Field Development Planning Optimization Using Reservoir

Simulation
C. C. Mezzomo and D. J. Schiozer
Universidade Estadual de Campinas
([email protected])

Abstract
Field development planning comp rises a great amount of investments and involves a high number of
parameters related to the geological and structural characteristics of the reservoir, to the operational scheduling and the
economic scenario. The importance of this problem demands the elaboration of methodologies that can help in the
management decision making process, leading to better recovery strategies that increase both reserves and profitability
of reservoirs.
This paper presents a methodology applied to the field development planning process. This methodology was
elaborated concerning different types of reservoir and can be adapted to attend different geological characteristics and
the operational constraints. It requires the use of reservoir simulation to obtain reliable predictions for production and
injection performance that will be used in the optimization process.
Some examples were selected for the validation of the proposed methodology and the results are presented. It
can be shown that this methodology is fast and requires a low number of simulation runs. It provides good quality
solutions that will be analyzed by the management that will define a field development strategy according to the
objectives of the project and the technical and financial resources available.

Introduction
Production strategy optimization has a great importance in the oil industry and must be applied to achieve
different objectives. Sometimes, the main purpose of this process is to select an adequate production strategy to be
applied in the reservoir development planning. In other instances, the objective is to utilize a detailed optimization
procedure in order to obtain accurate results to support complex decisions. Another possible objective can be the
optimization of mature fields in order to increase profitability or to adequate the production to a new economic
scenario. The objectives are established by the management regarding the importance of the project and the technical
and economic resources available and the decision making process must lead to lucrative results and high revenues,
considering the physical and operational restrictions for each particular project. Hence, it is very important to develop
new procedures to minimize risks and maximize profits in recovery strategy arrangements.
The use of reservoir simulation is very important to provide reliable production/injection forecast and correct
predictions for field recovery potential. However, during the initial field development phase the amount of available
information for the reservoir is very restricted and it is very difficult to obtain a correct reservoir model. Therefore, the
use of simplified simulation models provides more appropriate and lead to better results.
This work proposes a methodology including a robust optimization procedure that uses the production/injection
forecasts generated by reservoir simulation for the evaluation of an objective-function (NPV). This methodology helps
in the decision making process granting a correct evaluation of relevant parameters in field recovery planning and it
provides adequate solutions using a small number of simulation runs. Some examples based on different offshore
fields were selected in order to validate the methodology and the results are presented. It can be shown the importance
of reservoir simulation in field development planning to determine an adequate amount of producer and injector wells
and propose a suitable scheduling. The procedure can be refined to increase the accuracy of the solutions and can also
be adapted to define production strategies for field development under uncertainty. In this case different strategies are
proposed for each geological model generated.

