Predicting of Oil Water Contact Level using Material Balance Modeling

of a Multi-tank Reservoir

Muslim Abdurrahman1 , Bop Duana Afrireksa2 Hyundon Shin2 , Adi Novriansyah1,3

1 Petroleum
Engineering Department, Universitas Islam Riau, Pekanbaru, Indonesia
2 Department
of Energy Resources Engineering, Inha University, Incheon, South Korea
3 Department of Energy and Mineral Resources Engineering, Sejong University, Seoul, South Korea

Keywords: Oil Water Contact, Material Balance, Tank Model, Sand Production, Prediction, Reservoir Modeling.

Abstract: Nowadays, the increase in water production becomes a problem in the oil and gas industry. Besides being a
problem, it also becomes extra energy to produce oil and gas. OWC is one of the keys for water production
determination for each layer. If the perforation at production well is at OWC or below OWC, the production
will be 100% water. In general, the log is used to determine OWC. Besides with log, tank modeling from
the material balance equation is also used to determine OWC. WH field located 15 km from Bangko Field in
Riau. This primary field has high water production with 97% water cut. Before tank modeling starts, each
layer needs to be analyzed based on its reserves, production cumulative and remaining reserves to determine
the productive layer, which can be developed in the future. Prediction can be done when history matching and
calibration process for both historical data and simulated data by software. Prediction ends in August 2021,
which is the end of development contract in WH field. From the results, it can be determined that from C sand,
the OOWC and COWC are at 2922 ft and 2883 ft with the cumulative oil production is 6.78 MMSTB. From
E sand also can be determined the OOWC at 2368 ft and COWC at 2325 ft with the cumulative oil production
is 14.57 MMSTB. From K sand, the OOWC is at 2002 ft and COWC at 1911 ft with the cumulative oil
production is 13.5 MMSTB. L sand the OOWC is at 2243 ft and COWC at 2191 ft with the cumulative oil
production is 29.17 MMSTB. From the analysis, K sand has the most significant OWC movement, which is
91 ft and it is also validated with the current log data. This sand needs more care to maintain water production.

1 INTRODUCTION (Noordin, 2009). Water shut off method can be done

by using a mechanical method (packer), cementing
Water production is one of the common problems of (squeeze), or using chemical mixtures. These meth-
the past few years (Hudiman and Permadi, 2016). ods can be used in order to maintain water production
Water production is also one of the dilemmas in oil so it will increase oil production with low expendi-
and gas industries, on the other side water is known tures (Stashin, 1989).
as an energy source in reservoir flow (Daneshy, 2006). Oil water contact is the key to determine water
Production well at the beginning of development has production when the production reaches 100% water
a bigger oil production than water does. As time goes cut, OWC must be at or above the perforation. Log-
by, oil production will decrease because of several ging is the common method to determine OWC po-
things, there are formation damage, pump mechanical sition either the original one (OOWC) or even cur-
failure, etc. This also caused by the increase in wa- rent position of OWC (COWC). Besides that, there
ter production (increasing of water cut), where water are several methods to determine OWC position, there
movement is faster than oil. With this water produc- are RFT, DST, and other good tests. The following
tion, it can decrease production efficiency and profit methods including logging data are costly and have
for the oil and gas company. some limitations especially in certain reservoir issue
The method that has been used to maintain wa- (Ghahri et al., 2013). Material balance is a low-cost
ter production is by doing workover jobs, one of the approach for determining OOWC or even COWC po-
jobs is by closing the zone, which is not productive sitions (Nwaokorie and Ukakuku, 2012). By material
or it has 100% water cut which called water shut off balance also we can study the movement of OWC it-

331
Abdurrahman, M., Afrireksa, B., Shin, H. and Novriansyah, A.
Predicting of Oil Water Contact Level using Material Balance Modeling of a Multi-tank Reservoir.
DOI: 10.5220/0009404603310336
In Proceedings of the Second International Conference on Science, Engineering and Technology (ICoSET 2019), pages 331-336
ISBN: 978-989-758-463-3
Copyright c 2020 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
ICoSET 2019 - The Second International Conference on Science, Engineering and Technology

