SPE 121982

Fluid Pressure Gradients in Tight Formations

J.A.Larsen, SPE, StatoilHydro Consultant, H. Urkedal SPE and A. Lønøy, StatoilHydro ASA

Copyright 2009, Society of Petroleum Engineers

This paper was prepared for presentation at the 2009 SPE EUROPEC/EAGE Annual Conference and Exhibition held in Amsterdam, The Netherlands, 8–11 June 2009.

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-permeable reservoirs have long been recognized as a challenge for economical production. Characterization of complex
carbonate reservoirs exhibits some special challenges. Pay zones may be poorly defined due to marginal reservoir properties,
i.e., uncertainties regarding the distribution of oil on the pore scale and the ability of oil to flow at high initial water
saturation. Hence, the oil water contact may not be a precise level in the reservoir and the vertical span of the oil-water
transition zone may be greater than 100 meters in such reservoirs. The traditional reservoir engineering approach defines free
water level and hence the water- and oil-zones from the oil- and water pressure gradients. Pressure gradients are, however,
difficult to obtain in low-permeable media and it has been reported that supercharging effects may be the result of mobile oil
and water at the same place in the reservoir.

A case study is made for an offshore exploration well in a complex and heterogeneous carbonate oil reservoir. The well is not
production tested and gives a clear water gradient. However, there is indication of relatively high oil saturation with possible
live oil properties. This is not typical for a reservoir that has been completely water washed during historical time.

Different initial fluid distributions are studied in simulation of filtrate invasion in order to explain the observations in the
exploration well. The results are of general interest, with application to many low-permeable reservoirs. This work gives new
insight into the important interplay between type of drilling mud, reservoir wettability and interpretation of fluid gradients.
The novel implication of the results is that a water-gradient is measured in the transition zone if an oil-wet reservoir is drilled
with water-based mud. A water gradient can therefore exist even though the oil saturation is high. The shift between the
water gradient and oil gradient occurs when mobile oil is present in water-wet pores giving a positive oil-water capillary
pressure.

Introduction
The fluid distribution in oil reservoirs reflects the accumulation history and hence chemical, biological and geological
processes during millions of years. The complexity of such processes seems to be significantly and the outcome difficult to
predict. However, by experience it is known that some simple physical laws may adequately explain today’s water and oil
saturation without detailed knowledge of the accumulation history during millions of years.

The main assumptions made for practical engineering purpose are i) the equilibrium between gravity and capillarity; and ii)
the oil has originally migrated into a water-wet environment. The initial conditions in a homogenous reservoir with high
permeability can therefore most often be described easily from the primary drainage process and the reservoir divides into
clearly defined oil and water zones. Hence, water is the only mobile phase in the water zone and oil is the only mobile phase
in the oil zone. The initial fluid mobilities are thus a matter of single phase flow with possible modifications for the connate
water saturation or the residual oil saturation (if paleo oil is present). Fluid gradients will reflect the mobile phase and
capillary pressure effects on the gradient can be neglected. The most common reservoir engineering interpretation of fluid
gradients is accordingly that an oil gradient implies producible oil and a water gradient implies either 100% water saturation
or in some special cases water and residual oil saturation (paleo zone). This is most often supported by the petrophysical
evaluation with a rapid change in resistivity at the oil-water contact (OWC). The free water level (FWL) is per definition the
position where the oil-water capillary pressure is zero and is estimated to be where the extrapolated oil and water gradients do
cross.
2 SPE 121982

The FWL is located at the OWC or slightly below for the high permeable reservoir accumulated according to the primary
drainage process in an initial water-wet reservoir. The shift between FWL and OWC comes directly from the capillary entry
pressure between oil and water given from the primary drainage capillary pressure function (Desbrandes et. al., 1988).

Many carbonate reservoirs is neither high permeable nor homogenous. The following factors may give significant challenges
with respect to defining the fluid zones:
 low permeability and heterogeneous reservoir
 small density difference between oil and water
 charging without water-wet conditions
 re-distribution by water encroachment

In all these cases vertical equilibrium between gravity and capillarity may imply that a large diffuse zone exists between oil
and water, and this zone is frequently abbreviated “oil-water transition zone” or “capillary transition zone”.

The motivation for this work was observations from a low permeable and heterogeneous carbonate reservoir with possible
multiple fillings, evidence of historical upward movements of oil through cap rock and hold-up of mobile oil below low
permeable layers. The fluid gradient from the Wireline Formation Tester (WFT) measurements showed a clear water gradient
except some outlying points that were related to high resistivity on the saturation log. These points were initially interpreted
as being supercharged. The 130m vertical section was evaluated to have oil saturation generally higher than expected for
residual oil saturation, although, the heterogeneous nature of the reservoir prohibited any direct conclusions.

The main objective was therefore, from a conceptual approach, either to verify a consistent relationship between water
gradient and residual oil (paleo zone) or if no such conclusion could be made, try to quantify the mobility to oil. Only wells
drilled with water-based mud is considered both since the well in question was drilled with water-based mud and that only a
minority of wells are drilled with oil-based mud (Carnegie, 2006)

Initial fluid distribution in the transition zone

The capillary transition zone can be defined as areas in the reservoir where both oil and water are present and where the water
saturation is higher than the connate saturation and the oil saturation is higher than the residual saturation, i.e. both phases are
mobile. Such areas need special attention both for characterizing the initial conditions, but also for designing special core
analysis (SCAL) programs for generation of input data to dynamic simulations. The precision of production forecasting for
primary, secondary or tertiary recovery methods will rely on proper description of the initial conditions in the reservoir.

Carbonate reservoirs are often found to contain pore-types of completely different capillary characteristics even at the same
level of porosity as shown in Figure 1a. The heterogeneity effect from shifting pore-types between lateral layers may
therefore result in areas containing oil with only connate water (oil zone) surrounded by mobile water and mobile oil
(transition zone) areas both above and below the layer in question. The classical text book example of the oil-water transition
zone with a general depth dependence on saturation can therefore be difficult to obtain in practice. J-functions for eight
distinct pore-types (Lønøy, 2006) have been used for calculation of the initial water saturation in figure 1c. The distribution
reflects core plug measurements with respect to permeability, porosity and pore-type. The FWL is taken to be at
approximately at 5.100 meters and the saturation distribution depends on the porosity, permeability, the J-function for the
pore types and depth in the reservoir. The depth dependence in the interval between 4.850 m and 5.030 m is difficult to
interpret if the observed saturation profile had been from a well log without detail a priori knowledge of the pore-type
systems and FWL

It has for several decades been recognized that the reservoir wettability is closely linked to the initial water saturation. The
likelihood for polar components in the oil, such as resins and asphaltenes, to reach the rock surface is increasing as the water
saturation is decreasing. The dependence of water saturation on wettability results in non-Archie effects for resistivity
measurements at low water saturations (Jing et. al, 1993, Fleury, 2002). The characterization of wettability and multiphase
flow behaviour from a few random core plugs will accordingly imply large uncertainties in the heterogeneous carbonate
reservoir. Understanding the results from wireline formation tester (WFT) may therefore give valuable additional information
in planning of core plug locations for conventional core analysis (CCA) and SCAL programs.

