SPE 113464

Experiments to Investigate Steam Injection in Light Oil Fractured

Carbonates

Marco Verlaan, Shell International Exploration and Production, Rijswijk, The Netherlands; Paul Boerrigter, Shell
International Exploration and Production, Rijswijk, The Netherlands; Shell Technology Oman, Muscat, Sultanate of
Oman;. Sjaam Oedai, Shell International Exploration and Production, Rijswijk, The Netherlands; and Johan van
Dorp, Shell Technology Oman, Muscat, Sultanate of Oman
Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE/DOE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, U.S.A., 1923April2008.

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
Conventional displacement methods such as water flooding do not work effectively in densely fractured reservoirs. In such
reservoirs one has to rely on recovery mechanisms like capillary imbibition or gravity to recover oil from the reservoir rock
matrix. In oil-wet or mixed-wet fractured carbonates only gravity drainage remains a feasible process. However, low
permeabilities result in low gravity drainage production rates with high remaining oil saturation.
EOR methods have the potential to improve GOGD drainage rate and ultimate recovery. Especially for shallow fractured
reservoirs, it may be attractive to inject steam to improve oil rate and recovery. Heating of the matrix will result in oil
expansion, reduction of viscosity, solution gas drive and steam stripping of intermediate hydrocarbon components. Solution
gas drive and steam stripping effects potentially become more important than the viscosity reduction.
We experimentally investigated the physical mechanisms involved. We present the results of a laboratory study in which
reservoir core with light crude oil at reservoir conditions is heated to steam temperature. From these experiments and separate
PVT measurements the effects of thermal expansion of oil, gas liberation and initial water saturation are investigated. The
experiments are interpreted numerically by detailed modelling of the observed production. The results show that connate water
has a big impact on the gas drive and distillation process, and as a consequence enhances the oil recovery.

Introduction
The connected fracture network in densely fractured reservoirs has a strong impact on reservoir displacement mechanisms.
Conventional displacement methods such as water flooding do not work effectively: due to the high fracture permeability it is
not possible to establish significant pressure differentials across oil bearing matrix blocks to drive oil from matrix rock into the
fracture system. In densely fractured reservoirs one relies on mechanisms like capillary imbibition or gravity to recover oil
from the matrix reservoir rock. Fractured carbonate reservoirs are commonly oil wet or mixed wet, and the main production
mechanism is gravity. Once a gas cap is established in the fracture system, the oil will drain down the matrix rock driven by
gravity and into the fracture system at flow barriers. In the fracture system the oil forms a (thin) rim that can be produced.
Production rates achieved with this GOGD (Gas Oil Gravity Drainage) process are often low due to low matrix rock
permeability, capillary hold-up and re-imbibition effects. Capillary hold-up also reduces ultimate recovery.
Both miscible gas injection and steam injection are feasible EOR processes to accelerate the production and increase
recovery. Steam injection in heavy oil reservoirs is common practice and recently receives more attention in naturally fractured
reservoirs
1-6
. Steam injection in light oil reservoirs is not common, although there are some examples of steam flooding non-
fractured or sparsely fractured reservoirs
7
. Thermally assisted gas-oil gravity drainage (TA-GOGD) in light oil has not been
done before. Burger
8
already suggested that the increase in temperature in light oil naturally fractured reservoirs would lead to
oil expulsion of significant quantities of oil from the matrix blocks into the fracture. The recovery mechanisms that play a role
are very similar to those of a light oil steam flood
9
:

1. Viscosity reduction
2. Distillation
3. Gas drive
2 SPE 113464
4. Thermal expansion
5. IFT change
6. Wettability alteration