self. OOIP is 184.457 MMSTB. In February 2017, the av-

Material balance is one of several methods used erage water cut of this field reached 97%. High water
estimating reserves for oil and gas reservoir and thus cut becomes a dilemma in this field.
allows for making the critical decisions concerning The purpose of this paper is making the tank
development plans and strategies regarding the reser- model of each most productive layer from WH field
voir. It is also the simplest way to express the conser- by using IPM – MBAL software and predict the OWC
vation of mass in a reservoir. The material balance is movement until August 2021, which is the end of the
zero-dimensional, meaning that it is based on a tank contract for the WH field development. The predic-
model and does not take into account the geometry of tion is used to determine the sand, which has a sig-
the reservoir, the drainage areas, the position, and ori- nificant movement of OWC. The log data is needed
entation of the wells. The other uses of this concept to validate the OWC movement for each productive
are to determine the size of an aquifer, encroachment sand.
angle of the aquifer, estimate the depth of fluid con-
tact, etc (Dake, 1983).
The material balance equation mathematically de-
fines the different producing mechanisms which ef- 2 GEOLOGY AND RESERVOIR
fectively relates the reservoir fluid and rock expansion CONDITION
to the substance of fluid withdrawal. Several methods
have been developed and published applying the ma-
terial balance equation to the various types of reser- WH is located at Central Sumatera Basin, Indone-
voirs and solving the equation to obtain the initial oil sia, at Bangko Area in Riau Province. This for-
in place (N) and the ratio of the initial gas to oil (m) mation consist of Brown Shale Formation at Pe-
in the reservoir (Havlena and Odeh, 1963). For wa- matang Valley as the source rock. The lithofacies of
ter drive reservoir diagnostic plot, Campbell plot is Brown Shale Formation is carbonaceous and algal-
used to determine the energy of the aquifer and the amorphous (Katz and Mertani, 1989). Where algal-
OOIP itself by using F/Eowf vs Np plot (Campbell amorphous is oil prone at the upper and middle part
and Campbell, 1978). of Brown Shale Formation (Aman, Kamba, and Ran-
The general material balance equation for an oil gau). Carbonaceous is the gas and light condensate
reservoir is expressed as: prone, which located at Kiri, Aman, Kamba, and Ran-
gau. The transition facies between algal-amorphous
F = NEt +We (1) and carbonaceous is also located at Aman, Kamba,
and Rangau. Pematang group (fine and medium sand-
Where the underground withdrawal F equals to the
stone from Upper Red Formation) and Sihapas Group
production of oil, water, and gas corrected to reservoir
come as reservoir rock after the primary migration to
condition:
the hinge margin basin caused by the Pematang to-
pography, which is asymmetric. The result is, reser-
F = N p (Bo − Bg ∗ Rs ) + Bg ∗ (G p − Gi ) voir rocks along steep fault scarp margin and hinge
(2)
+ (Wp −Wi ) ∗ Bw margin, which formed Telisa, Duri, Bekasap, Bangko,
Pematang, and Petani formation with a total of thick-
And the original oil in place is N stock tank barrels
ness reached 3300 ft.
and E is the unit per unit expansion of oil (and its dis-
solved gas), connate water, pore volume compaction,
and the gas cap:

E = (Bo − Boi ) + (Rsi − Rs ) ∗ Bg + m

 
Bg
∗ Boi −1
Bgi
+ (1 + m) ∗ Boi (3)
 
SW c ∗Cw +C f
∗
1 − Swc
∗ (Pi − P)
WH field is a primary field, which located in Riau
Province. This field discovered in July 1972 with the Figure 1: WH Field Map

332
 Predicting of Oil Water Contact Level using Material Balance Modeling of a Multi-tank Reservoir

WH field reservoir properties from the log data, 3.2 Sand Selection
core, single well-tracer, and volumetric data are as
follows: Sand selection is needed to filter which sand is suit-
Table 1: WH Field Reservoir Properties able to model and develop in the future. The screen-
ing criteria of this section initial volumetric OOIP,
Formation GOR, SCF/STB 26.4
production cumulative, and remaining reserves. In
Oil Gravity, API 34.5
this case, when the remaining reserves are too low for
Gas Gravity, sp. Gravity 0.8
a layer, it will not profit to develop. C, E, K, and L
Water Salinity, ppm 20000
are the selected sand based on these screening crite-
Connate Water Saturation, % 21 ria, which are suitable to model and develop.
Porosity, % 25
3.3 Material Balance Model

