SPE-175883-MS

Application of Using Fuzzy Logic as an Artificial Intelligence Technique in

the Screening Criteria of the EOR Technologies
Mohamed Nageh, Gpc; Mahmoud Abu El Ela, El Sayed El Tayeb, and Helmy Sayyouh, Cairo University

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE North Africa Technical Conference and Exhibition held in Cairo, Egypt, 14 –16 September 2015.

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
Screening criteria stage for the EOR applications is useful for many candidate reservoirs before expensive
reservoir descriptions and economic evaluations are done. This paper presents the application of using
fuzzy logic as an artificial intelligence technique in the screening criteria of the EOR technologies. EOR
screening criteria have been developed based on field results. The database of 347 successful EOR
projects worldwide is used to carry out statistical analysis which resulted in determination of four critical
values (minimum, maximum, r1, r2) for each fluid and rock property.
The determined values identify the suitable range of the rock and fluid properties and the screening
criteria for each EOR method. In addition, these ranges can rate the properties of any field under study
from 0 to 1 (through fuzzy logic membership functions). The minimum and the maximum values represent
the successful boundaries of the fluid or rock properties for the studied EOR method; while r1 and r2
values represent the ranges with maximum rating (best conditions) for the fluid or the rock properties.
Not all fields are amenable to EOR processes. Effective screening practices must be employed to
identify suitable candidates. In the developed screening tool, there are 15 inputs representing EOR
screening criteria of the reservoir rock and fluid properties including: API gravity, oil viscosity, reservoir
depth, rock permeability, oil saturation, reservoir temperature, formation water salinity, net pay thickness,
permeability variation coefficient (representing heterogeneity index), formation type, minimum miscibil-
ity pressure, initial reservoir pressure, current reservoir pressure, fracture pressure and an option showing
if the reservoir is dip or not.
The developed screening tool is characterized by an easily and friendly interface with additional
options comparing to the other existing software or expert systems. The proposed tool can be used to
support the decision making during the critical technology selection phase. Such study is an original
contribution to achieve successful EOR applications.
Introduction
Planning of Enhanced Oil Recovery methods is a complex problem that requires an integrated approach
for its solution. Although EOR is specific for specific reservoir, but EOR screening could be used as a
guide or a first step in the implementation of a new EOR project. Prospects that pass this screening are
candidates for further engineering and economic study. Most of early studies in the EOR selection were
2 SPE-175883-MS

to establish the technical screening criteria of each EOR method (El-Batanony et al., 1987; Sayyouh et.
al., 1990 and 1993; Shokir et al., 2002; Goda, et al. 2002; Lee et al., 2011). Based on laboratory
experiments and field experiences, the applicable ranges of the reservoir rock and fluid properties were
presented in these studies.
The large number of EOR projects (which have been utilized in many oil fields around the world) has
led to an increased understanding of the selection and applicability of certain oil recovery methods, and
that has ensured degree of technical and economical success. Aladasani proposed a new EOR selection
criterion after a thorough analysis of EOR projects reported from 1998 through 2008 (Aladasani and Bai,
2010). This criterion was based on dataset distribution, by detecting the ranges within which the majority
of the EOR projects fall. That will lead to more accurate results than using the average values of the
reservoir and oil properties as a guide in EOR screening. The work in this paper used the new
methodology of Aladasani to generate the fuzzy logic parameters required for the new developed
screening tool.
Model Structure
The developed model used Fuzzy Logic (Fuzzy Inference System – FIS) as an Artificial Intelligence
Technique in the EOR screening. The FIS of the developed model consists of three major sections as
shown in Figure 1 (Inputs Section, Rule Editor Section, and Outputs Section).