Figure 1 illustrates how the most significant mechanisms impact the production rate from a GOGD reservoir. The presence of
connate water or reservoir brine will have a significant impact on the produced oil properties and the recovery mechanisms due
to the oil distillation and gas formation. When solution gas forms with the rise in temperature, the water will also evaporate
into this gas phase, well below the water boiling temperature. This will lead to larger gas volumes generated. As a result, the
vapour pressure of the volatile hydrocarbon components is reduced. Consequently the oil sees a much lower pressure and more
and heavier components of the oil will be distilled to increase the hydrocarbon partial vapour pressure. This distillation will
not only lead to larger gas volumes, the oil properties will also change as a result of the difference in oil composition. The
presence of water can significantly impact all contributing recovery mechanisms. In our experiments, we tested the oil
response with and without brine present.
This paper focuses on the first four mechanisms indicated above. IFT reduction and wettability alteration are not addressed
in this paper. The outline of the paper is as follows. First we will discuss the experiments and procedures. The results will then
be presented for each of the four mechanisms. The interpretation of the gas drive experiments is not straightforward and
requires simulation modelling. Finally we will deduce the relative contribution of the various mechanisms to the enhanced
recovery process.

Experimental setup and procedures
Two different experiments were conducted to investigate the various mechanisms. Oil thermal expansion, viscosity reduction
and distillation in the presence of brine are fluid properties and have been measured with experiments in a PVT cell. To
measure the gas drive mechanism a special setup was used where a core plug saturated with live reservoir oil was heated.
Experiments were conducted with and without connate water (brine). The fluid used in the experiments is light oil recombined
from actual tank oil and synthetic gas to obtain the required reservoir live oil composition. The main fluid properties are shown
in table 1.

Differential Thermal Depletion Experiments
In order to simulate the reservoir mechanism in a PVT experiment we conducted differential thermal depletion (DTD)
experiments. These experiments are similar to regular differential liberation experiments with the difference that instead of
lowering the pressure, the temperature is increased in several steps (see figure 2). After every heating step the gas and oil
volumes, phase compositions and phase properties have been measured. The gas phase is then removed and the remaining oil
is heated in the next step. The recombined oil sample is heated in 5 steps from the initial reservoir temperature to a temperature
of 240C, which is just above the pure water steam temperature at 30 bar (233.85C). A second experiment has been conducted
with oil and synthetic reservoir brine. In this experiment the maximum temperature could not be reached because the gas
volume generated exceeded the PVT cells volume. This experiment was stopped at 222C.

Gas drive experiments
The gas drive experiments were conducted in a setup as shown in Figure 3. A 2-inch core-holder is connected to a piston pump
and the whole system is submerged in a thermostatic oil bath. The piston pump is controlled at a constant pressure and collects
all produced fluids during a heating step. The temperature is measured both in the oil bath and with a thermocouple inside the
core-holder. In addition the volume increase in the piston pump is recorded.
Plugs were drilled from a carbonate reservoir core and checked with the CT scan for homogeneity. The core is prepared in
the following way:
1. Cleaning
2. Standard determination of density, porosity and permeability
3. Fixing in a core holder
4 Flooding with recombined live crude at bubble point pressure
5. Aging for 4 weeks

The plug with connate water was flooded with synthetic brine to 100% saturation after step 3 and in step 4 the core was de-
saturated to connate water saturation with dead crude. After ageing the dead crude was displaced with live crude. The
properties of the two cores are given in table 2.

Experimental procedure
The concept of the experiment is to heat the core with stepwise temperature increments and to collect all produced fluids for
each temperature step in the production pump (figure 3), while maintaining a pressure of 30 bar in the core holder at all times.
The production pump reads the cumulative total production at reservoir conditions. Once temperature equilibrium is reached,
the connection between core and pump is closed, and the contents of the pump are collected through a separate connection in
the production vessel (at vacuum). The empty pump is then brought to elevated pressure and connected to the core again. A
SPE 113464 3
new temperature level of the heating oil is set and the procedure is repeated. The production vessels for each temperature level
are analyzed in the laboratory by diluting the produced fluids with pentane, draining the contents into a graduate tube and
reading off the remaining oil volume after the pentane has evaporated. Produced gas, dissolved water and dissolved gas are
lost by this procedure, and the determined volumes represent produced stock tank oil.

