Reservoir simulation results may be presented in terms of asset net present value. This may be used to make decisions regarding alternative production strategies or in-fill drilling programs based upon economic criteria. If the simulation and value assessment is performed for a number of possible realizations of the reservoir geological model an estimate is also obtained of the risk involved in each production decision made. Ideally, optimal strategies for all realizations should be compared, rather than one tuned to a single possibility.

Such a system implies a fast reservoir flow simulation engine. One possibility is to use a streamline solver. These yield simulation results quickly, but introduce significant approximations in the treatment of gravity and viscosity changes. An alternative is to use a fast finite difference or finite element simulator. We compare the results obtained using these methods, in economic terms as well as in terms of production profiles.


Optimization of a petroleum asset commonly requires the use of reservoir simulation in order to investigate the effects on asset value of various production strategies. The investigative process involves the construction and numerical simulation of many separate reservoir models, with each one being designed to illuminate some potential strategy. The output model production profiles are convolved with an appropriate economics model to infer the economic properties of each strategy.

Using conventional finite difference and finite element reservoir simulation programs such a process is obviously very numerically intensive. When the models involved must also account for multiple underlying geological realizations, the associated cost in computer and man time can make such studies prohibitive.

Streamline methods of reservoir simulation offer the potential of significantly speeding up optimization processes, albeit at the expense of accuracy due to an incomplete treatment of the effects of gravity and viscosity changes. Some success in their use has been reported in the literature in production data integration and geostatistical ranking 1–7. It is natural to consider then whether or not the speed gain of a streamline simulator may be used to alleviate the numerical intensity of an asset optimization process without compromising the accuracy of decisions made.

This paper seeks to address the above issue by comparing the results of an in-fill well location algorithm using a streamline solver with those obtained using a finite difference simulator 8. The well location algorithm used is a relatively simple iterative genetic algorithm from the literature 9. Using a straightforward economic analysis, the algorithm is used to place additional in-fill wells into a base case scenario so as to optimize net present value. The procedure is applied to a number of stochastically generated realizations, so that the cumulative positions inferred over all realizations may be used to assess the risk associated with each suggested in-fill position. The results obtained using the two simulators are compared for both the single and the multiple realization cases.

Finite Difference and Streamline Simulators

It is possible to apply the streamline method to the solution of reservoir simulation problems. Although the core technology is quite well established, issues such as the incorporation of gravity effects, dynamic well controls and viscosity changes during the production process remain the topic of current research.

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