Research Article

Energy Exploration & Exploitation

2020, Vol. 38(6) 2413–2435
Data-driven modeling to ! The Author(s) 2020
DOI: 10.1177/0144598720927470
optimize the injection well journals.sagepub.com/home/eea

placement for waterflooding

in heterogeneous reservoirs
applying artificial neural
networks and reducing
observation cost

Xinwei Xiong and Kyung Jae Lee

Abstract
Secondary recovery methods such as waterflooding are often applied to depleted reservoirs for
enhancing oil and gas production. Given that a large number of discretized elements are required
in the numerical simulations of heterogeneous reservoirs, it is not feasible to run multiple full-
physics simulations. In this regard, we propose a data-driven modeling approach to efficiently
predict the hydrocarbon production and greatly reduce the computational and observation cost
in such problems. We predict the fluid productions as a function of heterogeneity and injection
well placement by applying artificial neural network with small number of training dataset, which
are obtained with full-physics simulation models. To improve the accuracy of predictions, we
utilize well data at producer and injector to achieve economic and efficient prediction without
requiring any geological information on reservoir. The suggested artificial neural network model-
ing approach only utilizing well data enables the efficient decision making with reduced compu-
tational and observation cost.

Keywords
Data-driven modeling, well placement optimization, waterflooding, heterogeneous reservoirs,
artificial neural network, reducing observation cost

Department of Petroleum Engineering, University of Houston, Houston, USA

Corresponding author:
Kyung Jae Lee, Department of Petroleum Engineering, University of Houston, 5000 Gulf Freeway, Houston, TX 77204,
USA.
Email: kjlee6@central.uh.edu

Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution
4.0 License (https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and
distribution of the work without further permission provided the original work is attributed as specified on the SAGE and
Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
2414 Energy Exploration & Exploitation 38(6)

Introduction
Waterflooding is one of the most widely used secondary recovery methods to improve the oil
and gas production. When reservoir pressure significantly decreases after the primary pro-
duction, external energy is needed to drive the remaining oil to the production well. The
basic idea of waterflooding is to inject water into the reservoirs, to increase the formation
pressure and maintain it around the initial value. Thus, the performance of waterflooding is
controlled by the sweep efficiency, which is closely related to the reservoir heterogeneity.
Prediction of waterflooding performance is complex and challenging in heterogeneous
reservoirs, due to the uncertainty and variability. Uncertainty refers to the insufficiency of
knowledge and information on geological properties; variability refers to the reservoir het-
erogeneity such as spatial differences of porosity and permeability in the reservoir (Siirila
et al., 2012). Because of these challenges, numerical simulations of heterogeneous reservoirs
usually require a large number of discretized grid blocks in simulation models. Thus, it is
expensive and time-consuming to run full-physics simulations of heterogeneous reservoirs
for every possible condition of operation.
To improve the efficiency of the simulation of heterogeneous reservoirs, machine learning
methods have been introduced to significantly decrease the simulation time while reliably
predicting the fluid production. Machine learning is a data analysis method that can auto-
matically learn the relationships of input and output from the sampled training data (Michie
et al., 1994). Statistical learning methods and artificial neural network (ANN) are widely
used machine learning technologies in various scientific and engineering problems.
Statistical learning methods are based on relationship formulation between input and
output datasets by applying mathematical, logical, and statistic methods (James et al., 2013).
There are various statistical learning methods used to build proxy models of reservoir
simulations, such as support vector regressions, which utilize hyperplanes to maximize the
margin of tolerance (Smola and Sch€ olkopf, 2004). On the other hand, ANN models build
the computing systems that imitate the biological neural networks (Jain et al., 1996). These
neural networks can efficiently combine the multiple algorithms and process the complicated
input data. Each neuron contains multiple parameters relating input and output data, and
the prediction results can be improved by optimizing these parameters. As neural networks
become larger, the more complex functions can be processed (LeCun et al., 2015). If there is
a nonlinearity between input and output datasets, and the sizes of input and output datasets
are huge, ANN shows better performance than statistical learning methods (Meyfroidt
et al., 2009).
By their remarkable applicability in datasets with nonlinearity and complexity, ANN
models are actively applied in the fields of heterogeneous reservoir simulation in recent
years. Gaganis and Varotsis (2012) developed a non-iterative method to determine the
phase equilibrium of reservoir simulation using ANN methods. Here, molar fractions of
C1 , nC4 , and nC10 , pressure, and temperature were trained as input data, while equilibrium
coefficients of components were applied as output data. This ANN model greatly reduced
the simulation time of phase equilibrium calculations. Bansal et al. (2013) developed ANN
models to predict the production performance in a discontinuous tight oil reservoir, where
well log data, well location, and completion data were applied as input data; maximum
production rate, average production rate, and decline curve parameters were used as output
data. Guerillot and Bruyelle presented ANN models to forecast the hydrocarbon produc-
tion considering uncertainty in oil production by utilizing geological properties such as
Xiong and Lee 2415

