Cost-Effective PC-based Reservoir Simulation and Management

Modified from: 1998, Gerlach, P, and S. Bhattacharya, T. R. Carr, Application of Cost-Effective PC-based Reservoir Simulation and Management - Schaben Field (Mississippian), Ness County, Kansas: 1998 AAPG Meeting, Salt Lake City, Utah.

In today's competitive economic climate, cost-effective production technology is required by producers of marginal petroleum reservoirs to continue to survive and prosper. Field management based on reservoir characterization and simulation studies can assist the producer in efficient exploitation of hydrocarbon reserves in marginal fields. In the past reservoir simulation and management was restricted to large oil companies and to producing fields considered "core assets". Today, PC-based reservoir simulation is economically and technically feasible for the small independent producer.

The objective of the Class 2 project was to characterize and simulate a typical oil field producing from a Mississippian reservoir by using tools that are modern and cost-effective for small independent producers operating mature fields. General application of PC-based simulators such as BOAST3 to large-scale or full-field simulation has been restricted by hardware and software limitations. Recent advances in the computational speed and memory capabilities have drastically reduced the simulation run time. The development of powerful and "user-friendly" spreadsheet, relational database, girding and mapping software have provided the front and back-end tools to efficiently assemble and manipulate simulation input data and generate useful maps and charts.

Integrated reservoir characterization forms the foundation for the development of a descriptive reservoir model and provides the framework for simulation. The descriptive reservoir model integrated existing and newly acquired well. Simulation input parameters were generated from the reservoir model and used to simulate the reservoir performance of the Schaben field from discovery to 1996. Analysis of the reservoir performance and the distribution of the remaining mobile oil in place led to the identification of regions with potential for incremental oil recovery. The simulator was used to predict the performance of potential infill wells drilled in these areas. It is hoped that this study will provide a model for improving field management of similar reservoirs in Kansas and in the mid-continent.

Reservoir Model and Volumetric Calculations. Descriptive reservoir characterization entailed integration and creative application of existing data and new data from three wells. New core and log data provided insight into fundamental reservoir parameters (e.g., core plug NMR analysis to determine effective porosity). Integrated analyses of welllogs, core data and field mapping provided a better understanding of the complexities of an extremely heterogeneity of the reservoir. Determination of pore and throat size, irreducible water saturation, permeability, effective porosity, and movable oil are part of an integrated reservoir characterization. The descriptive reservoir model developed for Schaben Field provided a major component of the input data for reservoir simulation. (Carr and others, 1996a, 1996b; Carr and others, 1997; Guy and others, 1996)

Prior to the start of reservoir simulation, a volumetric study of the Schaben simulation study area was completed on a grid-by-grid basis. The volumetric calculations were performed to check if the different reservoir parameters such as effective porosity, net pay thickness, and water saturation in the effective porosity were able to support the observed historic production volumes. The resultant oil saturation values in the grid cells of the reservoir layer indicate the combination of reservoir parameters can with the historical production figures for the Schaben field.

BOAST 3 Simulation. The major premise of this simulation study was to enter eleven years of historical data and have the simulator predict and match the next 23 years of known field production data. At the field level, a good match between simulated and observed was obtained for both oil and water production rates during the 34 years encompassed by the historical and predictive periods (Figure 2.16).

A good match was also obtained for the simulated and observed cumulative oil and water production for the field from 1963 to 1995. After matches were obtained within acceptable tolerances for both oil and water at a field scale, attention was focused on the performance of the individual wells. The mismatch of water production in some of the wells may be due to inaccurate description of the reservoir properties surrounding these wells. The vertical permeability in the reservoir and aquifer layers plays an important role in controlling the water production at each well. Several simulation runs were carried out with varied (decreased and increased) vertical permeabilities in the cells of the reservoir and aquifer layer surrounding wells with a poor history match. The results drastically improved the history match for water production. This process of local adjustment of the vertical permeability is now being applied on a well by well basis and should result in an acceptable individual well history match for the entire field.

Cost-Effective Reservoir Management. Oil saturation maps from simulation output at the end of 1973 (field life of 10 years) show areas of low oil saturation (<40%) have developed around most wells. The poor areal sweep efficiency of the reservoir, due to its heterogeneity, is demonstrated by the area between wells which have high oil saturation (>60%). At the end of 1996 (field life of 33 years) the simulation shows oil saturation around most wells to just above the irreducible oil saturation (between 31%-35%). However, significant pockets of high oil saturation (>60%) are still left unswept in between the drainage areas of surrounding wells.

