See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/287314602

An Integrated Model for History Matching and Predicting Reservoir

Performance of Gas/Condensate Wells

Article  in  SPE Reservoir Evaluation & Engineering · November 2013

DOI: 10.2118/151869-PA

CITATION READS

1 51

3 authors, including:

Ahmed Ibrahim Ahmed El-Banbi

Texas A&M University The American University in Cairo
49 PUBLICATIONS   148 CITATIONS    103 PUBLICATIONS   430 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

CO2 foam for Hydraulic fracturing and EOR applications View project

Waterflood Performance View project

All content following this page was uploaded by Ahmed Ibrahim on 11 September 2018.

The user has requested enhancement of the downloaded file.

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 1 Total Pages: 11

An Integrated Model for History Matching

and Predicting Reservoir Performance of
Gas/Condensate Wells
A.M. Farid, Ahmed H. El-Banbi, SPE, and A.A. Abdelwaly, Cairo University

Summary The CMBE assumes constant reservoir volume, complete equi-

The depletion performance of gas/condensate reservoirs is highly librium between gas and oil throughout the entire reservoir, and
influenced by changes in fluid composition below the dewpoint. immobilizes the oil phase. CMBE can be used to predict the retro-
The long-term prediction of condensate/gas reservoir behavior is grade gas/condensate-reservoir behavior (Exxon 1974). CMBE
therefore difficult because of the complexity of both composition simulates the reservoir performance by use of constant-volume-
variation and two-phase-flow effects. In this paper, an integrated depletion (CVD) -test data with liquid/vapor equilibrium ratios to
model was developed to simulate gas-condensate reservoir/well calculate liquid condensation in the reservoir during pressure
behavior. The model couples the compositional material balance depletion.
or the generalized material-balance equations for reservoir behav- This paper describes a method to match the production data on
ior, the two-phase pseudo integral pressure for near-wellbore the basis of coupling a CMBE or the GMBE for reservoir behav-
behavior, and outflow correlations for wellbore behavior. ior of gas/condensate wells, the two-phase pseudopressure inte-
An optimization algorithm was also used with the integrated gral for near-wellbore behavior, and outflow correlations for
model so it can be used in history-matching mode to estimate wellbore behavior to estimate OGIP, OOIP, and the analytical
original gas in place (OGIP), original oil in place (OOIP), and deliverability constants (A, B). The method can be used as a for-
productivity-index (PI) parameters for gas/condensate wells. The ward model to predict the gas/condensate well performance.
model also can be used to predict the production performance for
variable tubinghead pressure (THP) and variable production rate. Model Development
The model runs fast and requires minimal input.
The developed model was validated by use of different simula- To satisfy the objective of being able to run the coupled model to
tion cases generated with a commercial compositional reservoir match production performance (rate/time performance), or to run it
simulator for a variety of reservoir and well conditions. The results in forward mode to predict the production performance of gas/con-
show a good agreement between the simulation cases and the inte- densate wells, it was necessary to describe the reservoir and well
grated model. After validating the integrated model against the behaviors with several models. The following sections describe the
simulated cases, the model was used to analyze production data for pressure/volume/temperature (PVT) methods, material balance,
a rich-gas/condensate field (initial condensate/gas ratio of 180 bbl/ flow equations, tubing flow correlations, and the optimization rou-
MMscf). THP data for four wells were used along with basic reser- tines that were coupled to achieve the objective.
voir and production data to obtain original fluids in place and PIs of
the wells. The estimated parameters were then used to forecast the Fluid-Property Calculations
gas and condensate production above and below the dewpoint. The Standard PVT correlations and techniques are used to calculate
model is also capable of predicting reservoir pressure, bottomhole the PVT properties required in the CMBE, GMBE, flow equation,
flowing pressure, and THP and can account for completion changes and vertical-lift performance. Gas pseudocritical temperature and
when they occur. pressure are calculated with either Sutton correlations or Kay’s
rules (Tarek 2007). The Dranchuk and Abou-Kasem (1975) corre-
Introduction lation is used to calculate the single-gas-phase z-factor, and the
Rayes et al. (1992) correlation is used to calculate the two-phase
It is a common problem in reservoir engineering that volumetric z-factor.
methods are not adequate for estimating fluid reserves. Also, Modified black-oil PVT properties that are used in the GMBE
buildup tests may be impractical in tight (low permeability) reser- to predict the gas/condensate behavior are the oil and gas forma-
voirs because of the long time required for the pressure to stabi- tion volume factors (Bo, Bg), solution gas/oil ratio (Rs), and the
lize. The conventional material-balance equation cannot be used vaporized oil/gas ratio (Rv). These properties are generated with
to predict gas/condensate-reservoir performance because it does the Walsh and Towler method (Walsh 1995). This method
not account for the change in fluid-flow composition below the requires only data normally available from standard constant-vol-
dewpoint. In these situations, analyzing and history matching pro- ume depletion such as cumulative produced gas, cumulative pro-
duction data seem to be the most practical tools for reserves esti- duced oil, single-phase and two-phase compressibility factors,
mation. Compositional material-balance equations (CMBEs) or and the volume fraction of liquid condensate in the PVT cell dur-
the generalized material-balance equations (GMBEs) should be ing pressure-depletion steps. The method uses mass-balance equa-
used in this situation to predict the reservoir performance. tions and the ideal-gas law to calculate the required four PVT
The GMBE differs from other forms of the conventional mate- properties as functions of pressure at the CVD pressure steps.
rial-balance equation because it includes a term (Rv ) that accounts Vapor/liquid equilibrium ratios are calculated by use of the
for the amount of vaporized oil in the gas phase. Walsh (1994, Whitson and Torp (1983) procedure. Whitson and Torp modified
1995) presented a GMBE that is applicable to the full range of the Wilson (1968) equation to accommodate the compositional
reservoir fluids including gas condensates. effects at high pressures by incorporating the convergence pres-
sure pk (psia). The convergence pressure ( pk) concept is based on
the observation that if we hold a hydrocarbon mixture of a certain
Copyright V
C 2013 Society of Petroleum Engineers
composition at a constant temperature and increase the pressure,
This paper (SPE 151869) was accepted for presentation at the North Africa Technical then the equilibrium values for all its components converge to-
Conference and Exhibition, Cairo, 2022 February 2012, and revised for publication.
Original manuscript received for review 7 March 2012. Revised manuscript received for
ward a common value of unity at a certain pressure called the con-
review 20 February 2013. Paper peer approved 23 April 2013. vergence pressure.

