Dikova J. Subsystems of diagnostics and forecasting of information-measuring systems of coal mines

The thesis is devoted to solving the tasks of improving the quality of serving coal mine production processes by creating highly efficient special-purpose subsystems that will significantly expand the functionality of the IMS used in the coal industry. To solve this problem, within the framework of a unified approach, a special-purpose subsystem was created based on the methods of forecasting and monitoring the state of the equipment and the mine atmosphere. A subsystem for analysis and construction of routes was also proposed, which solves the problem of finding the optimal transportation of mine materials. The proposed methods underlying the developed subsystems are based on artificial neural networks, fuzzy logic and meta-algorithms. The paper proposes methods for monitoring the state of the mining equipment and the mine atmosphere based on a fuzzy neural network and a high-order network. The created subsystem for analyzing the state of the mining equipment and the mine atmosphere allows for a comprehensive assessment of the state, taking into account simultaneously the required number of input parameters. The accuracy of the diagnostic results obtained is 93%, which is 13% more compared to existing methods. Methods were also proposed for predicting the state of the mining equipment and the mine atmosphere based on the NARX and NARMAX artificial neural networks, allowing to take into account the influence of external factors on the forecast of the main indicator. The proposed neural network approaches for analyzing and predicting the state are extensible with the required number of input parameters. Were considered metaheuristic methods of searching for the optimal transportation of materials, allowing to take into account the restrictions on the type of vehicle, the cost of the route and the landscape features of the mine sections. Using solutions to find the optimal route allows you to reduce transportation time by more than 40%, while taking into account restrictions on the type of vehicle use and the cost of the route. Keywords: information and computing system, state monitoring, complex state forecast, fuzzy neural network, high-order neural network, NARMAX, NARX, exogenous factors, optimal transportation.