porosity and permeability of reservoir as input data, and reservoir simulation solutions such
as cumulative oil production and oil production rate as output data, respectively. This ANN
model could increase the accuracy of simulation results by minimizing the effect of uncer-
tainty and reduce the simulation time as well (Guerillot and Bruyelle, 2017). BuKhamseen
and Ertekin proposed ANN models to predict the properties of hydraulic fractures in res-
ervoir simulations, where fracture half-length, effective permeability, and productivity index
were utilized as input data, and production data as output data, respectively. This ANN
model predicted the fracture half-length and effective permeability with a good accuracy
(BuKhamseen and Ertekin, 2017). Nwachukwu et al. (2018) proposed a tree-based learning
method to estimate the optimized injection well placement, by utilizing connectivity between
injection and production wells as input data and net present value as output data.
As stated, most of the previous works on data-driven models as proxy models to solve
the complex reservoir problems utilized the geological data of reservoirs as input parame-
ters, such as connectivity between wells, porosity and permeability fields, seismic data,
time of flight, and radius of investigation, which could cause significant monitoring cost
by drilling new wells for data observation. In this research, data-driven models based on
ANN are proposed by utilizing data obtained at production and injection wells only, to
efficiently predict the productivity in heterogeneous reservoirs where waterflooding process-
es are applied.
For the efficient prediction of productivity in heterogeneous reservoirs using ANN
models, production history as an input dataset obtained with relatively short observation
time is favored. However, it is challenging to reliably predict the ultimate productivity
with production history obtained with short observation time. In order to improve the
prediction performance of ANN models still utilizing short observation time of input,
we propose an inclusion of additional informative monitoring data obtained at production
and injection wells.
In this regard, this paper is structured as follows. In the following Methodology section,
a two-dimensional (2D) numerical simulation model with different injector positions,
which has been built to generate training data, is introduced. In the numerical simulation
and data-driven modeling approaches, homogeneous and heterogeneous reservoir models
are presented, where ANN approaches are applied to predict the fluid productions. In the
training of ANN data-driven models, optimization by Levenberg–Marquardt backpropa-
gation algorithm is introduced. In the Results and discussion section, the results of numer-
ical simulation of waterflooding and the performance of ANN models are presented.
Here, the improvement of ANN prediction performance by utilizing informative input
data is discussed, and two criteria to decide the optimized injector position are proposed.
In the Conclusions, summarizing remarks are addressed to provide insights for relevant
applications.

Methodology
Numerical model for reservoir simulation
In this study, reservoir simulation code of TOUGH2 was used to generate training data of
ANN models (Pruess et al., 1999). TOUGH2 is designed for simulating multiphase and
multicomponent fluid flow in porous and fractured media under nonisothermal conditions.
In this study, black oil and water systems (two-phase and two-component conditions) were
2416 Energy Exploration & Exploitation 38(6)

simulated using Equation of State Module 8 in TOUGH2. As the results of TOUGH2

simulations, system behavior including productions of fluid phases was obtained, by solving
the discretized equations of mass balance.
A nine-spot waterflooding 2D simulation model was created as presented in Figure 1.
This model included 51
51 cells, where reservoir length and width were both 1020 m.
Each injector was evaluated to find the best oil production performance, where only one
injector and one producer were open for each case of simulation runs. Reservoir properties
and initial conditions are presented in Table 1.

Reservoir permeability
In order to take account for the uncertainty caused by sparse information of permeability,
multiple equi-probable realizations of permeability fields were considered. Using the per-
meability values at the nine wells as in Table 2, 100 different permeability fields for each
injector case were created; and the total of 800 different cases were simulated (¼8 injector
cases
100 permeability fields). Sequential Gaussian simulation methods were utilized to
create the different permeability fields, by using iTOUGH2 GSLIB (Finsterle and
Kowalsky, 2007). Here, ordinary kriging method was applied to generate the modification

Figure 1. Reservoir geometry of 2D model.

