The paper presents a methodology for training neural networks for vision tasks on synthesized 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.

Big data analysis is becoming a daily task for companies all over the world as well as for Russian companies. With advances in technology and reduced storage costs, companies today can collect and store large amounts of heterogeneous data. The important step of extracting knowledge and value from such data is a challenge that will ultimately be faced by all companies seeking to maintain their competitiveness and place in the market. An approach to the study of metallurgical processes using the analysis of a large array of operational control data is considered. Using the example of steel rolling production, the development of a predictive model based on processing a large array of operational control data is considered. The aim of the work is to develop a predictive model of rolling mill roll wear based on a large array of operational control data containing information about the time of filling and unloading of rolls, rolled assortment, roll material, and time during which the roll is in operation. Preliminary preparation of data for modeling was carried out, which includes the removal of outliers, uncharacteristic and random measurement results (misses), as well as data gaps. Correlation analysis of the data showed that the dimensions and grades of rolled steel sheets, as well as the material from which the rolls are made, have the greatest influence on the wear of rolling mill rolls. Based on the processing of a large array of operational control data, various predictive models of the technological process were designed. The adequacy of the models was assessed by the value of the mean square error (MSE), the coefficient of determination (R²), and the value of the Pearson correlation coefficient (R) between the calculated and experimental values of the mill roll wear. In addition, the adequacy of the models was assessed by the symmetry of the values predicted by the model relative to the straight line Ypredicted = Yactual. Linear models constructed using the least squares method and cross-validation turned out to be inadequate (the coefficient of determination R² does not exceed 0.3) to the research object. The following regressions were built on the basis of the same operational control database: Linear Regression multivariate, Lasso multivariate, Ridge multivariate, and ElasticNet multivariate. However, these models also turned out to be inadequate to the object of the research. Testing these models for symmetry showed that, in all cases, there is an underestimation of the predicted values. Models using algorithm composition have also been built. The methods of random forest and gradient boosting are considered. Both methods were found to be adequate for the object of the research (for the random forest model, the coefficient of determination is R² = 0.798; for the gradient boosting model, the coefficient of determination is R² = 0.847). However, the gradient boosting algorithm is recognized as preferable thanks to its high accuracy compared with the random forest algorithm. Control data for symmetry in reference to the straight line Ypredicted = Yactual showed that, in the case of developing the random forest model, there is a tendency to underestimate the predicted values (the calculated values are located below the straight line). In the case of developing a gradient boosting model, the predicted values are located symmetrically regarding the straight line Ypredicted = Yactual. Therefore, the gradient boosting model is preferred. The predictive model of mill roll wear will allow rational use of rolls in terms of minimizing overall roll wear. Thus, the proposed model will make it possible to redistribute the existing work rolls between the stands in order to reduce the total wear of the rolls.

This paper presents studies on the possibility of utilization of technogenic waste from the metallurgical industry by the method of complex processing in order to reduce the anthropogenic load on the environment of the region with the example of the zinc silicate-magnetite-carbon system. The selected sample of clinker dump from welting was subjected to chemical and scanning electron microscopic analyses and thermodynamic modeling. Thermodynamic studies were carried out in the temperature range 1600–2200 K and pressure p = 0.1 MPa, modeling the process of electric melting of clinker from welting in an arc furnace using the software application Astra 4 developed at the Bauman Moscow State Technical University (Moscow, Russian Federation). As a result of the thermodynamic modeling, the optimal temperature range was established, which was 1800–1900 K. Thermodynamic studies established that it is possible to drive away zinc from the system under study by 99–100% in the entire temperature range under study. The maximum degree of silicon extraction (α_Si) in the alloy is up to 69.44% at T = 1900 K, and the degree of iron extraction (α_Fe) in the alloy is up to 99.996%. In particular, it was determined and proved that clinker waste from welting can act as a secondary technogenic raw material when it is processed as a mono mixture to produce iron silicides with a silicon content of 18 to 28%.

The paper studies the properties of brass workpieces for antifriction rings under severe plastic deformation by high-pressure torsion. The evolution of microstructure and mechanical properties of deformed workpieces after six cycles of deformation by high-pressure torsion at 500 °C have been studied. All metallographic studies were performed using modern methods: transmission electron microscopy (TEM) and analysis electron back scatter diffraction patterns (EBSD). The deformation resulted in an ultrafine grained structure with a large number of large-angle boundaries. The strength properties of brass increased compared to the initial state almost by three times, the microhardness also increases by three times, i.e., increased from 820 MPa in the initial state to 2115 MPa after deformation. In this case, the greatest increase in strength properties occurs in the first two cycles of deformation.

Copper-sheathed steel wires combine the conductivity of copper and the traction resistance of steel, which makes a bimetallic wire an ideal material for the construction of power lines. Currently, there is a small number of studies devoted to the change in the microstructure of steel-copper wire during its strain. Since steel and copper have different mechanical properties, these metals at the interface can be deformed in different ways. Therefore, the present research is devoted to the study of ECAP-drawing process impacts on the properties of bimetallic steel-copper wire. During the conducted studies, the possibility and efficiency of using the combined strain technology for the formation of ultrafine grained structure and increased strength properties of steel-copper wire have been proved.

In the study, experimental and theoretical studies were carried out to assess the influence of the shapes of dents in the tank wall on the stress-strain state of the defect zone. By testing fragments of a cylindrical tank, it was found that the most appropriate expression is (5), which could take into account the leaching of the tank wall, resulting in a decrease in the stress concentration index. At the same time, during theoretical studies in this paper, it was found that polynomials determined the stress concentration coefficient, where the obtained analytical expression data were compared with the data determined numerically in the ANSYS program, and it was found that the spread was from 2% to 10%. According to the results of a numerical study of the stress-strain state of the dent zone in the tank wall, graphical dependences of the stress concentration coefficient on the dimensionless depth of the dent for various values of the dimensionless radius of the dents and do not exceed 2% of the indicators that are obtained. At the conclusion of the experimental and numerical studies, a conclusion was made about the degree of influence of the geometric dimensions of the dents on the stress concentration index.