Sunteți pe pagina 1din 18

This template is provided to give authors a basic shell for preparing your

manuscript for submittal to a meeting or event. Styles have been included to


give you a basic idea of how your finalized paper will look before it is
published. All manuscripts submitted will be extracted from this template and
tagged into an XML format; standardized styles and fonts will be used when
laying out the final manuscript. Links will be added to your manuscript for
references, tables, and equations. Figures and tables should be placed
directly after the first paragraph they are mentioned in. The content of your
paper WILL NOT be changed.

OTC-28367-MS

New Inflow Performance Relationship for a Horizontal Well in a Naturally


Fractured Solution Gas Drive Reservoirs using Artificial Intelligence
Technique.
Zeeshan Tariq, Abdulazeez Abdulraheem, Mohammad Rasheed Khan, and Ahmed Sadeed, King Fahd University
of Petroleum & Minerals

Copyright 2018, Offshore Technology Conference

This paper was prepared for presentation at the Offshore Technology Conference Asia held in Kuala Lumpur, Malaysia, 20-23 March 2018.

This paper was selected for presentation by an OTC 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 Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Offshore Technology Conference, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the
written consent of the Offshore Technology Conference 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 OTC copyright.

Abstract
The idea of continuous assessment of individual well performance is imperative globally to E&P
organizations when it comes to production optimization regimes and increasing profitability from each
barrel of oil present. One of the most effective ways to assess this performance is through the use of
Inflow-Performance Relationship (IPR) & Outflow-Performance Relationship (OPR) curves.
Consequently, the use of the pertinent IPR which is representative of the performance is essential.
Vogel’s IPR has been employed in the industry more like a standard when it comes to conventional
reservoirs well performance. Moreover, Vogel’s IPR model can successfully model IPR for vertical wells
in a homogeneous reservoir producing from solution gas drive mechanism. However, current IPR models
for horizontal wells are only valid for single porosity reservoirs, and their applicability to dual
porosity/dual permeability reservoirs is questionable. The complexity in such reservoirs arise due to the
combined flow between the distinct systems of the matrix and fracture, and as a result, it is imperative to
develop a new IPR model that incorporates the impact of the fracture parameters.
This work focuses to inspect effects of the complex flow behavior on the inflow curves, concentrating
on horizontal wells in NFR’s. Foundation of this work is based on the development of a base case black
oil computational model incorporating typical reservoir and PVT properties. Finite number of data points
linking oil rate to flowing bottom-hole pressures were used to generate the dimensionless IPR curves.
Wide-ranging PVT and particular reservoir properties concerning two systems of porosity and
permeability were used which included; Inter-porosity flow co-efficient, storativity ratio, normalized
horizontal well length, reservoir thickness and saturation pressure. As expected, it was concluded that the
NFR parameters of storativity and inter-porosity flow coefficient, along with the normalized horizontal
well length had substantial control on dimensionless IPR curve. To further augment the results, effort is
expended to congregate outcomes into one unpretentious model using a combination of support vector
2 OTC-28367-MS

machine and non-linear regression techniques along with the particle swarm optimizer to formulate a new
empirical IPR equation.
The newly suggested pragmatic IPR model produces results within acceptable absolute error range
of 2%, when compared with the actual data. Accordingly, this proves that the developed correlation is
very accurate and can prove to be a vital tool for production/reservoir engineers concerned with the
production optimization/enhancement of horizontal wells in naturally fractured dual porosity-dual
permeability reservoirs.

1. Introduction
An engineering tool used in petroleum production to evaluate the performance of a well by plotting
the oil production rate with the well bottom-hole pressure (BHP) typically termed as Inflow Performance
relationship (IPR). An IPR evaluates how much barrels of oil or gas can be produced per psi of drawdown.
These relationships are very vital not only to monitor field/reservoir life, but also to optimize it. Moreover,
they are important tools to assist in designing, calculating and estimating various parameters such as in
artificial lift, well completion equipment, workover evaluations, etc. In general Oil & Gas Engineers make
use of this throughout the exploration and production lifecycle of a reservoir; the most common of which
is to predict the effect of various operational situations on the flow profile of the well.
An Inflow towards the well is related to pressure drop concerning the well bore and reservoir and can
be represented by linear relationship for single phase at under-saturated above bubble point conditions.
However, this relationship is no longer linear when two phases start flowing in saturated oil systems.
During last many decades’ engineers and scientists have developed many empirical correlations and
analytical models to capture the non-linear behavior between well production rate and well bottom-hole
pressure below bubble point. But these correlations and models are sensitive to specific well type or
reservoir conditions and it is highly recommended by previous researchers to use specific model or
correlation for specific reservoir. A brief review of generally used IPR models is described below
Typically for slightly compressible fluid the productivity index can be found from Darcy’s equation
given by Eq. 1
0.00708Kh K ro (1)
J= ( )
re 3 μ β
ln (r ) − 4 + S o o
w

