Intelligent Systems Can Design Optimum Fracturing Jobs

Abstract

ABSTRACT Many computer software programs have been developed to assist petroleum engineers and scientists in designing hydraulic fractures. These programs use analytical, numerical or empirical methods (or a combination of two or all three methods) to model fracture propagation in the reservoir. The user - which is usually experienced petroleum professional - provides the necessary input parameters to the model. These input parameters include reservoir characteristics, fluid and proppant characteristics and treatment schedule. Usually among the major outputs of these models are the fracture geometry and conductivity. The goal of this study is to provide a tool that allows the engineer to solve the inverse problem in a fast and efficient manner. Using the tool developed in this study the engineer identifies the desired fracture geometry that she/he would like to achieve and enters this value along with the reservoir characteristics into the intelligent software tool. The software tool would then solve the inverse problem and provides the engineer with fluid, proppant and the treatment schedule that would produce the desired fracture in that particular reservoir. This new tool has been tested and shows to be capable of designing hydraulic fractures replicating those designed by expert engineers. An innovative hybridization of two virtual intelligence paradigms has been used to develop the methodology that is being introduced in this paper. An industry standard three-dimensional fracturing model (FRACPRO) has been used as the data set generator for the study.