Experimental Results

Thermal expansion and viscosity reduction
The thermal expansion of the oil is determined in the PVT experiment by the measurement of the density at every temperature
step. The oil expansion between 61C and 240C is 20% (see figure 4). The oil expansion can be quite significant for part of
the reservoir where oil saturations are high and the oil can be expelled directly into the fracture system. In de-saturated zones it
will be less important, and only lead to an increase in oil saturation.
The viscosity at 240C is reduced to 20% of its original value at reservoir conditions (figure 5). Since oil viscosity is a key
fluid property in gravity drainage
11
, and inverse proportional to the production rate, a fivefold reduction of the viscosity results
in 5 times faster recovery of the oil.
The effect of the presence of brine is also shown in figures 4 and 5. At temperatures above 150C the effect of the
distillation on the oil properties become important. Distillation of the light/intermediate components will result in higher oil
density and higher oil viscosity. Consequently it will reduce the expansion effect and viscosity reduction.

Distillation
The heating of the oil results in gas formation and evaporation of light components from the oil phase to the gas phase. The
volume fraction of gas after each temperature step is shown in figure 6. When only oil is heated, the gas volume fraction
ranges between 6 to 11 vol. percent for each step, however, when the oil is heated in the presence of brine the amount of gas
formed increases enormously. In the last step the gas volume was 6 times larger than the oil volume. This shows that the
presence of connate water has a tremendous effect on the gas drive mechanism, even well below the boiling temperature of the
brine (connate water).
The evaporation of the water has also a large impact on the distillation of hydrocarbon components. The presence of water
vapour in the gas phase will reduce the hydrocarbon vapour pressures and as a result more and heavier components will be
stripped from the oil. This is clearly illustrated in figure 7, where the composition of the oil after heating is compared to the
original fluid composition. To be able to compare, the composition after heating is rescaled on the C29 component mol
fraction. Without brine the oil composition at 240C shows reduction in components up to C10, and only reduction by more
than 50% for components lighter than C4. In the presence of brine however, the oil is stripped of components up to C16. The
fraction lighter than C4 is completely removed and between C4 and C8 reduced with more than 50%.
The impact of the distillation on the oil properties (density and viscosity) has been shown above. Another consequence is
that the gas phase will contain a fraction of the condensable hydrocarbons (C4-C16) that can be recovered as condensate on
surface or drop out at cooler areas in the reservoir. The mass fraction of the components that were stripped from the oil phase
in these 4 temperature steps was approximately 15% and therefore could contribute significantly to the recovery process.

Gas drive
The objective of the gas drive experiments is to quantify the effect of the various mechanisms on the oil production from a
plug of reservoir core. Contrary to the PVT experiments only few parameters are measured: the total produced volume at
pressure and temperature (continuous) and the volume of the stock tank oil (at the end of each temperature increment).
Compositional change due to distillation and or condensation could therefore not be isolated. Experiment M is performed with
100% live oil in the pore space. The temperature and total volume produced are shown in figure 8. The total and stock tank oil
volumes recovered are given in Table 3. The total volume (gas + oil at P,T) produced is 65% of the pore volume and the oil
recovery factor was 17%. The first temperature increment shows an anomalous low oil recovery, which is possibly caused by a
small leak of the pump seal. In experiment P with connate water present, the total produced volume was 3 times more than in
the experiment without connate water and the oil recovery factor increased to 33%. The extra recovery can be attributed to the
enhanced gas drive mechanism and distillation by the connate water. Interpretation of these results is difficult due to the sparse
data and the complexity of the process. To get more insight in the role of the different mechanisms a simulation model was
constructed and history matched (discussed below).

Gas drive interpretation

PVT model
To interpret the gas drive experiments a simulation model was constructed using a thermal k-value reservoir simulator with 7
components (5 hydrocarbon, H
2
O and salt). A 5-component Peng-Robinson EOS model has been tuned to the experimental
PVT data from the DTD and additional CCE experiments. A considerable effort has been made to match the phase volumes of
the DTD experiments, to be able to match the gas drive volumes of the experiments. This EOS model is then used to generate
K-value input data for the simulator. The 5 components are chosen such that the distillation effect up to C18 could be captured
4 SPE 113464
(see table 4). The salt content of the water has a significant effect on the water evaporation, reducing its vapour pressure. To
capture this behaviour a salt component was added to the water, which reduced the water vapour pressure by Raoults law.
Because the reservoir brine consisted mainly of Calcium Chloride we incorporated only this component. The non-ideal vapour
pressure reduction and precipitation were extrapolated from the data of Conde
12
.

