Development of a framework for parallel reservoir simulation

Date

2017-09-15

Authors

Barrios Molano, Hector Emilio

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Parallel reservoir simulation is a topic of special interest to reservoir engineers and reservoir simulator developers. Parallel reservoir simulators provides several advantages over non-parallel reservoir simulators, such as

• Capability to run bigger models. • Capability to have simulation results faster by using several processing units at once. • Not limited to single computer memory. Memory available increases as more computers are used.

All these are compelling reasons for reservoir engineers. However, for reservoir simulator developers, the creation of a parallel reservoir simulator is a more complex task than non-parallel simulators. Problems related to parallel implementation such as parallel communication, model division among processors, and the management of data distributed among processors, among others should be addressed and solved on top of the already complex task of simulator development. Hence, development time for parallel reservoir simulators is more time intensive than the traditional development on single processor computers.

The objective of this work is to separate the development focus of parallel reservoir simulators in two: parallel development and reservoir simulator development. To achieve such separation, a parallel framework was developed. The framework developed in this work implements and handles the parallel complexity and provides easy to use programming interfaces to accelerate the development of new parallel reservoir simulators or the parallelization of existing ones.

The University of Texas Compositional Simulator (UTCOMP) was used with the framework to create a new parallel reservoir simulator. Several cases were used to verify accuracy, to assert usability and to test parallel performance on our new parallel reservoir simulator. The parallel reservoir simulator developed in this work has all of UTCOMP's features and is able to run models with up to 102.4 million cells using up to 1024 processors.

Description

LCSH Subject Headings

Citation