Table 1. Reservoir properties and initial conditions as a synthetic problem.

Reservoir properties Values

Reservoir temperature 60 C

Bulk density of formation 2650 kg/m3
Grid size (Dx, Dy, Dz) 20 m
Porosity 0.2
Water injection rate 0.5520 kg/s (¼300 bbl/day)
Initial pressure 2.07
107 Pa (¼3000 psia)
Bottomhole pressure 2.0
107 Pa (¼2900 psia)
Waterflooding and production period 10 years
Xiong and Lee 2417

coefficient of permeability, which were log-normally distributed. Permeabilities for every

grid block can be calculated by the following equation (Jung et al., 2011)

kðxÞ ¼ kðxÞ
fðxÞ (1)

where f is the modification coefficient; k is the mean value of the permeability distribution.
Figure 2 shows the samples of generated permeability fields of heterogeneous
reservoirs. For the comparison with the heterogeneous cases, 100 homogeneous fields
were also considered. Only injectors 1 and 2 were considered in the simulations of homo-
geneous cases, by the principle of symmetry. The range of permeabilities in the homoge-
neous cases was from 1014 m2 to 1011 m2 (¼10 mD–10 D), which were evenly divided into
100 intervals in log-scale.

Artificial neural network

In order to build proxy models to predict the oil and water productions of stated
reservoir simulation problems, ANN models were developed. ANNs are composed
with a number of neurons, where every neuron includes three parameters: weights (wÞ,
biases ðbÞ, and activation functions ( f Þ (Demuth and Beale, 1993). These parameters can
optimize the input signals received from the previous neural layer and send the optimized
signals to the next neural layer as an output signal. Figure 3 shows the structure of ANN
models used in this study (Lee, 2019). We used two-layer feed-forward network composed
with one hidden layer and one output layer. The numbers of neurons in the hidden and
output layers were chosen to avoid the overfitting problem, by applying optimized sugges-
tions (Heaton, 2008).
The relationships of the three parameters in each neuron are described as follows

X
n
vi ¼ bi þ wi xi (2)
i

yi ¼ fðvi Þ (3)

Table 2. Permeability at injection and production wells in a synthetic

reservoir model.

Well Permeability

Injector 1 2.000
10–11 m2 (¼20 D)

Injector 2 5.023
10–12 m2 (¼5090 mD)
Injector 3 2.000
10–12 m2 (¼2 D)
Injector 4 3.169
10–12 m2 (¼321 mD)
Injector 5 1.262
10–12 m2 (¼1279 mD)
Injector 6 3.990
10–13 m2 (¼404 mD)
Injector 7 1.002
10–12 m2 (¼1016 mD)
Injector 8 7.962
10–13 m2 (¼807 mD)
Producer 6.324
10–13 m2 (¼641 mD)
2418 Energy Exploration & Exploitation 38(6)

Figure 2. Permeability fields in heterogeneous reservoirs (log k(m2)). (a) Field 11, (b) field 21, (c) field 31,
(d) field 41, (e) field 51, (f) field 61, (g) field 71, (h) field 81.
Xiong and Lee 2419

Figure 3. Structure of ANN models (Lee, 2019).

where bi , wi , and f are the bias at neuron i, weight at neuron i, and the activation function,
respectively (Cybenko,1989).
Activation functions need to be selected carefully, because the failure to choose a proper
activation function may lead to improper optimization of input data. In our ANN models,
sigmoid function and linear function were utilized in hidden layer and output layer, respec-
tively, considering both stability and efficiency (Agostinelli et al., 2015). Sigmoid curve is the
special type of logistic function given by the following equation (Demuth and Beale, 1993)

1
fðxÞ ¼ (4)
1 þ ex

In our ANN models, input datasets included the cumulative oil and water productions
obtained from 0 to 1500 days, with the observation frequency of 1 day. Output datasets
included the cumulative oil and water productions after 10 years of waterflooding and pro-
duction. MATLAB Neural Network ToolboxTM was used to build ANN models in this
study (Demuth and Beale, 1993). Here, the datasets were divided into three different cate-
gories—training, validation, and test data. In the training stage, the neural network was
adjusted with respect to the evolution of training error. In the validation stage, the perfor-
mance and improvement of neural network were checked to determine when the training
procedure needed to be stopped. In the test stage, there was no alteration in the neural
networks, and only the network performance was measured. In this study, 70%, 15%, and
15% of data were used for training, validation, and test, respectively.