The parameters of Darcy’s equation except well parameters are pressure dependent and they are
approximately constant above bubble point, so the IPR above the saturation pressure can be found out by
simple linear relation given by Eq. 2

Q = J(PR − Pwf ) (2)


However, as the pressure dips below the saturation point, due to liberation of lighter molecules from the
oil and the emergence of two phase flow, the linear relationship between well flow-rate and well bottom-
hole pressure is no longer valid. (Vogel, 1968) proposed a general curve fitted equation for solution gas
drive reservoirs by executing a classical numerical study. But his correlation is only valid for vertical wells
drilled in a homogeneous reservoir producing from solution gas drive mechanisms. Since then, several
other IPR models for vertical wells and horizontal wells have been developed. Some of the vertical wells
IPR are (Eghbali and Gerami, 2013; Kilns and Clark, 1993; Klins and Majcher, 1992; Retnanto and
Economides, 1998; Sadeghi et al., 2013; Tariq et al., 2016b; Wiggins, 1993) and horizontal wells are
(Bendakhlia and Aziz, 1989; Cheng, 1990; Khalid et al., 2014). The breif description of each mentioned
IPR’s are given in Table 1.
OTC-28367-MS 3

Table 1. Summary of few IPR correlation over the years.

Model Names IPR Equations Description

Valid for solution gas


qo Pwf Pwf 2
(Vogel, 1968) = 1 − 0.2 ( ) − 0.8 ( ) drive homogenous
qmax Pr Pr
reservoirs.
Valid for solution gas
(Richardson and qo Pwf Pwf 2
= 1− V( ) − (1 − V) ( ) drive homogenous
Shaw, 1982) qmax Pr Pr
reservoirs.
(Bendakhlia and n Valid for horizontal wells
qo Pwf Pwf 2
= (1 − V ( ) − (1 − V) ( ) )
Aziz, 1989) qmax Pr Pr solution gas drive.

qo Pwf Pwf 2 Valid for slanted wells


(Cheng, 1990) = 0.985 − 0.2055 ( ) − 1.1818 ( )
qmax Pr Pr solution gas drive.
qo Pwf Pwf N
= 1 − 0.295 ( ) − 0.705 ( ) Considered the effect of
qmax Pr Pr
(Klins and Majcher,
where, bubble point pressure
1992)
Pr valid for vertical wells.
N = (0.28 + ) ∗ (1.235 + 0.001 ∗ Pb)
Pb

(Kilns and Clark, qo PWf PWf 2


= 1 − 0.1225 ( ) − 0.8775 ( ) Valid for horizontal well
1993) qmax Pr Pr

qo PWf PWf 2 Valid for three phase


(Wiggins, 1993) = 1 − 0.519167 ( ) − 0.481092 ( )
qmax Pr Pr systems, vertical wells.

qo PWf PWf n
= 0.985 − 0.25 ( ) − 0.75 ( )
qmax Pr Pr
(Retnanto and Valid for two phase
where,
Economides, 1998) systems, horizontal wells.
P P
n = [−0.27 + 1.46 ( ) − 0.96 ( )] (4 + 1.6E − 3Pb )
Pb Pb

(Sadeghi et al., 0.00708kf h(Pr −Pwf ) kr f,o Analytical IPR valid for
q= re x( )
[ln( )+s−0.75] μo βo pav
2013) rw vertical well.
Valid for dual porosity
(Eghbali and qo PWf PWf 2
= 1 − 0.63 ( ) − 0.37 ( ) single permeability
Gerami, 2013) qmax Pr Pr
reservoirs, vertical well.
qo PWf Valid for solution gas
= 1 − (0.673 − 0.0278γ) ( ) − (0.0278γ +
qmax Pr
(Khalid et al., 2014) drive multilayered
PWf 2
0.36212) ( ) horizontal wells.
Pr

qo PWf PWf 2.76


= 1− α( )− β( )
qmax PR PR
Valid for dual porosity
where,
(Tariq et al., 2016b) 0.715∗λ0.04
dual permeability
α=
ω0.027 reservoirs, vertical well.
0.388 ∗ ω0.11
β=
λ0.030
4 OTC-28367-MS