Three different types of initial fluid distributions (Phelps et. al., 1984, Desbrandes et. al, 1988) can be defined and classified
according to their distinct behaviour. In all cases the assumption of initial equilibrium everywhere implies that the water
saturation can be inversely estimated from the equation:
Pc   o w gh (1)
SPE 121982 3

0 0.5 1
1000
4850

100
4900

10
4950
Pc

Depth [m]
1
5000

0.1

5050

0.01
0 0.2 0.4 0.6 0.8 1
5100
Water saturation Sw

a) b) c)
Fig. 1: Carbonate pore-types may have different capillary characteristics even at the approximate same porosity value (a). Thin
section micrograph showing mudstone microporosity (top), macromoldic porosity and intercrystaline macroporosity (b). The large
variation in capillary pressure results in a strong dependence on pore-type in the initial water saturation distribution (c).

The main challenge is to select the correct oil-water capillary pressure function for deriving the water saturation and secondly
to find the correct position of the FWL. The definition of the OWC will in all practical cases, except one, be imprecise and
reference should therefore always be made to the FWL. The FWL may be located both above and below the OWC given the
common definition of the OWC as the deepest depth of mobile oil. Fluid distributions and fluid contacts for non-equilibrium
systems have also been discussed in the literature (Siddiqui, 1999), but this is not the topic of this work.

Type I distribution (illustrated in Figure 2) is the most common fluid distribution in producing reservoirs. The initial charging
of the reservoir will be according to the primary drainage capillary pressure and thus the same process as oil invading a
strongly water-wet rock. The wettability may change from the initial water-wet state towards intermediate or even oil-wet
character without any significant change in the local water saturation. This insignificant change of water saturation with
alteration of wettability during historical times can be explained from the water located in nooks and crannies become
immobile when films of water rupture as the wettability is changing (Kaminsky et. al., 1997). The main bulk of water is
therefore contained in the pore system similar to the original water-wet system since thin films of water contribute negligible
to the saturation. This water is most often called the connate water saturation. Layers in the transition zone with high water
saturation will, in many cases, maintain the water-wet characteristics and the mobile water is therefore still bounded in these
layers by capillary pressure as it was during the initial charging. The majority of producing reservoirs will therefore have an
initial water saturation distribution associated with the original water-wet conditions but be classified as intermediate-wet or
even oil-wet with respect to the dynamical situation. Accordingly, the possible effect on fluid gradients measured from WFT
needs to be investigated for type I distribution for different wettability scenarios as shown later.

The position of the FWL will be in the water zone with 100% water saturation and the initial water saturation can be found
above the FWL by applying the primary drainage oil-water capillary pressure either from centrifuge or porous plate
measurement using oil and water or proper scaling of Hg-air capillary pressure. In all cases proper scaling of IFT and
possible also contact angle  should be done. The OWC is located above the FWL with a depth difference proportional to the
capillary entry pressure for oil (Desbrandes et. al., 1988). The initial mobility of oil and water will depend on the respective
primary drainage relative permeability curves since the initial fluid distribution in the reservoir has been generated by the
primary drainage process. The primary drainage oil relative permeability is often found to be S-shaped which implies that
high oil mobility and low water cut can be obtained even at relatively high initial water saturations. For many clastic
reservoirs, negligible water-cut and no significant reduction in oil mobility can be expected for initial water saturation up to
50%. Capillary forces will also prevent mobilisation of water as long as higher water saturated layers have maintained their
water-wet characteristics. The residual oil saturation is also expected to be reduced with lower initial oil saturation (Land,
1968, Larsen et. al., 2000). It is also reported from carbonate reservoirs that oil reserves may be underestimated in the
transition zone due to higher oil mobility than expected (Chenggang et. al., 2006).

Type II distribution occurs when water encroachment has occurred in an earlier type I distribution. The initial distribution
will therefore shift from a primary drainage based to an imbibition based. This can happen if the contact has tilted either due
to movement of the reservoir itself after charging or by pressure variations in the supporting aquifer. Other scenarios involve
leakage from cap rock after or during initial charging. Muscat (Muscat, 1948) suggested that an imbibition distribution
always will be found in the lower part of the transition zone if the reservoir is charged faster then water could be drained
away from the upper part of the reservoir. This distribution has also been somewhat misleading been called the mixed-wet
4 SPE 121982

distribution (Phelps, 1984). Since water encroachment has occurred into a state where any increase in water saturation must
respond to the dynamic reservoir wettability, typically mixed-wet, the new distribution will be according to the reservoir
wettability after the first charging. The bounding imbibition curve should be used for predicting the water saturation in areas
where water has invaded into layers containing only connate water saturation. Imbibition scanning curves must be used for
describing the water saturation where water has invaded a historical transition zone as shown in Figure 2. The new
distribution in a heterogeneous and low permeable reservoir will therefore require use of imbibition scanning-curves,
imbibition bounding curves and knowledge about the historical type I distribution including the associated FWL from the
initial charging. Solving this problem will therefore be a difficult task even if a comprehensive database is available for the
reservoir in question.

The FWL for the type II distribution will be found at the level where Pc=0 after water invasion. Special pore-level
distributions of the oil and water can be interpreted for the mixed-wet reservoir containing both water-wet and oil-wet pores.
Oil will be mobile in both water-wet and oil-wet pores above the FWL corresponding to the positive part of the imbibition
oil-water capillary pressure curve. Oil will only be mobile in oil-wet pores below the FWL and trapped in water-wet pores
according to the negative imbibition oil-water capillary pressure curve. Hence, defining the OWC as the deepest level of
mobile oil may result in a much deeper OWC than the FWL for the type II distribution. The type II distribution should not be
confused with paleo oil zones where the water washing has been complete and only residual oil is left. A complete water
washing is less likely in a heterogeneous reservoir where both reservoir quality and wettability effects may give rise to
membrane effects. Membrane effects occur when tight layers are not pressure sealing but have too large capillary entry
pressure for oil to flow through the layer.

It should be expected that oil and water mobility will be similar to the type I distribution above the FWL. The oil mobility
should be significantly reduced and high water cut should be expected below the FWL and down to the OWC. Below the
OWC only trapped oil in water-wet pores and thin oil films in oil-wet pores could be expected. The volume of oil between
the FWL and OWC will be determined by the negative part of the oil-water capillary pressure imbibition curve or the
combination of series of imbibition scanning curves for the capillary pressure (Figure 2). This volume may be large for a low
permeable reservoir with an oil-wet nature and the volume is usually never regarded as reserves for reasons explained later.
Production from type II distribution is, however, reported in the literature (Marwah et. al, 2002, Gupta et. al., 2005). The
highest oil saturation should be found at low porosity that has maintained the oil-wet preference since the oil are bounded
(but not trapped without mobility) in oil-wet pores by capillarity. Low permeable and water-wet layers, not originally a
strategraphic trap, could therefore accumulate significant oil below it due to the potential high entry pressure for the oil.
Table 1 gives an estimate of oil columns that could be hold below a layer of a certain carbonate pore type at a given
permeability. The significant spread in values for different pore-types at the same permeability reflects the heterogeneous
nature of the carbonate system. The data are based on a large in-house database (Lønøy, 2006). Density difference between
oil and water is taken to be 300 kg/m3.