Literature Review
The planning of adequate recovery strategies for petroleum reservoirs has a great economic importance in oil
industry and several studies have been performed in order to develop efficient procedures for this optimization
problem.
Arps et alli [1967] participated of a study, organized by the American Petroleum Institute, with the objective of
developing equations for the assessment of recovery factors for petroleum fields. The well spacing was one of the most
studied parameters. However, from the analysis of 312 different reservoirs they concluded that there was no
mathematical relationship between recovery factor and well spacing. The purpose of their work was to develop a
methodology for field recovery planning through an analysis preceding the stage involving simulation, but it was not
possible to obtain satisfactory results.
Davis and Shepler [1969] verified that the well spacing initially used to develop a petroleum field, in general,
isn’t the most adequate spacing. The ideal well spacing depends on characteristics of each reservoir. Thus being
necessary to take into account the uncertainties related to the geological model and the dynamic behavior of the
economic and technological scenario.
Reservoir simulation became an important tool for the development and management of petroleum reservoirs.
Accurate reservoir performance predictions can be obtained through numerical simulation using a previously built
geological model that comprises several parameters obtained through reservoir characterization. The simulation model
is the most important tool for the evaluation of an objective-function that represents the global objective of the project.
Using numerical simulation, Nystad [1985], Damsleth et alli [1992], Beckner and Song [1995] among others
authors developed methods for optimization problems related to the development and management of petroleum
reservoirs. These works presented the following common features: They required some previously established
simplifications and the number of simulations runs performed and evaluated parameters was small. The objectives of
such works were the evaluation of the most important parameters in the assessment of the objective-function and their
optimization.
Bittencourt [1997] developed a hybrid algorithm based on direct methods, like genetic algorithm, Polytope
method and Tabu Search, for optimization problems related to petroleum field development. The simulator was used
as a data generator for the evaluation of the objective function, which involved an analysis of cash flows resulting
from production predictions obtained from reservoir simulation.
Pedroso and Schiozer [2000] developed a methodology for the optimization of the number of producer wells
and their location in a reservoir in development stage.
Cruz [1999] introduced the concept of “quality map”, which is a bi-dimensional representation for reservoir
performance and its uncertainties, mainly related to the geological model. This concept can be applied to compare
reservoirs, to classify stochastic realizations and to include geological uncertainties in the decision making process for
recovery strategy planning.
Mezzomo and Schiozer [2000] developed a procedure for primary recovery strategies optimization, comprising
only vertical producer wells. In order to expand the scope of that procedure, compassing a greater number of
reservoirs, this work developed a more flexible and adaptable methodology, including water injection with producer
and injector vertical and horizontal wells.

Methodology
In order to develop a methodology to the optimization problem of recovery strategy planning for different
reservoirs, it is necessary to evaluate several parameters mainly related to the geological model, operational conditions
and economic scenario.
The objective of this work is to present a new methodology to support managers in the decision making process
for water injection planning optimization for fields in development stage and under operational and economic
restrictions.

Methodology Description
The general methodology proposed in this work is organized in several steps that are presented in Figure 1 and
will be described in the following paragraphs.
Depending on the objectives defined for the project and the time available to the decision making process, some
of these steps can be simplified or discarded.

Step 1: Recovery Strategy Assessment

In this first step, a study based on field data is performed for the assessment of relevant geological and physical
reservoir parameter, that will be used for the definition of basic important parameters related to recovery strategy like
well type (producer or injector) and geometry (vertical, horizontal or deviated). This study will also evaluate the
necessity of a secondary recovery (water and/or gas injection) or an enhanced recovery method (thermal, chemical,
etc).
 Step 1: Recovery Strategy Definition

Step 2: Production/Injection Patterns Evaluation

Step 3: Algorithm: Assessment of Total Number of wells

Step 4: Obligatory Optimization

Step 5: Economic Sensitivity Analysis

Fig. 1: Methodology flowchart.

Step 2: Production/injection Patterns Evaluation

The second step comprises an evaluation of several production/injection patterns proposed according to the
recovery method and the well type and geometry established in the previous step. The production/injection patterns are
defined based on field characteristics the technical and financial resources available.
During this stage, simulation runs are performed to access the recovery potential for each pattern proposed for
the field. At this stage all wells comprised by the defined patterns must be opened simultaneously at the initial time of
the simulation runs in order to obtain a correct evaluation of their production/injection performance. The best patterns
are retained and submitted to the optimization procedure in the next step.

Step 3: Estimation of the Approximate Number of Wells

As stated before, the patterns retained in the previous stage are then submitted to an optimization procedure.
The optimization procedure will provide the ideal number of wells for these selected recovery patterns and it must take
into account the relevant reservoir parameters, the number of required simulation runs and the desired quality for the
results
As in the previous stage, throughout this optimization procedure all wells must start to produce/inject
simultaneously in order to obtain a correct evaluation of the potential of each well and a accurate value for their
objective-function. This optimization procedure must be adapted to attend the operational and economic restrictions
imposed to the project. Figure 2 presents a flowchart that summarizes this optimization procedure

Step 4: Obligatory optimization

The optimization process is always restricted to operational restrictions that are mainly defined by the
investments and costs established for the project. These conditions always lead to a mandatory optimization and this
stage defines important parameters, like well scheduling, in order to present feasible solutions that can lead to good
results.
The conditions applied to this obligatory optimization may be proposed according to the characteristics of the
project and the technical resources available.

