SPE

Society of Petroleum Engineers

SPE 23616 Methodological Approach for Reservoir Simulation
N. Sanchez, C. Martinez* and A. Rattia*, Maraven, SA
*

SPE Members

Copyright 1992, Society of Petroleum Engineers, Inc This paper was presented at the Second Latin American Petroleum Engineering Conference, II LAPEC, of the Society of Petroleum Engineers held in Caracas, Venezuela, March 8·11, 1992

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the II LAPEC or the SPE and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the II LAPEC or the SPE. its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper is presented. Write Public.tions Manager, SPE, P.O.Box 833836 Richardson TX 75083·3836 U.S.A. Telex 730989 SPEDAL.

INTRODUCTION
The reservoir simulation is a technology of generalized use in the petroleum industry to evaluate actual and future reservoir performance. However, a general methodology for reservoir simulation has not been completed defined in the literature (Refs. 1,2,3,4). In view that reservoir simulation is a basic tool of reservoir management, Maraven S.A decided to standardize the procedures involved into a reservoir simulation study, developing a methodology aimed to eliminate engineering and computing time waste during the process of model initialization, history match, and predictions, which represent the main part of the reservoir simulation processes. Consecuently, work processes were stablished and documented including flow diagrams, and forms were prepared to obtain information, in order to keep track of the time used during the execution of each stage of the processes, allowing to use statistic control to monitore each process performance. The application of the developed methodology allowed to identify the main time consuming activities, optimizing the studies execution process by reducing the risk of overlooking key aspects and variables, thus increasing the level of confidence of the simulation results. Finally, the preliminary statistic control results have allowed to identify the following actions. 1) Guaranty the use of flow diagrams and measures of the time used in the execution of each process stage, 2)Develope automatic data processing and administration systems to accomplish the simulator requirements, and 3) Continuous trainnig in computarized tools.

the process of inicialization, history, match, and predictions. The aim was to provide new reservoir simulator users with a detailed description of the activities (a series of actions and changes) involved in the execution of each of the main reservoir simulation processes. Figures 1,2 and 3, shown the flow diagrams of each process. Detailed discussion of the different satages involved in each process wiil be presented next.

PROCESS FOR INICIALIZATION RESERVOIR SIMULATION MODEL

OF

THE

The inicialization process consist of the reservoir model validation through the calculation of the original fluid in place volumens . The model initialization allow to stablish the initial fluid saturation and pressure distribution within the reservoir. The flow diagram presented in Figure I, show the procedure to carry out the initialization process for the hypothetical general case of a reservoir with a gas cap and water zone. A detail discussion of the different stages involved is presented next.

1. Preliminary
The first order to pressure. volumes obtained.

initialization

run

run of of the inicialization data is conducted in check data entry format errors. As result the fluids saturation distribution. and fluid in place for the different fluids inside the reservoir are i.e.• the simulator interpreted inicialization data.

2. Data entry verification
Normally. the first inicialization run of a simulation model may abort due. to errors in data entry formats. inconsistency of relative permeabilities and/or PVT data. The errors of inicialization data are normally of the following types. Parametric problems. fluid properties tables. saturation tables, equilibrium condition tables. aquifer definition. and array generation. 35

GENERAL METHODOLOGY
The general methodology for reservoir simulation presented in this work, consisted of structuring into stages References and illustrations at the end of paper