Table 1: Calculation of oil columns based on capillary entry-pressure for various carbonate pore-types.
Permeability Pore type A Pore type B Pore type C Pore type D
0.01 mD 22 m 120 m 7.6 m 362 m
0.1 mD 11 m 46 m 4.1 m 114 m
1 mD 5.4 m 13 m 2.3 m 36 m
10 mD 2.7 m 3.7 m 1.3 m 11 m
100 mD 1.3 m 1.4 m 0.7 m 3.6 m

Type III distribution may develop from three different starting points. The first one is rare and seldom documented. It occurs
when oil migrate up-wards to an initially oil-wet trap or in the case where the reservoir being its own source rock (Phelps,
1984). The other occurrence is actually a special case of type II distribution where the reservoir has become completely oil-
wet by exposure to oil components during historical times so that a small increase in water saturation will immediately shift
the oil-water capillary pressure to a negative value. The third scenario is that all mobile oil is removed from water-wet pores
during water encroachment and the remaining mobile oil is therefore only distributed in oil-wet pores.

In all cases it will imply that the FWL is located at the top of the reservoir with the OWC far down in the reservoir with
initial water- and oil saturation distributed according to a negative oil-water capillary pressure curve. The type III distribution
will result in initial conditions and flow dynamics that both relates to a strictly oil-wet system. The mobility to oil is expected
to be low according to the wetting-phase relative permeability
SPE 121982 5

Pc Pc Sw
Type I

Depth relative to respective FWL

Type I Type I Type II
Re-distribution to type III
OWC (I)
Re-distribution to type II
Sw Pc=0
 FWL
Type III
OWC (II)
Type II

OWC (III)
Type III

Figure 2. Re-distribution from Type I to Type II and Type III. The oil-water capillary pressure is reduced when the FWL is shifted up-
wards due to water encroachment. Each point on the original primary drainage curve follows a distinct scanning-curve and a new
equilibrium distribution occurs with the new shallow FWL. The Type II distribution has both positive and negative capillary pressure
and oil is therefore mobile in water-wet pores above the FWL. The Type III distribution has negative capillary pressure everywhere;
oil is therefore only mobile in oil-wet pores.

Characterization of capillary transition zones

The fluid distribution in the transition zone and the capillary correction necessary for interpretation of WFT measurements
have been a topic of research for two decades. In the author’s opinion, more work is needed for bridging the gap between the
publications covering capillary pressure hysteresis and transition zone flow dynamics on one side and interpretation of WFT
results on the other side. It should also be noted that the reported transition zone behaviour is mostly related to producing
reservoirs and therefore to cases where oil is recovered at zero or negligible water-cut (Chenggang, et. al., 2006). It is
believed that such observations are biased by the fact that reservoirs with potential oil production at initially high water-cut
are not produced from the transition zone or possible abandoned without proper long term well testing. These reservoirs may
still contain significant volumes of oil and therefore be possible future prospects in a high oil price scenario. Oil production at
high water-cut is reported from the Hunton formation (Marwah et. al., 2002), a fractured and heterogeneous carbonate
reservoir. The initial water-cut is reported to be higher than 90% for some wells and decrease with time. Many of the earlier
wells were abandoned due to the high water-cut. The oil and gas well production were, however, found to increase with time
giving more economical prospects. It is proposed (Gupta et. al., 2005) that the reservoir has been re-distributed due to water
influx during historical times, however, some continuity should be expected in the hydrocarbon phase since the oil today is
found to be at the bubble-point. Economical oil production is taken place with depletion and a special case of solution gas
drive.

In the literature numerous studies have discussed how the imbibition process will depend on the variable initial saturation for
either the wetting-phase (Land 1968, Killogh 1976) or in more general wettability perspective the water-phase (Skjaeveland
et. al., 1998, Kjosavik et. al., 2002). General for all hysteresis models for capillary pressure is that some scheme is developed
for calculation of imbibition scanning curves that originates from the primary drainage capillary pressure curve. These
models can therefore be used to calculate the shift from type I distribution to type II distribution, although the original
development were targeted against flow dynamics in reservoir simulation. The most common approach is to constrain
imbibition scanning curves to the so-called Land relation:

1 1
 C. (2)
S oi S or

The Land relation was originally developed on empirical data for trapped gas saturation and has later been used for
connecting the residual oil saturation (Sor) to the initial oil saturation (Soi). However, the relevance of the relation may be
seriously questioned for wettability scenarios other then strongly wetting behaviour, i.e. the most frequent conditions that
occur when type I shifts to type II. Several resent experimental studies have also discussed the relationship between initial
saturation and trapped saturation (Masalmeh and Oedai 2000, Masalmeh 2000, and Larsen et. al. 2000). Numerical
simulation of the transition from type I distribution to type II distribution using both the Killough model and the Skjaeveland
model was first reported by Eigestad (Eigestad et. al. 2000). Water encroachment was modelled at different velocities from
oil migration to production settings. It was concluded that the two models may give different results even when the bounding
6 SPE 121982

drainage and imbibition curves are similar and that both models are based on the Land relation. It was also shown that the
final type II fluid distribution will depend on whether viscous effects are present or not during water encroachment. The work
from Mashalmeh et. al. 2007 is recommended for more detailed description of the characterization of transition zones

Pressure measurements from wireline formation tester (WFT)

Phelp et al. were among the first authors to comment on the capillary correction necessary for estimating the correct pressure
from WFT in low permeable reservoirs. Their main discovery was that since the pressure is actually measured in the filtrate
the true formation pressure must be corrected for a Pc, called the capillary pressure correction factor, in the oil zone to
obtain the correct pressure when the filtrate is not miscible with the oil. It was also argued that pressure is measured in the
filtrate (water-phase) since the usual drawdown used to extract fluids from the formation for pressure measurements usually
exceeds the capillary pressure holding filtrate to the water-wet formation. In addition any hydrocarbon saturation will be
close to residual in the flushed zone. The second important finding was that supercharging is more likely when both oil and
water are mobile since the total mobility of the system then will be less than the case where only one phase is mobile. The
time for equilibrium to occur between the filtration affected near-wellbore area and the unaffected far pressure field will
depend on the total mobility away from the invaded zone. Areas where the initial saturation gives the minimum of Krw+Kro,
i.e. where the total mobility is lowest, will need longer time to equilibrium and is therefore more likely to be observed as
supercharged. The work from Phelps et. al., 1984, assumed water-wet conditions in a type I reservoir.

Elshahawi (Elshahawi et al., 1999, Elshahawi et. al., 2000) discussed the capillary correction for different mud types, water
or oil based, in the strongly water-wet and oil-wet cases. The dynamics of the invasion process was not evaluated and the
capillary pressure correction was taken directly from wettability consideration and imbibition capillary pressure behaviour.
For water-based filtrate in the oil zone, two different scenarios may occur when the pressure is measured in the filtrate. The
oil saturation will in both cases be close to residual at the sand face. The capillary correction is therefore small, but positive if
the formation is water-wet and the measured gradient is shifted to the left of the true gradient. The capillary correction is
large and negative if the formation is oil-wet and the gradient is shiftet to the right of the true gradient. The FWL is
accordingly in the water-wet case interpreted to be too shallow based on the measured gradients without correction and too
deep for the oil-wet formation drilled with water-based mud.

