ISSN: 0374-8588

Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Vishesh Bhadariya*, Fenil Khatri*, Shindlay Nath*, Musa Saleem*

*
School of Chemical Engineering and Physical Sciences, Lovely Professional University,
Phagwara, Punjab, India

Abstract

Reservoir performance prediction is an important aspect of oil and gas field development planning
and reserves estimation that reflects future reservoir behavior. Reservoir production success is
dependent on accurate exemplification of reservoir rock properties, reservoir fluid properties,
rock-fluid properties, and reservoir flow performance. Petroleum engineers must have knowledge
about reservoir characteristics, production operation optimization and, more importantly, to
develop an analytical model that will adequately describe the physical processes occurring in the
reservoir. Reservoir performance prediction based on the material balance equation described by
several authors such as Muskat, Craft and Hawkins, Tarner, Havlena & Odeh, Tracy and
Schilthuis.
This paper compares the estimation of reserves using dynamic simulations in MBAL software and
predictive physical equilibrium method after matching the history of both of these models. The
results of this paper demonstrate the functionality of MBAL in terms of history matching and
performance prediction. The purpose of this paper is to establish the water drive reservoir model
and algorithms for each technique are presented and validated with an example. PVT analysis,
collecting well data for production and quality checks, quality checking based on determining
discrepancies between data and physical reality with the help of correlations. Also, this paper
shows the history of the original oil matching the place and size of the aquifer. The conclusions
obtained from the various plots between different parameters indicate the history matching data,
simulation results, and future performance of the reservoir system and observations of these results
that represent the results of similar simulations and future predicted plots.

Keywords: Performance Prediction, Water Drive Reservoir, MBAL, Reservoir Simulation, History
matching
______________________________________________________________________________

792
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Introduction
The objective of reservoir engineering is to optimize and balance the optimal benefits from a
prospect area. The engineer or engineering team must identify and define all individual reservoirs
and their physical properties, deducing the performance of each reservoir, preventing the drilling
of unnecessary wells, initiating operational control at the appropriate time and all-important
economic factors should be considered. Early and accurate identification and definition of the
reservoir system are essential for well-reasoned engineering. Traditional geologic techniques rarely
provide sufficient data to identify and define individual reservoir; Geological studies should be
complemented with engineering data and tests to provide the necessary information to the engineer.
Reservoir management requires an in-depth knowledge of the reservoir that can only be achieved
through its characterization by the process of acquiring, processing, and integrating many basic
data. [1]
In this paper, a software suite called Integrated Production Modeling (IPM) was developed by
petroleum experts to predict the performance of a hydrocarbon reservoir based on material balance,
while the simulation method of prediction is very complex, requiring a geologic model. It is to
populate the model with rock properties, PVT data, historical production data and all events
occurring in the reservoir. Though the traditional simulation and modeling techniques are explicit,
it requires rigorous exercise for performing material balance on each grid block in the reservoir.
Tarek stated that the material balance equation is an analytical tool for reservoir engineers to
interpret and predict the performance of a given hydrocarbon system. When MBE is properly
applied, it can be used to estimate initial hydrocarbons in place, predict future performance and
recovery of ultimate hydrocarbons for several driving mechanisms. [2] The technical description
is considered a tank model for a given hydrocarbon system with a frame of reference to the material
balance model. These models have certain benefits where they are used to obtain average reservoir
pressure for given quantities of production and water influx from initial quantiles and pressure. [3]
Havlena and Odeh described techniques for interpreting material balance as a straight-line equation
that has the ability to estimate three unknowns, like the original oil, cumulative water flow, gas cap
size compared to the size of the oil field and driving mechanism. This linear solution required to
plot a variant group against another variant group determining the actual drive mechanism. If actual
plots turn out to be non-linear then deviation could itself be selected based on the plot and it is
observed diagnostic to estimating actual drive mechanism. [4] Once the linearity is achieved, the
production and history data plots select the suitable mathematical model to determine the future
reservoir performance.
The initial reservoir model may be found to be inaccurate before the history is matched due to an
unaccounted energy-water influx that helps maintain system pressure. The determination of water
influx requires a mathematical model that relies on aquifer properties. Correct matching with field
history data requires the creation of a correct Aquifer model and to 'spot' it. There are various
models such as Pot-Aquifer, van Everdingen, and Hearst. The Van Everdingen & Hurst are