Figure 1—Fuzzy inference system editor

The first section is the inputs section. In the developed screening tool, there are 15 inputs representing
the EOR screening criteria of the reservoir rock and fluid properties. These inputs are: API gravity, oil
SPE-175883-MS 3

viscosity, reservoir depth, rock permeability, oil saturation, reservoir temperature, formation water
salinity, net pay thickness, permeability variation coefficient (representing heterogeneity index), forma-
tion type, minimum miscibility pressure, initial reservoir pressure, current reservoir pressure, fracture
pressure and an option showing if the reservoir is dip or not. Each input is represented by Membership
Function (MF) which reflects its effect on each EOR method. Each input could be evaluated through its
membership function. A membership function (associated with a given fuzzy set) maps an input value to
its appropriate membership value. This section in the fuzzy inference system allows the user to rate each
input and output through the membership functions and it is the main idea in this work. Through these
membership functions, each input could be rated from 0 to 1 based on the relation between the exact value
of the rock or fluid property and the main values of the membership function (feet and shoulder of the
membership function). There are different types of membership functions each type is characterized by
its shape and its parameters.
Four types of MF are used in the new developed screening tool; Trimf, Trapmf, Smf, and Zmf. Trimf
refers to Triangular Shaped MF. Figure 2 shows the typical shape of the Trimf. In this work this MF is
used in the inputs with only one optimum value. The inputs ⬙Formation Type⬙ and ⬙Permeability
Variation⬙ are examples of this MF. This MF is used also for the inputs which its values extracted from
literature (Aladasani and Bai, 2010; Taber, Martin and Seright, 1997), as the values of these inputs are:
minimum value, maximum value and one average value. Trapmf refers to Trapezoidal shaped MF. Figure
3 shows the typical shape of the Trapmf. It is the most common MF in this work. The parameters ⬙a⬙ and
⬙d⬙ locate the ⬙feet⬙ of the trapezoid with least rating. The parameters ⬙b⬙ and ⬙c⬙ locate the ⬙shoulders⬙
with maximum rating; they represent the range through which the maximum number of EOR projects had
been applied (best conditions). The inputs API, viscosity, depth, permeability, oil saturation and temper-
ature are examples of using Trapmf. Figure 4 and Figure 5 show the typical shapes of both Smf and Zmf.
Smf refers to the S Shaped MF. It could be used when all values above certain limit result in a rate equal
to 1. The input ⬙Net Pay Thickness⬙ is example of this MF. Zmf refers to the Z shaped MF. It is opposite
to Smf, that all values below certain limit result in best rate (rate equal to 1). The input “Salinity” is
example of this MF.

Figure 2—Triangular shaped membership function (Trimf)

4 SPE-175883-MS

Figure 3—Trapezoidal shaped membership function (Trapmf)

Figure 4 —S shaped membership function (Smf)

Figure 5—Z shaped membership function (Zmf)

The second section is the Rule Editor section. It controls the FIS by rules. These rules link between the
inputs (the rock and fluid properties of the reservoir under study), and the outputs (the EOR methods).
There are 24 rules in this FIS. These rules form conditions through mathematical relations between the
different inputs which affect the outputs. Graphical User Interface (GUI) is constructed by use of “Matlab,
version 2008” to facilitate the process of entering the inputs and reading the outputs, to make the
developed tool easier and friendly user. Figure 6 shows the final interface of the new developed screening
tool.
SPE-175883-MS 5

Figure 6 —Final Interface of the developed screening tool

The third section is the output section. The model outputs are represented in 14 EOR methods as
follows: CO2 miscible flooding, hydrocarbon miscible flooding, WAG (Water Alternative Gas) miscible
flooding, N2 miscible flooding, N2 immiscible flooding, CO2 immiscible flooding, hydrocarbon immis-
cible flooding, immiscible flooding of hydrocarbon in addition to WAG, polymer flooding, ASP
(Alkaline-Surfactant-Polymer) flooding, insitu combustion, steam injection, hot water injection, and
microbial enhanced oil recovery. Because of the sensitivity of each EOR method toward specific and
critical parameters, the weights of the inputs in the evaluation are not the same in all EOR methods (for
example the depth is critical for steam injection). Those critical parameters for each EOR method have
a greater weight than other parameters (not critical parameters). The process of adding weight for each
critical input at the corresponding EOR method is implemented through the rule editor section.