Levenberg–Marquardt backpropagation algorithm

In the ANN modeling with backpropagation for training the datasets, optimization algo-
rithms need to be applied to compute optimal values of weights and biases of neurons.
In this study, Levenberg–Marquardt backpropagation algorithm was applied to predict the
output results. Levenberg–Marquardt backpropagation algorithm has been widely used to
provide numerical solution of a non-linear function (More, 1978). Levenberg–Marquardt
2420 Energy Exploration & Exploitation 38(6)

backpropagation algorithm combines the Steepest descent algorithm and Gauss–Newton

algorithm (Hagan and Menhaj, 1994).
Steepest descent algorithm provides stable prediction results, but takes long time to con-
verge because of the small step sizes (Widrow and McCool, 1976). On the other hand, Gauss–
Newton algorithm takes second-order derivatives of error functions, automatically finds the
proper step size, and converges rapidly (Wedderburn, 1974). However, Gauss–Newton algo-
rithm can be applied only when the error functions are approximated as quadratic functions.
In this regard, Levenberg–Marquardt backpropagation algorithm has been developed to auto-
matically analyze the error function and selects more suitable algorithm for ANN models.
Here, Steepest descent algorithm is applied when complex curvatures exist; Gauss–Newton
algorithm is utilized when the quadratic approximations of error functions are reasonable.

Results and discussion

Production behavior of homogeneous and heterogeneous reservoirs
Simulation results of homogeneous reservoir cases are provided in Figures 4 and 5.
Cumulative oil and water productions increased with increasing average permeability.
Injector 2 cases showed shorter water breakthrough times than injector 1 cases, because

Figure 4. Production rates of oil and water in homogeneous reservoirs utilizing each injector. (a) Oil
production rate in injector 1 cases, (b) oil production rate in injector 2 cases, (c) water production rate in
injector 1 cases, (d) water production rate in injector 2 cases.
Xiong and Lee 2421

the distance between injector and producer was shorter in the injector 2 cases. Injector 1
cases showed higher cumulative oil production, by the higher sweep efficiency. From the
production rates and cumulative productions of oil and water, both reservoir permeability
and the distance between injector and producer were influential for production behavior.
This implies that it will be critical to include the information on the reservoir permeability
(or geological property of reservoir affecting fluid flow) and the injector responses as an
input dataset for the reliable prediction by ANN models.
Simulation results of heterogeneous reservoir cases are provided in Figures 6 to 9.
Irregular patterns of oil and water productions and water breakthrough were observed by
the heterogeneity-induced complex connectivity between injector and producer. From the
production behavior, it is expected that the inclusion of informative input dataset on the
reservoir connectivity affecting sweep efficiency in ANN modeling will be more critical in
heterogeneous reservoir cases than homogeneous reservoir cases.

ANN prediction performance using only production history as an input dataset

As mentioned in the Introduction, the objective of this study is to successfully predict the
long-term performance of production by using the input data obtained at wells only during
short monitoring time. In this regard, the observation time of 1500 days with

Figure 5. Cumulative productions of oil and water in homogeneous reservoirs utilizing each injector. (a)
Oil production rate in injector 1 cases, (b) oil production rate in injector 2 cases, (c) water production rate
in injector 1 cases, (d) water production rate in injector 2 cases.
2422 Energy Exploration & Exploitation 38(6)

Figure 6. Oil production rates in heterogeneous reservoirs utilizing each injector. (a) Injector 1 cases,
(b) injector 3 cases, (c) injector 6 cases, (d) injector 8 cases.

1 day-observation frequency was applied to predict the cumulative oil and water produc-
tions after 10 years of waterflooding processes.
ANN prediction results of the cumulative oil and water productions after waterflooding
processes, obtained with only production history as an input dataset, are shown in
Figures 10 (homogeneous reservoirs) and 11 (heterogeneous reservoirs). Injector 2 cases
of homogeneous reservoirs and injector 1 cases of heterogeneous reservoirs are presented
representatively. Both in homogeneous and heterogeneous reservoirs, ANN prediction per-
formances were not satisfactory. It indicates that the inclusion of input datasets implying the
information of influential factors on production behavior will be necessary to improve the
ANN prediction performances.