The understanding of building a material balance

3 METHODOLOGY model for each productive layer is needed to make a
sand predictive model in material balance. It requires
In this section, the methodology, which applied in basic and fundamental knowledge related to the reser-
this paper will be discussed in order to build the sand voir structure, type, and the aquifer effect to the reser-
predictive material balance equation models by using voir itself. Several analytical models of the aquifer
IPM – MBAL software. were tested in a bid to model the geometry of the
reservoir. Carter Stacy, Van Everdingen, Van Everdin-
gen modified, Hurst-Van Everdingen modified, etc
are the available aquifer models at the software. Af-
ter aquifer model selection (in this case, Hurst-Van
Everdingen modified model was selected), the model
already established to connect the reservoir volume.
The predicted OOIP which generated by the software
can be compared with the volumetric OOIP. In this
Figure 2: General IPM - MBAL Workflow case, the generated OOIP is matched to the volumetric
OOIP for all layers (see Fig 2 for initialization model
plots).
3.1 Data Gathering
3.4 History Matching
Proper data acquisition has to be carried out in or-
der to build a good material balance equation model
With the aquifer model being the key of uncertainty,
or MBAL model. Most of these data are acquired
encroachment angle, ReD, aquifer permeability, and
at the early phase of field development. Either us-
inner/outer ratio were regressed upon the reservoir
ing well tests (RFT, MDT, Swab, PBU, etc) or core
pressure history matching process and production
test (RCAL or SCAL) data acquired are, Pressure,
data assuming reservoir volume reproduced to stock
Production data, PVT, Rock properties, OOIP from
tank condition. The regression needs to be done re-
the volumetric calculation, and PV fraction vs depth.
peatedly until the deviation is lower than 5. It needs
Porosity, permeability, and water connate saturated
to be done in order to validate the model due to the
also are obtained from existing well logs and core
aquifer model uncertainties.
data. Original oil in place (OOIP) obtained by calcu-
lating the rock properties (porosity, water connate sat-
uration, formation volume factor) and net pay thick- 3.5 Simulation
ness and area from well-logs to get the OOIP math-
ematically. Effort should be made in order to under- At this part, reservoir pressure over time is simu-
stand the uncertainties related to the reservoir param- lated from the production history data. This simulated
eters, which used to calculate OOIP. In cases when the reservoir pressure is compared to the measured reser-
MBAL initialize volumes are different from the vol- voir pressure at the field from the input data to see
umetric calculated volumes, basically due to the high the MBAL model could replicate the actual or current
uncertainty of the MBAL data which is used in the reservoir pressure which is given by the same reser-
simulation. voir energy and properties (see Fig 3 to Fig 6). Sim-

333
ICoSET 2019 - The Second International Conference on Science, Engineering and Technology

ulated OWC from the MBAL were calibrated with

logged OWC for modeled sands (Fig 7).

Figure 7: Pressure and Cumulative Production History

Match from C Sand

Figure 3: IPM – MBAL Initialization Output starting point in order to make prediction more vali-
dated. In this section, pseudo-prediction will be gen-
erated by using the prediction tool. Since the goal is to
predict using tank model, a well prediction model was
not used in this case. For the constraint, history pro-
duction rate and time will be used to generate pseudo-
prediction to calibrate the model. Once both points
matched, prediction can be generated next.

3.7 Prediction
Figure 4: Pressure and Cumulative Production History After the model already matched and validated, the
Match from K Sand
next thing is the prediction of the field performance.
Prediction generated until the end of contract of this
field development (August 2021). The models were
further calibrated by running pseudo-prediction for
existing sands. Results were compared with the out-
come from another method in determining the height
of OWC as shown in Fig 8.

Figure 5: Pressure and Cumulative Production History

Match from L Sand

Figure 8: OWC Prediction

4 RESULT
Figure 6: Pressure and Cumulative Production History
Match from E Sand Various results were discussed during the study which
involved saturation reservoir with concurrently oil
production from the oil rim. Well logs will be adopted
3.6 Calibration to verify results from MBAL models. Table 3 shown
material balance results the OWC from MBAL has
Material balance model calibration is needed to match compared well with the log data. For the production
the end of history matching point with the prediction forecast, it predicted using no well prediction which