The choice of location for infill (increased density) wells for efficient oil recovery in accordance with a cost-effective reservoir management plan requires the consideration of current oil saturation and pay height in the reservoir layer. Due to the difficulty of producing oil from zones with low oil saturation (<40%) and thin pay height (<20 ft.), those areas of the field at low oil saturation and thin pay height can be eliminated from consideration for infill drilling. Those areas of the field with the highest predicted infill drilling potential can be identified with a saturation-feet map. All grid cells on the reservoir layer with an oil saturation less than 40% or with a net pay thickness less than 20 feet were set to zero. The oil saturation layer and the pay height layer were then multiplied in a grid to grid operation to produce a saturation-feet map showing those areas of the field with best infill potential.

Based on the infill potential map and in consultation with Ritchie Exploration, operator of Schaben field, three infill-drilling sites were chosen. Subsequent simulation runs covering the ten year period of 1996 to 2006 were performed to predict the production rates for each of the three well locations and the effect on the oil saturation of the reservoir layer. The three new wells were simulated to produce with a flowing bottom hole pressure equal to that of the nearest well at the end of 1995. The daily production rate simulated for the Moore BCP #3 is calculated to produce a total of 47,200 bbls of oil and 227,600 bbls of water over a period of ten years (Figure 2.17). The simulator also predicts daily oil production above 10 bopd during the first 5 years. Predicted daily field production rate of oil and water with the addition of the three new wells indicates the addition of significant additional oil production.

The following conclusions were drawn based on the results of the Class 2 project:

  1. Practical application of cost-effective technologies in reservoir simulation enables the small independent producer to map remaining hydrocarbon reserves in marginal fields.
  2. Simulation results allow proper field management by targeting infill drilling in areas of best potential.
  3. PC-based reservoir simulation is a practical reality for small independent producers with limited resources.
  4. Procedures demonstrated in this study provide a guide for geologic modeling, simulating, and managing similar reservoirs in Kansas and in the mid-continent.

Material Balance Calculations

The volumetric estimate of original -oil-in-place (OOIP) for the Schaben Field was calculated to be 37.8 MMSTB (Carr and others, 1997). The reservoir at Schaben Demonstration Site has been in production since 1963. Initial reservoir pressure was approximated at 1370 psi by using the DST pressure recordings from the early wells (Carr and others, 1997). PVT properties were generated by using standard correlations. All wells in the Schaben Field produce under artificial lift. The current fluid columns in most wells indicate that the reservoir is producing significantly above the bubble point pressure (calculated at 225 psi). Reported gas production has been negligible and the reservoir is assumed to have no gas cap and or significant dissolved gas. The absence of gas is common in reservoirs of central Kansas (Walters, 1958). The main source of energy driving the production from the reservoir comes from the strong natural water drive.

For a reservoir with no gas cap and being driven by an aquifer, the generalized material balance equation gets simplified as:

where

F denotes the underground withdrawal of fluids from the reservoir, Eo represents the change in volume of the oil and the dissolved gas, Efw stands for the connate water expansion and the reduction in pore volume, and We stands for the reservoir volume of water that influxes from the aquifer. The initial volume of oil in the reservoir is defined as N. This simplified material balance equation appears as a straight line, with a unit slope, when F/E is plotted against We/E and the Y-axis intercept (i.e. N) of this line estimate the OOIP (Figure 2.18). This estimate of the OOIP should be comparable to that obtained from volumetric calculations if assumptions about the drive mechanism and in the calculation of the aquifer water influx are reasonable. The material balance OOIP is considered to represent the oil volume that contributes to the production and pressure history of the field (Dake, 1994). This is often referred to as the "active" or "effective" initial oil in place in the reservoir. Because the OOIP determined by volumetric calculations includes immobile oil, it is generally higher than determined by material balance. A difference of less than 10% in calculated OOIP is regarded as an acceptable in the industry (Dake, 1994). If the OOIP is significantly different between volumetric and mass balance calculations one may need to reevaluate reservoir parameters (e.g., dimensions, petrophysical properties and cut-offs).

Aquifer Description. Water influx calculations are based on the geological and petrophysical parameters of the aquifer. Incorrect choices of aquifer parameters will result in deviation of the data from the straight line when F/E is plotted against We/E. Modifications of the aquifer parameters through a process of "aquifer fitting" can improve the match of observed pressure and production data with the reservoir characterization. Aquifer fitting assumes importance in situations where, as at Schaben Field, little is known about the aquifer geometry and petrophysics. At Schaben Field, very few wells are drilled into the aquifer.

Water influx from very small aquifers can be calculated by time-independent material balance equations. However, for large reservoirs the aquifer boundary takes a finite time to respond to reservoir pressure changes and thus time dependent models such as developed by Hurst and van Everdingen, Fetkovitch, Carter and Tracy, or Allerd and Chen are used to calculate the water influx, We (Dake, 1994).