Pam (Pam et al., 2005) pointed out the Pc should not been considered as a constant correction factor as applied by
Elsahawier et. al. The correction factor depends on whether mobile water is present together with mobile oil or not. The
apparent gradient will therefore not have the same steepness as the true oil gradient since Pc also will be a function of depth
in the transition zone. They also performed simulation of the invasion process and showed that a water gradient where
generated in areas above the OWC with mobile oil. The detail set-up of the simulation model with respect to initial fluid
distribution and wettability scenarios were not given.

Carnegie (Carnegie, 2006) showed practical examples of fluid gradients from Middle East reservoirs. He classified the
reservoirs into two types depending on the fluid gradient behaviour without any a priori relationship to the initial fluid
distribution. All simulation examples were, however, related to a type I distribution. “Type A” was said to have the most
general form of transition zone gradient for homogenous limestones in the 2mD to 20 mD range. In the upper levels where an
oil zone is located, the gradient was found to be consistent with the actual oil density. In the upper parts of the transition zone
oil is significantly lighter than what really will be accounted. Further down in the transition zone an inflexion with sharp
curvature occur to the left before a true water gradient is measured even above the OWC and in the water zone (similar to
Figure 7 below). It is specially mentioned that significant amount of oil can be produced from the inflexion area as well as
where a water gradient is measured above the OWC. Static and dynamic filtration was simulated using a type I distribution
and different ways of representing the imbibition process during filtration. It is not completely clear how the model was
initialized and thus how the shift from the initial fluid distribution to the filtration process was done.

The water-wet case with only one positive bounding imbibition capillary pressure curve could neither explain the inflexion
nor the light oil gradient effect. The mixed-wet bounding imbibition capillary pressure could explain the inflection, but not
the light oil effect. Only the case with Killogh hysteresis (Killough, 1976) describing the more consistent approach of
scanning curves could explain both the light oil effect and the inflexion effect. It was concluded that the lowering of residual
oil saturation with lower initial saturation found in the transition zone was causing the apparent light oil effect on the
gradient. The “Type B” behaviour covered cases where a water gradient is measured, but still oil production occurred at low
water-cut. The “Type B” behaviour was not discussed in any details. Chenggang (Chenggang et. al., 2006) showed fluid
gradients from a tight carbonate reservoir in the Middle East supporting the “Type A” behaviour with one case of strong
inflexion and another case with apparent light oil behaviour below the true oil gradient. Depth dependent fluid property
variation has also been reported (Jakson et. al., 2007) which make the interpretation of fluid gradients even more
complicated. More details regarding formation testing challenges and methods in low permeable reservoirs can be found
SPE 121982 7

elsewhere (Parkers, et. al. 1998 and Weinberg et. al. 2008). Proett et. al. (Proett et. al., 2001) discusses details regarding the
fluid invasion process and mud cake growth model.

The summary of previous work reveals some important issues:

 Details regarding which assumption have been made for the initial fluid distribution in the transition zone is often
missing.
 It is demonstrated both from a practical and theoretical point of view that oil production at low water-cut can occur
even if a water gradient is measured.
 Type I distribution seems to be dominating in the reporting fields studies and the main reason may be that fields with
type II or type III behaviour have been abandoned at an early stage due to oil production at high water-cut.

Simulation model of filtrate invasion

A simulation model has been prepared to test the dynamical behaviour of the filtrate pressure during dynamic and static
filtration. The radial well-model has constant porosity and permeability except from the grid cell closest to the well where the
permeability is set to 0.002 mD in order to model the mud cake. The uniform matrix permeability is set to 1 mD. Figure 3
shows the grid cells where the radial size is doubled away from the well. More details of the simulation model are given in
appendix. One large cell is added outside the near well-bore area so that undisturbed conditions are maintained in the outer
part of the reservoir for all parameter sensitivities. The simulation approach taken in this study is similar to that of Carnegie
(Carnegie, 2006). The density of the filtration water is set to be the same as for the formation water. This is a simplification
assuming that a fast mixing occurs due to dispersion and diffusion effects in the near well-bore area.

The pressure is always taken in the water-phase except from the reference solution giving the oil gradient where mobile oil is
present. The reference solution is the fluid gradients in a reservoir that is not influenced by capillary effect and this is shown
in the following Figures as a green oil gradient and a blue water gradient. Several authors have pointed out that the actual
pressure measured by the WFT is the filtrate pressure (Phelp et. al. 1984, Elshahawi et. al. 1999). The simulation results
show also that it will take weeks or months of static filtration for the saturation in the near-wellbore area to equilibrate.
Table 2 gives the different capillary pressure assumptions made for the initial fluid distribution and for simulation of the
dynamic and static filtration process.

331 m

2m

200 m

So

a b

Figure 3: Half of the simulation model (a) and near-well bore region after filtrate invasion (b).

The typical pressure and saturation behaviour at the sand face close to the wellbore during dynamic and static filtration is
given in Figure 4. The case shown is with type I initial fluid distribution and Killough capillary pressure hysteresis with
water-wet conditions. The near-wellbore water pressure is pertubated with approximately 20 bars during 10 days of dynamic
filtration and a pressure gradient is maintained until dynamic filtration is stopped. The pressure at the sand face will then
immediately start to equilibrate with the pressure in the undisturbed parts outside the near well-bore area.
8 SPE 121982

The effect of different compressibility between oil and water can also be seen in Figure 4. The water zone has low
compressibility compared to the transition zone and a pressure change of nearly 5 bars occurs in the water zone just one
minute after dynamic filtration has stopped. The water pressure reaches fast a semi stable state that can be observed in this
case from 3 hours to approximately 1 month after dynamic filtration has stopped. The time for arriving in this semi stable
state is longest at the depth where the total fluid mobility is at its lowest value in the undisturbed transition zone. This occurs
at approximately 50% water saturation in the simulation example using generic relative permeability to oil and water. The
observation is consistent with earlier studies (Phelp et. al., 1984) and may give rise to supercharging effects. The water
pressure must align to the capillary pressure perturbation in the near-wellbore area in this period and a true water gradient can
therefore only be observed below the OWC where the capillary pressure is not affected. The OWC is in the following defined
at the level where oil saturation starts to increase from zero. The segregation of the filtrate will change the capillary pressure
in the invaded zone, but this effect take place several weeks after dynamic filtration has stopped.

Both the capillary pressure and saturation will after a long period of time return to the original values when the filtrate has
segregated away from the near-wellbore area. It takes more than 5 months before the water pressure is back in the original
equilibrium situation in the simulation example given in Figure 4. Sensitivities where run for different reservoir permeability
mud cake permeability, fluid viscosities, relative permeabilities, capillary pressure, oil compressibilities and pressure
pertubations. The details are too much to be included here, however, all cases seem to verify the observations in Figure 4
with relatively fast fallback in water pressure to the semi stable state except when extremely high oil compressibility gives
supercharging effects that can last for longer periods. The water pressure at one day of static filtration has been taken as the
reference solution for measuring the fluid gradients for all the simulation results presented in this work.