this guaranty that the model will be adecuated to predict the reservoir performance. The second stage consist in the reproduction of the reservoir performance at individual well level. the first step within a match should be to make sure that this .Reviw data affecting STOIlP Verify that the value of STOIIP calculated by the model is in line with stimated values by volumetric calculations and material balance. In the following a detail description of the most common correction actions to achieve the average pressure match arc presented. The number and size of the blocks used to reproduce the aquifer. and its interrelation with the completion intervals and net sand thickness. The first stage consist in the reproduction of the observed performance at total reservoir level. Once the required adjustments have been made. Once the necessary adjustment have been conducted. numerical stability errors will be prevented. the inicialization run should be repited and the calculated value for the STOUP should verified. tacking into account the convergency rates and tolerances within the material balance (see Figure 2a). and the water cut.Review aquifer size The initial ratio of the aquifer water volume and reservoir volume.2 METIlOOOLOGICALAPPROACH FOR RESERVOIR SIMULATION SPE 23616 In case of errors. Review PVT data During the development of a reservoir simulation model. and gas-oil contact location. change. for the first two rows of blocks in the aquifer. net sand thickness in the free gas zone. e) Errors in net sand thickness. If there is not gas cap within the reservoir go to step 6. Review gas cap size The initial ratio of the gas volume iside the gas cap and the reservoir oil volume should match the volumetric and material balance calculation. a) High/low porosity values (data entry format error) b) Misplace fluid contacts (gas-oil and/or water-oil) c) Inclusion/exclusion of grid blocks that belong or not to the reservoir model. This ratio has strong effect on the reservoir energy level. 7. 2. structure elevations or total sand thickness. In fact recent developed simulators have build in citerias to select. In order to achieve the match. and by major reservoir regions formed by groups of wells with similar characteristics. If the calculated value is too high/low. Review initial fluid pressures Verify that the initial pressure distribution at the reference depth (datum) match the average reported value. This process is repited until achieve the match. The flow diagram presented in Figures 2a. This procedure is carry out simulating the last five years of history without input the historical oil production rates values. A detail discussion of the different stages involved is presented next 1. fluid properties (fluids specific gravity). it is of main concern to verify that the calculated initial solution gas-oil ratio reproduce satisfactory the value stablished in the PVT analysis characterization. a) Review oil production rates. the parameters affecting the gas cap dimensions must be reviewed. 3. S. the porosity. should match the calculated value using material balance. 2b and 2c show the history match process. normally is necessary to perform modifications on certain reservoir parameters (geologic or from fluids). After complete the match of the reservoir performance the model should be validated. The solution methods do not have to be fixed along all the simulation period. If not. correction of data entry formats should be performed and the run must be repited until a clean data code is obtained. the gas-oil ratio. porosity values and distribution. 36 PROCESS FOR HISTORY MATCH OF THE RESERVOIR PERFORMANCE A very common way to reproduce a reservoir historical performance. greater time intervals. net sand thickness. If there are significative differences the fluid data should be review. The runs for this part of the match should be conducted using time periods of approximatelly 20 % of the available production history and including restart in each run. If the simulation calculated oil production rates are reproduced satisfactory. This procedure is repited until a match of the STOIIP is achieved. Solution method selection In order to select the method that will allow to obtain smaller run times and greater run stability (smaller number of iterations. Consecuently. d) High/low values in the capilar pressure curves. consist on input the historical data for the oil production and fluid injection rates. and assigning only the constraings impossed by the field production system. i.5 to 2 times the size of the blocks in the oil zone.e. this is normally due to errors of the following type. and water-oil contact location. if the size of the blocks used to represent the aquifer are increased gradually (between 1. 6. review the assigned parameters to the equilibrium tables. well completion intervals and net sand thickness A very common error source is found on the production rates.. If not. Average pressures match Generally the history match process is carry out in two stages. consecuently. and leave the model to simulate the pressure performance. Alterations of the water-oil transition zone size and its saturations distributions could be prevent. instead of making local changes affecting only individual blocks which are useful only when the match of the individual well behavior is been performed . In order to modify the aquifer size the following parameters have to be considered. smaller material balance errors or greater tendency to improve it) the model must be run for a small time period using the simulator availables solution methods. the inicialization run must be repited and the STOIIP value has to be verify againg. making sensitivities to changes on the required variables until obtain a satisfactory match. 4. or alternate solution methods during a simulation run. The history match of the complete reservoir pressure and major regions is obtained using an interactive process that involve global changes affecting regions or groups of blocks where the wells have observed similar behavior..