Simulation model
The experiment is history matched using a K-value thermal version of MoReS, an in-house reservoir simulator. The radial
symmetric simulation grid is shown in figure 10 where every part of the setup was explicitly modelled. The grid resolution is
sufficient to capture the thermal processes and this was checked with a 4x4 grid refinement. The pump is modelled with a grid
block with an extremely large compressibility. This ensured the constant pressure in the system. The core is heated from the
outside, top and bottom by heater blocks that reproduce the temperature profile in the oil bath over time. The heater blocks are
separated from the core by a no-flow boundary. The heat is transported to the core by conduction only. The pipe grid blocks
model the connection tube between the top of the core and the pump. These blocks contain the dead volume between the core
and the pump (1.1cc) and have a 1000 times higher absolute permeability than the core. Furthermore straight-line relative
permeabilities and zero capillary pressure are assumed in the pipe. The core is modelled to be homogeneous and all grid cells
have the same porosity and permeability as given by table 2. The SCAL parameters (relative permeability and capillary
pressure) were not measured on these plugs but were taken as from other cores located in the same reservoir zone and with
similar porosity & permeability. The oil and gas relative permeability functions are matching parameters in fitting the
experimental data.

History match
Both experiments were matched with the model described above in two steps. First the total produced volume was matched
and secondly the stock tank oil volumes. The total volume produced is mainly determined by the PVT behaviour. In
experiment M the simulated volumes are very close to the data and no adjustment was made. In experiment P the total
produced volume in the first increment was much lower than expected with the same PVT as in experiment M. A sound
explanation comes from the different core preparation, mentioned above. The core was aged with dead crude and the dead
crude was displaced by live crude before the experiment. An incomplete displacement results in a higher bubble point
temperature and less gas liberation in the first temperature step. To match the total produced volume 8% dead crude had to be
mixed in with the live crude. The composition of this mixture is used in all simulations of experiment P.

Experiment M without connate water
Figure 11 shows the produced volume vs. time for experiment M. The simulation shows a good match for the absolute
volumes and the dynamic behaviour. The only significant deviation is at the end of the first temperature step, where the
produced volume decreases at constant temperature. The mismatch with the model at the end is 0.35cc and can be explained
by a small leak. All other temperature increments did not show any signs of leaks, even after 15-20 hours of equilibrium.
The stock tank oil volumes are matched using a Monte-Carlo approach. The parameters of the Corey type relative
permeability curves for oil and gas, except for Kg, were varied over a wide range for 1000 realisations. The best fit was
obtained by the minimum of a least squares objective function. Because of the abnormal low value for the first increment, two
best fits were obtained; one with and one without the first data point (mod.). The parameter values of the best fits are shown in
table 5, and the values for the oil recovery are given and compared to the data points in table 6.
The best regular fit (all three points) shows considerable deviations for the stock tank oil recovery of the first two steps of
8% and -13%. Other realizations either fit the first point well and enormously underestimate the second point, or fit the second
point well and then the first point was overestimated. There is no combination of relative permeability functions that could
explain both data points within measurement error. This is again a strong indication that the first oil recovery point is incorrect.
The modified fit matched the data (step 2+3) within 1% and predicts almost double the recovery for the first point. The main
difference between the models is the higher critical gas saturation and the higher gas Corey exponent resulting in lower gas
mobility at low gas saturations.
Actually there are many realizations that fit the results within the measurement accuracy of 10%. This due to the fact that
the amount of data is sparse and only covers a small range of the saturation range. The range of parameters that fit the data
within the accuracy is also shown in table 5. A wide range demonstrates that the fit is not sensitive to that parameter. This is
the case for the oil Corey exponent (Nog), oil relative permeability endpoint (Krog) and the oil residual saturation to gas
(Sorg). The ranges of Ng and Sgc are smaller indicating that these are more constraint by the data. The low values for Ng and
Sgc are comparable with reported critical gas saturations
13
and low gas mobility
14
for solution gas drive by depressurization.

