Application of Methods of Fuzzy Logic and Neural Networks for Automation of Technological Processes in Oil and Gas Engineering and Increasing the Efficiency of Oil Production

Abstract

This paper discusses the issue of improving the efficiency of pumping systems, as the most energy-intensive part of an oil and gas field. The relevance of the research topic for the oil and gas engineering industry is considered, the main goal of the research is formulated, which is to digitalize the processes under consideration and create domestic automatic control systems using fuzzy logic algorithms and neural networks. Methods of constructing modern control systems are considered, their advantages and disadvantages are analyzed. The features of various approaches to the construction of automatic control systems for technological objects during oil production and transportation are considered. The most common are direct digital control or feedback. A classic description of automation and telemechanization objects based on system parameters is presented. The main characteristics of the considered technological processes, such as oil production, preparation and transportation, which do not allow for achieving maximum efficiency in the existing approach are given. The most important factors for efficient systems of automatic data objects are identified. The experimental data obtained have shown that the parameters of the technological process vary within significant limits from the nominal values, which leads to a low quality of operation of the regulators. The accuracy of the system identification models based on linear autoregressive methods is no more than 30%. It is concluded that it is necessary to use for control of nonlinear objects with inherent uncertainties based on neuro-fuzzy and fuzzy controllers with discrete terms.