Approach and Methodology

This section shows the approach and the methodology used to generate the fuzzy logic parameters of
each EOR method in the developed screening tool. Database comprising of rock and fluid properties for
347 EOR projects is built to carry out statistical analysis. The dataset was collected from the 2004, 2010,
2012, and 2014 Worldwide EOR Survey of the Oil & Gas Journal (Moritis, 2004; Moritis, 2010;
Koottungal, 2012; Koottungal, 2014). The EOR projects in this database are classified into seven
categories, namely, steam injection, insitu combustion, CO2 miscible flooding, CO2 immiscible flooding,
hydrocarbon miscible flooding, chemical EOR methods, and others.
The data related to each EOR method were screened to remove the inconsistent data and the data of the
unsuccessful projects. Then, the data of the successful projects related to each EOR method were used to
generate the fuzzy logic parameters.
The EOR screening parameters in the developed screening tool are very sensitive to the results of these
analyses, so checking the quality of the data is very important to avoid misleading results. The problems
6 SPE-175883-MS

encountered in the EOR survey data are: missing data, inconsistent data, data related to unsuccessful and
discouraging projects, and finally data related to the new projects that can’t be evaluated at the current
time.
The next step of the analysis is representing the distribution of the reservoir and fluid properties for each
EOR method. Then, the minimum and the maximum values for each property against each EOR method
are identified. Also, the property range (through which the maximum number of the successful EOR
projects had been applied – r1&r2) for each EOR method is identified. Finally, the results of the current
work are compared to the results of Aladasani (Aladasani and Bai, 2010) and Taber (Taber, Martin and
Seright, 1997) to be validated. Aladasani summarized his results by presenting only r1 and r2 values.
Taber summarized his results by presenting three values: minimum value, maximum value and one
average value. Taber work was updated by Aladasani in 2010 (Aladasani and Bai, 2010) to include the
data of the EOR survey reports submitted from 1998 through 2008, and to include also the new categories
and sub categories of EOR methods.
The Fuzzy Logic parameters for the EOR methods which either has no data or its available data are not
sufficient are extracted from literature (Taber, Martin and Seright, 1997; Aladasani and Bai, 2010).
Table 1 shows the summary of the fuzzy logic parameters for all of the EOR methods used in the new
developed screening tool. The number which is presented in brackets (red colored) is the number of the
successful EOR projects used in the analyses of the current work. However, the number without brackets
is the number of the projects which are used from the literature (Updated Taber Screening Criteria by
Aladasani). The same for the values of the parameters: the values between brackets represent the values
which are generated from the current work, while the values of the parameters which are presented
without brackets represent the values which were collected from the literature (Taber and Updated Taber
Screening Criteria).
SPE-175883-MS 7

Table 1—Summary of fuzzy logic parameters used in the new developed screening tool

Model Verification and Validation

To test the reliability and consistency of the new developed screening tool, the results of the tool is
compared with the results of commercial EOR software (EORgui). The properties of some of the Egyptian
Oil Reservoirs (which are summarized in Table 2) are used as inputs for the two EOR screening tools
(EORgui and the developed tool). The results of the screening for both of them are summarized in Table
3.
8 SPE-175883-MS

Table 2—Rock and fluid properties for some of the Egyptian Oil Reservoirs

Table 3—Summary of the screening results

Reservoir EORgui New Developed Screening Tool

X1 IMM (83%), ST (80%), COMB (67%). COMB (83.81%), ST (80.85%)

