PARALLEL IMPLEMENTATION OF MUSKAT-LEVERTT EQUATION USING CUDA

Abstract

Capillary pressure plays a crucial role in waterflooding by influencing the displacement of oil by water in reservoir rocks. It is influenced by factors such as pore size distribution, wettability, and pore connectivity. Understanding and accounting for capillary pressure in the design and implementation of waterflooding operations can lead to improved oil recovery from reservoirs. In this work, to investigate the effects of capillary pressure in the waterflooding process in porous media, a one-dimensional numer-ical model is proposed, and the execution time of the serial model is computed. In the serial model, the absolute permeability, water and oil viscosity are considered as constant. In order to speed up the execu-tion time of the serial model, the high-performance computing technology CUDA is used, and the re-sults (execution time and speedup) on different threads are calculated. The results of serial and CUDA parallel models for the effects of capillary pressure are presented and analyzed.