Addition of input datasets to improve ANN prediction performances

In order to improve the prediction performance of ANN models by still utilizing the short
observation time, highly informative data needed to be included as an input of ANN
models. We selected the five additional data as follows.

1. Ratio of oil–water production rates

qo
R¼ (5)
qw
Xiong and Lee 2423

Figure 7. Water production rates in heterogeneous reservoirs utilizing each injector. (a) Injector 1 cases,
(b) injector 3 cases, (c) injector 6 cases, (d) injector 8 cases.

where R is the ratio of oil–water production rates; qo is the oil production rate at any time
during observation; qw is the water production rate at the same time to qo measurement.
2. Ratio of cumulative oil–water productions

Qo
RC ¼ (6)
Qw

where Rc is the ratio of cumulative oil–water productions; Qo is the cumulative oil produc-
tion at any time during observation; Qw is the cumulative water production at the same time
to Qo measurement.
3. Injectivity

qinj
I¼ (7)
DP

where I is the injectivity; qinj is the water injection rate; DP is the difference of injector
pressure between any time during observation and the initial time.
2424 Energy Exploration & Exploitation 38(6)

Figure 8. Cumulative oil productions in heterogeneous reservoirs utilizing each injector. (a) Injector 1
cases, (b) injector 3 cases, (c) injector 6 cases, (d) injector 8 cases.

4. Pressure transient data at injector

DP ¼ Pinj;i  Pinj;0 (8)

@P Pinj;i  Pinj;i1
DP0 ¼ ¼ (9)
@ðlnðtÞÞ lnðti Þ  lnðti1 Þ

where DP is the pressure difference; DP0 is the pressure difference derivative; Pinj;i is the
injector pressure at time i; Pinj;0 is the injector pressure at initial time; Pinj;i1 is the injector
pressure at time i  1 (Lee et al., 2003).
Utilization of oil–water ratios as an additional input dataset was proposed, because it was
the direct indicator of relative productivity of oil and water. Utilization of pressure transient
data at injector as an additional input dataset was proposed, because it included the infor-
mation of reservoir connectivity (Lee et al., 2003). Pressure transient data at injector gives
the information of permeability around the injector, which significantly affects the sweep
efficiency of waterflooding. By applying the dimensionless variables of radial distance from
Xiong and Lee 2425

Figure 9. Cumulative water productions in heterogeneous reservoirs utilizing each injector. (a) Injector 1
cases, (b) injector 3 cases, (c) injector 6 cases, (d) injector 8 cases.

Figure 10. Prediction performances of ANN models in homogeneous reservoirs with injector 2.
(a) Cumulative oil productions at 10 years, (b) cumulative water productions at 10 years.
2426 Energy Exploration & Exploitation 38(6)

Figure 11. Prediction performances of ANN models in heterogeneous reservoirs with injector 1.
(a) Cumulative oil productions at 10 years, (b) cumulative water productions at 10 years.

Figure 12. Improved prediction performances of ANN models in homogeneous reservoirs utilizing each
injector, after applying additional input datasets of injector. (a) Cumulative oil productions at 10 years
(injector 1 cases), (b) cumulative water productions at 10 years (injector 1 cases), (c) cumulative oil
productions at 10 years (injector 2 cases), (d) cumulative water productions at 10 years (injector 2 cases).
Xiong and Lee 2427

wellbore, time, and pressure, as defined in following equations (10) to (12), we have a
dimensionless pressure solution as shown in equation (13) (Lee et al., 2003)

r
rD ¼ (10)
rw

kt
tD ¼ (11)
1lCt r2w

2pkh
PD ¼ ðP  Pi Þ (12)
ql
 
1 4tD
PD ðrD ; tD Þ ¼ ln c 2 (13)
2 e rD

Figure 13. Error histograms of ANN models in homogeneous reservoirs utilizing each injector, after
applying additional input datasets of injector. (a) Cumulative oil productions at 10 years (injector 1 cases),
(b) cumulative water productions at 10 years (injector 1 cases), (c) cumulative oil productions at 10 years
(injector 2 cases), (d) cumulative water productions at 10 years (injector 2 cases).
2428 Energy Exploration & Exploitation 38(6)

where c is the Euler’s constant (¼0.577216); rD , tD , and PD are the dimensionless variables
for radial distance from wellbore, time, and pressure, respectively; rw is the wellbore radius;
k is the reservoir permeability; h is the reservoir thickness; q is the injection rate; and l is the
fluid viscosity. By combining equations (9), (12), and (13), we can obtain the following
relationship for reservoir permeability and pressure difference derivative, DP0

q
k/ (14)
4phDP0

As can be seen from equation (14), the inclusion of pressure difference derivative data
(DP0 ) will contain the information of permeability around the injector.
The improved prediction performances of ANN models applying these additional input
datasets are provided in Figures 12 to 17. Most of the data points fell on the dashed line in
the homogeneous reservoir cases, which indicated the good performance of ANN predic-
tions. In the heterogeneous reservoir cases, the additional input datasets made a significant