2013 SPE Reservoir Evaluation & Engineering 1

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 2 Total Pages: 11

Material-Balance Equation DmðpÞ ¼ Aqw þ Bq2w ; . . . . . . . . . . . . . . . . . . . . . . . . ð7Þ

The objective of the material-balance equation is to predict gas/
condensate performance with pressure depletion by predicting oil where qw is the wellstream rate that accounts for both gas and oil
saturation in the reservoir, and incremental oil and gas production rate and DmðpÞ is the two-phase pseudopressure integral (Fevang
with pressure drop. Two material-balance equations can be used and Whitson 1995) defined by
to achieve this for gas/condensate wells. !
ðp
Kro Krg
DmðpÞ ¼ Rs þ dp . . . . . . . . . . . . . . ð8Þ
CMBE pwf lo Bo lg Bg
Material balance is made on each component in the fluid system.
The composition and total production of the wellstream are calcu-
lated at various stages of pressure depletion. For further pressure Outflow Two-Phase-Flow Correlations
drop below the dewpoint, the reservoir behavior is described by In case of gas/condensate flow in the tubing, two-phase-flow cor-
relations can be used to calculate pressure drop along the produc-
Xi¼n Xi¼n nt Z  0:5DnGwp yi;ð j1Þ
y ¼  i;ð j1Þ  ¼1 tion tubing. The Hagedorn and Brown (1965) correlation and
i¼1 i; j i¼1 1 Beggs and Brill (1973) correlation are used in the model to calcu-
nL; j  1 þ nt; j1  0:5DnGwp
Ki; j late pressure drop in the production tubing.
Hagedorn and Brown (1965) developed their method by
                   ð1Þ
obtaining experimental pressure drop and flow-rate data from a
10:73znv T nL MWo 1,500-ft-deep instrumented well. Pressures were measured for
VR ¼ þ . . . . . . . . . . . . . . . . . . . . . ð2Þ flow in tubing sizes ranging from a 1 1/4- to 2 7/8-in. outside di-
P qo
ameter. A wide range of liquid rates and gas/liquid ratios was
where DnGwp is the number of wet-gas-produced moles for pres- included, and the effects of liquid viscosity were studied with
sure drop Dp and VR is the reservoir-fluid volume at dewpoint. water and oil as the liquid phase. The oils used had viscosities at
The solution of Eqs. 1 and 2 requires a trial-and-error process to stock-tank conditions of 10, 35, and 110 cp.
calculate DnGwp ; an incremental number of moles of wet gas pro- Beggs and Brill (1973) developed their correlation from exper-
duced for a certain Dp and nL ; and a number of moles in liquid imental data obtained in a small-scale test facility. The facility
phase in the reservoir at pressure step. The production calculated consisted of 1- and 1.5-in. sections of acrylic pipe 90 ft long. The
by Eqs. 1 and 2 is then converted to stock-tank-oil and gas pro- pipe could be inclined at any angle. The parameters studied and
duction by flashing the produced fluid through field-separator con- their ranges of variation were
ditions to stock-tank conditions. Oil saturation in the reservoir can  Gas-flow rate (0 to 300 Mscf/D)
be calculated by use of Eq. 3. These calculations are explained in  Liquid-flow rate (0 to 30 gal/min)
more detail in Appendix A.  Average system pressure (35 to 95 psia)
 Pipe diameter (l and 1.5 in.)
nL MWo
 Liquid holdup (0 to 0.870)
qo  Pressure gradient (0 to 0.8 psi/ft)
so ¼ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð3Þ
VR  Inclination angle (90 to þ90 )
Fluids used were air and water. For each pipe size, liquid and gas
rates were varied so that all flow patterns were observed when the
GMBE (Beggs and Brill 1973) pipe was horizontal. For a particular set of flow rates, the angle of
pipe was varied through the range of angles so that the effect of
Material balance is made on each phase in the gas/condensate sys-
angle on holdup and pressure gradient could be observed. Liquid
tem. The GMBE for volumetric reservoirs can be written as
holdup and pressure gradient were measured at angles from hori-
Np ½Bo ð1Rp  Rs Þ þ Bg ðRp  Rs Þ ¼ N½Rsi ðBg  Bo Rv Þ zontal of 0, plus and minus 5, 10, 15, 20, 35, 55, 75, and 908. The
correlations were developed from 584 measured tests. The Begg-
þ ðBo  Bg Rs Þ  Boi ð1  Rs Rv Þ;               ð4Þ and-Brill method requires the determination of the flow pattern
that would exist in the pipeline if the pipe were horizontal. Differ-
where Rp is the cumulative produced gas/oil ratio. ent equations are used to calculate liquid holdup for each flow
The oil saturation as a function of the fractional oil recovery pattern.
Np =N and the instantaneous gas/oil ratio R are given by Eqs. 5
and 6, respectively:
Automatic History Matching
2  3
Np The model is coupled with an optimization routine to estimate
1 Bo Bg  Boi Bo Rv
6 N 7 OGIP, OOIP, and PI parameters for gas/condensate reservoirs that
So ¼ ð1  Swi Þ6
4
7 . . . . . . . . . ð5Þ
5 give the best match with the observed well-production history.
Boi ðBg  Bo Rv Þ
The analysis is carried out by assuming values for OGIP, OOIP,
and the analytical deliverability constants (A, B) and calculating
Bo lo Krg the total performance (gas rate and oil rate vs. time) for the system.
þ Rs
Bg lg Kro The error in the model performance can then be quantified by
R¼ . . . . . . . . . . . . . . . . . . . . . . . . . ð6Þ     
Bo lo Krg
Rv þ 1 1 Xn qgobserved  qgmodel  Xn qoobserved  qomodel 
Bg lg Kro Error ¼ 1  þ 1  ;
2n qgobserved qoobserved
Eqs. 4, 5, and 6 can be solved by use of Tarner’s solution pro-                    ð9Þ
cedures (Tarner 1944) to estimate Np , Gp , and So . The calculation
procedure is detailed in Appendix A. where n is the number of total data points used in the calibration
period.
This error can be treated as an objective function for a multi-
Gas/Condensate-Flow Equation variable optimization routine. The optimization process proposed
The integrated model uses the flow equation given by Eq. 7. It in this model is based on a genetic algorithm. Genetic algorithms
represents the boundary-dominated analytical solution of the dif- are one of the most common artificial-intelligence techniques for
fusivity equation for gas flow in a reservoir that is producing optimization. This algorithm searches for the values of OGIP,
against constant bottomhole flowing pressure: OOIP, and the analytical deliverability constants (A, B) respecting