Experiment P with connate water
The total volume match in experiment P (see figure 12) is also good for the fist two steps, where the presence of the water has
little influence. During the third step the water also starts to evaporate in significant amounts and generates much more gas
volume. The exact amount of gas generated is mainly determined by the salt concentration in the model. Fresh water
overestimates the gas volume by 23%; while the applied salt model
12
underestimates the total volume produced by 10%. The
salt concentration was not used as a fit parameter, as it is accurately known. It is possible that the salt model is not adequate
SPE 113464 5
because it was extrapolated beyond its valid temperature limit of 100C. Another possibility is CO
2
gas generation due to
brine-mineral reactions
2, 10
. This could not be confirmed with quantitative gas analysis, however CO
2
isotope analysis of a
produced gas sample shows CO
2
of different origin than the oil. It has to be noted that the generated volumes in this core
experiments are much smaller than the volumes in the DTD PVT experiments. This is a result of continuous removal of the
gas phase from the core, comparable to the effect of a multistage flash. To accurately predict the gas drive the coupled effect of
transport mechanisms and PVT have to be solved simultaneously.
The results and parameters of the stock tank oil match are given in table 5 and 7. The deviation of the data is within the
estimated measurement accuracy of 10%. The parameters of this simulation are comparable with those of experiment M. This
indicates the cores have fairly identical properties. It is then the presence of connate water that causes the increase in oil
recovery of experiment M and P. Looking at the parameter ranges for the fitted results within the measurement error, the main
difference with experiment M. is the smaller range of Ng. This can be explained by the larger gas drive recovery and hence
lower oil saturations in the core, reducing the uncertainty in the gas relative permeability.

Recovery Mechanisms
The simulation model can be used to isolate the contributions of the various mechanisms to the oil recovery in the heating
experiment. The three mechanisms considered are thermal expansion, solution gas drive and distillation. Viscosity reduction is
ignored as a separate recovery mechanism because there is no gravity drainage occurring in the experiment, but it will still
have an effect on the solution gas drive. The contribution of the solution gas drive is determined by running the same
simulation with a reduced gas viscosity (factor 1000). The difference between this run and the simulation containing all
mechanisms quantifies the contribution by gas drive. The contribution of thermal expansion can be calculated by mass balance
from the produced fluid composition, because the heaviest component (CMP5) does not evaporate in significant amounts. The
remaining oil then comes from distillation. Figure 13 and 14 show the cumulative oil recovery for each temperature step for
simulation of experiment M (mod.) and P. The main recovery mechanism in experiment M is thermal expansion (13.0% of
STOIIP). Only 6.3% recovery comes from gas drive and 0.7% from distillation. For the experiment with connate water, the
thermal expansion recovery is lower (10.1%). This difference occurs in the third increment and is a result of the impact of
distillation on oil density. The majority of the recovery however is by gas drive (18.7%). The recovery by distillation is
considerable (4.5%), but smaller than the 15% recovered in the PVT experiment. Just as for the gas volumes this is the result
of the coupling of PVT behaviour with fluid transport.

Conclusions
1. PVT properties of a 33 API light oil with and without reservoir brine are quantified up to the steam temperature
using a Differential Thermal Depletion experiment. Thermal expansion of 20% and viscosity reduction of 80% at
steam temperature will greatly enhance the recovery potential for this light oil.
2. Presence of water (brine) has a large positive effect on gas generation and distillation, increasing the contribution of
these recovery mechanisms. The negative effect of the removal of light components (<C16) on oil density and
viscosity are small.
3. The gas volumes generated in a core heating experiment are much smaller than in a PVT experiment, because the gas
phase is continuously removed in the gas drive experiment. The PVT behaviour and transport problem has to be
solved simultaneously to obtain correct results.
4. Salt dissolved in the connate water decreases the contribution of water to the gas drive and distillation due to vapour
pressure lowering. It is essential to take this into account in both experiments and modelling.
5. The gas drive mechanism can be the major recovery mechanism, when connate water is present.
6. The distillation mechanism also contributes significantly to the enhanced oil recovery

Acknowledgements
We gratefully acknowledge Erwin Dufour, Andrey Bychkov, Wim Scherpenisse, Jos Maas, Dan Yang and Arnold de Vries for
their contributions to the experiments, modelling and discussions. We also would like to thank Shell Exploration & Production
for their support for the project and for permission to publish this paper.