Figure 14. Improved prediction performances of ANN models in heterogeneous reservoirs utilizing each
injector, after applying additional input datasets of injector. (a) Cumulative oil productions at 10 years
(injector 1 cases), (b) cumulative oil productions at 10 years (injector 3 cases), (c) cumulative oil productions
at 10 years (injector 6 cases), (d) cumulative oil productions at 10 years (injector 8 cases).
Xiong and Lee 2429

contribution of improving prediction performances as well. In both problems, error histo-

grams show that the ANN models were built well, with the most frequent errors concen-
trated around zero.

Selection of optimized injection well placement

In order to design the optimized waterflooding operations in heterogeneous reservoirs,
injector well placement with high cumulative oil production and relatively low cumulative
water production was considered. In our selection, two criteria were utilized—P90 and P50.
P90 was the cumulative oil production which was higher than 90% of whole cases; P50 was
the cumulative oil production which was higher than 50% of whole cases. Figure 18 indi-
cates the P90 and P50 lines of cumulative productions at injector 1. The cases of other
injectors applied the same method to select the P90 and P50 productions.
Tables 3 and 4 present the P90 and P50 results from ANN models and actual simulation
results, respectively. To select the optimized injection well placement, we first consider the

Figure 15. Error histograms of ANN models in heterogeneous reservoirs utilizing each injector, after
applying additional input datasets of injector. (a) Cumulative oil productions at 10 years (injector 1 cases), (b)
cumulative oil productions at 10 years (injector 3 cases), (c) cumulative oil productions at 10 years (injector
6 cases), (d) cumulative oil productions at 10 years (injector 8 cases).
2430 Energy Exploration & Exploitation 38(6)

Figure 16. Prediction performances of ANN models in heterogeneous reservoirs utilizing each injector,
after applying additional input datasets of injector. (a) Cumulative water productions at 10 years (injector 1
cases), (b) cumulative water productions at 10 years (injector 3 cases), (c) cumulative water productions at
10 years (injector 6 cases), (d) cumulative water productions at 10 years (injector 8 cases).

highest cumulative oil production. However, if multiple injection well placements show very
close cumulative oil production results, we selected the case with the least cumulative water
production as the optimized injection well placement. More specifically, we considered the
cumulative water production as additional criteria for the selection of optimized injection
well placement, if the differences of cumulative oil productions in different cases were
smaller than 0:15
107 kg.
As can be seen from the tables, injector 6 and injector 3 were the optimized injector well
placement from ANN models using P90 and P50 criteria, respectively. Since we aim at
higher cumulative oil production and lower cumulative water production, injector 8 and
injector 3 were the optimized injector well placement from actual simulation results using
P90 and P50 criteria, respectively. By using the P90 criteria, the optimized injection well
placement was different in the ANN models and actual simulation results, which was caused
by the very close cumulative oil productions in the cases with injectors 6 and 8. By consid-
ering P50 criteria, the optimized injection well placement by the ANN models showed
conformity with the actual simulation results.
Xiong and Lee 2431

Figure 17. Error histograms of ANN models in heterogeneous reservoirs utilizing each injector, after
applying additional input datasets of injector. (a) Cumulative water productions at 10 years (injector 1 cases),
(b) cumulative water productions at 10 years (injector 3 cases), (c) cumulative water productions at 10 years
(injector 6 cases), (d) cumulative water productions at 10 years (injector 8 cases).