Water pressure (bar) Water saturation

120 130 140 150 160 170 0 0.2 0.4 0.6 0.8 1
1450
initial stop dynamic 1 min 1 hr initial stop dynamic 1 min 1 hr
3 hrs 1 day 30 days 5 months 3 hrs 1 day 30 days 5 months
1500

1550
Depth (m)

1600

1650 lowest total mobility

1700 OWC

1750 start dynamic filtration stop dynamic filtration

start dynamic filtration stop dynamic filtration

Figure 4: Water pressure and saturation development during static and dynamic filtration. The pressure equilibrates in a short time
with the pressure outside the invaded region (< 3 hours in this example). The saturation will be in equilibrium when the filtrate is
removed from the near- wellbore area and this process is slow (> 5 months in this example). This behaviour implies that WFT
measurements, most likely, are taken in the filtrate while the water pressure is influenced by the shifted capillary pressure; Pc. It is
observed that Pc is relatively constant in the period from 3 hrs to 30 days (blue curves) due to the slow changes in saturation and
this period is therefore called the semi stable state. Note that the saturation profile for 1 min., 1 hr, 3 hrs, and 1 day after stop
dynamic filtration is similar to the saturation at stop dynamic filtration.

The impact of both smaller and larger pressure perturbation on the time before semi stable state occurred where also tested by
adjusting the bottom hole pressure in the well during the dynamic filtration period. Again, no significantly influence on the
main result occurred and this indicates that it is likely that formation pressure measurements are taken in the semi stable
period in most practical situations. The rather simple simulation model may therefore adequately describe the capillary
pressure influence on formation pressure measurements without modelling the filtration process in detail. Special drilling
techniques, as underbalanced drilling, with little loss to the formation may although have a different behaviour not considered
here.

Simulation scenarios
Eight different scenarios were investigated with basis in initial fluid distribution and how capillary pressure and relative
permeability where modelled during dynamic filtration as explained in Table 2. Four different cases where tested for the
type I distribution. The first case uses the same primary drainage capillary pressure curve for both the initial condition and
SPE 121982 9

during dynamic filtration. The second case is honouring the process-dependent behaviour for the capillary pressure in the
water-wet case. The third case is split into cases 3a and 3b. Both cases use an oil-wet capillary pressure bounding curve. The
scanning curves will accordingly be of a mixed-wet nature since the Killough model interpolates between the positive
primary drainage curve and the negative bounding imbibition curve. The residual oil saturation is taken to be constant
everywhere in the transition zone in case 3a. In case 3b the residual oil saturation is assumed to depend on the initial oil
saturation according to Equation 2. Typical capillary pressure scanning curves for 3a and 3b are given in Figure 5.

Table 2: Simulation cases

CASE Initial fluid distribution Dynamic / Static filtration
1 Type I Primary drainage
2 Type I Water-wet bounding imbibition, Pc scanning curves from Killough
3a Type I Oil-wet bounding imbibition, Pc scanning curves from Killough
3b Type I Oil-wet bounding imbibition, Pc and Kr scanning curves from Killough
4a Type II Mixed-wet bounding imbibition, high Sorw
4b Type II Mixed-wet bounding imbibition, low Sorw
5a Type III Oil-wet bounding imbibition, high Sorw
5b Type III Oil-wet bounding imbibition, low Sorw

For the type II and type III distributions it is assumed that the capillary pressure used for initial condition is the same as
during dynamic conditions, which obviously is a simplification, made for modelling purpose. Two different cases for oil
mobility are made for the mixed-wet type II distribution and the oil-wet type III distribution as explained later.

Pc hysteresis and fixed Sorw Pc hysteresis and Sorw=f(Soi)

5 5
Primary drainage curve for initialization

3 3
Pc (bars)

Pc (bars)

1 1

-1 -1

-3 -3
Imbibition scanning curves

-5 -5
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

Sw Sw

Figure 5. Imbibition scanning curves for capillary pressure generated by the Killough for three different grid cells. The impact on
dynamic capillary pressure scanning curves during filtration are shown when the oil relative permeability is modelled with either
constant residual oil saturation (left) or the residual oil saturation depends on the initial oil saturation (right). The bounding capillary
pressure imbibition curve is in both cases strictly oil-wet (negative capillary pressure). The two different approaches for modelling
oil relative permeability impacts the trend of negative capillary pressure with depth after filtrate invasion, and therefore the shape of
the measured fluid gradients.

Results from Type I distribution

The results from the simulation cases 1, 2, 3a and 3b are summarized below. The capillary pressure versus depth is given at
different times in Figure 6 for case 1, case 2 and case 3a. One observation is that the capillary pressure after 10 days of static
filtration still is very similar to the values just after dynamic filtration has stopped. Case 1 gives a slightly positive capillary
pressure at the sand face after dynamic filtration. In case 2, capillary pressure hysteresis according to Killough is used in the
water-wet situation. The capillary pressure is therefore close to zero at the sand face since the oil saturation is close to the
residual value. The capillary pressure is in both cases nearly constant with depth above the OWC and a water gradient is
measured in the water zone and an oil gradient is measured at the top of the transition zone. The oil gradient is shifted to the
left for case 1 due to the positive Pc. The shift for case 2 is also to the left, but much smaller due to the small value for Pc.
A shift from oil gradient to water gradient occurs in the lower part of the transition zone. The shift is found to be largest for
case 1 where Pc is largest.

The oil-wet system, case 3a, gives a different behaviour then the two other type I cases. First of all, an oil gradient can be
observed at the top of the transition zone. However, the gradient is now shifted to the right of the true oil gradient due to the
10 SPE 121982

negative contribution from Pc. The shift from oil gradient to water gradient shows a clear inflexion as also reported by
Carnegie (Carnegie, 2006), and a water gradient can be observed above the OWC. The residual oil saturation is taken to be
constant with depth in case 3a. The capillary pressure at the sand face will in the case with constant residual oil saturation
have the largest negative values at the top of the transition zone after dynamic filtration has stopped. The negative capillary
pressure closer to the OWC is therefore less than higher up in the transition zone, this can also be observed in Figure 6. No
change in capillary pressure during dynamic filtration can take place below the position where Soi = Sor (approx. 15 meters
above the OWC), and the filtrate gradient will be according to the water gradient below this value.

Pc (bars) Pc (bars) Pc (bars)

-2 0 2 4 6 8 -2 0 2 4 6 8 -2 0 2 4 6 8
1500

1550

1600
Depth (m)

1650

1700
water-wet PD water-wet Killough
oil-wet Killough
1750 initial stop dynamic 10 days 60 days 2000 days

Pressure (bars) Pressure (bars) Pressure (bars)

125 130 135 140 145 125 130 135 140 145
125 130 135 140 145
1500

1550 oil gradient

oil gradient
Depth (m)

1600 oil gradient

shift to water gradient

1650
shift to water gradient

shift to water gradient water gradient above OWC

1700

1750 Case 1 Case 2 Case 3a

Figure 6. Sand face capillary pressure (above) and the effect on the pressure gradient (below) for type I distribution.

