Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back System of Automatic Selection of ESP Equipment to Well Based on Machine Learning V.A. KONNOV, E.A. MURAVYOVA, M.I. SHARIPOV Currently, the largest oil companies in Russia are facing the problem of depletion of operated oil wells, which causes an increase in the cost of extracted raw materials. This stimulates the need to introduce better tools to improve the efficiency of the downhole equipment of electric centrifugal pumps. Taking into account these problems, it seems appropriate to develop software systems for selecting the optimal characteristics of electric centrifugal pumps to wells to ensure their reliable operation and reduce the cost of oil extraction. In this paper, the prospect of constructing an automated software package for selecting the characteristics of an electric centrifugal pump based on artificial neural networks is considered. The structure of the program using neural networks is described. The choice of architecture and training of the neural network on a sample of data was made. For the formation of a training sample and further analysis, data from the producing wells of the Vankor field were taken. The data sample includes variables such as: production well flow rate, pump supply, water content, oil density, water density, depth of upper perforation holes, bottom hole depth, depth of descent of tubing, dynamic level, reservoir pressure, wellhead pressure, fluid viscosity, feed ratio, outer diameter of tubing, roughness of tubing, pipe wall thickness. As a result of the testing of the program on the data on the production of wells of the Vankor field, the possibility of increasing the economic feasibility of oil production in impoverished wells was established by replacing the electric center pump with a more optimal model selected by the program. Keywords: well flow rate, production well, downhole pressure, neural network, machine learning, injection well, algorithm, reservoir pressure, development. DOI: 10.25791/pribor.11.2021.1302 Pp. 17-22.