2 2013 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 3 Total Pages: 11

Start

Read reservoir fluid properties,

production data, and deliverability data

Estimate OGIP, OOIP, A, B

Better estimate through

Run the performance prediction model
genetic algorithm subroutine

Match the
Y measured N
data

Run the performance prediction model

for prediction

End

Fig. 1—Flow diagram for the procedure of estimation of OOIP, OGIP, and analytical deliverability constants.

the constraints imposed that are the maximal and minimal values. ignores the transient flow, which is introduced by the variation in
The constraints are defined as input variables to the genetic algo- production rate and/or bottomhole pressure. Practically, and for
rithm that explore the best combinations that minimize the error many gas reservoirs, those transients are usually small unless the
between the observed data and the rates calculated by the model. permeability of the reservoir is very low, so this methodology is
The procedure of the genetic algorithms is presented in Appendix applicable in high-permeability gas/condensate wells and reser-
B. Fig. 1 shows a flow diagram for the procedure of estimation of voirs. Also, in applying the model, we recommend the use of a
OOIP, OGIP, and analytical deliverability constants (A, B). calibration period [similar to the suggestion made by El-Banbi
This automatic-history-matching method differs from the auto- and Wattenbarger (1996)] for history matching after transient
matic-history-matching techniques for material balance per- effects are no longer important.
formed by Havlena and Odeh, Caret, and Sills (1996). Other The integrated model used in this work assumes that each well
methods are based on a straight-line solution for the material-bal- is treated individually, similar to the assumption in most decline-
ance equation and will require static-pressure measurements in curve-analysis procedures. This assumption comes from the use
addition to production data. The method presented here uses pro- of the chosen flow equation that assumes single-well flow. The
duction data and will not require estimates of static reservoir pres- derived parameters for the material-balance equation (OGIP and
sures to arrive at OGIP and OOIP. Also, other methods determine OOIP) will therefore give the values for the well’s drainage area.
the fluid in place and then need to use some form of an inflow per- Interference effects are ignored in this calculation.
formance relationship (IPR) relation or flow equation to provide a
production forecast. This method, however, can provide the pro-
Hypothetical-Cases Applications
duction forecast directly after the original fluids in place and flow
coefficients are known by use of the same calculation procedure To validate the integrated model, it was tested against the produc-
as in history matching. tion data of several hypothetical cases generated by a commercial
compositional reservoir simulator. The model was tested against a
variety of cases, and two of these cases are illustrated next, to val-
Model Applicability idate the model. The reservoir properties and PVT data that were
This model methodology is based on a boundary-dominated flow used to construct the hypothetical cases are presented in Tables 1
equation to describe the fluid flow in the reservoir. The equation and 2, respectively. Fig. 2 shows the gas and oil relative perme-
ability, and Fig. 3 shows the change of K-values with pressure.
TABLE 1—RESERVOIR PROPERTIES FOR THE
HYPOTHETICAL CASES Hypothetical Case 1
The verification was conducted with bottomhole flowing pressure
Data Value Unit and oil-, and gas-production rates vs. time from simulator results
as input data for the integrated model. The integrated model was
Initial reservoir pressure 5,800 psia run in history-matching mode to obtain the reservoir parameters.
Dewpoint pressure 5,450 psi Finally, these parameters were used to predict the future perform-
Reservoir temperature 215 8F ance. In this case, the well was assumed to produce at a constant
Reservoir depth 12,800 ft gas-production rate of 3.5 MMscf/D. Table 3 presents the initial
Stock-tank-oil gravity 36 8API guess for the reservoir parameters to start history matching. Figs.
Reservoir permeability 100 md 4 through 6 show the differences between the simulator and the
Reservoir porosity 0.30 fraction model results by use of initial guess for the reservoir parameters.
The optimization technique then was used to minimize the error
Reservoir thickness 30 ft
between the history data and the model. Fig. 7 shows the evolu-
Reservoir area 368 acres
tion of cumulative error function along the history-matching

2013 SPE Reservoir Evaluation & Engineering 3

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 4 Total Pages: 11

TABLE 2—THE PVT DATA USED IN THE HYPOTHETICAL CASES

P, psi Rv , STB/scf Rs , scf/STB Bo , bbl/STB Bg , cf/scf lo , cp lg , cp

5,450 0.000166 6,042 4.468 0.004149 0.0587 0.0587

5,420 0.000164 2,795 2.378 0.004155 0.135 0.0581
5,300 0.000157 2,750 2.366 0.004172 0.1338 0.0554
4,800 0.000114 2,128 2.032 0.004256 0.1826 0.0436
4,300 0.000089 1,730 1.828 0.004458 0.2354 0.0368
3,800 6.5210–5 1,422 1.674 0.004795 0.3001 0.0308
3,300 4.8310–5 1,177 1.554 0.005317 0.3764 0.0261
2,800 0.000036 960 1.448 0.00612 0.4781 0.0222
2,300 0.000025 776 1.36 0.007372 0.6041 0.0191
1,800 0.000019 607 1.279 0.009416 0.7746 0.0166
1,300 0.000015 443 1.2 0.013004 1.0295 0.0148
800 1.3510–5 293 1.131 0.020747 1.358 0.0135

kro krg K-c1 K-c2 K-c3 K-ic4 K-nc4 K-ic5

1.2 K-nc5 K-c6 K-N2 K-CO2 K-H2S K-c7+
10

0.8
1
K value
0.6

0.4 0.1

0.2

0.01
0 0 1,000 2,000 3,000 4,000 5,000 6,000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Pressure (psi)
Oil Saturation
Fig. 3—The change of K-values with pressure.
Fig. 2—Gas and oil relative permeability.

