Neural network modeling and optimization of operational parameters for thermal impact on a bituminous reservoir

Abstract

It is known that the extraction of oil reserves from bituminous sandstones using open-pit mine workings is unprofitable due to the large depth of the productive formation. Therefore, special attention is paid to the development of methods for in-situ oil production from tar sands. One of the promising methods of in-situ production of bituminous oil is the heating of the reservoir in the «injection» well to the temperature of oil self-ignition. Due to the chemical reaction of the interaction of the oxidizer with the fuel in combustible materials, the combustion front is maintained and the advancement and subsequent extraction of oil through the production well is carried out. In this case, the stimulation of the formation can take place at various operational parameters of the lubricants influence technology. To date, one of the most effective methods for studying complex systems is computer mathematical modeling. Therefore, a promising way to solve difficult problems is to use the universal computing potentialities of modern software tools based on artificial neural networks, which have the widest capabilities. They allow, based on empirical experience alone, to build neural network models that help extract knowledge from data and identify previously unknown patterns, as well as use them to solve specific practical problems. In this work, using modern computer technologies, an application software module has been developed that allows, based on the developed algorithm using a neural network model, to determine the optimal operating parameters of the thermal technology for influencing a bituminous reservoir.