Fluid Injection Optimization Using Modified Global Dynamic Harmony Search

Abstract

One of the mostly used enhanced oil recovery methods is the injection of water or gas under pressure to maintain or reverse the declining pressure in a reservoir. Several parameters should be optimized in a fluid injection process. The usual optimizing methods evaluate several scenarios to find the best solution. Since it is required to run the reservoir simulator hundreds of times, the process is very time consuming and cumbersome. In this study a new intelligent method of optimization, called “global dynamic harmony search" is used with some modifications in combination with a commercial reservoir simulator (ECLIPSE^®) to determine the optimum solution for fluid injection problem unknowns. Net present value (NPV) is used as objective function to be maximized. First a simple homogeneous reservoir model is used for validating the developed method and then the new optimization method is applied to a real model of one of the Iran oil reservoirs. Three strategies, including gas injection, water injection, and well placement are considered. Comparing the values of NPV and field oil efficiency (FOE) of gas injection and water injection strategies, it is concluded that water injection strategy surpasses its rival. Considering water injection to be the base case, a well placement optimization is also done and best locations for water injection wells are proposed. The results show the satisfying performance of the algorithm regarding its low iterations.