),
Enhanced Waterflood for Carbonate Reservoirs Impact of Injection Water Composition, SPE 142668.
Hadia, N., Lenhe, H.H., Kumar, K.G., Selboe, K.A., Stensen, F.A., Torsater, O. (2011), Laboratory
Investigation of Low Salinity Waterflooding on Reservoir Rock Samples from the Fry field, SPE
141114
Hogg, A.J.C. (2003), The Montrose, Arbroath and Arkwright Fields, Blocks 22/17. 22/23a, UK North
Sea, Geological Society, London, Memoirs 2003, v.20; p611616
Idowu, J., Somerville, J., Adebari, D., Meshioye, O. (2011), Effect of Salinity Changes of the
Injected Water on Water Flooding Performance in Carbonate Reservoirs, SPE 150816
Jerauld, G.R., Webb, K.J., Lin, C.-Y., Seccombe, J.C. (2008), Modelling Low-Salinity Waterflooding, SPE Annual Technical Conference and Exhibition, 24th-27th September, SPE 102239
Jerauld, G.R., Webb, K.J., Lin, C.-Y., Seccombe, J.C. (2008), Modelling Low-Salinity Waterflooding, SPE Reservoir Evaluation & Engineering, SPE 102239
Kristensen, M.R., Ayan, C., Ramamoorthy, R., Cig, K. (2011), Feasibility of an EOR MicoPilot for
Low-Salinity Water Flooding, IPTC 14507
Kunka, J.M., Williams, G., Cullen, B., Boyd-Gorst, J., Dyer, G.R., Garnham, J.A., Warnock, A.,
Wardell, J., Davis, A. & Lynes, P. (2003), The Nelson Field, Blocks 22/11, 22/6a, 22/7, 22/12a, UK
North Sea, Geological Society of London, Memoirs 2003, V.20; p617646
Lager, A., Webb, K.J. and Black, C.J.J (2007), Impact of Brine Chemistry on Oil Recovery, 14th
European Symposium on Improved Oil Recovery, EAGE.
Mahani, H., Sorop, T., Ligthelm D.J., Brooks, D., Vledder, P., Mozahem, F., Ali, Y. (2011), Analysis
of field responses to low-salinity waterflooding in secondary and tertiary mode in Syria, SPE 142960
Mair, C. (2010) Clair Ridge LoSal EOR Case Study: Laboratory Measurement to Front End
Engineering Design, BP Exploration & Production, IEA Workshop & Symposium, Aberdeen, 18th
20th October 2010.
McGuire, P.L., Chatham, J.R., Paskvan, F.K., Sommer, D.M., Carini, F.H. (2005), Low Salinity Oil
Recovery: An Exciting New EOR Oppurtunity for Alaskas North Slope, SPE 93903.
Morrow, N.R., Tang, G., Valat, M. and Xie, X. (1998), Prospects of Improved Oil Recovery Related
to Wettability and Brine Composition, J Pet. Sci. Eng. 20(3 4), pp267276
Omekeh, A., Friis, H.A., Fjelde, I. Evje, S. (2012), Modeling of Ion-Exchange and Solubility in Low
Salinity Waterflooding, SPE 154144

16

SPE-170725-MS

RezaeiDoust, A., Puntervold, T., Austad, T. (2010), A Discussion of the Low Salinity EOR Potential
for a North Sea Sandstone Field, SPE 134459
Rivet, S.M. (2009), Core Flooding Oil Displacements with Low Salinity Brine MSc Thesis,
University of Texas
Schleidegger, A.E. (1974), The Physics of Flow in Porous Media, Monograph
Seccombe, J., Lager, A., Jerauld, G., Jhaveri, B., Buikema, T., Bassler, S., Denis, J., Webb, K.,
Cockin, A., Feug, E. (2010), Demonstration of Low-Salinity EOR at Interwell Scale, Endicott Field,
Alaska, SPE 129692
Sorbie, K.S. and Collins, I.R. (2010), A proposed pore-scale mechanism for how low salinity
waterflooding works, SPE 129833
Spangenberg, D et alet al.. (2008), Low Salinity Waterflooding: Opportunities and Challenges for
Field Pilot Tests, presentation at the FORCE seminar on Low Salinity, NPD, Stavanger.
STARS User Guide 2013, Computer Modelling Group, Calgary, Canada
Tang, G. and Morrow, N. (1999), Influence of Bring Composition and Fines Migration on Crude
Oil/Brine/Rock Interactions and Oil Recovery, Proceedings of the 5th International Symposium on
Evaluation of Reservoir Wettability and its Effect on Oil Recovery, Trondheim, Norway, June 1998 and
published in J. Pet. Sci. Eng, 24: 99 111
Verma, S.K., Adibhatla, B., Leahy-Dios, A., Willingham, T. (2009), Modeling Improved Recovery
Methods in an Unstructured Grid Simulator, IPTC 13920
Webb, K. (2008), The LoSal EOR process: From Laboratory to Field, presentation at the FORCE
seminar on Low Salinity, NPD, Stavanger
Wills, J.M. (1991), The Forties Field, Block 21/10, 22/6a, UK North Sea, United Kingdom Oil and
Gas Fields, 25 years Commerative Volume, Geological Society Memoir No. 14, pp 301308.
Wu, Y.S. and Bai, B. (2009), Efficient Simulation for Low Salinity Waterflooding in Porous and
Fractured Reservoirs SPE 118830.
Vledder, P., Gonzalez, I.E., Fonseca, J.C.C., Wells, T., Ligthelm, D.J. (2010), Low Salinity Water
Flooding: Proof of Wettability Alteration On A Field Wide Scale, SPE 129564

You might also like