Assisted History Matching Using Pattern Recognition Technology

Abstract

This study examines the application of pattern recognition technologies to improve the time and effort required for completing successful history matching projects. The pattern recognition capabilities of artificial intelligence and data mining techniques are used to develop a surrogate reservoir model (SRM), which is then employed to perform the assisted history matching process. A well-known reservoir model, PUNQ-S3, was selected to study the potentials of the SRM in an assisted history matching process. The SRM is a prototype of a full-field reservoir simulation model that demands a low development cost and has a high implementation pace. SRMs are built based on a spatio-temporal database, which includes different types of data extracted from a few realisations of the simulation model. The SRM was coupled with the differential evolution optimisation method to construct an automated history matching workflow. The results of this study prove the SRMs' capability in assisting history matching processes. [Received: December 3, 2015; Accepted: June 17, 2016]