From this extensive literature survey, it is evident that many empirical correlations exist to estimate
the productivity of vertical and horizontal wells below the bubble point pressure but these correlations are
specific and applicable to the case for which they are developed for. An IPR for horizontal well in a
naturally fractured dual porosity-dual permeability solution gas drive reservoirs is not deeply investigated
yet. So, the objective of this research work is to investigate the parameters associated with naturally
fractured reservoirs and their impact on dimensionless IPR curve to formulate new pragmatic IPR
correlation for horizontal wells using non-linear regression and support vector machine, optimized by
particle swarm optimizer.

1.1 Brief Overview of Naturally Fractured Reservoirs


Naturally Fractured Reservoirs (NFR) are represented by two types of porous and permeable media:
fractures & matrix (Guo and Liu, 2014; Warren and Root, 1963). In such reservoir systems the fluid flow
is from matrix and fractures both, while in dual porosity single permeability reservoirs the fluid flow is
from fractures only. Because of the unique conductivity and fluid storage characteristic of fractures and
matrix, these reservoirs are usually called as dual-porosity/dual permeability reservoirs (Kazemi et al.,
1969; Warren and Root, 1963). Major storage is provided by the matrix while fractures provides the
principal passage for the fluid flow.
Fractured reservoir systems vary vastly in comparison to simple single-porosity systems, as not only
the dual porosity/dual permeability continuum should be taken into consideration, but even the processes
involved in the transfer between the two systems is important in understanding these complex structures.
(Dean and Lo, 1988). Considering these complexities, efforts to model this have been made several years
ago. The amount of heterogeneity and anisotropy incorporated in NFR’s present a challenge starting right
from reservoir characterization and including the production strategies; hence it is imperative to know the
future performance from such reservoirs to plan all the activities from drilling to completion and facilities
design. (Bourbiaux et al., 2002; Gang and Kelkar, 2006).
Warren and Root worked in great depth to investigate NFR, and used a fundamental assumption that
the matrix although contributed to the main porosity, however, its contribution towards the flow capacity
is insignificant. Consequently, their study produced an idealized concept to classify such reservoirs.
(Warren and Root, 1963). An important conclusion of this work was the defining of two parameters to
characterize the unconventionality of NFRs from single-porosity systems; inter-porosity flow coefficient
which is the extent of fluid capacitance of matrix porosity and the other parameter was the storativity ratio,
being the gauge of heterogeneity of the rock system.
Inter-porosity flow coefficient ‘λ’ which defined the flowing ability in the fracture system and is
given by Eq. 3
Km (3)
λ = σrw 2
Kf
Km and Kf are the matrix and fracture permeabilities in milidarcies, rw is the wellbore radius in feet and σ
is the shape factor which is a matrix to fracture interface area per unit volume measured in length-2.

(Kazemi et al., 1976) proposed the subsequent formula to find σ. Eq. 4


1 1 1 (4)
𝜎 = 4 [ 2 + 2 + 2]
lx ly lz

where lx, ly and lz are 𝑥, 𝑦 and 𝑧 dimensions of the blocks of the material that makes up the matrix volume.

The storativity ratio ‘ω’ which describes the storing ability of matrix system and is given by Eq. 5
OTC-28367-MS 5

(ϕVCt )f (5)
ω=
(ϕVCt )m + (ϕVCt )f

ϕ is the porosity, ct is the total compressibility and V is the ratio of total volume of one medium to the
total volume of the whole system.

1.2 Support Vector Machine (SVM)


SVM is a supervised learning technique which can be used for classification, regression and data
reduction problems. SVM belongs to a group of a principle linear classifiers. A linear classifier makes
classification decision based on the value of a linear combination of the characteristics. SVM has a special
feature of optimum hyperplane which increases the space to training examples in a great dimensional
feature space (Nooruddin et al., 2013). SVM depends on a statistical learning theory in which the risk is
minimized by decreasing the generalization error of the upper bound instead of decreasing the error of
training. SVM works in way of solving linearly quadratic programming problems with a distinctive,
optimum solution that is improbable to be local minima (El-Sebakhy et al., 2007). SVM’s performance
depends on many parameters that should be selected carefully to achieve a perfect prediction model
(Bennett and Blue, 1998; Elkatatny et al., 2018, 2017; Tariq et al., 2017).

