Developing neural network analysis technology for the resolution of issues in oil and gas geophysics

Abstract

   Neural network analysis represents a promising avenue for enhancing the efficacy of petroleum geophysics and the oil and gas industry. The analysis of the obtained experience of using the available neural network analysis methods and software packages in solving problems of oil and gas geophysics shows the absence of a significant (breakthrough) effect. In order to achieve a significant effect, it is proposed to move from methods to neural network analysis technologies. The article presents a fundamental framework for neural network analysis technology in the context of oil and gas geophysics. This includes a neural network designer, a subsystem for training geophysicists in the field of neural network analysis, a digital polygon, and a knowledge base comprising tasks, neural network analysis methods, techniques, and experience in solving applied problems. The elements of the proposed technology and their interrelation are discussed in detail. The pilot version of the proposed technology, which includes its principal elements, is initially described in terms of its orientation towards the training of specialists. The results of the pilot version's approval have demonstrated the efficacy of the proposed technology. The scientific and technological priorities of the proposed technology development have been delineated.