700
Model Results
TABLE 3—INITIAL GUESS FOR HISTORY MATCHING
600 Actual Data

Parameter Value
Oil Rate (STB/day)

500

OGIP, Bscf 9.6 400

OOIP,million STB 1.59
A 6237 300

B 0.0423
200

100

0
4,000 Model Results 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000
Actual Data Time (day)
3,500
Gas Rate (Mscf/day)

3,000 Fig. 4—Oil rate vs. time (initial guess) for Hypothetical Case 1.
2,500

2,000 iterations. Table 4 presents the final results after history match
and comparison with the correct values used in simulation. The
1,500 simulator and the program results (both history and prediction)
1,000 are plotted together in Figs. 8 through 10, showing that a good
match has been achieved.
500

0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 Hypothetical Case 2
Time (day) This case is used to investigate the effect of transient-flow periods
that were induced from varying the producing conditions. The
Fig. 5—Gas rate vs. time (initial guess) for Hypothetical Case 1. well was assumed to start production at a constant gas-production

4 2013 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 5 Total Pages: 11

7,000 1.2
Model Results

Bottomhole Flowing Pressure (psi)

6,000 Actual Data 1

5,000
0.8

Error Function
4,000
0.6
3,000
0.4
2,000

0.2
1,000

0 0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 1 2 3 4 5 6 7 8 9 10

Time (day) Trail Number

Fig. 6—Bottomhole flowing pressure vs. time (initial guess) for Fig. 7—The evolution of error function along the history
Hypothetical Case 1. matching.

700
TABLE 4—HISTORY-MATCHING RESULTS FOR CASE 1 Model Results
600 Actual Data
Parameter Actual Model Error %

Oil Rate (STB/day)

500
OGIP, Bscf 35.37 35.8 1.2
OOIP, million STB 5.854 5.925 1.2 400

A 2678 2724 1.7 300

B 0.0184 0.01866 1.4
200

100

4,000 Model Results 0

Actual Data 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000
3,500
Time (day)
3,000
Gas Rate (Mscf/day)

2,500
Fig. 8—Oil-production rate vs. time (history matching and pre-
diction), Case 1.
2,000

1,500
7,000
Bottomhole Flowing Pressure (psi)

1,000
Model Results
500 6,000 Actual Data

0 5,000
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000
Time (day) 4,000

Fig. 9—Gas-production rate vs. time (history matching and pre- 3,000
diction), Case 1.
2,000

1,000
4,500
Actual Data
0
Model Results (GMBE)
Gas-Production Rate (Mscf/day)

4,000 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000
Model Results (CMBE)
3,500 Time (day)

3,000 Fig. 10—Bottomhole flowing pressure vs. time (history match-

2,500
ing and prediction), Case 1.

2,000

1,500 rate of 3.5 MMscf/D for 2,000 days; then the production rate was
1,000 lowered to 1.5 MMscf/D. Similar to the previous case, the simula-
tor and the model results were plotted together in Figs. 11
500 through 18, showing that a good match was achieved. As shown
0 in these figures, the program forecasts the well performance very
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 well and was able to handle the transient-flow periods. Slight dif-
Time (day) ferences were observed in the case of CMBE, and they are
expected to be a result of the error arising from the use of correla-
Fig. 11—Gas-production rate vs. time (Case 2). tions to compute K-values.

2013 SPE Reservoir Evaluation & Engineering 5

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 6 Total Pages: 11

700 0.3
Actual Data Actual Data
Model Results (GMBE) Model Results (GMBE)
Oil-Production Rate (STB/day) 600 Model Results (CMBE) Model Results (CMBE)
0.25

Oil Saturation (fraction)

500
0.2
400

0.15
300

0.1
200

100 0.05

0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000
Time (day)
Time (day)
Fig. 12—Oil-production rate vs. time (Case 2). Fig. 13—Oil saturation vs. time (Case 2).

40,000,000 2,000,000
Actual Data Actual Data
Cumulative Gas Production (Mscf)

Cumulative Oil Production (STB)

Model Results (GMBE) 1,800,000 Model Results (GMBE)
35,000,000
Model Results (CMBE) Model Results (CMBE)
1,600,000
30,000,000
1,400,000
25,000,000
1,200,000

20,000,000 1,000,000

800,000
15,000,000
600,000
10,000,000
400,000
5,000,000
200,000

0 0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000
Time (day) Time (day)

Fig. 14—Cumulative gas production vs. time (Case 2). Fig. 15—Cumulative oil production vs. time (Case 2).

Field-Case Applications field data. The field has several fault blocks, with two blocks exhib-
Field Description. A high-pressure high-temperature offshore gas iting a large variation in fluid properties. Three wells were produced
condensate field producing with high condensate/gas ratio (initial from one block, and the last one is produced from the second block.
CGR of 180 bbl/MMscf) was used in this work (El-Banbi 2010). Two fluid samples were collected and analyzed and show a large
Four wells were drilled from one platform to delineate and produce difference in PVT properties in the different fault blocks. Table 6
the field. Gas, condensate, and water are separated on the platform shows the initial fluid composition for the two samples. Wells X-1,
where they were metered and then mixed together and shipped X-2, and X-3 produce from the first block with fluid represented by
through a pipeline to a gas plant for further processing. The fluids the first sample. Well X-4 produces from the second block with fluid
were separated once again at the gas plant, and the gas was proc- represented by the second sample. A gas/water contact was
essed to strip out natural-gas liquids. Table 5 shows some basic observed in Well X-3, and field performance confirmed the contact.