Figure 18. Selection of P90 line and P50 line at injector 1 using simulation results.
2432 Energy Exploration & Exploitation 38(6)

Table 3. P90 and P50 for each injector case using ANN results.
Mass (kg)a Inj 1 Inj 2 Inj 3 Inj 4 Inj 5 Inj 6 Inj 7 Inj 8

P90 oil 6.62
107 5.67
107 6.62
107 5.55
107 5.54
107 6.91
107 5.33
107 6.24
107
P90 water 5.45
107 7.75
107 6.90
107 8.10
107 9.42
107 7.60
107 9.77
107 8.51
107
P50 oil 5.01
107 4.21
107 5.25
107 4.02
107 3.84
107 5.38
107 4.17
107 4.99
107
P50 water 1.02
108 1.08
108 8.25
107 1.33
108 9.44
107 1.37
108 1.14
108 6.31
107

1 STB oil ¼ 135.13 kg oil, 1 STB water ¼ 158.98 kg water.

a

Table 4. P90 and P50 for each injector case using simulation results.
Mass (kg)a Inj 1 Inj 2 Inj 3 Inj 4 Inj 5 Inj 6 Inj 7 Inj 8

P90 oil 6.55
107 5.50
107 6.54
107 5.53
107 5.63
107 6.58
107 5.35
107 6.71
107
P90 water 7.03
107 9.15
107 8.59
107 1.04
108 9.34
107 1.00
108 9.77
107 8.74
107
P50 oil 4.96
107 4.11
107 5.11
107 4.11
107 3.82
107 4.98
107 4.10
107 5.12
107
P50 water 9.25
107 1.21
108 6.82
107 6.38
107 7.86
107 8.51
107 9.75
107 1.10
108

1 STB oil ¼ 135.13 kg oil, 1 STB water ¼ 158.98 kg water.

a

Conclusions
In this study, data-driven ANN models efficiently forecasting waterflooding production
performance in heterogeneous reservoirs have been developed, which successfully proposed
the optimized injector well placement in accordance with the actual simulation results.
Application of additional input data measured at the injector, which implied geological
information and sweep efficiency of reservoirs, made significant contributions to improve
the performance of ANN predictions. This suggested data selection method for the ANN
modeling enabled the efficient prediction of production performance with relatively short
observation time. The optimized ANN modeling approach is expected to contribute to
reduce the computational and monitoring cost, because the data only obtainable at already
drilled wells, such as production history and injector pressure transient data, were utilized as
an input of ANN models. This is a noticeable improvement from the relevant previous
works, where the majority of them required the geological properties at undrilled positions
as an input of data-driven models with significant cost for geological survey.
The envisioned work will include the application of suggested data-driven modeling
approach to actual field data, where different strategy needs to be taken. In the actual
field data application, there is no future production data, which would be used as output
data in the training of ANN models. In order to overcome the absence of future production
data, the production history data until current time can be divided into input data (early
time period production among measured data) and output data (late time period production
among measured data) for the training of ANN models, which will be used to predict the
future production. As soon as new production data are obtained, ANN models will be
updated by including newly measured data. As stated, even though the time periods
of input and output data for ANN training are different, the scientific methodology of
data-driven modeling is the same in the cases applying simulation-generated data and
actual field data.
Xiong and Lee 2433

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or
publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or
publication of this article: The author appreciates the financial support for this research from the
National Research University Fund of University of Houston.