334
 Predicting of Oil Water Contact Level using Material Balance Modeling of a Multi-tank Reservoir

assumpted the sand production rate is decline natu- REFERENCES

rally due to the pressure loss at the reservoir. Pre-
diction rate will be generated by software as long Petroleum Experts IPM-MBAL Manual.
the reservoir pressure and aquifer is enough to pro- Campbell, R. A. and Campbell, J. M. (1978). Mineral prop-
vide energies. From the result, K sand has significant erty economics. Petroleum Property Evaluation, 3.
movement of OWC, the contact moves from 2002 ft Dake, L. P. (1983). Fundamentals of reservoir engineering.
at 1973 to 1911 ft at 2012. This 91 ft movement Elsevier.
in 48 years from prediction makes this sand needs Daneshy, A. A. (2006). Selection and execution criteria for
more concern due to the water production mainte- water-control treatment. In SPE Symposium and Ex-
hibition on Formation Damage Control, Los Angeles.
nance. The other sand has a certain movement less
Ghahri, P., Berthereau, G., Milner, S., Orta, M. E., Sikan-
than 55 ft in 48 years. dar, A. S., et al. (2013). Estimated fluid contact using
Table 2: Predicted OOWC vs Log OOWC material balance technique and volumetric calculation
improves reservoir management plan. In SPE Offshore
MBAL Log Europe Oil and Gas Conference and Exhibition. Soci-
Sand OOWC OOWC Error (%) ety of Petroleum Engineers.
(ft) (ft) Havlena, D. and Odeh, A. S. (1963). The material balance
C 2922 2925 0.103 as an equation of a straight line. Journal of Petroleum
E 2368 2366 0.085 Technology, pages 896–900.
K 2002 2002 0.000 Hudiman, A. and Permadi, B. Y. (2016). Analisa penentuan
L 2243 2246 0.134 laju air produksi yang optimum untuk memperlambat
water coning di lapisan tipis. JTMGB, 10(1):17–22.
Katz, B. J. and Mertani, B. (1989). Central sumatra — a
Table 3: Predicted COWC vs Log COWC geochemical paradox. In Proc 18th Indon Pet Assoc
Ann Con, volume 1, pages 403–425, Jakarta.
MBAL MBAL Noordin, F. M., e. a. (2009). Case study: Water shut off
Log COWC
Sand COWC in COWC in mechanism in small, remote platform-process & chal-
in 2014 (ft)
2021 (ft) 2014(ft) lenge. In SPE European Formation Damage Confer-
C 2922 2925 0.103 ence, pages 27–29, Netherlands.
E 2368 2366 0.085 Nwaokorie, C. and Ukakuku, I. (2012). Well predictive ma-
K 2002 2002 0.000 terial balance evaluation: A quick tool for reservoir
performance analysis. In SPE Nigerian Annual Inter-
L 2243 2246 0.134 national Conference and Exhibition, Abuja.
Stashin, K. (1989). An analytical approach to determining
oil/water contact rise at utikuma field. In 40th Annual
Technical Meeting of The Petroleum Society, Banff.
Petro Society of CIM.
5 CONCLUSION AND
RECOMMENDATION
• Sand predictive Material Balance Models have
been proved to be a quick alternative tool to de-
termine OWC movement as reservoir simulation
in the sand analysis.
• Good surveillance acquisition data is needed to
provide input data. The accuracy of each data
needs to be concerned as pre-requisite to make
validate models.
• Sand K has the most significant move of OWC
due to water production maintenance. It reached
91 ft in 48 years of prediction. The other sands
have certain movement below 55 ft.
• Lift tables are needed and also validated to make
well predictive models.

335
ICoSET 2019 - The Second International Conference on Science, Engineering and Technology

APPENDIX
API : American Petroleum Institute
Bo : Current oil volume factor
Boi : Initial oil volume factor
Bg : Current gas volume factor
Bw : Current water volume factor
Cf : Formation compressibility
COWC : Current Oil Water Contact
Cw : Water compressibility
DST : Drill Stem Test
Et : Total expansion of fluid
F : Fahrenheit
FT : Feet
Gi : Cumulative gas injection
Gp : Cumulative gas production
IOIP : Initial Oil in Place
IPM : Integrated Production Modeling
M : Gas oil Ratio
MBAL : Material Balance Modeling
Software
MSTB : Thousand Stock Tank Barrel
MMSTB : Million Stock Tank Barrel
N : Initial Oil in Place
OOIP : Original Oil in Place
OOWC : Original Oil Water Contact
OWC : Oil Water Contact
PBU : Pressure Build-Up Test
ppm : Part per Million
PSIG : Pound Square Inch Gauge
PV : Pore Volume
PVT : Pressure Volume Temperature
RCAL : Routine Core Analysis
RFT : Repeat Formation Test
Rs : Current solution gas oil ratio
Rsi : Initial solution gas oil ratio
SCAL : Special Core Analysis
SCF : Standard Cubic Feet
STB : Stock Tank Barrel
Swc : Connate water saturation
We : Water influx
Wi : Cumulative water injection
Wp : Cumulative water production

336