Step 5: Economic Sensitivity Analysis

The economic scenario presents a very dynamic behavior and is strongly influenced by technical and political
circumstances. Therefore, parameters like oil price tend to show significant variations throughout the project life.
An economic sensitivity analysis is very important for a recovery strategy planning. The use of reservoir
simulation allows an accurate evaluation of the potential of each alternative comprised in the set of recovery patterns
presented by the methodology regarding different feasible economic scenarios, and the objective at this stage of the
methodology is to identify the alternatives that present low financial risk and that are less influenced by economic
scenario alterations.

Initialization: Assessment of the Initial

Well Spacing for the Selected Patterns

Simulation Run

Objective-function Evaluation –
Producer/Injector wells Ranking

Total NPV
increases ?
y
Determination of the new total
n number of wells

End

Fig. 2: Optimization Algorithm

Case Study
Two examples were proposed for the validation of this methodology, based of offshore fields. The first one
presented a solution gas drive mechanism and the second one comprised a strong bottom aquifer. Considering that
they were on initial development stage, the simulation models used were simple due to the restricted amount of
geological and physical data available. They comprised a black-oil system containing three phases: oil, gas and water.
On important aspect of this work is that it is not necessary to perform all the simulation runs executed. They
were performed to validate the methodology.

Results

Selection of the Objective -Function

The net present value (NPV) was defined as the objective-function to be maximized. Two other indicators,
cumulative oil production (Np) and rate of return (RR), defined as NPV/investment, were also analyzed for a more
adequate evaluation of the set of alternatives of water injection patterns. The objective was to identify from graphs,
strategies that allow the simultaneous maximization of oil production, return coefficient and Net Present Value for the
two selected reservoirs.

Step 1:
Considering that the purpose of this work that was to develop a methodology for optimization of recovery
strategies including water injection, it was assumed that a geological study indicated recovery strategies comprising
vertical producer and injector wells.

Step 2:
The water injection patterns evaluated in for the two reservoirs selected for the study case were the most
frequently used for secondary recovery of offshore fields: peripheral (with different ratios between producer and
injector wells), five-spot (5spot), seven spot (7spot), nine spot (9spot), direct linedrive, alternate and combined. The
recovery performance and the values obtained for the objective-function and for the rate of return for the first reservoir
are presented in Figures 3, 4, 5 and 6. The same graphs are presented for the second reservoir in Figures 7, 8, 9 and 10.
 900 900
Well Spacing 450m Well Spacing 450m
800 800 Well Spacing 600m
Well Spacing 600m
Well Spacing 750m
700 Well Spacing 750m 700
NPV (US$ millions)

600

NPV (US$ millions)

600

500
500
400
400
300
300
200
200
100
100
0
0
0 2 4 6 8 10
25 30 35 40 45 50 55
Nproducer/Ninjector
Np (millions m3)

Fig 3: Case A – Step 2 – NPV vs ratio between producer and Fig. 4: Case A: Step 2 – NPV vs Cumulative Oil Production
injector wells

900 900

800 800

700 700
NPV (US$ millions)
NPV (US$ millions)

600 600

500 500

400 400

300 300
Well Spacing 450m
200 Well Spacing 600m 200 Well Spacing 450m
Well Spacing 750m Well Spacing 600m
100 100 Well Spacing 750m

0 0
0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 0,2 0,25 0,3 0,35 0,4 0,45
RR Wp/Wi

Fig. 5: Case A: Step 2 – NPV vs Rate of Return Fig. 6: Case A – Step 2 – NPV vs ratio between Cumulative
Produced Water and Cumulative Injected Water