An aquifer model that can match reservoir pressure and production data is generally determined through a process of trial and error. However aquifer models are not unique and problems may persist despite all efforts at aquifer fitting because of incorrect identification of the reservoir drive mechanism. Initial assumptions about the reservoir drive mechanism are indirect, and are based on the pressure and production performance profiles of the reservoir. Identification of reservoir drive mechanism is important to determine aquifer description and definition and also estimate the size of the initial gas cap. Aquifer parameters were not available for Schaben field and were inferred from reservoir parameters (e.g., porosity, permeability, thickness, rock and fluid compressibility). The small numbers of available logs that penetrate the aquifer in the vicinity of Schaben Field were used to estimate the height of the aquifer. The reservoir radius at Schaben was calculated volumetrically and was found to be 7000 feet Carr and others, 1997). The Carter-Tracy method was used for water influx calculations because it is the time-dependent aquifer modeling option available within the reservoir simulator BOAST3.

OOIP. Material balance calculations require adequate field pressure and production profiles along with the PVT data of reservoir fluids. One method to determine the average field pressure is by volume weighting the shut-in pressures within the drainage area of each well. Regular recording of reservoir pressure at each well form the basis of material balance calculations. Unfortunately at Schaben Field, a recorded history of pressure measurements carried out at individual wells is not available. Only current operating water column heights are available for most of the wells. With limited pressure data, it is impossible to obtain the average reservoir pressure through the life of the field. Thus, the material balance calculations were used to generate the average reservoir pressure profile through the life of the field, and to check if the aquifer description and assumed drive mechanism was adequate to support the reported field performance data. As a result the OOIP determined from volumetric calculations was accepted as correct.

The first nine years of production data from Schaben Field were used to generate yearly oil (Np) and water (Wp) production data along with the calculations for the underground volume withdrawal (F) of fluids. The Carter-Tracy formulation was used to calculate the water influx (We) from an infinite aquifer. Initial aquifer parameters were varied within geologic and engineering limits. The resulting plot between F/E versus We/E showed a straight line with unit slope and an intercept showing an OOIP value that is lower but within 10% volumetric OOIP (Figure 2.18). The average reservoir pressure (for the first 9 years) as a result of this match is plotted as the "base case" profile (Figure 2.19).

Sensitivity calculations were carried out by varying aquifer and reservoir parameters (e.g., aquifer height, reservoir radius, aquifer permeability, and aquifer porosity). In each case, the value of only one of the above parameters was changed. In each case the average reservoir pressure profile was generated, so that the resultant F/E versus We/E plot was a straight line with unit slope and its intercept read an OOIP value that was within acceptable tolerances. The plotted pressure profiles show the effects of varying different aquifer and reservoir parameters (Figure 2.19). Available fluid level data from the Schaben Demonstration Area indicate that the majority of the wells are currently producing against a backpressure of 400 to 1100 psi. The "Best case" scenario, incorporating knowledge of Schaben Field performance (e.g., strong water drive and a significant backpressure), was modified over a period of 34 years. Due to the rapid development of the field during the first 9 years, the average reservoir pressure profile shows a rapid decline from 1370 psi to 1000 psi (Figure 2.20). Subsequently, reservoir pressure stabilized near 1000 psi for the next 14 years and then gradually declined to 880 psi over the next 11 years. The plot of F/E versus We/E for the "best case" scenario remains a straight line with unit slope and the OOIP value at the intercept is within acceptable tolerances (Figure 2.21).

Summary.--The material balance study at the Schaben Demonstration Site confirms that the volumetric description of the reservoir-aquifer system together with the natural water drive mechanism is capable of supporting the reported fluid production history of the field. A process of "aquifer-fitting" was used to fine tune selected aquifer parameters (e.g., height, porosity, permeability, and effective compressibility). In addition a better understanding of the radius of the reservoir at Schaben was obtained.

As typical of older fields in the mid-continent the average reservoir pressure profile for the Schaben field is not available and the mass balance calculations can not be used to validate of the volumetric description of the reservoir. As a consequence the volumetric OOIP is assumed correct and used to calculate an average reservoir pressure profile. The reservoir pressure profile controls the PVT properties of the reservoir fluids and hence the mobility ratios operating during the production life of the field. Changes in average reservoir pressure are indicative of the amount of change occurring in the fluid viscosities.

The material balance calculations were used to check the consistency among different aspects of reservoir description. These calculations tie together the geologic reservoir characterization of the reservoir, the PVT data, production data and available pressure data. The material balance calculations were used to confirm the reservoir drive mechanism and aquifer parameters. The results of the material balance calculations confirm the initial reservoir characterization and refine important input parameters that can be used to revise the full-field reservoir simulation.

Next Page Previous Page Top of Report


This page maintained by the Kansas Geological Survey.
Updated June 1999
Comments to webadmin@kgs.ku.edu
URL=http://www.kgs.ku.edu/PRS/publication/99_22/resmgmt.html