Application of momentum backpropagation algorithm (MOBP) in identification of low-resistivity pay zones in sandstones

Abstract

Low-resistivity pay zones with substantial reserves in many oilfields around the world are drawing more attention than ever before. Through analyzing the features of the logging data in low-resistivity pay zones, a fast model for identification of low-resistivity pay zones was developed in this paper. Momentum backpropagation algorithm was used in the model development. Indicators that can amplify the characteristics of low-resistivity pay zones were designed. The proposed model can be used for reevaluating old wells using conventional logging data. Validations through field examples demonstrated the capability of the model in accurate identification of low-resistivity pay zones.