1000
Well Spacing 450m 1000
900 Well Spacing 600m Well Spacing 450m
900 Well Spacing 600m
800
800
700
NPV (US$ milhões)

700
NPV (US$ millions)

600
600
500
500
400
400
300
300
200
200
100

0
100
0 2 4 6 8 10 0
Nproducer/Ninjector 30 35 40 45 50 55
Np (millions m3)

Fig. 7: Case B: Step 2 – NPV vs Cumulative Oil Production Fig. 8: Case B: Step 2 – NPV vs Cumulative Oil Production
 1000 1000
900 900
800 800
700
NPV (US$ millions)

700

NPV (US$ millions)

600 600
500 500
400 400
300 300
200 Well Spacing 450m
200
Well Spacing 600m
Well Spacing 450m
100 100 Well Spacing 600m
0 0
0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 0,2 0,25 0,3 0,35 0,4 0,45 0,5 0,55
RR Wp/Wi

Fig. 9: Case B: Step 2 – NPV vs Rate of Return Fig. 10: Case B – Step 2 – NPV vs ratio between Cumulative
Produced Water and Cumulative Injected Water

These figures indicate the efficiency of the injection patterns for the selected field, with different well spacing
and they also allow the comparison of objective-function values obtained with their application. This is very important
in the decision making process for the planning of a recovery strategy.
From these graphs it can be shown that for both reservoirs, the most efficient patterns presented a ratio between
producer and injector wells close to 1. The initial well spacing of 450m provided the highest values for cumulative oil
production and NPV while the well spacing of 600m provided the best values for the rate of return.

Step 3:
The best patterns with the initial well spacing of 450m for both reservoirs were submitted to the optimization
algorithm for the assessment of an estimative of the total number of producer and injector wells.
It was necessary to perform 8 simulations runs to estimate the amount of wells for each of these selected
patterns. It is also important to observe that the objective-function, due to its characteristics of non-linearity, presents
local optimal values.
The results obtained throughout this step for the selected patterns of both reservoirs are presented in Figures 11,
12, 13 and 14. It can be noted a reduction in the cumulative oil production and a significant increase in the objective-
function values.
This optimization procedure is simple and fast. The quality of the results are mostly affected by the initial well
spacing previously defined and on the criteria applied to the wells during the simulation runs performed to achieve a
good estimate for the number of wells.
1050 1050
Stagg. Linedrive Stagg. Linedrive
Vert. Linedrive Vert. Linedrive
1000 Alternate Alternate
1000
Horiz. Linedrive Horiz. Linedrive
NPV (US$ millions)

NPV (US$ millions)

950 950

900 900

850 850

800 800
40 50 60 70 80 90 100 110 120 42 43 44 45 46 47 48 49
Total Number of Wells Np (millions m3)

Fig. 11: Case A – Step 3 – NPV vs Total Amount of Wells Fig. 12: Case A – Step 3 – NPV vs Cumulative Oil Production
 1100 1100

1050 1050

NPV (US$ millions)

NPV (US$ millions)

1000 1000

950 950

900 900

Vert. Linedrive Vert. Linedrive

850 Combined 850
Combined
Alternate
Alternate
800 800
40 50 60 70 80 90 100 110 120 45,5 46 46,5 47 47,5 48 48,5 49
Total Number of Wells Np (millions m3))

Fig. 13: Case B – Step 3 – NPV vs Total Amount of Wells Fig. 14: Case B – Step 3 – NPV vs Cumulative Oil Production

Step 4:
This step was performed for the assessment of an ideal well scheduling for the water injection patterns
optimized in the previous stage using reservoir simulation and considering the production and injection performance
presented by the well.
Three well scheduling were proposed for the selected patterns, with an interval of 2 months according to the
established operational constraints.
Figures 15 and 16 indicate the results obtained for the objective-function and for the cumulative oil production
(Np) for each of the selected schemes.
560 675
Stagg. Linedrive Vert. Linedrive
550 Vert. Linedrive 670 Combined
Alternate Alternate