X2 IMM (83%), COMB (83%). COMB (83.99%), CO2IMMF (81.93%), ST
(80.82%), N2IMMF (80.5%).
X3 IMM (83%), SP/ASP (73%). COMB (83.98%), POL (83.75%), ASP
(83.23%), ST (80.74%), N2IMMF (80.5%)
X4 IMM (83%), COMB (58%). COMB (83.75%), CO2IMMF (80.5%).
X5 IMM (83%, COMB (58%). COMB (83.99%), CO2IMMF (81.68%).
X6 IMM (83%), COMB (58%). COMB (83.75%), CO2IMMF (80.65%).
X7 SP/ASP (91%), IMM (83%), COMB (75%), POL (83.85%), ASP (82.6%), N2IMMF
HCMF (70%), CO2MF (67%). (80.5%).
X8 IMM (83%), HCMF (70%), CO2MF (67%). CO2MF (85.27%), HCMF (84.42%).
X9 SP/ASP (91%), COMB (75%), CO2MF (67%), POL (83.71%), ASP (81%), N2IMMF (80.5%).
IMM (67%), HCMF (50%).
X10 IMM (100%), SP/ASP (100%), COMB (83%), COMB (83.75%), POL (83.72%), ASP (81%),
CO2MF (67%). N2IMMF (80.5%).
X11 SP/ASP (91%), IMM (83%), COMB (75%), POL (83.47%), ASP (80.56%), N2IMMF
CO2MF (67%), HCMF (60%). (80.5%), COMB (68.5%).
X12 IMM (83%), CO2MF (78%), HCMF (70%). POL (83.6%), N2IMMF (80.5%).
X13 IMM (83%), CO2MF (67%), HCMF (60%). POL (83%), ASP (80.5%), N2IMMF (80.5%).
X14 IMM (83%), COMB (75%), CO2MF (67%), POL (80.79%), N2IMMF (80.5%), COMB
HCMF (60%). (65.5%).
X15 CO2MF (67%), IMM (67%), HCMF (50%). POL (83.6%), N2IMMF (80.5%), COMB
(68.5%).

IMM: Immiscible EOR methodST: Steam InjectionCOMB: In-Situ CombustionCO2IMMF: CO2 Immiscible FloodingN2IMMF: N2 Immiscible FloodingSP: Surfactant-
PolymerASP: Alkaline-Surfactant-PolymerPOL: Polymer FloodingHCMF: Hydrocarbon Miscible FloodingCO2MF: CO2 Miscible Flooding

The differences between the results of the two EOR screening tools are mainly due to the updated
screening criteria of the EOR methods. EORgui uses the screening criteria of Taber (Taber, Martin and
SPE-175883-MS 9