Nomenclature
Sgc : Critical gas saturation
Krg : Gas endpoint relative permeability
Krog : Oil endpoint relative permeability to gas
Ng : Gas Corey exponent
Nog : Oil Corey exponent to gas
Sorg : Residual oil saturation to gas
PV : Pore volume
RF : Recovery factor
IFT : Interfacial tension
GOR : Gas oil ratio
6 SPE 113464
FVF : Formation volume factor
Swi : Residual water saturation

References

[1] Ikwumonu, A., Rawnsley, K.Habsi, M., and Penny, R., Key learnings From History Matching a Thermally Assisted Gas Oil Gravity
Drainage Pilot in Fractired Reservoirs, IPTC paper 11411, presented at the 2007 International Petroleum Technology Conference,
Dubai, UAE.
[2] Mollaei, A., Maini, B., and Jalilavi, M., Investigation of Steam Flooding in Naturally Fractured Reservoirs, IPTC paper 11087,
presented at the 2007 International Petroleum Technology Conference, Dubai, UAE.
[3] Mateo Hernandez, J.A., Trevisan, O.V., Heavy-Oil Recovery Mechanisms During Steam Injection in Naturally Fractured Reservoirs,
SPE paper 107372, presented at the 2007 SPE Latin American and Caribbean Petroleum Engineering Conference, Buenos Aires,
Argentina.
[4] Das, Swapan, Application of Thermal Recovery Processes in Heavy Oil Carbonate Reservoirs, SPE paper 105392, presented at the
2007 SPE Middle East Oil & Gas Show and Conference, Kingdom of Bahrain.
[5] Shahin, G.T., Moosa, R., Kharusi, B., and Chilek, G., The Physics of Steam Injection in Fractured Carbonate Reservoirs: Engineering
Development Options That Minimize Risk, SPE paper 102186, presented at the 2006 SPE Annual Technical Conference and
Exhibition, San Antonio, TX, USA.
[6] Penny, R., Moosa, R., Shahin, G.T., Hadrami, F., Kok, A., Engen, G., van Ravesteijn, O., Rawnsley, K., and Kharusi, B., Steam
Injection in Fractured Carbonate Reservoirs: Starting a New Trend in EOR, IPTC paper 10727, presented at the 2005 International
Petroleum Technology Conference, Doha, Qatar.
[7] Kaveh Dehghani, Robert Ehrlich, Evaluation of the Steam-Injection Process in Light-Oil Reservoirs, SPE Reservoir Evaluation &
Engineering 4 (5), October 2001, p.395-405.
[8] Burger, J., Sourieau, P., Combarnous, M.,Thermal Methods of Oil Recovery, Editions Technip, Paris, France (1985).
[9] B. Todd Hoffman, Anthony R. Kovscek, Light-Oil Steamdrive in Fractured Low-Permeability Reservoirs, SPE paper 83491,
presented at the 2003 SPE Western Regional Meeting, Long Beach, Ca., USA.
[10] Reis, J.C., Oil Recovery Mechanisms in Fractured Reservoirs During Steam Injection, SPE paper 20204, presented at the 1990 SPE
Symposium on Enhanced Oil Recovery, Tulsa, OK, USA.
[11] Hagoort, J., Oil Recovery by Gravity Drainage, SPE Journal 20 (3), June 1980, P.139-150.
[12] Manuel R. Conde, Properties of aqueous solutions of lithium and calcium chlorides: formulations for use in air conditioning
equipment design, International Journal of Thermal Sciences 43 (2004), P. 367382
[13] Boge, R., Lien, S.K., Gjesdal, A., and Hansen, A.G., Turning a North Sea Oil Giant into a Gas Field Depressurization of the
Statfjord Field, SPE paper 96403, presented at Offshore Europe 2005, Aberdeen, Scotland, U.K.
[14] P. Egermann, O. Vizika, Critical Gas Saturation and Relative Permeability During Depressurization in the Far Field and the Near-
Wellbore Region, SPE paper 63149, presented at the 2000 SPE Annual Technical Conference and Exhibition, Dallas, TX, USA.