540 665
NPV (US$ millions)

NPV (US$ millions)

660
530
655
520
650
510
645
500
640

490 635
43000 44000 45000 46000 47000 48000 45500 46000 46500 47000 47500
Np (millions m3) Np (millions m3)

Fig. 15: Case A – Step 4 – NPV vs Cumulative Oil Production Fig. 16: Case B – Step 4 – NPV vs Cumulative Oil Production

Step 5:
Considering the objective of this work and the type of recovery method planned for both reservoirs, it was
assumed that the results obtained after applying the four previous steps were satisfactory and it was not required the
application of Step 5.

Discussion
The methodology and the optimization procedure presented a low computational cost and provided good
quality solutions for both reservoirs studied in this work. It allowed a proper evaluation of the objective-function
behavior for the different alternatives proposed for recovery strategy with water injection using vertical wells and in
order to achieve this it is very important to define carefully the constraints for the optimization procedure and to
execute an individual analysis for the behavior of each well throughout the simulation runs performed in steps 3 and 4.
The procedure can be easily automated in order to reduce the global time required for the process. However, at
the end of the procedure, it is necessary to execute a manual analysis of each well performance to account for details
that may be not evaluated by the automatic procedure.
 Conclusions
1. The methodology developed in this work is adequate to help in the decision making process for recovery strategy
planning including oil production and water injection. It provides a set of good alternatives that must be analyzed
regarding the objectives and the resources available previously defined for the project.
2. The methodology is flexible and presents a low computational cost. Depending on the characteristics of the
project some steps can be simplified or discarded.
3. Steps 3 and 4 must be adapted to attend the different characteristics of each natural drive and recovery method. It
is necessary to account for the constraints associated to each project for the obligatory optimization.
4. For the examples studied in this work, the well spacing and the well scheduling were parameters that strongly
affected the results generated by the methodology.
5. The best water injection patterns for the reservoirs studied in this work presented a ratio between producer and
injector wells close to 1.
6. In the procedure used in this work, as the steps are executed, production strategies alternatives are eliminated and
the number of required simulations runs is reduced. At the end of the procedure, a manual analysis of each well
performance is recommended to account for details that may be not evaluated by the automatic procedure.

References
1. Bittencourt, A. C.: Reservoir Development and Design Optimization, SPE 38895, October 1997.
2. Beckner, B. L. and Song, X.: Field Development Using Simulated Annealing – Optimal Economical Well
Scheduling and Placement, paper SPE 38895, SPE Annual Technical Conference and Exhibition, San Antonio,
TX, Oct. 5-8, 1997.
3. Clark, S. K., Tomlinson, C. W., Royds, J. S.: Well Spacing – Its Effects on Recoveries and Profits, AAPG,
February 1994.
4. Crawford, P. B.: Factors affecting Waterflood Pattern Performance and Selection,” J. Pet. Tech, December 1960.
5. Damsleth et alli: Maximum Information at Minimum Costs, Journal of Petroleum Technology, 1992, pag. 1349-
1356.
6. Mezzomo, C. C., Schiozer, D. J.: Production System Optimization for a Petroleum Field – Paper presented at 8th
Brazilian Congress of Thermal Science and Engineering – ENCIT, October, 2000.
7. Nystad, A. N.: Reservoir Economic Optimization, SPE 13775, fevereiro de 1985.
8. Pedroso JR., C. and Schiozer, D. J., Optimizing locations of wells in Field Development using Reservoir
Simulation and Parallel Computing (PVM), RIO OIL AND GAS, 2000, Rio de Janeiro.
9. Willhite, G. P.: Waterflooding, SPE textbook Series Vol. 3, Second Printing, Richardson, Texas, 1986.

View publication stats