Seright, 1997); while the new developed EOR screening tool uses the updated screening criteria of Taber
(Aladasani and Bai, 2010) which are represented in Table 1. For example, Steam Injection is seen to be
feasible in Reservoirs X2 and X3 by the new developed tool, while it is not feasible by EORgui. That is
related to that the EORgui excludes the Steam Injection methods when the reservoir permeability is less
than 200 md. The new developed tool shows that the Steam Injection is feasible because many recent and
successful EOR projects are implemented with reservoir permeability below this limit. This indicates that
the screening criteria are updated to include wide range of the successful EOR projects.
The same for the In-Situ Combustion and Polymer flooding results for Reservoir X3. After updating
the permeability range for the In-Situ Combustion and the oil saturation range for the Polymer flooding,
both of them will be feasible for Reservoir X3.
The EORgui doesn’t take into consideration the Minimum Miscibility Pressure (MMP), so it shows that
miscible injection methods are feasible for all of the reservoirs in Group 2 in Tables 2 and 3. The new
developed tool shows that the miscible injection methods are not feasible for most of these reservoirs
because the miscibility conditions are not achieved. However, water injection can be used to increase the
reservoir pressure to achieve the miscibility before CO2 is injected. The new developed tool shows that
Polymer flooding is feasible for all of the reservoirs of Group 2. ASP is feasible for Reservoir X13, and
In-Situ Combustion is feasible for Reservoir X15. However, EORgui excludes these EOR methods from
the results of those reservoirs. That is due to that the new developed tool used wide range for the screening
criteria of the parameters (based on the recent successful EOR projects). Although EORgui shows that
In-Situ Combustion is feasible for Reservoirs X7 and X9, the new developed tool shows that In-Situ
Combustion is not feasible for these reservoirs. This result may be attributed to that the oil viscosity. The
value of the oil viscosity is out of range for the updated screening criteria in the new developed tool.
Case Studies
Table 2 presents the data of some of the Egyptian oil reservoirs. The MMP values in Table 2 are calculated
by correlations (Abu El Ela et al., 2014; Yellig, 1980). Group 1 presents reservoirs in the Gulf of Suez
and Group 2 presents reservoirs in the Western Desert. The developed EOR screening tool was used to
investigate the applicability of applying EOR in these reservoirs. The results of the screening for the
reservoirs under study are summarized in Table 3. Based on the available data and the screening results,
the following results are illustrated:
Steam injection method is not seen to be feasible in the Western Desert reservoirs because the oil
viscosity, formation thickness or/and API gravity are out of range in the studied reservoirs. However,
Steam injection is feasible in some of the Gulf of Suez reservoirs.
In-Situ Combustion is the most suitable EOR method for Gulf of Suez reservoirs (Group 1). However,
it is not feasible (Except Reservoirs X11, X14, X15) in some of the Western Desert reservoirs (Group 2)
because of the reservoir thickness. In-Situ Combustion is the most suitable EOR method for Reservoir
X10 in Group 2.
The results indicate that the chemical EOR methods are feasible in most of the Western Desert
reservoirs (Group 2), but they are not seen to be feasible in most of the reservoirs in the Gulf of Suez
(Group 1) due to the high formation water salinity, and the heterogeneous carbonate formations.
Egyptian oil reservoirs are good candidate for immiscible gas injection. Nitrogen Immiscible Injection
(N2IMMF) is feasible in most of the Western Desert reservoirs and in some of the Gulf of Suez reservoirs.
Carbon Dioxide Immiscible Injection (CO2IMMF) seems to be feasible for most of the Gulf of Suez
reservoirs, but there is a restriction on the source of CO2 there.
Miscible methods are not seen to feasible in most of the Egyptian oil reservoirs, because the MMP is
greater than the current reservoir pressure (there is a great difference), and most of these reservoirs are
depleted. Both of Miscible Carbon Dioxide injection (CO2MF) and Miscible Hydrocarbon Injection
(HCMF) are seen to be feasible for Reservoir X8 in the Western Desert (Group 2).
10 SPE-175883-MS

Conclusions
Based on the available data, the following conclusions can be drawn from this work:
1. Fuzzy logic as an artificial intelligence technique was used to develop EOR screening criteria
model.
2. The new screening methodology was generated from statistical analysis of database of 347
successful EOR projects worldwide.
3. The new developed tool can execute screening for fourteen EOR methods.
4. The screening results for some of the Egyptian oil reservoirs are presented and analyzed.
5. Although Steam injection is feasible in some of the Gulf of Suez reservoirs, the method is not seen
to be feasible in the Western Desert reservoirs. In-Situ Combustion is the most suitable method for
Gulf of Suez reservoirs. The chemical methods are feasible in most of the Western Desert
reservoirs, but they are not seen to be feasible in most of the reservoirs in the Gulf of Suez.
Egyptian oil reservoirs are good candidate for immiscible gas injection, however, miscible
methods are not seen to feasible in most of these oil reservoirs.

