Resumen
The modern oil industry is characterized by a strong trend towards the digitalization of all technological processes. At the same time, during the transition of oil fields to the later stages of development, the issues of optimizing the consumed electricity become relevant. The purpose of this work is to develop a digital automated system for distributed control of production wells using elements of machine learning. The structure of information exchange within the framework of the automated system being created, consisting of three levels of automation, is proposed. Management of the extractive fund is supposed to be based on the work of four modules. The ?Complications? module analyzes the operation of oil wells and peripheral equipment and, according to the embedded algorithms, evaluates the cause of the deviation, ways to eliminate it and the effectiveness of each method based on historical data. The ?Power Consumption Optimization? module allows integrating algorithms into the well control system to reduce energy consumption by maintaining the most energy-efficient operation of pumping equipment or optimizing its operation time. The module ?Ensuring the well flow rate? allows you to analyze and determine the reasons for the decrease in production rate, taking into account the parameters of the operation of adjacent wells. The Equipment Anomaly Prediction module is based on machine learning and helps reduce equipment downtime by predicting and automatically responding to potential deviations. As a result of using the proposed system, many goals of the oil company are achieved: specific energy consumption, oil shortages, and accident rate are reduced, while reducing the labor costs of engineering and technological personnel for processing the operation parameters of all process equipment.