Table 1 Light oil physical properties at initial reservoir conditions.)
API GOR
[m
3
/m
3
]
FVF
[-]
Viscosity
[cP]
33.0 22.3 1.095 2.776

Table 2 Core properties for carbonate plug M and P.
Length
(cm)
Diameter
(cm)
pore vol.
(cc)
Porosity
(%)
perm.
(mD)
Swi
(%)
Plug M 6.03 5.09 50.84 41.4 4.5 0.0
Plug P 6.04 5.07 46.53 38.1 6.8 10.1

Table 3 In place and produced volumes in experiment M and P.
Experiment

PV
(cc)
Vol. Produced
(cc r.c.)
STOIIP
(cc s.c.)
ST oil produced
(cc s.c.)
RF
(%)
Exp. M 50.84 32.7 46.8 7.95
*
17
*
Exp. P 46.53 107.0 38.2 12.47 33
* part of measurement shows anomalous low recovery
SPE 113464 7
Table 4 Lumping scheme of the 5cmp PVT model
Pseudo Component Lumped Components
CMP1 C1
CMP2 N2, CO2, C2-C4
CMP3 C5-C9
CMP4 C10-C18
CMP5 C19+

Table 5 Simulation History match input parameters
Parameter Exp. M
Best fit
Exp. M
Range (<10% error)
Exp. M (Mod.)
Best fit
Exp. M (Mod.)
Range (<10% error)
Exp. P
Best fit
Exp. P
Range (<10% error)
Sgc 0.01 0.00 - 0.05 0.04 0.00 - 0.05 0.01 0.01 - 0.04
Nog 4.4 2.33 5.61 3.43 2.14 5.92 3.05 2.41 - 5.65
Ng 2.19 1.57 3.12 2.5 1.60 3.24 2.65 2.50 - 3.28
Krog 0.55 0.36 - 0.93 0.59 0.32 - 0.98 0.81 0.53 - 0.92
Sorg 0.49 0.22 - 0.51 0.48 0.18 - 0.52 0.21 0.19 - 0.37

Table 6 Experiment M stock tank oil volumes produced. Experiment and simulation results.
Temperature
(Celsius)
Exp. M
(cc s.c)
Sim. M
(cc s.c.)
Dev.
(%)
Sim. M (Mod.)
(cc s.c.)
Dev (Mod).
(%)
95 1.57
*
1.70 8.6 3.10 97.2
*
153 3.28 2.85 -13.2 3.30 0.52
225 3.10 3.20 3.2 3.11 0.46
* First temperature step shows anomalous result

Table 7 Experiment P stock tank oil volumes produced. Experiment and simulation results.
Temperature
(Celsius)
Exp. P
(cc s.c.)
Sim. P
(cc s.c.)
Dev.
(%)
100 3.26 3.5 7.36
150 3.78 3.54 -6.36
223 5.43 5.93 9.19
8 SPE 113464

S t e a m
V i s c o s i t y r e d u c t i o n
I F T / W e t t a b i l i t y c h a n g e
E x p a n s i o n
S o l u t i o n G a s d r i v e
D i s t i l l a t i o n
S t e a m
w a t e r
g a s
o i l
S t e a m
V i s c o s i t y r e d u c t i o n
I F T / W e t t a b i l i t y c h a n g e
E x p a n s i o n
S o l u t i o n G a s d r i v e
D i s t i l l a t i o n
S t e a m
w a t e r w a t e r
g a s g a s
o i l o i l

T=61 oC
P=30 bar
Gas
Oil
Remove gas
for analysis
Remove gas
for analysis
T=100 oC
P=30 bar
Gas
Oil
Remove gasfor
analysis
T=150 oC
P=30 bar
Gas
Oil
H2O
H2O
H2O
H2O + Oil
Next stage
H2O + Oil
Next stage
Etc.
T=61 oC
P=30 bar
Gas
Oil
Remove gas
for analysis
Remove gas
for analysis
T=100 oC
P=30 bar
Gas
Oil
Remove gasfor
analysis
T=150 oC
P=30 bar
Gas
Oil
H2O
H2O
H2O
H2O + Oil
Next stage
H2O + Oil
Next stage
Etc.

