Numerical reservoir simulation demands very high computing performance. Vectorization has been widely used as a performance enhancing technique. Parallel computing can also be used to satisfy the increasing requirements of more sophisticated models. This paper presents the implementation of a novel method of parallelizing an already vectorized reservoir simulator, which has the significant advantage of reducing not only the elapsed CPU-time, but also the total CPU-time used. The method relies on the physical independence of isolated geological structures, coupled through the wells, a feature present in many of the hydrocarbon reservoirs found in the North Sea. The vectorized and parallelized code has already been in production a period of time. Results are shown for a realistic case.
CITATION STYLE
Kårstad, T., Henriquez, A., & Korsell, K. (1988). Parallelization of a reservoir simulator. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 297 LNCS, pp. 863–872). Springer Verlag. https://doi.org/10.1007/3-540-18991-2_50
Mendeley helps you to discover research relevant for your work.