Artificial neural networks using for solving of the tribological problems

• Main Authors: Uvarov V.M., Bespalov S.A.
• Format: Article
• Language: English
• Published: National Academy of Sciences of Ukraine. Physico- Technological Institute of Metals and Alloys 2019-09-01
• Series: Металознавство та обробка металів
• Subjects: structural steel; hardening temperature; wear resistance; contact interaction surface; modeling; artificial neural networks.
• Online Access: https://momjournal.com.ua/sites/default/files/1_10.pdf

The sclerometric studies of steel 40X after quenching from 860, 1050 ° C and high tempering showed the average microhardness increases with an increasing quenching temperature. On the scratches two types of microhardness maximum value are revealed. The first one (T1) are closest to each other. The average distance between them does not change practically depending on the temperature of quenching. The second (T2) are the maximums with a longer period, which increases and more clearly manifests itself with increasing quenching temperature. It is noted, that the determining factor of the microhardness cyclical changes with a T1 period are the grains and martensite packets boundaries, and with T2 period are the uneven carbon distribution, which increases with increasing quenching temperature. The influence of hardening temperature on the wear resistance of 40X steel after improvement is investigated. It was revealed that quenching from 1050 ° C and high tempering increase its tribological characteristics, as well as reduce the wear rate of the counterbody compared with standard heat treatment. The nature of the destruction of contact surfaces is studied. It has been shown, that in improved samples hardened from 860 ° C, it occurs according to the cleaving and smooth delamination mechanisms with plastic deformation. The nature destruction of the contact interaction surface is changes with the temperature quenching increasing to 1050 ° C and with high tempering. The microstructure’s areas with greater then surrounding volume fracture resistance during friction are identified. The structural-geometric parameters characterizing the roughness and bearing capacity of the contact interaction surface are analyzed. The temperature quenching increasing to 1050 ° C reduce its roughness and increase the surface reference curve, which characterizing its bearing capacity. The possibility of using artificial neural networks to predict the tribological properties of structural steels are considered. Based on the results of modeling the structural and geometric parameters of the surface an analysis is made of the bearing capacity of the contact surface of 40X steel samples depending on the hardening temperature.