Nomenclature
AI ⫽ Artificial Intelligence.
ASP ⫽ Alkaline-Surfactant-Polymer.
CO2IMMF ⫽ CO2 Immiscible Flooding.
CO2MF ⫽ CO2 Miscible Flooding.
COMB ⫽ In-Situ Combustion.
EOR ⫽ Enhanced Oil Recovery.
FIS ⫽ Fuzzy Inference System.
GUI ⫽ Graphical User Interface.
HCMF ⫽ Hydrocarbon Miscible Flooding.
IMM ⫽ Immiscible EOR Method.
MF ⫽ Membership Function.
MMP ⫽ Minimum Miscibility Pressure.
N2IMMF ⫽ N2 Immiscible Flooding.
POL ⫽ Polymer Flooding.
R1 ⫽ the lower boundary of the range through which the maximum number of successful EOR
projects had been applied.
R2 ⫽ the upper boundary of the range through which the maximum number of successful EOR
projects had been applied.
Smf ⫽ S Shape Membership Function.
SP ⫽ Surfactant-Polymer.
ST ⫽ Steam Injection.
Trapmf ⫽ Trapezoidal Membership Function.
Trimf ⫽ Triangular Membership Function.
WAG ⫽ Water Alternative Gas.
Zmf ⫽ Z Shape Membership Function.

References
Abu El Ela, M. et al. 2014. An Integrated Approach for the Application of the Enhanced Oil Recovery
Projects. Journal of Petroleum Science Research, 3 (04): 184 –185.
SPE-175883-MS 11

Aladasani, A., and Bai, B. 2010. Recent Developments and Updated Screening Criteria of Enhanced
Oil Recovery Techniques. Paper presented at the CPS/SPE International Oil & Gas Conference
and Exhibition, Beijing, China, June 8 –10. SPE 130726
El-Batanony, M. H., Sayyouh, M. H., Mahgoub, I. S. and Abdel-Waly, A. 1987. Screening guides for
enhanced oil recovery methods under Egyptian reservoir conditions. Journal of Egyptian Society
of Engineers, 26 (01): 60 –66, Egypt.
Goda, H. M., Shokir, E. M., Fattah, K., and Sayyouh, M. H. 2002. Neural network modeling approach
for EOR method selection and evaluation. Nafta journal, Zagreb, CROATIE. 53 (09): 327–330.
Koottungal, L. 2012. 2012 Worldwide EOR Survey. Oil & Gas Journal, 110 (4).
Koottungal, L. 2014. 2014 Worldwide EOR Survey. Oil & Gas Journal, 112 (5).
Lee, J. Y., Shin, H. J., and Lim, J. S. 2011. Selection and evaluation of enhanced oil recovery method
us-ing artificial neural network. Geosystem Engineering, 14 (4): 157–164.
Moritis, G. 2004. 2004 Worldwide EOR Survey. Oil & Gas Journal, 102 (14): 53.
Moritis, G. 2010. 2010 Worldwide EOR Survey. Oil & Gas Journal, 108 (14): 41.
Sayyouh, M. H, and Al-Blehed, M. S. 1990. Screening criteria for enhanced recovery of Saudi crude
oils. Int. J. Energy Sources, 12 (01): 71–82, USA.
Sayyouh, M. H., Al Blehed, M. S. and Hemieda, A. 1993. Possible application of microbial enhanced
oil recovery to Arab oil fields. J. King Saud University (Engineering Science), 5 (02), Saudi
Arabia.
Shokir, E. M. El-M., Goda, H. M., Sayyouh, M. H., and Al-Fattah, K. 2002. Selection and Evaluation
EOR Method Using Artificial Intelligent. Paper presented at the 26th Annual SPE International
Technical Conference and Exhibition in Abuja, Nigeria, August 5–7. SPE 79163.
Taber, J. J., Martin, F. D. and Seright, R. S. 1997. EOR screening criteria revisited: Part 1:
Introduction to screening criteria and enhanced recovery field projects. SPE Reservoir En-
gineering, 12 (3): 189 –198.
Taber, J. J., Martin, F. D.; and Seright, R. S. 1997. EOR screening criteria revisited: Part 2:
Applications and impact of oil prices. SPE Reservoir Engineering, 12 (3): 199 –206.
Yellig, W. F., Metcalfe, R. S. 1980. Determination and Prediction of CO2 Minimum Misci-bility
Pressures. SPE J. Petrol. Technol., 32 (1): 160 –168.