Pc (bars) Pressure (bars)

-5 -3 -1 1 3 5 7 125 130 135 140 145
1500 1500

oil
1550 1550

1600 1600
Depth (m)
Depth (m)

apparent light oil or gas gradient

1650 1650 shift to water gradient

water gradient above OWC

1700 1700

1750 1750

Figure 7: Case 3b with Sor as function of depth.

The next step was to allow the residual oil saturation to depend on the initial oil saturation, i.e. the residual oil saturation will
go to zero at the OWC. Thus, oil is mobile from the top of the transition zone down to the OWC in case 3b. The resulting
fluid gradients are given in Figure 7. The largest negative capillary pressure after dynamic filtration is now observed closest
to the OWC due to the increased oil mobility (for a given oil saturation) with depth. This is also seen in Figure 5 and shows
how the interplay between capillary pressure and oil relative permeability scanning-curves determine the value for the
capillary correction, Pc. For case 3b, an oil gradient is observed at the top of the transition zone. An apparent light oil or gas
gradient is observed below the oil gradient before a sharp inflexion to the water gradient occurs. The water gradient is
observed up to 40 meters above the OWC. The inflexion is sharper in case 3b than in case 3a, but the interpretation of the
SPE 121982 11

apparent light oil or gas gradient below the oil gradient can be made in both cases. The conclusion from Carnegie (Carnegie,
2006) that the variable Sor with depth is necessary for obtaining the light oil or gas gradient seems therefore not to be strictly
valid for all cases. It is more the combination of capillary pressure and oil relative permeability that gives rise to the apparent
light oil or gas gradient. Mixed- or oil-wet conditions are a necessary condition for the apparent light oil or gas gradient since
the oil gradient needed to be shifted to the right of the true gradient. The variable Sor with depth was also tested for the water-
wet case 2 without any changes in result for that case. This can easily be explained by the Pc ~ 0 independent on how oil
relative permeability is modelled.

Results from Type II and Type III distributions

Sw
The type II, mixed-wet, distribution has the FWL at
1.655m. The initial capillary pressure is positive above 0.0 0.2 0.4 0.6 0.8 1.0
this level and negative below the FWL. The water 1500
saturation at the FWL (Pc=0) is 55% which is typical FWL for oil-wet distribution
for mixed-wet systems. The saturation distribution with
depth is given in Figure 8. Two different simulation 1550
cases where tested for this distribution. Case 4a had a
residual oil saturation of 15% which defines the level of
mobile oil to be at 1.670m. The other scenario, case 4b, 1600

Depth (m)
had a residual oil saturation equal to zero giving the
level of mobile oil at 1.702m. In both cases a water FWL for mixed-
wet distribution
gradient was observed in the filtrate up to 1.655m as 1650
shown in Figure 8. The water gradient is shifted to the
right of the similar water gradient for the type I
reference distribution. The reason for this is the 1700 mixed
negative capillary pressure below the FWL which also oil
define the shift of the water pressure where Sw = 1. An
oil gradient is only found at top of the reservoir for the 1750
mixed-wet case. Moreover, a long transition occurs
from the water gradient to the oil gradient. Figure 8. Saturation-depth for type II and type III distributions.

Pc (bars) Pc (bars) Pc (bars)

-2 0 2 4 6 -6 -4 -2 0 -6 -4 -2 0
1500
High Sorw – type II mixed-wet High Sorw – type III oil-wet Low Sorw- type III oil-wet

1550

1600
Depth (m)

1650

1700

1750 init ial stop dynamic 10 days 60 days 2000 days

Pressure (bars) Pressure (bars) Pressure (bars)

125 130 135 140 145 125 130 135 140 145 125 130 135 140 145
1500
High Sorw – type II mixed-wet High Sorw – type III oil-wet Low Sorw- type III oil-wet

1550

1600
Depth (m)

water gradient with mobile oil water gradient with mobile oil

water gradient below

1650 FWL in type II
distribution

1700

1750

Figure 9 Sand face capillary pressure (above) and the effect on the pressure gradient (below) for type II & III distributions.
12 SPE 121982

The saturation-depth plot for the oil-wet type III distribution shows a different shape than the type II distribution as seen in
Figure 8. The FWL is now found to be at the top of the reservoir and oil is distributed according to the negative capillary
pressure. The oil is mobile below the FWL and in an equilibrium situation where oil is hold by the capillary forces in oil-wet
pores balancing the buoyancy effect. Again, two different assumptions for residual oil saturation were done defining the
level of mobile oil at either 1.600m (case 5a) or at 1.702m (case 5b). The results gave in both cases a clear water gradient
from the bottom to the top of the reservoir.

Simulation cases 4a, 4b, 5a and 5b uses the same capillary pressure for initialization of the water saturation as for modelling
of dynamic filtration. The generation of the type II distribution comes from re-distribution of a type I distribution by water
encroachment and shifting of the FWL as illustrated in Figure 2. The dynamic modelling approach in this work will be
correct if just one mixed-wet imbibition bounding curve defines the new distribution. The other alternative is the type II
distribution defined from series of imbibition scanning curves. It can be argued that the approach taken for modelling of
capillary pressure, either one mixed-wet bounding imbibition curve or series of imbibition scanning curves, will not have
impacted on the results. The level where oil is assumed to be mobile is not controlling the top of the water gradient, but it is
observed that the water gradient is independent of the oil relative permeability. The water gradient is, however, connected to
where the capillary pressure is negative in the undisturbed parts of the reservoir away from the well-bore. This result is also
consistent with observing the water gradient everywhere in the oil-wet type III distribution.

Heterogeneous wettability distribution in the transition zone

The simulation cases presented so far has been simplified by assuming both homogenous reservoir properties and uniform
wettability. The reservoir properties like permeability and porosity as well as properties for the near well-bore damage zone
will have impacted on the dynamic results and time for the semi stable state to occur. The implication of reservoir
heterogeneity will be possible supercharging effects that are not studied in detail in this work. All the simulations gave stable
gradient behaviour after 1 day or less and this stable situation continued for several weeks. A special test was done where
some parts of the reservoir where assumed to be more oil-wet than other parts. A type I distribution was assumed with the
intervals 1.546-1.569m and 1.646 – 1.698m to be oil-wet while the rest of the reservoir was assumed to be water-wet.

The result is given in Figure 10 and shows that the gradient for the oil-wet areas will be shifted to the right of the water-wet
oil gradient. A natural conclusion, but a wrongly one, will be that the pressure points in the oil-wet areas are taken to be
supercharged. The pressure has everywhere reached the semi stable state and the observed shift is due to the heterogeneous
nature of wettability in the reservoir. The water-wet parts of the reservoir give an oil-gradient similar to what observed in
case 2.

Pc (bars) Pressure (bars)

-4 -2 0 2 4 6 125 130 135 140 145
1500

1550 oil-wet area

1600
Depth (m)

1650
oil-wet area

1700

init ial stop dynamic 10 days 60 days 2000 days

1750

Figure 10. Sand face capillary pressure (left) and the effect on the pressure gradient (right) for type I distribution with both water-wet
and oil-wet conditions.