793
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

unstable state models, based on the superposition principle [5]. There are several models such as
Carter & Tracy [6], Fetkovich [7] and Leung [8], which tried to eliminate the desired computing
power drawback, and therefore became an alternative solution in commercial flow simulators [9].
Methodology
An abstract idea about material balance was demonstrated by Schilthuis in 1941 is based on the
law of conservation of mass. Certain assumptions were made in this technique where the reservoir
was considered to be a homogenous tank model [10]. The material balance is used to match history
and past performance to predict the estimated volume of the reservoir and future performance.
Based on this concept using minimal data, the MBAL software reservoir engineer can use this tool
for reservoir analysis throughout the reservoir lifetime. Basic equation used in MBAL software
i.e.:
F - 𝑾𝒆 = N [𝑬𝒐 + m𝑬𝒈 + E𝒇𝒘 )

Workflow Procedure
The MBAL software in the Analytical Reservoir Engineering Tool Kit is a progressive option to
run simulations based on historical data that must be matched with the analytical model. MBAL
provides adequate matching facilities and aquifer modeling that advances research about pressure
support response and forecasting. Some steps were taken during this research study to forecast
performance for the water drive reservoir.
i. In system options a single tank model is selected and simple pvt model was opted. The
production history of the model is performed by well basis.
ii. PVT, Cumulative oil production and pressure reduction data were recorded into our model
while considering a tank and production model.
iii. Fluid correlation properties were used to fit the pvt laboratory data. These were done with
the various reservoir fluid data and with help of non-linear regression method proper
correlation were selected.
iv. Glaso correlation was selected for saturation pressure, Gas-oil ratio and Formation volume
factor; and beggs correlation selected for viscosity
v. The tank data included for the development of further reservoir models such as initial
pressure, filtrate, reservoir temperature, initial hydrocarbon, water saturation and
production start date data. Initial hydrocarbons in calculations based on geological data.
History matching requires the declaration of previous production data which shows the
declination of pressure.
vi. The analytical method shows the differences between the model and the historical data that
indicate unaccounted energy. There was a possibility of Aquifer's presence based on the
Campbell plot. Three graphical plots were developed that determined the reservoir and
aquifer parameters. The energy system was used to observe the driving mechanism and the

794
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Campbell plot was a diagnostic tool to identify the type of reservoir based on the indication
of pressure and production behaviour.
vii. The non-linear regression method was employed to evaluate the unknown aquifer potential
and the parameters of the reservoir then tuned the pressure and output-related data.
Different aquifer models are chosen and the best-fitted aquifer model is selected.
viii. The accuracy of the model was validated with historical data such as pressure, cumulative
oil and gas production data

Input Parameter
In MBAL model the input data are PVT properties, Oil and gas Production data, aquifer parameter,
relative permeability data, reservoir thickness, porosity and Permeability data. The latter two are
obtained from well logs. Correlation from the software is used to calculate the relative permeability.
A summary of input PVT data, Reservoir Tank data and relative permeability data shown in Table
1, Table 2 and Table 3 respectively. Additional Water influx data which is shown in Table 4
required during the Campbell plot shows variation which shows that an unaccounted energy source
was contributing to the historical production.
Table 1. Summary of input PVT data.
Parameter Input data
Temperature (oF) 160
Bubble point pressure, (Psig) 2725.3
Solution GOR (scf/stb) 650
Oil FVF 𝐵𝑂 @ 𝑃𝑏 (RB/STB) 1.404
Oil viscosity @ 𝑃𝑏 (cp) 0.4
Oil gravity (𝑜 𝐴𝑃𝐼) 40
Gas gravity (Specific gravity) 0.6
Mole percentage 𝐻2 𝑆 0
Water salinity (PPM) 1,40,000
Mole percentage 𝐶𝑂2 0
Mole percentage 𝑁2 0

795
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Table 2 Summary of Reservoir Tank input data

Parameter Input data
Tank fluid Type Oil
Temperature (oF) 160
Initial Tank Pressure (psig) 2725.3
Porosity (fraction) 0.25
Connate Water saturation (fraction) 0.15
Water Compressibility (1/psig) 3*ⅇ−6
Initial Gas-cap 0
Original Oil in Place (MMSTB) 365.949

Table 3. Relative Permeability data

Residual saturation End Point Exponent
𝐾𝑟𝑤 0.15 0.8 2.806
𝐾𝑟𝑜 0.2 0.09 2.33
𝐾𝑟𝑔 0.08 0.95 1.35

Table 4 Water influx data

Parameter Input data

Model Hurst-Van Everdingen-Dake
Reservoir Radius (ft) 9200
Reservoir Thickness (ft) 100
Outer Inner radius ratio 8
Encroachment angle (degree) 360
System Radial Aquifer
Aquifer Permeability (md) 35

Overview of the Reservoir

The pressure of the oil reservoir is at bubble point pressure of 2725.3 psig, oil API gravity was 40
and rock compressibility is 3* 10-6 psi-1.