100 7,000
Actual Data Actual Data
Average Reservoir Pressure (psi)

90 Model Results (GMBE) Model Results (GMBE)

Model Results (CMBE) 6,000 Model Results (CMBE)
80

70 5,000
GOR (Mscf/STB)

60
4,000
50
3,000
40

30 2,000
20
1,000
10

0 0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000
Time (day) Time (day)

Fig. 16—Instantaneous gas/oil ratio vs. time (Case 2). Fig. 17—Average reservoir pressure vs. time (Case 2).

6 2013 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 7 Total Pages: 11

7,000
Actual Data
TABLE 5—BASIC FIELD DATA (El-Banbi 2010)

Bottomhole Flowing Pressure (psi)

Model Results (GMBE)
6,000 Model Results (CMBE)
Lithology Sandstones
5,000 (shaly sand)

4,000 Depth 15,300 ft

Initial reservoir pressure 14,200 psia
3,000
Reservoir temperature 310 F
2,000 Porosity variation interpreted from logs less than 5% to 14%
Permeability 44 md
1,000 Average net-pay thickness 40 ft
observed on logs
0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000
Time (day)

Fig. 18—Bottomhole flowing pressure vs. time (Case 2). TABLE 7—HISTORY-MATCHING RESULTS FOR WELL X-1

Parameter Value

TABLE 6—INITIAL FLUID COMPOSITION FOR OGIP (Bscf) 71.75

FLUID SAMPLES OOIP (thousand STB) 12.3
A ¼ laminar-flow coefficient (psia2/cp)/ (MMscf/D) 5,824
The The
B ¼ turbulent-flow coefficient (psia2/cp)/(MMscf/D)2 0.033
First-Sample Second-Sample

Component Mole Fraction Mole Fraction

actual data and the integrated-model results. In Fig. 19, it is
C1 0.6673 0.4833 observed that a reasonable match was achieved between the actual
C2 0.1022 0.1812
and the program THP, except for the first period. It is expected
that some change in the completion occurred after the first 100
C3 0.059 0.109
days, but details about completion change were not available to
n-C4 0.021 0.021 the authors. Nevertheless, the match is excellent in the following
i-C4 0.0188 0.0188 period of time.
n-C5 0.0083 0.0083 In Well X-2, history-matching was performed for 19 days of
i-C5 0.0137 0.0137 production. Table 8 presents the final results after the history
C6 0.0156 0.0186 matching. These results may not be representative for a long-term
N2 0.0014 0.0014 performance of the reservoir because of a short available produc-
CO2 0.0279 0.0239 tion history.
C7þ 0.0648 0.1208 The constructed program is not applicable in the case of Well
C7þ molecular weight, lbm mol 152.2 160
X-3 because water production was observed.
In Well X-4, history-matching was performed for 180 days of
C7þ specific gravity 0.7986 0.81
production. Table 9 presents the final results after performing the
match. Figs. 22 through 24 show the match between the actual
Analysis of Production Data. Separate well models were data and the program results.
assumed for the three wells with production. The production data
amounted to almost 1 1/2 years of production from these wells. Discussion
The history match was achieved through specifying the gas-pro- The integration of reservoir, near wellbore, and tubing flow to
duction rate and matching condensate production for each well. predict the production performance of gas/condensate wells is im-
THP data were available for the wells and were also matched. portant. The model proposed in this paper was successfully used
Vertical-lift performance was calculated by use of the Hagedorn to predict both gas- and oil-production rates from gas/condensate
and Brown (1965) correlation. wells. The model can also take into consideration any variation of
In Well X-1, history matching was performed for 150 days of
production. Table 7 presents the final results after performing his-
70,000
tory matching. Figs. 19 through 21 show the match between the Program Results
Prediction
60,000 Production History
13,000
Production History
Gas Rate (Mscf/day)
Tubing Head Pressure (psi)

12,000 Program Results 50,000

11,000
40,000
10,000

9,000
30,000
8,000

7,000 20,000

6,000
10,000
5,000

4,000 0
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Time (day) Time (day)

Fig. 19—THP vs. time (Well X-1) Fig. 20—Gas rate vs. time (Well X-1).

2013 SPE Reservoir Evaluation & Engineering 7

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 8 Total Pages: 11

11,000 Program Results

Prediction TABLE 8—HISTORY-MATCHING RESULTS FOR WELL X-2
10,000
Production History
9,000
Parameter Value
8,000
OGIP (Bscf) 19.27
Qo (STB/day)

7,000
OOIP (thousand STB) 2.8
6,000
A ¼ laminar-flow coefficient (psia2/cp)/ (MMscf/D) 14,354
5,000
B ¼ turbulent-flow coefficient,(psia2/cp)/(MMscf/D)2 0.14
4,000

3,000

2,000
10,000 Program Results
1,000
Production History
9,000
0

Tubing Head Pressure (psi)

0 50 100 150 200 250 300 350 400 8,000
Time (day)
7,000

Fig. 21—Oil rate vs. time (Well X-1). 6,000

5,000

4,000

TABLE 9—HISTORY-MATCHING RESULTS FOR WELL X-4 3,000

2,000
Parameter Value
1,000
OGIP (Bscf) 24.32 0
OOIP (thousand STB) 13.19 0 20 40 60 80 100 120 140 160 180 200

A ¼ laminar-flow coefficient (psia2/cp)/(MMscf/D) 30,202 Time (day)

B ¼ turbulent-flow coefficient,(psia2/cp)/(MMscf/D)2 0.85
Fig. 22—THP vs. time (Well X-4).