In order to reduce this effect the slop of the relative permeability curves must be modify. MARTINEZ. 4. reductions of the vertical permeability allow to decrease this effect Contrariwise. This implicate a good knowledge of the gas or water 37 drive mechanism acting into the reservoir. 3. only is necessary to modify the parameters playing a roll in the aquifer influence funtion. In this case. the completion of the well located on the problem grid-block or the values assigned to the porosity. b) Adjust energy level If the calculated pressure values are considerable far from the observed performance. it will be necessary to make the required corrections and to repit the run until the model is able to reproduce the observed production rates. and keep running the model using restart of the recurren data. Therefore. the adjust consist of modify the dimension and/or porosity of the blocks representing the aquifer. if the gas in the system is segregating too fast. This stage should be repited until complete the match of the reservoir production history. pore volume. it will not be adecuatelly reflected on the reservoir due to high/low reservoir-aquifer comunication. and net oil sand must be reviewed. i. as well as the aquifer activity and/or the fluid and rock compresibilities. becouse of the main flow channals and sand energy levels distribution definition. c) Vertical permeability distribution.. increase the history period to be matched during the next set of runs in 20 % of the available history. In this case the opposite to the former case must be done. It is possible that some of the runs made during this stage may abort due to large block pressure changes. the water breakthrough is mainly controled by sedimentologic parameters. the slop of the relative permeability curves could look reazonable good. This can be controled by modifying the vertical permeability in those model regions where faster segregation is required in order to reduce the gas production. the individual well performance match follows. manipulate the slop of the relative permeability curves until a higher ratio of krg/kro is achieved. which cause the block pressure to get out of the PVT table pressure range. modifications of the transmisibilities along the blocks describing the faults system will allow to modify the water influx in any desired direction. i. . d) Shape of relative permeability curves. Match of the historical GOR and water cut performance Frequently. Also. The problem is quit difficult in stratified reservoirs where there is vertical comunication between layers. the match can not be achieved if the gas and water front are not model adecuately. although. the secondary effect could be an increase of the reservoir pressure. Another problem that may arise. A general rul is to match the maximun number of wells. This normally takes place in very large size reservoirs and/or complex geological model. The following procedure should be used for the individual well match (see Figure 2b) . is an early gas flow in the model. this means that there is not enough or too mutch energy inside the model compared to the real reservoir condition. Individual well match After the reservoir and region match have been completed. In this case the sedimentological reservoir description will play a very important roll on the water front advance. The required actions to adjust the aquifer activity are presented in the following. and try separately at the end the not matchable wells. at region level (group of wells with similar characteristics) considerable discrepacies are observed compared to the historical information. The procedure to adjust the model energy level consist in review the size and pore volumes distribution inside the gas and oil zones. The increase in the gas production could create a second factor that is reflected as a lower reservoir pressure. In this case an increase in the critical gas saturation coul stop the gas production.SPE 23616 N.For aquifers represented by model grid blocks.For aquifers represented analiticaly. It is possible that the aquifer will be strong enough. RATIIA 3 information is correct.e. Therefore. it is necessary to obtain a pressure match during that period before decide to iniciate the match of the GOR and water cut (Figure 2b). and capacity. If not.e. which must be trated as mentioned previously. In this case small adjustment to the transmisibilities in the regions closed to the aquifer will help to correct these discrepancies. The opposite must be applied in the case of a delaied gas production. The parameters that should be modify in order to adjust the GOR and water cut are presented next a) The already mentioned for the pressure match. The GOR is controled by the degree of segregation allowed within the model. The variation of this values are due to the gas and water advance into the reservoir. Also. although. mainly good producers and wells with large production history using the same reservoir parameters. if the GOR is very low. SANCHEZ. the keg curve must be make less monotony growing and the kro curve more monotony growing. . If this is the case. b) Relative prmeability curves for each layer. this indicate that there is not enough gas within the system. The regions selection must be carry out based on pressure analysis. in some cases modification of the slope of the relative permeability curves or critical gas saturation permit to obtain the same effect A very high GOR could indicate a very optimistic ratio of krg/kro. thus reducing the reservoir depletion. C. c) Adjust fluid flow between major regions It is possible that the simulation calculated average reservoir pressure is in line with the observed value. In some cases the permeability in some directions could be increased (axis of lithoral bars or flow channals for example) in order to reproduce the regional water saturation distribution.. the GOR and water cut start to show significative variations after some production period. The pressure match of the different regions is obtained modifying the fluid flow between then performing appropiate transmisibility changes. the transmisibilities between the aquifer and the reservoir zone or from the main flow channals must be modified. but. After complete the match of the average pressure. or water front advance observed in the reservoir. In the case of the water cut. Also. AND A.