Case Definition
For described water drive reservoir a single tank and two well model was built. The unaccounted
energy-Aquifer influx was then selected based on the trend and best result. The well production
history data is being input and based on that, the tank production history is calculated. The models

796
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

used to match with the help of Regression method where pressure and Production data are
regressed.
MBAL Model
To build a reservoir model, knowledge of the reservoir structure, aquifer support and gas cap was
necessary. Those were identified based on available data which used to ensure good engineering
judgement.

History Matching

The idea behind history matching was that the model input is adjusted to match the field pressure
and production history data. Procedure should be a way of systematic where adjusting the reservoir
model to agree with the field operation data. Encroachment angle and radius were regressed to
obtain history matches on pressure and production data which ensure a satisfactory pressure match
for the tank.
Aquifer model

In this study, different aquifer models were used in the reservoir model to obtain a quality history
match and their parameters were reconstructed. The best-fit Aquifer model ie. Hurst-Van-
Evendingen-Dake is being used. This was observed when the model and field pressure and
production data were matched together in a graphical representation.

Simulation and Prediction

For history matching of pressure and production performance is best useful in material balance but
has disadvantage when it comes to prediction, which is strongly related to numerical reservoir
simulation modelling. s a good history match was achieved in this case, the MBAL model showed
a good pressure and production match for the tank with historical production and pressure data.
Further research predicts the performance of reservoir tanks.

Results and Discussion

The Energy of the system

Different driving mechanism accounts to drive the reservoir fluid to the wellbore and thus to the
surface, providing enough energy for the system. After running history matching, it was necessary
to select the aquifer and for better history matching we select suitable aquifer. The relative
contribution of the different driving mechanism energy and aquifer system to the recovery from
the reservoir were discovered with certainty. In figure 1, three driving mechanisms with fluid
expansion, pore volume and water influx were presented. Pore volume compressibility can be
ignored due to insufficient involvement to provide energy to the reservoir system. Water flow is
identified as a dominated energy in the system and contributes about 70% of the total energy system

797
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

while the other two contribute the remaining energy. The relative importance of each driving
mechanism was detected in the energy plot.

Figure 1 Drive Mechanism for Reservoir

Analytical Plot

The Analytical plot represent the cumulative oil production as a function of reservoir pressure
decline in Figure 2. From analytical plot it was observed that with the current aquifer model, the
model was predicting the cumulative oil production higher than those observed without considering
water influx initially when there was uncertainty of possible energy system. Plot shows
considerable deviation between history matching data and history matched simulation model result.

798
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Figure 2: Analytical Plot reservoir pressure vs cumulative oil production

Analytical Plot after regression

Regression played a constructive role for eliminating the deviation between the simulated model
and historical behavior in terms of production and pressure data in figure 3.

Figure
3 Analytical plot after regression analysis

799
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Campbell Plot

Basic material balance equation for oil reservoir

F = N𝑬𝒕 + 𝑾𝒆

Rearranging the equation, we got

𝑭 − 𝑾𝒆
=𝑵
𝑬𝒕

Now, if (F - 𝑊𝑒 / 𝐸𝑡 ) vs F is plotted, a horizontal line with Y intercept equal to N should be obtained.

We can clearly see in figure 4 history points deviates from the horizontal, it indicates the model is
not able to predict the response as seen from the reservoir. The input data must be reviewed in this
case.

Figure 4: Campbell Plot with aquifer (F - We / Et ) vs F

Best way to get the straight horizontal line in Campbell plot was to activate regression analysis,
where aquifer parameter water influx along with the Original Oil in Place (OOIP) could be
changed. Figure 5 represent intercept which is equal to OOIP 364.106 MMSTB. Campbell plot
after regression provide straight horizontal line where OOIP closed to the volumetric OOIP after
simulation result.

800
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Figure 5: After Regression Campbell Plot with aquifer (F - We / Et ) vs F

Reservoir Simulation

Reservoir simulation considered as a tool for overall field development planning and used to
perform reverse calculations. Simulation study in figure 6 revealed that if simulated model has
been properly history matched, there should be no variance between predicted reservoir pressure
as a function of time from simulation result and historical measured pressure result. After result in
this case simulated and historical data matched together.