1.3 Particle Swarm Optimization (PSO)


PSO is a stochastic population based evolutionary algorithm motivated by the societal attributes of
fish schooling and birds clustering (Abido, 2002; Kennedy, 1997; Shi and Eberhart, 1998). PSO represents
population of random solutions in the search space as particles assigning random velocities to them and
iteratively tuning the fitness of the particles until the best solution, called global best is achieved. Every
particle in PSO algorithm is assigned with a fitness value that measures the particle’s potential to gain the
required purpose. At the beginning, the particles are positioned randomly in the solution space and the
swarm moves within the solution space to grasp the value of optimal fitness. Every particle’s velocity and
position are randomly initialized. At each epoch, the fitness values of the particles are computed and
required tunings are made to the velocity/place of every particle to shift them to the value of optimal
fitness (Catalao et al., 2010; Chatterjee et al., 2017; Tariq et al., 2016a; Vasumathi and Moorthi, 2012;
Wang et al., 2015).

2. Methodology
2.1 Simulation Model Description
A commercial simulator Eclipse is used to create the two-phase three-dimensional reservoir
simulation model having single horizontal well. The reservoir type is naturally fractured, having two
porous and permeable regions, matrix and fractures respectively. Reservoir drive mechanism is the
solution gas drive. Horizontal well is placed at the midpoint of the reservoir. Fig. 1 and Fig. 2 shows the
top and front view of simulation model having single well in cartesian coordinate system. Around the
horizontal well local grid refinement is used to capture the flow inside the well.
6 OTC-28367-MS

Figure 1. Top view of reservoir simulation model showing local grid refinement around the vicinity of horizontal well.

Dual Porosity-Dual permeability model was built through different combinations of matrix & fracture
porosities and permeabilities (Km, Kf, øf and øm respectively). All cases were run at constant bottom hole
pressures. The base case was modeled as a single well producing from a reservoir having five layers of
constant thickness of 20ft each. A horizontal well of diameter 0.375 ft was located at the middle of the
reservoir. A constant connate water saturation of about 12% was assigned to all grids. A bubble point
pressure of 4000psi was used in base case. The base case data used for IPR development is reported in
Table 2. Fig. 3 shows the PVT properties of fluids used in the simulation model.

The assumptions used for the development of simulation model is given below

• Three phase 3D model is implemented.


• The reservoir is assumed to be at bubble point condition.
• Well stands positioned at the middle of the reservoir.
• Horizontal well is perforated vertically in center of reservoir section.
• Assumed neither damage nor stimulation.
• Neglected the gravity effects, non-darcy flow effects and interfacial tension.
OTC-28367-MS 7

Figure 2. Side view of reservoir simulation model showing horizontal well location.

Table 2. Base Case Parameters of Reservoir Simulation Model.


Reservoir Parameters Values Units
Grid System Cartesian
Reservoir Dimensions 20 x 20 x 5
Grid Size in X-Direction 50 ft.
Grid Size in Y -Direction 50 Degrees
Grid Size in Z-Direction 10 ft.
Reservoir Area 23 acres
Reservoir Thickness 50 ft.
Reservoir Depth 4000 ft.
Reservoir Pressure 4000 psia
Bubble Point Pressure 4000 psia
Reservoir Lithology Limestone
Rock Compressibility 3x10-6 psi-1
Porosity Varying fraction
Reservoir Temperature 212o F
Oil Gravity 40o API
Average Permeability Varying md
No. of production Well 1
Well diameter 0.375 ft.
8 OTC-28367-MS

Figure 3. PVT properties of fluids used in the simulation model.

