Integration of genetic algorithm and a coactive neuro-fuzzy inference system for permeability prediction from well logs data

Abstract

Permeability is one of the reservoir fundamental properties, which relate to the amount of fluid contained in a reservoir and its ability to flow. These properties have a significant impact on petroleum fields operations and reservoir management. The most reliable data of local permeability are taken from laboratory analysis of cores. Extensive coring is very expensive and this expense becomes reasonable in very limited cases. Thus, the proper determination of the permeability is of paramount importance because it affects the economy of the whole venture of development and operation of a field. In this study, we introduce a new hybrid network based on Coactive Neuro-Fuzzy Inference System (CANFIS). CANFIS is a dependable and robust network that developed to identify a non-linear relationship and mapping between petrophysical data and core samples. Then to improve the system performance, genetic algorithm (GA) was integrated in order to search of optimal network parameters and decrease of noisy data in training samples. An Iranian offshore gas field is located in the Persian Gulf, has been selected as the study area in this paper. Well log data are available on substantial number of wells. Core samples are also available from a few wells. It was shown that the new proposed strategy is an effective method in predicting permeability from well logs.