Development of a framework for optimization of reservoir simulation studies

Abstract

We have developed a framework that distributes multiple reservoir simulations on a cluster of CPUs for fast and efficient process optimization studies. This platform utilizes several commercial reservoir simulators for flow simulations, an experimental design and a Monte Carlo algorithm with a global optimization search engine to identify the optimum combination of reservoir decision factors under uncertainty. This approach is applied to a well placement design for a field-scale development exercise. The uncertainties considered are in the fault structure, porosity and permeability, PVT, and relative permeabilities. The results indicate that the approach is practical and efficient for performing reservoir optimization studies.