Application of the results to an offshore exploration well

The gradient behaviour from the type I reservoir has been documented from several papers with reference to Middle East
reservoirs. The basis for this work was an exploration well drilled in an area with possible leakage scenario. The MDT
formation pressure gave a clear water gradient as shown in Figure 11. The interpretation of the CPI logs were regarded as
uncertain in estimates of oil saturation, however, there where indications of oil saturations in excess of 50% in some parts of
SPE 121982 13

the well. Only the top of the well were cored. The cored section of the well was dominated by carbonate breccia clasts and
inter-clast carbonate matrix, showing very variable degrees of oil staining. The staining is often patchy, reflecting local
differences in capillary entry pressure which is controlled by porosity and pore type. The lower intervals in the well were
classified based on drill cuttings and sidewall core plugs. An OWC was assumed at about 1410m MD based on few shows
below this depth. A 15m interval at about 1220m MD, just below the cored section, had very poor reservoir properties and a
pore type characterized by high capillary entry pressure. High oil saturation with possible live oil properties were accounted
below this interval. The sealing capacity of the low-permeable interval was calculated based on pore-type distribution,
porosity and a large in-house carbonate database on capillary entry pressures. The results are given in Table 3, and a thin-
section micrograph from the low-permeable interval is shown in Figure 11. The entry pressure is converted from Hg-curves
assuming water-oil IFT equal to 30 mN/m and contact angle of either 0 (optimistic) or 60 degrees (pessimistic). The entry-
pressure was then converted to the oil column that could be hold below the low permeable interval by assuming 800 kg/m3
oil density and 1137 kg /m3 water density.
Table 3. Calculation of sealing capacity in low permeable layer
Depth Pore type Porosity Permeability Oil column 1 Oil column 2
(%) (mD) (m) (m)
1215 Intercrystalline, patchy microporosity 6.3 NMP 117 59
1219.1 Intercrystalline, patchy microporosity 9.0 0.02 74 37
1220.5 Intercrystalline, patchy microporosity 7.5 0.015 93 47
1225.9 Intercrystalline, homogenously distributed 9.4 NMP 70 35
microporosity

The calculations show that the most optimistic case could give an oil column of 117m below the low-permeable layer (Table
3). This is nearly 100m shallower than the possible OWC at 1410m, and can be explained by a possible paleo FWL within
the reservoir. At some time a leakage through the cap rock has occurred, and the FWL has moved upwards until the oil
column has been small enough to be below the sealing capacity of the low-permeable layer. The low-permeable layer, once
not sealing the reservoir, will then become a trap for further movement of oil up-wards.

(1-Sor)
Sw
C1 Flair GCMS
1130

1150
Low C1  water Sor 
1170
washed residual “dead-oil”
1190
oil
1210
Low permeable layer
1230
FWL (2)
Depth (MD)

High C1  “live” oil

1250

1270
below low permeable So>Sor  mobile oil
Depth

1290
layer in oil-wet pores
water gradient even 1310

though high So from 1330

CPI 1350

1370

1390

1410

1430
0 10 20 30 40 50 60 70 80 90 100 FWL (1)
Normalised composition C1-nC5 (%)

a) b) c)

Figure 11. Measurements and schematic sketch explaining the complex observations in an offshore exploration well. The water
gradient (a) is consistent with formation water density and is observed throughout the well. Some apparent supercharged points
can also be an indication of heterogeneous wettability distribution. The methane concentration is low above a layer of particular
poor reservoir quality (b) and increases through the layer and below the layer. The low permeable layer has been estimated to hold
an oil column in the range of 35 meters to 117 meters based on entry pressure and carbonate pore-type found in sidewall cores.
The most likely scenario is that a re-distribution from Type I to Type III has occurred during historical times (c) with the low
permeable layer controlling the present oil column. The measured water gradient can be explained by the discovered gradient
behavior for the Type III reservoir and do not necessary imply residual oil saturation everywhere.

The rising water level will trap oil in water-wet pores and leave mobile oil behind in oil-wet pores. The fluid distribution will
therefore shift from a type I distribution to a type III distribution. According to the previous simulation results, the fluid
gradient should therefore be a water gradient everywhere, even though the oil saturation from the CPI’s is too high for typical
residual oil saturation. This was also the result from the MDT formation pressure measurements given in Figure 10. A clear
14 SPE 121982

water gradient consistent with the density of the formation water was measured in the well. Some points where interpreted as
supercharged, but these points compared with high resistivity values and could therefore also be the result of very strong
negative capillary pressure (more oil-wet character) at the location of the measurements.

The natural question regarding the ability to produce the oil in a type II or type III reservoirs is not the topic of this paper.
However, parallels to the production behavior of the Hunton formation can be made (Gupta et. al., 2005). It is likely that the
initial mobility of the oil is small and that depletion will results in extremely high water cuts difficult to handle in an off-
shore situation. A combination of fluid expansion and additives that can change the wettability and thereby mobilize oil from
oil-wet pores seems to be attractive recovery strategies for type II and type III reservoirs if the in place volumes are large
enough.

Conclusions

1. Pressure gradient behavior in tight formation has been classified according to the initial reservoir condition with
respect to fluid distribution and capillary pressure. Simulation models has been used to account for the capillary
pressure effect on fluid gradients in order to explain the observed fluid gradients for three main types of reservoir:
a. Type I is in a status similar to the first charging of the reservoir given from a primary drainage process.
b. Type II is a redistributed type I where the initial oil and water saturation are given from both positive and
negative capillary pressure where the FWL is located where the capillary pressure shifts from a positive to
a negative value. Mobile oil will be found both above and below the FWL.
c. Type III distribution contains only mobile oil in oil-wet pores and the initial oil and water saturation is
therefore given from one or several negative capillary pressure curves.
2. The pressure gradient is impacted by capillary pressure and the measured pressure may therefore not be the true pore
fluid pressure when for example an oil zone is drilled with water-based mud. The capillary pressure effect results in
a shift in the measured pressure compared to the actual pore fluid pressure away from the invaded zone.
3. The fluid gradient behavior for the type II reservoir is closely related to the FWL with a water gradient below this
level.
4. The type III reservoirs give a water gradient everywhere even though the oil is mobile below the FWL.
5. An example is made for an offshore exploration well where the sum of observations best can be explained by a type
III distribution.
6. Further work should include more sophisticated hysteresis models in capillary pressure and relative permeability
then used in this study.