bottomhole pressure (or surface THP) and variation of production lent gas produced and its variations) assumes that the condensate
rates. produced on surface comes solely from the gas in the reservoir,
The flow equation uses the integral form of the pseudopressure and, therefore, computes a modified-gas specific gravity for the
(Fevang and Whitson 1996). The assumption in the model is that reservoir gas. It then uses the z-factor on the basis of the reser-
the condensate forms a condensate bank around the wellbore, and voir-gas specific gravity. The gas-equivalent method does not
it does not move. The condensate produced is mainly coming account for the condensation of liquids in the gas/condensate res-
from the vaporized condensate in the gas phase. This assumption ervoir below the dewpoint.
seems to be reasonable for the vast majority of gas/condensate Although not tested here, the model is expected to be able to
reservoirs (Tarek 2007). handle cases with water influx in the reservoir if the material-bal-
In the hypothetical cases used for testing the model, we ance equation uses the form with water influx. In this situation,
observed that the GMBE model was consistently better than the additional model-to-computer water influx at every pressure step
CMBE model in matching the simulated data. This is expected to will be used to calculate the water influx at every timestep (on the
be because the CMBE model uses correlations to estimate K-val- basis of pressure drop in the reservoir), and water influx will be
ues, whereas the generated data from the compositional simulator added in the material-balance calculation. This extension of the
used an equation of state (EOS). We recommend the use of an model is beyond the scope of this paper and was not tested.
EOS to generate K-values to obtain better results with the CMBE
model. The use of EOS-generated K-values was not tested in this
work. Conclusions
The model suggested here is different from the use of the gas- The following conclusions were derived from this work:
equivalent method that we normally use for wet-gas and light-gas • A practical model coupling CMBE or GMBE for reservoir
condensate reservoirs. The gas-equivalent method (p/z vs. equiva- behavior, the two-phase pseudopressure integral for near-

20,000 11,000 Program Results

Program Results
18,000 10,000
Production History
Production History
9,000
16,000
8,000
Gas Rate (Mscf/day)

Qo (STB/day)

14,000
7,000
12,000
6,000
10,000
5,000
8,000
4,000
6,000 3,000
4,000 2,000
2,000 1,000

0 0
0 20 40 60 80 100 120 140 160 180 200 0 20 40 60 80 100 120 140 160 180 200
Time (day) Time (day)

Fig. 23—Gas rate vs. time (Well X-4). Fig. 24—Oil rate vs. time (Well X-4).

