Dynamic design of technical and economic indicators of open-pit mining with the help of neuron network technologies

Purpose. The dynamic management conditions of mining enterprises require rapid adaptation to changes in the external environment, which is seen as possible through the use of neural network technologies. In this context, the purpose of this work is to study the experience of using these technologies and develop a neural network model for forecasting the technical and economic indicators of the enterprise, based on historical data of its functioning. The methods. The paper uses methods of analysis and synthesis of literary sources of information when studying the experience of using neural network technologies in the design of open-pit development; retrospective analysis of project decisions and technical and economic performance indicators of mining enterprises; neural network modeling – for predicting the cost of mineral extraction. Findings. The analysis of scientific sources and project solutions was performed. In the TensorFlow environment, a neural network model for predicting the cost of ore mining was obtained, which demonstrated a smaller RMSE deviation than the "naive" model, which allows to talk about real predictive performance. The resulting model made it possible to predict the design values of the cost of goods for large iron ore open-pits of Kryvbas, comparable to detailed calculations of development projects. The originality. The conducted retrospective and engineering analysis of project solutions made it possible to identify the most promising scientific approaches to the design of open-pit mining, in particular, those based on neural network technologies. A regression model for forecasting technical and economic indicators of open-pit mining operations was obtained and the accuracy of its operation was assessed. Practical implementation. As a result of the performed research, the values of the cost of extracting a useful mineral for the deep steep deposits of Kryvbas were predicted. The obtained results can be used by design organizations and mining enterprises when designing the open-pit mining.