3. New IPR Model Development


3.1 Data Generation and Sensitivity Analysis
Production rates and well bottom-hole pressures is required for the generation of IPR curve. Oil
production rate divided by the maximum oil well rate at absolute open flow potential (AOFP) condition,
corresponding to 100% pressure drawdown condition and well bottom-hole pressure divided by the
average reservoir pressure, results in the generation of dimensionless IPR curves. Dimensionless IPR
curves were generated to compare their curvatures by the sensitivity of several reservoir and fluid
parameters to see the proportion of alteration of oil production rate by flowing bottom-hole pressure.
These parameters included reservoir thickness, bubble point pressure, normalized horizontal well length,
storativity ratio and interprosity flow coefficient. Normalized horizontal well length is the ratio of well
length from toe to heel to the length of the reservoir in which the well is drilled horizontally.
Fig. 4 shows that changing the reservoir thickness from 25ft to 75ft does not have any effect on the
curvature of dimensionless curve. Fig 5 shows that changing the bubble point pressure from 2000 psi to
4000 psi does not impact the dimensionless IPR curve. Fig. 6 indicates that changing normalized
horizontal well length LeL shifts the curvature of dimensionless IPR curve. Fig. 7 to Fig. 8 shows that
changing inter-porosity flow coefficient has a significant impact on dimensionless curves. Similarly, Fig.
9 to Fig. 10 shows that changing the storativity ratio has also a vital impact on dimensionless curves. To
obtain sufficient amount of data to generate new empirical IPR correlation for dual porosity dual
permeability system several simulation instances were run with diverse data values of LeL, λ and ω. The
typical ranges of λ is from 10-1 to 10-9 and of ω is 0.5 to 0.001 (Cinco-Ley and Samaniego-V., 1981). So
in this study the values of λ and ω varied in the ranges defined by (Cinco-Ley and Samaniego-V., 1981).
In all the cases run, the initial reservoir pressure was assumed to be 4000 psi which was at the bubble point
pressure. For every case with the different combination of (LeL, λ and ω), the well flowing bottom-hole
pressure (Pwf) varied from slightly lower than initial reservoir pressure to atmospheric pressure i.e. (3950,
3500, 3200…14.7psi) in 20 steps. The set of data was obtained by coupling MATLAB with Eclipse.
Eclipse simulator was run by MATLAB thru changing one parameter at a time with in a range of data
defined in the MATLAB code. The total number of simulation runs were calculated by the product of
number of variables changing steps. Table 3 shows the steps of each parameter changes and their ranges;
total simulation runs were 10,000. After generating significant amount of data, the data set was divided in
OTC-28367-MS 9

to two parts with ratio of 0.7:0.3, that means 70% of the data generated was used to make the correlation
while 30% was used for testing of the developed correlation.

Table 3. Total number of simulation runs.


Parameters Range Steps
Normalized Horizontal Well length, LeL 0.1 - 0.6 5
Interporosity flow coefficient, λ 10E-9 - 0.01 10
Storativity ratio, ω 0.001 - 0.45 10
Pwf 14.7 - 3950 20
Total Number of Simulations 5 * 10 * 10 * 20 10,000

Figure 4. Effect of reservoir thickness on dimensionless IPR curves.


10 OTC-28367-MS

Figure 5. Effect of bubble point pressure on dimensionless IPR curves.

Figure 6. Effect of normalized well length at ω = 0.200 and λ = 0.006 on dimensionless IPR curves.
OTC-28367-MS 11

Figure 7. Effect of λ at ω = 0.200 and LeL =0.55 on dimensionless IPR curves.

Figure 8. Effect of λ at ω = 0.1429 and LeL =0.55 on dimensionless IPR curves.


12 OTC-28367-MS

Figure 9. Effect of ω at λ = 0.006, LeL = 0.55 on dimensionless IPR curves.

Figure 10. Effect of of ω at λ = 0.003, LeL = 0.55 on dimensionless IPR curves.


OTC-28367-MS 13

3.2 Implementation of Support Vector Machine and Particle Swarm Optimization


From the Eclipse simulator a total 10,000 data points were generated. To reduce number of data points
and select only the representative ones, support vector machine is used. This algorithm, transfers each data
point in an ‘n’- dimensional space. Next, for the classification and selection of the most representative
points, hyper-plane is sought after which differentiates the good points from the bad ones. However, it
must be noted that deciding on the right hyper-plane is essential to reaching a good quality segregation of
the data points. In SVM, it is simpler to have straight hyper-planes, but when the data points are scattered
in such a way that linearity does not work, the kernel option of the SVM algorithm is used which basically
takes the low dimensional points and converts it into higher dimensional separable data points. Therefore,
here to reduce the number of data points, enormously intricate data conversions employing SVM is carried
out to come up with high quality representative data which would lead to a reliable correlation.
Once the data has been filtered out, non-linear regression is used to find a non-linear model of the
relationship between the dependent variable (Dimensionless flowrate) and set of independent variables
(Normalized horizontal well length, storativity ratio, interporosity flow coefficient and dimensionless
pressure). This helped to estimate model for our case with arbitrary associations amongst the variables;
we have achieved this by employing iterative assessment procedures. To improve the coefficient of non-
linear regression model, Particle swarm optimizer is used by setting average absolute percentage error
(AAPE) between actual and predicted values as an objective function to be minimized. Fig. 11 summarizes
the workflow adopted to reach the optimum model.