the first run usually denominated base case. The reservoir characteristics. Plot the real and simulated water cut performance versus time for each well. the well index of the affected wells must be adjust at the time of the event execution. and the strategic program will determine the exploitation options that could be carry out in the reservoir. . Repitstep 1 and 2 until obtain the match. Water cut match In this case the procedure is as indicated in the following. Model valiadation After the reservoir and individual well performance match have been completed. The next step consist in validate the reservoir model leaving it run without any constrains to see if the model is able to match the last five years of observed history. and can be corrected modifying the relative permeability curves in the gas saturation ranges existing in the well block location during the not matched period. If there is a secondary recovery project taking place in the reservoir. 3. only the maximun bolton hole flowing pressure have to be defined. Study the feasibility of additional recovery by using enhance oil recovery process. a model suitable to predict the reservoir performance is available. The predictions of the reservoir simulation study will provide the base for the economic analysis that will allow to stablish the optimun exploitation scheme selection. well recompletions. a model able to predict the future reservoir performance is available. Evaluate the advantage of a flank water injection project versus a pauern injection project. If that is the case. and assigning well index values to each active well to reproduce its real oil production potencial at the bigining of the period. b) Match of the GOR The GOR depend to a high degree from the krg/kro ratio. or review the saturation pressure around the well in those reservoirs where there are variations of saturation pressure with depth. the krg and kro curves could be modify using a procedure similar to that used during the reservoir GOR match. 6. The reservoir model must be validate going back the last five years of the available production history. When the model does not reproduce the gas-oil ratio. and reduce the slop of the krw curve in the range of water saturations in which the simulated water cut is greater than the real. and identified not matched regions. 2c). 1. 2. 5. Plot the well block water saturation versus time. The model validation will be verifyied observing if there is shifting of the well production performance curve above or bellow its observed behavior. It is common that relative permeability curves describe the well performance only during a certain period. and achieve a good representation of the botton hole flowing pressure (Fig. because at saturations smaller than that the krw value must be closed to cero. artificial lift curves must be developed and included into the model to simulate the vertical flow performance. and neighbor blocks if necessary. Also. and recomendations about additional information required should be point out if necessary. and the production rate must be matched. In this run the model only will be constrained by the assigned well head pressure in the artificial lift tables in case of wells producing by natural flow or by artificial 38 gas-lift. Determine the effect of well location and spacing. History match documentation. 4. review if there have been events that justify changes in the flow index (workovers. 6. If a satisfactory validation is not achieved. the limitations that avoided to obtain a beuer match must be indicated. This stage is aimed to indicate all barriers found during the process and how were solved. show the prediction behavior under the actual exploitation scheme. Therefore. or reducing the vertical permeability to achieve slower gas segregation and consecuently it can be produced at the well. and sleeves changes). If this is too high/low. In case of well producing by mechanical pumping. 7. Special concern must be given to the satuaration at wich the water production start out. 1. Investigate the effect of infill drilling on recovery. The following are the more common evaluated alternatives. the production guidlines. 3. it must be consider to increase/reduce the horizontal or vertical permeability (or both if it is the case) in the block where the well is located. and increase the krw value in the range of water saturation where the real water cut curve is greater than the simulated. reducing the critical gas saturation to allow an early gas movement. 2. The flow diagram of the process to carry out predictions of the future reservoir performance is shown in Figure 3. Evaluate the reservoir future performance under natural depletion. This performance will be use as reference to determine the benefits of any feasible modification of the actual exploitation scheme. 1. This is due to lack of gas availability in the region around the well.4 ME1lIODOLOGICAL APPROACH FOR RESERVOIR SIMULAnON SPE 23616 a) Adjust the well energy level The well pressure is related to the energy level in its neighborhood. or also. and gas or water injection. c. Stablish base case During the process of a simulation study predictions. PROCESS FOR PREDICTION OF THE RESERVOIR FUTURE PERFORMANCE After an adecuated match of the historical reservoir performance have been achieved. the engineer knowledge about the study area. In this case it is only necessary to execute a restart in order to initiate the reservoir predictions. If a satisfactory validation is achieved in this stage. and the run should be executed again. Define schemes of well completion in layered reservoirs. new adjust must be made to the relative permeability curves and return to stage 4 (see Table 2c). observing that model calculated GOR decrease while the observed field GOR is still increasing. the process efficiency must be evaluated to decide if is possible to improve the actual s. If this modification does not help. Details of the procedure arc discus in the following. Investigate the effect on recovery of production or injection rate variations.