801
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Figure 6: Tank Pressure vs Time Reservoir Simulation Model History Match

Performance Prediction/Forecasting

After acceptable history matched obtained, prediction of cumulative oil production, oil recovery
factor, water production and reservoir pressure decline are carried out.

Figure 7: Performance prediction of reservoir and History match data plotted

Figure 7 shows that from the beginning of the production in 01/03/1987 to continuous decline of
the production in plot till 01/03/2009. As we can clearly see that history model, simulated model

802
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

and performance prediction fall in same line. Further cumulative oil production after 18/07/2004
will follow as an according to performance prediction plot. This will provide the information
regarding water injection requirement to sustain the reservoir pressure and economic limit of the
reservoir.

Figure 8: Oil recovery factor with tank pressure decline versus Time
From the beginning of the production period to end of the production period, 01/03/1990, recovery
factor stood at 5.5%, while model forecasts the recovery factor to increase to 13.10% till
16/05/1998 further increment in recovery factor 20.15% in 01/08/2009.

To performed Forecasts with a history matched model, the total amount of water and gas production
needs to be predicted accurately.

803
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

Figure 9: Water Production along with pressure decline vs time

In figure 9 this was illustrated that the model is ready for prediction as we can see a good agreement
between the data and forecast in plot. Since we observed that reservoir most energy supplied by
water influx only. After certain period this is observed that water production visible after certain
time of oil production. Water cut raised from 11/05/1989 0% to 11.89% in 01/08/2009 of total
production of oil according to future prediction plot.
Conclusion and Future Scope
Results obtained from different graphs the following conclusion are drawn from this research work.

The reservoir engineering material balance equation is an important exploratory tool when time is
limited. It is a very sophisticated analytical tool for evaluating reserve quantities through historical
production. In this study, it has been proved that good data acquisition is required to do a reserve
quantity assessment with MBAL. The reservoir analysis tool MBAL is used to initiate, examine,
and benchmark the history mismatch.

The main source of energy in reservoir was from Water influx and rock and fluid expansion drive
mechanism(negligible). Water system energy provide about 70% of the reservoir energy which is
required for oil recovery. The Hurst-van Everdingen-Dake water influx model best describe the
reservoir.

The cumulative oil produced for the historical period was 63.49 MMSTB and 73.1 MMSTB for
forecast period, with recovery factors of 5.5% and 13.10% respectively. The total amount of
cumulative water production for the historical period was 1.88 MMSTB and 3.86 MMSTB for
forecasts period.

804
 ISSN: 0374-8588
Volume 21 Issue 8s, November 2019

_____________________________________________________________________________________

The material balance was used in reservoirs where sufficient enough data were available for history
matching and performance prediction. Finally, with the help of reservoir simulations, the central
objective of this paper was fulfilled to produce a future prediction that would lead to optimizing
the performance of the reservoir which meant that the reservoir evolved in a way that was highly
relevant to commercial business brings benefits.

References
[1] S. R. Sills, “Improved Material-Balance Regression Analysis for Waterdrive Oil and Gas
Reservoirs,” SPE Reserv. Eng., vol. 11, no. 02, pp. 127–134, 2007.
[2] T. Ahmed, Reservoir Engineering Handbook, vol. 27, no. 7. 2010.
[3] P. D. Mckinney and A. Tarek, Advanced Reservoir Engineering, no. c. 2005.
[4] D. Havlena and A. S. ODEH, “The material balance as an equation of a straight line.” p.
Part 1. Trans. AIME, 228, I-896, 1963.
[5] A. F. Van Everdingen and W. Hurst, “The Application of the Laplace Transformation to
Flow Problems in Reservoirs,” J. Pet. Technol., vol. 1, no. 12, pp. 305–324, 2012.
[6] G. W. Carter, R. D., & Tracy, “Improved Method for Calculating Water Influx. Society of
Petroleum Engineers.” 1960.
[7] M. J. Fetkovich, “Fetkovich Paper,” J. Pet. Technol., vol. 23, no. 07, pp. 814–828, 1971.
[8] W. F. Leung, “A Fast Convolution Method for Implementing Single-Porosity Finite/Infinite
Aquifer Models for Water-Influx Calculations,” SPE Reserv. Eng., vol. 1, no. 05, pp. 490–
510, 2007.
[9] J. Marques and O. Trevisan, “Classic Models of Calculation of Influx: A Comparative
Study,” 2007.
[10] M. Balance et al., “Chapter 3 material balance applied to oil reservoirs 3.1 introduction,”
vol. i, 1941.

805