Parameters
Simulation Model
Sensitivity Data Generation
Development
Analysis

Regression Model
Coefficient Non-Linear Data Reduction
optimization Regression using SVM
using PSO

Correlation Statistical Recommendation


Development Analysis for Future

Figure 11. Work flow of the proposed Model.


14 OTC-28367-MS

4. New Empirical IPR Correlation.


Using state of the art Support Vector Machine and Particle Swarm Optimization algorithms along
with non-linear regression technique, the new correlation given by Eq. 6 is developed. The new IPR for
horizontal well takes into account the effect of fracture parameters and horizontal well length.

qo Pwf Pwf 6
= 1 − α( ) − β( ) (6)
q max PR PR

Where

0.713 ∗ ω0.23
α= (7)
LeL 0.024 ∗ λ0.032

0.47 ∗ λ0.076
β= (8)
LeL 0.07 ∗ ω0.105

4.1 Steps to use New IPR Correlation

Step 1: If the values of ω and λ are not known then estimate the values of ω and λ from pressure draw
down or pressure build up test, detailed procedure to find these parameters is given in reference
(Najurieta, 1980)

Step 2: Calculate the values of α and ß by using Eq. 7 and Eq. 8

Step 3: Calculate qomax using Eq. 6 at any given test point. Eq. 6 can be written in form of qomax as
belows:
q o (test)
q max = (STB/Day)
Pwf Pwf 6
1 − α(P )− β(P )
R R

Step 4: Assume several values of Pwf starting from less than average reservoir pressure till AOFP
conditions and calculate corresponding qo values.

Pwf Pwf 6
q o = q max ∗ [1 − α ( ) − β ( ) ] (STB/Day)
PR PR

Step 5: Produce the future IPR curve by graphing qo vs Pwf.


OTC-28367-MS 15

5. Conclusion
In this study, several commonly used empirical IPR correlations were reviewed. The new IPR
correlation is formulated for horizontal wells in NFR. Based on the results we can conclude that;

i. The new model depends upon the parameters associated with the NFR, aimed at horizontal wells
producing hydrocarbons from solution-drive reservoirs.
ii. The new correlation requires three parameters only, namely; normalized horizontal well length,
inter-porosity flow coefficient and storativity ratio. Inter-porosity flow coefficient and storativity
ratio can be determined from well test interpretations (pressure buildup or draw down tests).
iii. The new correlation can be served as a very handy tool or a general reference curve for horizontal
wells in a dual porosity/dual permeability NFR solution-gas-drive reservoirs.

Acknowledgement
The authors would like to thank College of Petroleum & Geosciences, King Fahd University of
Petroleum & Minerals for providing research opportunities to produce this paper.

Nomenclature
AOFP = Absolute open flow potential, STB.
Bg = Gas formation volume factor, SCF/bbl.
Bo = Oil formation volume factor, bbl/STB.
ct = Total compressibility, psi-1
h = Reservoir thickness, ft.
IPR = Inflow performance relationship.
J = Productivity index, STB/psi.
K = Permeability, md.
Kro = Relative Permeability of oil.
Km = Matrix Permeability, md.
Kf = Fracture Permeability, md.
LeL = Normalized horizontal well length.
Pb = Bubble point pressure, psi
PR = Average reservoir pressure, psi
Pwf = Flowing bottomhole pressure, psi
qo = Oil flow rate, STB
qmax = Oil flow rate at AOFP, STB
re = Reservoir radius, ft.
rw = Well radius, ft.
S = Total skin.
V = Ratio of total volume of matrix of fracture to the total volume of the whole system.