Nomenclature
C Land Constant
g acceleration of gravity
Soi initial oil saturation
Sor residual oil saturation
 oil-water density
SPE 121982 15

References
Carnegie, A.J.G., 2006. Understanding the Pressure Gradients Improves Production from Oil/Water Transition Carbonate Zones, SPE
99240, presented at the 2006 SPE/DOE Symposium on Improved Oil Recovery, Tulsa, U.S.A., 22-26 April 2006.
Chenggang, X. Dawoud, A., Carnegie, A., Mahidi, A. E., Hajeri, A. H., Basioni, M., 2006. Identification and Characterization of Transition
Zones in Tight Carbonates by Downhole Fluid Analysis. SPE 101257, presented at the 2006 Abu Dhabi International Exhibition and
Conference, Abu Dhabi, U.A.E., 5-8 November 2006.
Desbrandes, R. and Gualdron, J., 1988. In Situ Rock Wettability Determination with Wireline Formation Tester Data. The Log Analysts
(1988), July-August, p. 244-251.
Eigestad, G.T. and Larsen, J.A., 2000. Numerical Modelling of Capillary Transition Zones. SPE 64373 presented at the SPE Asia Pacific
Oil and Gas Conference and Exhibition, Brisbane, Australia, 16-18 October, 2000.
Eslhahawi, H., Fathy, K. and Hiekal, S., 1999. Capillary Pressure and Rock Wettability Effects on Wireline Formation Tester
Measurements. SPE 56712, presented at the 1999 SPE ATCE, Houston, Texas, 3-6 October 1999.
Eslhahawi, H., Samir, M. and Fathy, K., 2000. Correcting for Wettability and Capillary Pressure Effects on Formation Tester
Measurements. SPE 63075, presented at the 2000 SPE ATCE, Dallas, Texas, 1-4 October 2000.
Fleury, M., 2002. Resistivity in Carbonates: New Insights. SPE 77719, presented at the SPE ATCE, San Antonio, Texas, 29 September – 3
October, 2002.
Gupta, M. Joshi, R. and Kelkar, M., 2005. Dewatering of Hunton Reservoir: What Makes it Work? SPE 94347, presented at the 2005 SPE
Production and operations Symposium, Oklahoma City, OK, U.S.A, 17-19 April 2005.
Jakson, R.R, Carnegie, A. and Dubost, F.X., 2007. Pressure Measurements and Pressure Gradient Analysis: How Reliable for Determining
Fluid Density and Compositional Gradients? SPE 111911, presented at the 31st Nigeria Annual International Conference and
Exibhition. Abuja, Nigeria, 6-8 August 2007.
Jing, X.D, Gillespie, A. and Trewin, B.M., 1993. Resistivity Index From Non-Equilibrium Measurements using Detailed In-Situ Saturation
Monitoring. SPE 26798, presented at the Offshore Europe Conference, Aberdeen, 7-10 September, 1993
Kaminsky, R. and Radke C.J. 1997. Asphaltenes, Water films, and Wettability Reversal. SPEJ, V.2, December 1997.
Killough, J.E., 1976. Reservoir Simulation with History Dependent Saturation Functions. SPEJ (Feb. 1976) 16, 37; Trans., AIME., 261.
Kjosavik, A., Ringen, J.K. and Skjaeveland, S.M., 2002. Relative Permeability Correlation for Mixed-Wet Reservoirs. SPEJ 7 (1): P. 49-
58. SPE 77328.
Land, C.S., 1968. Calculation of Imbibition Relative Permeability for Two- and Three-Phase Flow from Rock Properties. SPEJ 8 (2): 149-
156, 1968; Trans., AIME., 243.
Larsen, J.A., Thorsen, T., Haaskjold, G., 2000. Capillary Transition Zones from a Core Analysis Perspective. Paper SCA 2000-20
presented at the International Symposium of the Society of Core Analysts, Abu Dhabi, UAE, 18-22 October, 2000.
Lønøy, A., 2006. Making sense of carbonate pore systems. AAPG Bulletin, v. 90, no.9 (September 2006), 1381-1405.
Marwah, V. and Kelkar, M., 2002. Reservoir Mechanism for hunton Formation Production. SPE 75127, presented at SPE/DOE Improved
Recovery Symposium, Tulsa, OK, U.S.A. 13-17 April 2002.
Masalmeh, S.K. 2000. High Oil Recoveries from Transition Zones. Paper SPE 87291 presented at the Abu Dhabi International Petroleum
Exhibition and Conference, Abu Dhabi, UEA, 13-15 October.
Masalmeh, S.K. and Odai, S. 2000. Oil Mobility in Transition zone. Paper SCA 2000-02 presented at the international Symposium of the
Society of Core Analysts, Abu Dhabi, UAE, 18-22 October 2000.
Masalmeh, S.K. Abu Shiekah, I. and Jing, X.D., 2007. Improved Characterization and modelling of Capillary Transition Zones in
Carbonate Reservoirs. SPE Reservoir Evaluation & Engineering, 191-204, April 2007
Muscat, M. 1948. Calculation of Initial Fluid Distributions in Oil Reservoirs. Petroleum Tech., Tech. Pub. No. 2405. (July 1948), p. 119-
127.
Parkers, R., Carnegie, A. and Hookway, S., 1998. New Techniques in Wireline Formation Testing in Tight Reservoirs. SPE 50128,
presented at the 1998 SPE Asia and Pasific Oil & Gas Conference and Exhibition, Perth, Australia, 12-14 October 1998.
Pham, T.R., Al-Afaleg, N. I., Kelder, O. and Al-Otabi, U. F. 2005. Field Example of Capillary Pressure Effects on Wireline Formation
Tester Measurements and OWC Estimation in a Mixed-Wettability Oil Reservoir.
Phelps, G.D., Stewart, G., and Peden, J.M., 1984. The Effect of Filtrate Invasion and Formation Wettability om Repeat Formation tester
Measurements. SPE 12962 presented at the 1984 European Petroleum Conference, London, England, October 25-28, 1984.
Phelps, G.D., Stewart, G., and Peden, J.M., 1984. The Analysis of the Invaded Zone Characteristics and Their Influence on Wireline Log
and Well Test Interpretations. SPE 13287 presented at the 1984 ATCE, Houston, Texas, 16-19 September 1984.
Proett, M.A., Wilson, C.C., Manohar, M. Sigal, R. and Wu, J., 2001. Multiple Factors That Influence Wireline Formation Tester Pressure
Measurements and Fluid Contacts Estimaes. SPE 71566, presented at the 2001 SPE ATCE, New Orleans, Lousianna, 30 September –
3 October 2001.
Siddiqui, F.I., 1999. Defining Fluid Distribution and Fluid Contacts for Dynamically Charged Reservoirs. SPE 56513, presented at the
1999 SPE ATCE, Houston, Texas, 3-6 October 1999.
Skjaeveland, S.M., Siqveland, L.M., Kjosavik,A., Hammervold, W.L., and Virinovsky, G.A., 1998. Capillary Pressure Correlation for
Mixed-Wet Reservoirs. Paper SPE 39497, presented at the SPE India Oil and Gas Conference and Exhibition, New Dehli, India, 17-
19 February.
Weinberger, P., Boratko. E., Contreiras, K. D. and Van-Dunem, F., 2008. Best Practices for Formation Testing in Low Permeability
Reservoirs. SPE 115825 presented at the 2008 SPE ATCE, Denver, Colorado, U.S.A., 21-24 September 2008.
16 SPE 121982

Appendix: Reservoir data for simulation model.

Grid type: Radial
Grid size: 13*4*100
DRV (grid size radial direction in m.): 0.1, 0.03, 0.06, 0.12, 0.24, 0.48, 0.96, 1.92, 3.84, 7.68, 15.36, 300.72
DZ (meters): 2
PERMR (mD): 1
PERMZ(mD): 1
Permeability in invaded zone (mD) 0.002
Porosity (fractions) 0.1
Water Density (kg/m3) 1.035
Water Viscosity (cp) 0.3
Oil Density (kg/m3) 700
Oil Viscosity (cp) 0.3
Relative Permeability Corey type with No = 4 and Nw = 4, Krw,end = 0.5
+ various sensitivities and Sorw scenarios (Killough hysteresis)