Control of the dynamic stability of metal-cutting systems in the process of cutting based on the fractality of roughness of the machined surface

Abstract

The problem of increasing the efficiency of mechanical treatment within modern automated production is relevant for many branches of the processing industry. This problem requires a deep study of the physical processes occurring during cutting. The urgency of the problem increases even more with the development of digital production in our country. Today, in the presence of a wide range of products, enterprises are forced to create conditions for reducing the technological cycle when manufacturing a particular product. To carry out the study, an experiment was conducted in which the U8 carbon steel was used as the processed material, and the T15K6 alloy was used as the tool material. During the experiment, the authors observed a change in the roughness of the machined surface depending on the cutting speed. The paper considers the possibility of assessing the quality of the surface layer during cutting based on fractal and neural network modeling. It is identified that the fractal dimension shows the regularity of the reproduction of the machined surface roughness during cutting. The calculated fractal dimension of the machined surface roughness correlates well with the values of the machined surface roughness (correlation coefficient is 0.8–0.9). A neural network structure has been developed, which allows controlling the machined surface quality depending on the cutting conditions. The authors studied the possibility of using neural network models to control technological systems of cutting treatment. When creating digital twins, it is proposed to take into account factors affecting the quality of the treated surface and processing performance, which are poorly accounted for in modeling, as well as when conducting full-scale experiments during machining. Such factors are wear of the cutting tool, the process of plastic deformation, and cutting dynamics.

Keywords

• cutting process
• machined surface roughness
• neural network
• surface layer quality control