Greek Symbols
α = Prosed Model Parameter.
β = Prosed Model Parameter.
λ = Interporosity flow coefficient.
σ = Matrix to fracture shape factor.
ϕm = Matrix porosity.
ϕf = Fracture porosity.
ω = Storativity ratio
16 OTC-28367-MS

μ = Viscosity, cp
μo = Oil viscosity, cp
μg = Gas viscosity, cp

Subscripts
m = matrix.
f = fractures.
g = gas phase.
o = oil phase.
max = maximum.
OTC-28367-MS 17

References
Abido, M.A., 2002. Optimal design of power-system stabilizers using particle swarm optimization. IEEE
Trans. Energy Convers. 17, 406–413. doi:10.1109/TEC.2002.801992
Bendakhlia, H., Aziz, K., 1989. Inflow Performance Relationships for Solution-Gas Drive Horizontal
Wells, in: SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers.
doi:10.2118/19823-MS
Bennett, K.P., Blue, J.A., 1998. A support vector machine approach to decision trees, in: 1998 IEEE
International Joint Conference on Neural Networks Proceedings. IEEE World Congress on
Computational Intelligence (Cat. No.98CH36227). IEEE, pp. 2396–2401.
doi:10.1109/IJCNN.1998.687237
Catalao, J.P.S., Pousinho, H.M.I., Mendes, V.M.F., 2010. Hybrid Wavelet-PSO-ANFIS Approach for
Short-Term Wind Power Forecasting in Portugal. IEEE Trans. Sustain. Energy.
doi:10.1109/TSTE.2010.2076359
Chatterjee, S., Sarkar, S., Hore, S., Dey, N., Ashour, A.S., Balas, V.E., 2017. Particle swarm optimization
trained neural network for structural failure prediction of multistoried RC buildings. Neural Comput.
Appl. 28, 2005–2016. doi:10.1007/s00521-016-2190-2
Cheng, A.M., 1990. Inflow Performance Relationships for Solution-Gas-Drive Slanted/Horizontal Wells,
in: SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers.
doi:10.2118/20720-MS
Cinco-Ley, H., Samaniego-V., F., 1981. Transient Pressure Analysis: Finite Conductivity Fracture Case
Versus Damaged Fracture Case, in: SPE Annual Technical Conference and Exhibition. Society of
Petroleum Engineers. doi:10.2118/10179-MS
Dean, R.H., Lo, L.L., 1988. Simulations of Naturally Fractured Reservoirs. SPE Reserv. Eng. 3, 638–648.
doi:10.2118/14110-PA
Eghbali, S., Gerami, S., 2013. Modification of Vogel’s Inflow Performance Relationship (IPR) for Dual
Porosity Model. Pet. Sci. Technol. 31, 1633–1646. doi:10.1080/10916466.2010.551232
El-Sebakhy, E.A., Hadi, A.S., Faisal, K.A., 2007. Iterative Least Squares Functional Networks Classifier.
IEEE Trans. Neural Networks 18, 844–850. doi:10.1109/TNN.2007.891632
Elkatatny, S., Tariq, Z., Mahmoud, M., Abdulraheem, A., Mohamed, I., 2018. An integrated approach for
estimating static Young’s modulus using artificial intelligence tools. Neural Comput. Appl.
doi:10.1007/s00521-018-3344-1
Elkatatny, S.M., Tariq, Z., Mahmoud, M.A., Al-AbdulJabbar, A., 2017. Optimization of Rate of
Penetration using Artificial Intelligent Techniques, in: 51st US Rock Mechanics/geomechanics
Symposium. American Rock Mechanics Association., San Francisco.
Guo, J., Liu, Y., 2014. A comprehensive model for simulating fracturing fluid leakoff in natural fractures.
J. Nat. Gas Sci. Eng. 21, 977–985. doi:10.1016/j.jngse.2014.10.020
Kazemi, H., Merrill, L.S., Porterfield, K.L., Zeman, P.R., 1976. Numerical Simulation of Water-Oil Flow
in Naturally Fractured Reservoirs. Soc. Pet. Eng. J. 16, 317–326. doi:10.2118/5719-PA
Kazemi, H., Seth, M.S., Thomas, G.W., 1969. The Interpretation of Interference Tests in Naturally
Fractured Reservoirs with Uniform Fracture Distribution. Soc. Pet. Eng. J. 9, 463–472.
doi:10.2118/2156-B
Kennedy, J., 1997. The particle swarm: social adaptation of knowledge, in: Proceedings of 1997 IEEE
International Conference on Evolutionary Computation (ICEC ’97). IEEE, pp. 303–308.
doi:10.1109/ICEC.1997.592326
Khalid, M.A., Alnuaim, S., Rammay, M.H., 2014. Inflow Performance Relationship for Horizontal Wells
Producing from Multi-Layered Heterogeneous Solution Gas-Drive Reservoirs, in: Offshore
Technology Conference-Asia. Offshore Technology Conference. doi:10.2118/24757-MS
Kilns, M.A., Clark, J.W., 1993. An Improved Method To Predict Future IPR Curves. SPE Reserv. Eng.
8, 243–248. doi:10.2118/20724-PA
18 OTC-28367-MS

