FORECASTING OF THE PARAMETERS OF MICRO-FAULT OF COAL SEAMS USING ARTIFICIAL INTELLIGENCE

Abstract

Purpose. Research and development of a method for predicting micro-fault fractures in coal seams using artificial intelligence methods. Methodology. To carry out the study, a neural network of the sing structure was formed. The training of the network was based on the principles of natural selection, which is the strongest one. The implementation of this principle was designed to support the genetic algorithm. Each network was tested for fitness, proportional to its ability to predict, and the best one was used for forecasting. The forecast was made in such a way that from the base point the forecast was made for the entire project site. Results. The research was conducted to develop a method for predicting the coordinates and amplitudes of low-amplitude disturbances using neural networks and genetic algorithms.The method consists in training a neural network based on reliable data taken from mining plans and, after a certain number of trainings, forecasting micro-fault discontinuous disturbances on the project site. The parameters of the violations revealed by tunnelling and cleaning works are used as training data. Originality. The method of forecasting micro-fault discontinuous disturbances using artificial intelligence allows predicting the presence and probability of distribution of disturbances in the predicted area. The main advantage of this method is the minimal cost with sufficient reliability. This favourably distinguishes it from the known methods. The application of this method does not require the use of special equipment in mine conditions, which significantly reduces the labour-intensiveness of the forecast. The efficiency of this method allows you to quickly and in a short time make a forecast on a given site, which contributes to the fastest possible commissioning of new mining fields. Practical value. The proposed method can be used to predict micro-fault discontinuous disturbances in the extraction fields of mines in the Ukrainian Donbas to assess the reliability of working out the extraction pillars and ensure the stability of the preparatory works due to their rational location. At the same time, the forecasting technique is being improved, taking into account the main characteristics of the physical process of the genesis of micro-fault discontinuous disturbances.