This article is devoted to expanding the possibilities of assessing the technical state of the current consumption of asynchronous electric drives, as well as increasing the information capacity of diagnostic methods, in conditions of limited access to equipment and incompleteness of information. The method of spectral analysis of the electric drive current can be supplemented by an analysis of the components of the current of the Park's vector. The research of the hodograph evolution in the moment of appearance and development of defects was carried out using the example of current asymmetry in the phases of an induction motor. The result of the study is the new diagnostic parameters of the asynchronous electric drive. During the research, it was proved that the proposed diagnostic parameters allow determining the type and level of the defect. At the same time, there is no need to stop the equipment and taky it out of service for repair. Modern digital control and monitoring systems can use the proposed parameters based on the stator current of an electrical machine to improve the accuracy and reliability of obtaining diagnostic patterns and predicting their changes in order to improve the equipment maintenance systems. This approach can also be used in systems and objects where there are significant parasitic vibrations and unsteady loads. The extraction of useful information can be carried out in electric drive systems in the structure of which there is a power electric converter.

The results of experiments on determination of the cylindrical part mass loss of the cone-shaped strikers made of 38XM (AISI P20), У8 (AISI W1-A), and X12MФA (AISI D2) steels, which have been subjected to heat treatment applied by the striking tools manufacturers, and additional cold treatment, with their friction and abrasive wear on granite, are presented. It was found that the strikers made of steel X12MФA with all types of heat treatment show the greatest wear resistance under the wear conditions (static conditions), while the strikers made of 38XM have the lowest wear resistance. Cryogenic treatment and, to a lesser extent, cold treatment increase the wear resistance of all materials. The comparison of set values of wear resistance Ist and hardness HV of the strikers, as well as the analysis of their microstructure allows to conclude that wear resistance of a tool depends not only on its hardness, but also on the type, volume, shape of carbides forming the structure of material used for tool manufacturing. For this reason, after the cryogenic treatment, the tool made of X12MФA steel shows the greatest wear resistance, because its material has the highest content of carbon (~1.5%) and alloying elements (~13%), which provides up to 10.8% increase in the number of fine carbide particles in its structure, that ensure wear resistance increase, after the cold treatment. The results obtained for Ist are compared with the earlier obtained data on wear resistance of the same striker at striking upon granite (in dynamic conditions). The conclusion about similarity of the wear mechanisms of the analyzed strikers in dynamic and static conditions is made.

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.