Any technological process including metallurgical processes is supported by the control system operation, accompanied by large information flows to be formed. However, the most part of this information is not used by specialists due to restricted capabilities of a human being that is, in fact, incapable of processing such information flows. It has been demonstrated that process control is significantly influenced by the human factor. This paper contains a methodology to process production information that permits the process personnel to use its potential in a more rational way to control the process. An approach has been reviewed to studying metallurgical processes using the analysis of indirect indicators of process management, namely the spectral density and auto-correlation function of main process parameters. A method to separate useful signals from noise has been studied. A method has been given to check the ACS management efficiency using primary material flows. The adopted methodology for processing experimental data permits interpreting the obtained results for their further practical application to develop new algorithms for process control and improve the existing control system.

The paper presents a methodology for training neural networks for vision tasks on synthe-sized data on the example of steel defect recognition in automated production control systems. The article describes the process of dataset procedural generation of steel slab defects with a symmetrical distribution. The results of training two neural networks Unet and Xception on a generated data grid and testing them on real data are presented. The performance of these neural networks was assessed using real data from the Severstal: Steel Defect Detection set. In both cases, the neural networks showed good results in the classification and segmentation of surface defects of steel workpieces in the image. Dice score on synthetic data reaches 0.62, and accuracy—0.81.

The article covers the issues of increasing the efficiency of tunneling&cleaning combines “Ural”, widely applied at domestic potash mines. It is pointed out that it is possible to provide rational parameters of the process of destroying potash arrays with cutting tools of the executive bodies of mining combines. This assumption is based on the assessment of the drivers load which will reduce the specific energy consumption for ore extraction and will reduce the number of small unenriched classes in breakage products. The authors substantiate rational regime parameters of the operation of the planetary-disk executive bodies in the Ural combines. A method for automatic control of the tunnel&cleaning combine “Ural” is proposed, which ensures the implementation of the process of destruction of the potash massif with rational cutting parameters on the basis of information on the drivers load.