or drilling new wells on model identified poor drained zones.Oil. and the results obtained at the end of each stage. thus evaluating the following aspects. 3. Reconstruction of the additional recovery scheme.The effect of change the well spacing in none drained to stablish the process efficiency.The effect of infill drilling on potential and recovery. RATTIA 5 production/injection scheme. It must point out the used criteria. . i. .Variation of injection and production rates. Elaborate operational program The natural depletion predictions with and without additional development..Moddification of the number and distribution of injectors and producers. and by comparison with the predicted values. . . Additional development program The results of the predictions executed in stage 2 compared to the expected recoveries in reservoir with similar characteristics. biginin and end of injection. stratigraphy. in the opposite case go to stage 7. and well completions).When will be necessary to perform workovers and/or zone changes in each well. . Natural depletion predictions This predictions were aimed to evaluate alternatives to improve the actual natural depletion scheme (base case).The effect on production potential and recovery of drilling new wells. If there is not a secondary recovery project taking place in the reservoir. . using the optimun injection rate stablished previously. MARTINEZ. % of insitu mobile oil recovered by the project) . will allow to elaborate an operational program for the schemes considered to be further investigated. .Production-reserves ratio. . . the reasons for its use.Identify key operational parameters that will allow to guaranty the project succes (GaR. . predictions must be conducted . STATISTIC CONTROL PROCESS In order to monitor each process using statistic control. If the execution of the actions mentioned in the previous stage indicate that it is necessary to improve the actual additional recovery efficiency.e. the reservoir natural depletion should be used as reference to determine the benefits of any feasible change of the actual exploitation scheme. and sedimentology). it will be necessary to evaluate the effect of the following options. If the last is true go to next stage. AND A. Determine additional recovery process efficiency If there is an ongoing water and/or gas injection project in the simulated reservoir. From the former aspects analysis.Total fluid produced substitution factor. .Chronologie events secuence. the validity of the original reservoir predictions can be stablished. SANCHEZ. Feasibility evaluation of additional recovery The most important aspects that should be take into account to iniciate an additional recovery project are. C. data collection forms were developed to keep tracking of the time used during the execution of each stage involved in the described processes.Evaluate the effect on recovery of infill drilling.Well evaluation schemes. the additional recovery possibility must be investigated (stage 4). . If this is the case refer to stage 5.Volume of fluid injected at reservoir conditions. . water cut.Changes of injected fluids. . .Effect on well production of water or gas invaded zones isolation.Project interruption.Variation of injection pressures.Oil recovered (% of STOIW. which must be compared base in economic indicators to the best of stage 2 analyzed options. . and production mechanism will allow to decide if there is possibility for additional development. . and compare this results with the actual injection sheme. S.The annual total production of those wells and the cumulative production until total recovery. If this is not the case. In this case the most common questions are the following. . This program must contain: . .Number of well to drill and repair. 8. .Production/injection required production facilities. . . 4. and the additional recovery predictions. .Injection scheme considering geologic characteristics (structure. The organization of all this information should be clear enough to be used as guide for the reservoir operators during the execution of the stablished actions and to monitor the selected exploitation scheme. The selection of the best additional recovery program must take into account the following. 6.SPE 23616 N. After the required runs have been conducted go to stage 7. . and maximun injection pressure. The preliminary 39 areas. .Volume of injection required and well injection distribution. Wells and reservoir injection rates. or if it is necessary to modify the actual reservoir exploitation scheme or made additional predictions aimed to improved the reservoir recovery.Chronology or events secuence.Botton hole historical observed pressure.Changes of injection/production zones. water cut limit. . drilling and workover schedule.The analysis of the former presented factors will allow to stablish different natural depletion scenarios. gas and water production prediction. .Effect on recovery of change operational parameters as production GaR. Process documentation This stage must show a resume about the developed activity during the prediction process execution. 2. The indicated parameters will be used as reference in the economic evaluation. . 7. .The completion date of each new well and the initial assigned potential.Effect on production potential of change the well production method (artifitial lift versus natural depletion). .