8 2013 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 9 Total Pages: 11

wellbore behavior, and outflow correlations for wellbore behav- the North Africa Technical Conference and Exhibition, Cairo, Egypt,
ior to predict the gas/condensate well performance was 14–17 February. http://dx.doi.org/10.2118/128448-MS.
developed. Exxon. 1974. Reservoir Engineering Manual. Chapter 3. Houston, Texas:
• An optimization algorithm was also used with the integrated Production Research Company.
model so it can be used in history-matching mode to estimate Fevang, Ø. and Whitson, C.H. 1996. Modeling Gas-Condensate Well
OGIP, OOIP, and well productivity constants. Deliverability. SPE Res Eng 11 (4): 221–230. http://dx.doi.org/
• The model also can be used to estimate, on a time basis, future 10.2118/30714-PA.
gas-production rate, future oil-production rate, cumulative gas Hagedorn, A.R. and Brown, K.E. 1965. Experimental Study of Pressure
production, cumulative oil production, reservoir oil saturation, Gradients Occurring During Continuous Two-Phase Flow in Small Di-
average reservoir pressure, bottomhole pressure ( pwf ), and THP ameter Vertical Conduits. J. Pet Tech 17 (4): 475–484. http://
with different production control modes (constant production dx.doi.org/10.2118/940-PA.
rate, constant pwf , or constant THP) for every timestep. Havelena, D., Hudson’s Bay Oil & Gas Co. Ltd., Odeh, A.S., and Socony
• The integrated model can handle variation in bottomhole pres- Mobil Oil Co. 1963. The Material Balance as an Equation of a Straight
sure for moderately-high-permeability reservoirs in which tran- Line. J. Pet Tech 15 (8): 896-900. http://dx.doi.org/10.2118/559-PA.
sient effects can be ignored. Rayes, D.G., McCain, W.D. Jr., and Poston, S.W. 1992. Two-Phase Com-
• The application of the model on an actual field case shows a pressibility Factors for Retrograde Gases. SPE Form Eval 7 (1):
good match with the actual performance of three wells and pro- 87–92. http://dx.doi.org/10.2118/20055-PA.
vides an estimate of the reservoir gas in place, oil in place, and Sills, S.R. 1996. Improved Material-Balance Regression Analysis for
deliverability constants. Waterdrive Oil and Gas Reservoirs. SPE Res Eng 11 (2): 127–134.
http://dx.doi.org/10.2118/28630-PA.
Tarek, Ahmed. 2007. Equations of State and PVT Analysis. Chapter 2.
Nomenclature
Houston, Texas: Gulf Publishing Company.
A ¼ the laminar-flow coefficient, ðpsia2 =cpÞ=ðMMscf=DÞ Tarner, J. 1944. How Different Size Gas Caps and Pressure Maintenance
B ¼ the turbulent-flow coefficient, ðpsia2 =cpÞ=ðMMscf=DÞ2 Programs Affect Amount of Recoverable Oil. Oil Weekly 144: 32–34.
Bgi ¼ gas formation volume factor (FVF), bbl/Mscf Walsh, M.P. 1994. The New, Generalized Material Balance as an Equation
Boi ¼ oil FVF, bbl/STB of a Straight Line: Part 1—Applications to Undersaturated, Volumetric
Gi ¼ original gas in place, MMscf Reservoirs. Paper SPE 27684 presented at the SPE Permian Basin Oil
Gp ¼ cumulative dry gas at any pressure p, MMscf and Gas Recovery Conference, Midland, Texas, 16–18 March. http://
ki ¼ vapor/liquid equilibrium ratio of component i dx.doi.org/10.2118/27684-MS.
krg ¼ gas relative permeability Walsh, M.P. 1995. A Generalized Approach to Reservoir Material Balance
kro ¼ oil relative permeability Calculations. J. Cdn. Pet. Tech. 34 (1). http://dx.doi.org/10.2118/95-
Mwo ¼ molecular weight of the stock-tank liquid, lbm/lbm mol 01-07-PA.
N ¼ original oil in place, STB Whitson, C.H. and Torp, S.B. 1983. Evaluating Constant-Volume Depletion
nLj ¼ number of moles in reservoir liquid phase at any Data. J. Pet Tech 35 (3): 610–620. http://dx.doi.org/10.2118/10067-PA.
pressure Wilson, G. 1968. A Modified Redlich–Kwong EOS, Application Physical
Np ¼ cumulative oil produced at any pressure p, STB Data Calculation. Paper 15C presented at the annual AICHE National
nt ¼ total number of moles in reservoir fluid (gas plus liquid) Meeting, Cleveland, Ohio, 4–7 May.
phases at initial pressure
Pwf ¼ bottomhole flowing pressure, psi
qg ¼ flowing gas rate, Mscf/D Appendix A—Fluid-Flow Property Correlations
qo ¼ flowing oil rate, STB/D The gas pseudocritical temperature and pressure are calculated
Rp ¼ cumulative produced gas/oil ratio, Mscf/STB with standing (Tarek 2007) correlations.
Rs ¼ solution gas/oil ratio, Mscf/STB
Rsi ¼ initial solution gas/oil ratio, Mscf/STB TPC ¼ 187 þ 330cg  71:5c2g . . . . . . . . . . . . . . . . ðA-1Þ
Rv ¼ vaporized oil/gas ratio, STB/Mscf
Rvi ¼ initial vaporized oil/gas ratio, STB/Mscf PPC ¼ 706  51:7cg  11:1c2g . . . . . . . . . . . . . . . . ðA-2Þ
So ¼ oil saturation, fraction
Swi ¼ initial water saturation, fraction The Dranchuk and Abou-Kasem (1975) Correlation is used to
VR ¼ reservoir fluid volume at dewpoint calculate single gas-phase z-factor:
yi ¼ mole fraction of component i in reservoir gas phase, !
fraction A2 A3 A4 A5
Zi ¼ mole fraction of component i in reservoir fluid (gas plus Z ¼ 1 þ A1 þ þ 3 þ 4 þ 5 qr
liquid) phases, fraction Tpr Tpr Tpr Tpr
DnGwp ¼ number of wet gas-produced moles for pressure drop ! !
A7 A8 2 A7 A8
DP þ A6 þ þ 2 qr  A9 þ 2 q5r
lg ¼ gas viscosity, cp Tpr Tpr Tpr Tpr
lo ¼ oil viscosity, cp   q3  
þ A10 1 þ A11 q2r 3r exp A11 q2r . . . . . . . . . ðA-3Þ
Tpr
References
where A1 ¼ 0.3265, A4 ¼ 0.01569, A7 ¼ 0.7361, A10 ¼ 0.6134,
Beggs, H.D. and Brill, J.P. 1973. A Study of Two-Phase Flow in Inclined A2 ¼ 1.07, A5 ¼ 0.05165, A8 ¼ 0.1844, A11 ¼ 0.721, A3 ¼
Pipes. J. Pet Tech 25 (5): 607–617. http://dx.doi.org/10.2118/4007- 0.5339, A6 ¼ 0.5475, A9 ¼ 0.1056, cg ¼ gas SG, and qr ¼
PA. reduced pseudoreduced density, and the Rayes et al. (1992) corre-
Dranchuk, P.M. and Abou-Kasem, J.H. 1975. Calculation of Z-Factors for lation is used to calculate the two-phase z-factor:
Natural Gases Using Equations of State. J. Pet Tech 14 (3). http://
dx.doi.org/10.2118/75-03-03.
A2 A4 A5 PPr
El-Banbi, Ahmed H., and Wattenbarger, Robert A. 1996. Analysis of Z ¼ A0 þ A1 PPr þ þ A3 p2pr þ 2 þ ; . . . . . ðA-4Þ
Commingled Tight-Gas Reservoirs. Paper SPE 36736 presented at the Tpr Tpr Tpr
1996 SPE Annual Technical Conference and Exhibition, Denver, Col-
orado, 6–9 October. http://dx.doi.org/10.2118/36736-MS. where A0 ¼ 2.24353, A1 ¼ 0.0375281, A3 ¼ 0.000829231, A2
El-Banbi, A.H. 2010. Optimizing Simulation Studies for Gas Condensate ¼ 3.56539, A4 ¼ 1.53428, A5 ¼ 0.131987, PPr ¼
Field Development and Management. Paper SPE 128448 presented at pseudoreduced pressure, and Tpr ¼ pseudoreduced temperature.

2013 SPE Reservoir Evaluation & Engineering 9

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 10 Total Pages: 11

Outflow Calculation Equations 4. The total reservoir gas produced, DnGwp , is assumed; then
The general pressure-gradient equation, which will apply to the solve for nLj ,
flow of any fluid in a pipe at any inclination angle, was given as
Xi¼n Xi¼n nt Z  0:5DnGwp yi;ð j1Þ
y ¼  i;ð j1Þ  ¼ 1:
dp g f q v2 i¼1 i; j i¼1 1
¼ qm cos þ m m ; . . . . . . . . . . . . . . . . . . . ðA-5Þ nL; j  1 þ nt; j1  0:5DnGwp
dL gc 2gc d Ki; j
where qm ¼ two-phase density is calculated by use of the follow-                    ðA-13Þ
ing equation:
I. Assume initial guess for nL ; then calculate yi for each
qm ¼ qL HL þ qg ð1  HL Þ . . . . . . . . . . . . . . . . . . . ðA-6Þ component. Xi¼n
II. Check that y ¼ 1.
i¼1 i
and vm ¼ two-phase or mixture velocity can be calculated by the ai¼n
i¼1 yi 1
following equation: III. If not, then nL ¼ nL  ai¼n
 dy
.
i¼1 i

