This paper carries out the research on the continental mining of solid minerals and on the search of the alternative solutions of raw material’s procurement for leading industries. A value of conducting underwater operations for extraction of shelf zone’s ferromanganese nodules is specified. Schematic diagram of conducting mining operations is submitted. The construction of a mining complex for mining ferromanganese nodules of the Baltic Sea with the possibility of its separation partial enrichment on the seabed during the process of production is proposed. An analytical model of the system with bottom device and flexible traction element is given. The interaction between a flexible traction element and a bottom-mining unit allows to define the character of relative motion and also to describe the position of the flexible traction element as it moves in the aquatic environment. Specific conditions for steady operation of the complex for its inseparable moving on a seabed are presented.

Electromechanical equipment is widely used. It is used in industrial enterprises, in the spheres of public services, in everyday life, etc. Maintenance servicing of electromechanical equipment is an important part of its life cycle. High-quality and timely service can extend the life of the electromechanical equipment. The creation of special systems that simplify the process of servicing electromechanical equipment is an urgent task. Such systems can shorten the time for maintenance of electrical equipment, and, therefore, reduce the cost of maintenance in general. This article presents an analysis of information on the operation of service services for maintenance and repair of electromechanical equipment, identifies the list of services, and estimates the time required to perform basic service operations. The structure of the augmented reality system is presented, the ways of interaction of the augmented reality system with the automated control systems working at the enterprise are presented.

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.