Klins, M.A., Majcher, M.W., 1992. Inflow Performance Relationships for Damaged or Improved Wells
Producing Under Solution-Gas Drive. J. Pet. Technol. 44, 1357–1363. doi:10.2118/19852-PA
Najurieta, H.L., 1980. A Theory for Pressure Transient Analysis in Naturally Fractured Reservoirs. J. Pet.
Technol. 32, 1241–1250. doi:10.2118/6017-PA
Nooruddin, H.A., Anifowose, F., Abdulraheem, A., 2013. Applying Artificial Intelligence Techniques to
Develop Permeability Predictive Models using Mercury Injection Capillary-Pressure Data, in: SPE
Saudi Arabia Section Technical Symposium and Exhibition. Society of Petroleum Engineers.
doi:10.2118/168109-MS
Retnanto, A., Economides, M.J., 1998. Inflow Performance Relationships of Horizontal and
Multibranched Wells in a Solution- Gas-Drive Reservoir, in: European Petroleum Conference.
Society of Petroleum Engineers. doi:10.2118/50659-MS
Richardson, J.M., Shaw, A.H., 1982. Two-rate IPR Testinga Practical Production Tool. J. Can. Pet.
Technol. 21. doi:10.2118/82-02-01
Sadeghi, M., Shadizadeh, S.R., Ahmadi, M.A., 2013. Determination of Drainage Area and Shape Factor
of Vertical Wells in Naturally Fracture Reservoir with Help Well testing and Developed IPR Curve,
in: North Africa Technical Conference and Exhibition. Society of Petroleum Engineers.
doi:10.2118/164638-MS
Shi, Y., Eberhart, R.C., 1998. {P}arameter selection in particle swarm optimization, in: Evolutionary
Programming. pp. 591–600. doi:10.1007/BFb0040810
Tariq, Z., Al-Hashim, H.S., Sadeed, A., Janjua, A.N., 2016a. A Novel Methodology to Optimise the
Parameters of Hydraulic Fracturing in Gas Condensate Reservoirs, in: International Petroleum
Technology Conference. International Petroleum Technology Conference. doi:10.2523/18919-MS
Tariq, Z., Al-Nuaim, S., Abdulraheem, A., Khan, M.R., 2016b. New Methodology to Quantify
Productivity of Vertical Wells in Naturally Fractured Solution Gas Drive Reservoirs with Dual
Porosity and Dual Permeability, in: PAPG/SPE Pakistan Section Annual Technical Conference and
Exhibition. Society of Petroleum Engineers. doi:10.2118/185314-MS
Tariq, Z., Elkatatny, S., Mahmoud, M., Ali, A.Z., Abdulraheem, A., 2017. A New Technique to Develop
Rock Strength Correlation Using Artificial Intelligence Tools, in: SPE Reservoir Characterisation
and Simulation Conference and Exhibition. Society of Petroleum Engineers. doi:10.2118/186062-
MS
Vasumathi, B., Moorthi, S., 2012. Implementation of hybrid ANN–PSO algorithm on FPGA for harmonic
estimation. Eng. Appl. Artif. Intell. 25, 476–483. doi:10.1016/j.engappai.2011.12.005
Vogel, J.V., 1968. Inflow Performance Relationships for Solution-Gas Drive Wells. J. Pet. Technol. 20,
83–92. doi:10.2118/1476-PA
Wang, J., Zhou, Q., Jiang, H., Hou, R., 2015. Short-Term Wind Speed Forecasting Using Support Vector
Regression Optimized by Cuckoo Optimization Algorithm. Math. Probl. Eng. 2015, 1–13.
doi:10.1155/2015/619178
Warren, J.E., Root, P.J., 1963. The Behavior of Naturally Fractured Reservoirs. Soc. Pet. Eng. J. 3, 245–
255. doi:10.2118/426-PA
Wiggins, M.L., 1993. Generalized Inflow Performance Relationships for Three-Phase Flow, in: SPE
Production Operations Symposium. Society of Petroleum Engineers. doi:10.2118/25458-MS

S-ar putea să vă placă și