The purpose of reservoir simulation is to evaluate and analyze the future performance of a field under given operating conditions with the hope of being able to improve this. performance. Whilethe purpose of reservoir simulation has not changed since the beginning of the producing industry, the tools available for simulating a reservoir's performancehave increased significantly in sophistication. The combination of fast computers, improved numerical techniques, and a better understanding of the physics of fluid flow, has made it possible to develop extremely sophisticated reservoirsimulation models. However, the degree of sophistication which should be employed in redicting the future performance of a reservoir depends heavily upon the nature of the questions being asked regarding this performance and the dependence ofthis performance upon the reservoir rock and fluid description data.

The advanced technology in reservoir simulation has been a valuable tool forthe reservoir engineer but its use is not without problems. One of the important problems created by the availability of such a wide range of sophistication is that of selecting the proper tool for the job and then finding an adequate reservoir description to fit the chosen model. An important consideration in solving this problem is to obtain a consistency between the degree of sophistication of the simulation, the reservoir characterization and the problem being analyzed. While there is no simple formula to follow to obtain such a consistency, this paper attempts to il1u5trate with specific field examples how this should be accomplished.


Since the purpose of reservoir simulation is to predict the future reservoir performance under a variety of operating schemes, the utility of any simulationmodel depends on how well it can accomplish this objective. However, being able to pred1ct the future is an extremely broad concept and needs to be refined before anything, can be said about what degree of sophistication is need in either the model or the reservoir characterization. Therefore the first step in any type of reservoir simulation should be to decide what one would like to know about the future performance of a given reservoir and what can be done with this information if it is known.

Once a problem has been clearly defined and it is evident that being ableto predict certain information about a reservoir's future performance will be beneficial, one must determine whether or not these predictions are feasible. At this point one can begin to establish the degree of sophistication of the simulation model and the reservoir characterization necessary to accomplish the objectives. Here is where the complex problem of model selection und reservoir characterization begins.

The problem of model selection and reservoir characterization can only be solved by completely understanding the complex interaction between the problem being analyzed, the reservoir and fluid characterization, and the degree of sophistication of the reservoir simulation model. While a complete understanding of this interaction is beyond the scope of this paper we will attempt to consider some of the more important aspects of this problem.

This content is only available via PDF.
You can access this article if you purchase or spend a download.