qg þ qL IV. Repeat until obtaining reasonable accuracy,

vm ¼ ¼ vsl þ vsg ; . . . . . . . . . . . . . . . . . . . . ðA-7Þ where yi ¼ mole fraction of component i in reservoir gas phase
A at Pj , Zi ¼ mole fraction of component i in reservoir fluid (gas
where vsl ¼ liquid superficial velocity, vsg ¼ gas superficial veloc- plus liquid) phases at Pj1 , ntð j1Þ ¼ total number of moles in res-
ity, A ¼ pipe cross-sectional area, and HL ¼ liquid holdup can be ervoir fluid (gas plus liquid) phases at Pj1 , DnGwp ¼ number of
calculated by use of the Hagedorn and Brown (1965) and Beggs gas produced moles for pressure drop from Pj1 to Pj , nLj ¼ num-
and Brill (1973) correlations. ber of moles in reservoir liquid phase at Pj , and Ki; j ¼ equilibrium
ratio of component i at Pj .
Tarner’s Solution Procedures 5. With yi , calculate the corresponding xi :
1. Assume incremental reservoir pressure drop ¼ Dp. yi
xi ¼ ;
2. For the current reservoir pressure p, obtain the correspond- Ki
ing PVT data.
1. Assume incremental oil production (DNp ) corresponding to where xi ¼ mole fraction of component i in reservoir liquid phase
Dp, and calculate the current cumulative oil production, so at Pj .
6. For yi and xi and nL and VR , obtain the calculated DGwp ,
Npj ¼ Npj1 þ DNpj : . . . . . . . . . . . . . . . . . . . . . . . ðA-8Þ nLMWo
VR 
2. Calculate oil saturation So with Eq. 5. qo
DnGwpcal ¼ nt  nL  P ; . . . . . . . . . . . . . ðA-14Þ
3. Calculate instantaneous gas/oil ratio R. 10:73zT
 If So < Soc (critical oil saturation), the reservoir oil is not
mobile. Calculate instantaneous gas/oil ratio with the following where MWo ¼ molecular weight of condensed fluid at P (lbm/lbm
equation: mol), and qo ¼ density of condensed fluid at P (lbm/ft3 ).
7. Check if (jDnGwpcal  DnGwp assumedj) is within a certain
R ¼ 1=Rv : accuracy.
8. If not, then use the DnGwpcal and repeat Calculation Steps 4
If So > Soc (critical oil saturation), then reservoir oil is mobile, through 7.
and the instantaneous gas/oil ratio can be calculated with Eq. 6. 9. Calculate oil saturation on the reservoir (so ),
4. Calculate average gas/oil ratio (Ravg ) between the current
nL MWo
step and the preceding one.
5. Calculate incremental gas production (DGp ) corresponding qo
so ¼ :
to DNp , VR

DGp ¼ DNp Ravg ; . . . . . . . . . . . . . . . . . . . . . . . . . .ðA-9Þ 10. Calculate wellstream fluid composition (Xfsi ); assume it is
an average of the gas composition on two sequence pressures,
then calculate the current cumulative gas production, yi; j þ yi;ð j1Þ
Xfsi; j ¼ : . . . . . . . . . . . . . . . . . . . . . ðA-15Þ
2
Gpj ¼ Gpj1 þ DGpj : . . . . . . . . . . . . . . . . . . . . . . ðA-10Þ
11. Perform flash calculation for the wellstream fluid (DGwp
6. Calculate cumulative gas/oil ratio, and Xfsi ) to calculate separator recovery from oil and gas.
Gp
Rp ¼ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ðA-11Þ
Np Appendix B—Procedure of the Genetic
Algorithms
7. Calculate cumulative oil production Np with material bal-
ance in Eq. 4. Then compare it with the guessed value.  [Start] Generate random population of n chromosomes (suit-
8. Repeat the calculation until reasonable accuracy is obtained. able solutions for the problem).
 [Fitness] Evaluate the fitness f(x) of each Chromosome x in
the population.
Calculation of the CMBE  [New population] Create a new population by repeating the
1. From the reservoir volume (VR ), calculate the total number following steps until the new population is complete.
of moles in the reservoir at dewpoint. The total composition (Zi )  [Selection] Select two parent chromosomes from a popula-
at dewpoint is the same as the initial composition, tion according to their fitness (the better the fitness, the greater the
chance to be selected).
VR  Pi  [Crossover] With a crossover probability, cross over the
nt ¼ : . . . . . . . . . . . . . . . . . . . . . . . . . ðA-12Þ
10:73Zi Tf parents to form new offspring (children). If no crossover was per-
formed, offspring is the exact copy of parents.
2. Assume incremental pressure drop ¼ Dp, so P ¼ Pi  DP.  [Mutation] With a mutation probability, mutate new off-
3. For Pj and Zi and Tf , calculate the corresponding K-values. spring at each locus (position in chromosome).

10 2013 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032

 REE151869 DOI: 10.2118/151869-PA Date: 12-September-13 Stage: Page: 11 Total Pages: 11

 [Accepting] Place new offspring in the new population. Ahmed H. El-Banbi is currently a professor of petroleum engi-
 [Replace] Use new generated population for a further run of neering at Cairo University, where he teaches graduate and
the algorithm. undergraduate classes in petroleum engineering and runs an
 [Test] If the end condition is satisfied, stop and return the extensive research program. He has 21 years of diversified inter-
national experience including working for Schlumberger in a
best solution in current population. variety of technical and managerial positions in five countries.
 [Loop] Go to Fitness step. El-Banbi has considerable experience in managing multidisci-
plinary teams and performing integrated reservoir studies. He
has authored or coauthored more than 40 technical papers
Ahmed Farid is currently a teaching assistant in petroleum en-
and two book chapters, and he holds one patent. El-Banbi has
gineering at Cairo University, where he teaches undergradu-
been a member of numerous SPE committees and a technical
ate classes in petroleum engineering and from which he
reviewer for SPE Reservoir Engineering and Evaluation.
earned BS and MS degrees in petroleum engineering. Farid
has more than 4 years of diversified international experience A.A. Abdelwaly is currently a professor of petroleum engineer-
including working for Advantek-International Company in a ing at Cairo University, where he teaches graduate and
variety of technical positions. undergraduate classes.

2013 SPE Reservoir Evaluation & Engineering 11

View publication stats

ID: jaganm Time: 14:49 I Path: S:/3B2/REE#/Vol00000/130032/APPFile/SA-REE#130032