Abstract
This article discusses the development of an intelligent System for the operational maintenance of the level of oil and gas production as part of the implementation of the Strategy for the development of information technologies for data management and the Program for the development of digitalization of fields of JSC «NC Kazmunaigas». The advantage of the system is multitasking and using almost all the data coming from production facilities in real time. The main task of the system is to manage a group of wells taking into account their mutual influence to maximize oil production and reduce the negative impact of uncoordinated well operation without damaging the rational system of field development. A significant feature of the developed system is the creation of complex algorithms for predicting the main development indicators using artificial neural networks based on a combination of CRM (capacity resistance model), FFNN (neural network with direct communication), MBM (material balance model) and BFGS (Broyden-Fletcher-Goldfarb-Shanno algorithm) methods. During the pilot test on 10 wells, the modes were adjusted according to the recommendations issued by the system and the system confirmed its operability and effectiveness of application. Keywords: virtual flow meter; labor productivity; return distribution; machine learning; rational development system; neural networks.