ORCID iD
Kyung Jae Lee https://orcid.org/0000-0002-5631-9809

References
Agostinelli F, Hoffman M, Sadowski P, et al. (2015) Learning activation functions to improve deep
neural networks. arXiv preprint arXiv:1412.6830. Available at: https://arxiv.org/pdf/1412.6830.pdf
Bansal Y, Ertekin T, Karpyn Z, et al. (2013) Forecasting well performance in a discontinuous tight oil
reservoir using artificial neural networks. In: SPE unconventional resources conference-USA.
Woodlands, TX: Society of Petroleum Engineers.
BuKhamseen NY and Ertekin T (2017) Validating hydraulic fracturing properties in reservoir simu-
lation using artificial neural networks. In: SPE Kingdom of Saudi Arabia annual technical sympo-
sium and exhibition. Saudi Arabia: Society of Petroleum Engineers.
Cybenko G (1989) Approximation by superpositions of a sigmoidal function. Mathematics of Control,
Signals, and Systems 2(4): 303–314.
Demuth H and Beale M (1993) Neural Network Toolbox For Use with Matlab – User’s Guide. Version
3.0. Natick, MA: The MathWorks Inc.
Finsterle S and Kowalsky MB (2007) iTOUGH2-GSLIB User’s Guide. Berkeley, CA: University of
California.
Gaganis V and Varotsis N (2012) Machine learning methods to speed up compositional reservoir
simulation. In: SPE Europec/EAGE annual conference. Denmark: Society of Petroleum Engineers.
Guerillot D and Bruyelle J (2017) Uncertainty assessment in production forecast with an optimal
artificial neural network. SPE Middle East oil & gas show and conference. Kingdom of Bahrain:
Society of Petroleum Engineers.
Hagan MT and Menhaj MB (1994) Training feedforward networks with the Marquardt algorithm.
IEEE Transactions on Neural Networks 5(6): 989–993.
Heaton J (2008) Introduction to Neural Networks with Java. Washington: Heaton Research, Inc.
Jain AK, Mao J and Mohiuddin KM (1996) Artificial neural networks: A tutorial. Computer 29(3):
31–44.
James G, Witten D, Hastie T, et al. (2013) An Introduction to Statistical Learning. New York: Springer.
Jung Y, Imhoff P and Finsterle S (2011) Estimation of landfill gas generation rate and gas permeability
field of refuse using inverse modeling. Transport in Porous Media 90(1): 41–58.
LeCun Y, Bengio Y and Hinton G (2015) Deep learning. Nature 521(7553): 436–444.
Lee J, Rollins JB and Spivey JP (2003) Pressure Transient Testing. Richardson, TX: Society of
Petroleum Engineers.
Lee KJ (2019) Characterization of type and maturity of organic matter in source rock by in-situ
electrical heating and temperature transient analysis. In: SPE international conference on oilfield
chemistry. Galveston, TX: Society of Petroleum Engineers.
Meyfroidt G, Güiza F, Ramon J, et al. (2009) Machine learning techniques to examine large patient
databases. Best Practice & Research Clinical Anaesthesiology 23(1): 127–143.
2434 Energy Exploration & Exploitation 38(6)

Michie D, Spiegelhalter DJ and Taylor C (1994) Machine learning. Neural and Statistical
Classification 13.
More JJ (1978) The Levenberg-Marquardt algorithm: Implementation and theory. In: Watson GA
(ed.) Numerical Analysis. New York: Springer, pp.105–116.
Nwachukwu A, Jeong H, Pyrcz M, et al. (2018) Fast evaluation of well placements in heterogeneous
reservoir models using machine learning. Journal of Petroleum Science and Engineering 163:
463–475.
Pruess K, Oldenburg CM and Moridis G (1999) TOUGH2 User’s Guide. Version 2. Berkeley CA:
Lawrence Berkeley National.
Siirila ER, Navarre-Sitchler AK, Maxwell RM, et al. (2012) A quantitative methodology to assess the
risks to human health from CO2 leakage into groundwater. Advances in Water Resources 36:
146–164.
Smola AJ and Sch€ olkopf B (2004) A tutorial on support vector regression. Statistics and Computing
14(3): 199–222.
Wedderburn RW (1974) Quasi-likelihood functions, generalized linear models, and the Gauss–Newton
method. Biometrika 61(3): 439–447.
Widrow B and McCool J (1976) A comparison of adaptive algorithms based on the methods of
steepest descent and random search. IEEE Transactions on Antennas and Propagation 24(5):
615–637.

Appendix
Notation
bi bias value at neuron i, dimensionless
ct total compressibility, Pa1
h reservoir thickness, m
I injectivity, kg/Pas
k absolute permeability, m2
P reservoir pressure, Pa
PD dimensionless reservoir pressure, dimensionless
Pinj;0 injector pressure at initial time, Pa
Pinj;i1 injector pressure at time i  1, Pa
Pinj;i injector pressure at time i, Pa
qinj water injection rate, kg/s
qo oil production rate, kg/s
qw water production rate, kg/s
Qo cumulative oil production, kg
Qw cumulative water production, kg
rD dimensionless distance from wellbore, dimensionless
rw wellbore radius, m
R ratio of oil-water production rates, dimensionless
Rc ratio of cumulative oil-water productions, dimensionless
t time, s
tD dimensionless time, dimensionless
vi transformed input for activation function, dimensionless
wi weight value at neuron i, dimensionless
xi input at neuron i
yi output node of layer i
Xiong and Lee 2435

DP pressure difference, Pa
DP0 pressure difference derivative, Pa/ln(s)
/ reservoir porosity, dimensionless
c Euler’s constant, dimensionless
l reservoir fluid viscosity, Pas
f modification coefficient of permeability, dimensionless