Figure 1 Main mechanisms in thermally assisted gas oil
gravity drainage (TAGOGD)
Figure 2 Schematic of Differential Thermal Depletion
(DTD) experiment
P = 30 bar
P = 30 bar
T
ini
= 61 C
N
2
P
P = 30 bar
P = 30 bar
T
ini
= 61 C
N
2
PP

Figure 3 Schematic of the experimental setup of gas drive experiments.
SPE 113464 9
0.9
0.95
1
1.05
1.1
1.15
1.2
1.25
1.3
0 50 100 150 200 250 300
Temperature (Celcius)
O
i
l

E
x
p
a
n
s
i
o
n

f
a
c
t
o
r

(
-
)
Oil only
Oil + Brine
0
0.2
0.4
0.6
0.8
1
1.2
0 50 100 150 200 250 300
Temperature (Celcius)
R
e
l
a
t
i
v
e

O
i
l

V
i
s
c
o
s
i
t
y

(
-
)
Oil only
Oil + brine

Figure 4 Oil thermal expansion factor as function of
temperature
Figure 5 Viscosity reduction factor as function of
temperature
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
100 150 200 240/222
Temperature (Celcius)
G
a
s

f
r
a
c
t
i
o
n

(
m
3
/
m
3
)
Oil only
Oil+ brine
0
2
4
6
8
10
12
C
O
2
C
1
C
3
n
-C
4
n
-C
5
C
7
C
9
C
1
1
C
1
3
C
1
5
C
1
7
C
1
9
C
2
1
C
2
3
C
2
5
C
2
7
C
2
9
r
e
l
a
t
i
v
e

m
o
l

%
Composition at 61C
Comp. At 240C
Comp at 222C with brine

Figure 6 Gas fraction in PVT cell at each temperature
step (at 30 bar pressure).

Figure 7 Oil composition before and after heating.
Relative oil composition scaled to equal C29 mol%.
50
70
90
110
130
150
170
190
210
230
0 10 20 30 40 50 60 70 80
time (hrs)
T
c
o
r
e

(

C
)
0
5
10
15
20
25
30
35
p
r
o
d
u
c
t
i
o
n

v
o
l
u
m
e

p
u
m
p

[
c
c
]
T core
pump vol
50
70
90
110
130
150
170
190
210
230
0 10 20 30 40 50 60 70 80
time (hrs)
T
c
o
r
e

(

C
)
0
20
40
60
80
100
120
p
r
o
d
u
c
t
i
o
n

v
o
l
u
m
e

i
n

p
u
m
p

(
c
c
)
T core
pump vol

Figure 8 Temperature of the core plug and volumes
produced by the production pump as a function of time
for plug M.
Figure 9 Temperature of the core plug and volumes
produced by the production pump as a function of time
for plug P.
SPE 113464 10

C o r e
P i p e
P u m p
H e a t e r

Figure 10 Crossection of 2-D radial simulation model

0
5
10
15
20
25
30
35
0 10 20 30 40 50 60 70
Time (Hour)
P
u
m
p

V
o
l
u
m
e

(
c
c
)
Data Exp.M
Best Fit model
0
20
40
60
80
100
120
140
0 10 20 30 40 50 60 70
Time (Hour)
V
o
l
u
m
e

(
c
c
)
Data Exp. P
Simulation (Fresh water)
Simulation (118g/l brine)

Figure 11 Experiment M with simulation Figure 12 Experiment P with simulation
0%
5%
10%
15%
20%
25%
30%
35%
40%
95 153 225
Temperature step (Celsius)
C
u
m

%

o
i
l

r
e
c
o
v
e
r
d

Distillation
Solution gas drive
Thermal Expansion
0%
5%
10%
15%
20%
25%
30%
35%
40%
100 150 223
Temperature step (Celsius)
C
u
m

%

o
i
l

r
e
c
o
v
e
r
d

Distillation
Solution gas drive
Thermal Expansion

Figure 13 Simulated Stock tank oil recoveries for each
temperature step, split by mechanism (experiment M
(Mod.).
Figure 14 Simulated Stock tank oil recoveries for each
temperature step, split by mechanism (experiment P).