Neural network technologies in mining data on particle size distribution of muck pile rocks

INTRODUCTION. Current development of the mining industry increases the role of tools that support decision making and impact its speed. These trends have already affected almost all spheres of human activities and are expanding into the applied areas as artificial intelligence systems. The use of neural networks in the data mining technology is a current trend that is continuously evolving. The study solves the problem of defining the particle size distribution of muck pile rocks using the neural network technologies. The U-Net artificial neural network was used to solve the scientific and practical problem. This network had been trained to allow it to gain experience and adapt to the ongoing changes in the input data on the particle size distribution for different mineral deposits as it accumulates data. METHODS. A complex approach that included a system scientific analysis and generalization of previously published studies was used in addressing the tasks set. The U-Net architecture was used for preliminary assessment of the particle size distribution. RESULTS. The lump size parameters of the muck pile rocks were determined using the Russian-made Davtech equipment as the hardware and software support for experimental studies. CONCLUSIONS. The data obtained in the course of the study will make it possible to develop recommendations to optimize the control modes of an open-pit excavator, which will ultimately reduce the number of equipment failures and increase its service life.