The collected information during the execution of different studies was analyzed making run charts allowing to identify the main activities responsible for the time variations observed during the execution of each process stage (Ref. Mirabal. and identification of key efficiency variables as productivity measures in each process. kro Oil relativepermeability. fraction. Identification of common and special facts responsible for process alterations. Guaranty complete statistic control of each stage of the described processes.S. 3. 1. fraction. Finol who made helpful suggestions. 5.G and Toronyi.A. Special thanks to A. Aziz. the succes of the described methodology will depend of the continuous identification of corrective actions. M YSanchez N. In the following the used procedure will described. 3. Selection and confidence analysis of the information available. New Jersey. for permission to publish this paper. 4. Compagnie Francaise des Petroles. 2. 4. krg = Gas relative permeability. The developed general methodology guaranty an adecuate representation of the acting reservoir mechanism. REFERENCES 1. 2. Continuous trainning about the use of simulators and computer systems. Use of a statistic control system to interpret process variation. 5). 3.: " Modem Reservoir Engineering .:"Resultados de la aplicacion de calidad de gestion en simulacion de yacimientos". A general methodology for reservoir simulation has been defined. and identification of actions to improve the process. fraction. 1980. 2. guarantying the continuous methodology optimization. Finally. ACKNOWLEDGEMENTS We would like to express our apprecianon to the managements of Petroleos de Venezuela S. N. 4. 4.:"Reservoir Simulation Seminar".:"Engineering Control in Reservoir Simulation". Saler. M. Applied science publishers LTD. The analysis results and interpretation of the information collected allowed to identify the following corrective actions. 1985.and Maraven S.B. Criccklow. CONCLUSIONS 1. 1979. Maraven S. Guaranty the use of the flow diagrams stabJished for each process.. R. London.A Simulation Approach". 1. paper SPE 18305. and consecuentiy. The standardization of the procedures involved in reservoir simulation studies. Paris. Oct 1988. GOR = Solution gas-oil ratio. A. 3. NOMENCLATURE krw Water relative permeability. 2. thus increasing the confidence level of the simulation results. Inc. Selection of an automatized system for the statistic process control. Total . allow to obtain information about the time consumed in the execution of each activity of the procedure. Improved productivity measures visualization. H. giving a detailed step by step procedure to achieve adecuate simulation results. The success of the described methodology will depend of the continuous identification of corrective actions to optimize the quality and productivity of reservoir simulation studies. and which contribution is gratefully acknowledged. SCF/CTB STOUP = Stock tanck oil initial in place. Develop automatized systems for studies administration and data processing based in the simulator requirements. Prentice Hall. and prepare run charts to identify variation aspects using a statistic control process. Agosto 1991. K and Settary.A.:" Petroleum Reservoir Simulation".A. STB = = 40 .6 METIlOOOLOGICAL APPROACH RlR RESERVOIR SIMULATION SPE 23616 results analysis of the developed methodology application have allowed to identify the main engineering and computation time consuming activities. in order to optimize the processes and improve the confidence level of simulation results.. 5.

SEE 2361 6 WEll..tonelill relll!...oir perlormance .lory match of the hi. AND CouPL£1IOH he1 SAND IHT(RVAlS THICKNESS "-OJUST UOO[l [NlRGY ADJUST rlUIO rlOW BETWEEN WVQR REGIONS INCREASE 20% ri121!1~Pree ••• for hi..

Sign up to vote on this title
UsefulNot useful