Impact wear behavior of austenitic steel bucket teeth based on machine learning

The relationship between the material properties of bucket teeth, the material properties of coal, the dynamic parameters of bucket teeth, the wear depth of bucket teeth, and the pressure on the surface of bucket teeth was investigated using the enhanced discrete element simulation software EDEM in conjunction with the modified Archard wear model. The objective was to determine the impact of specific parameters on the wear of bucket teeth made from austenitic steel. The combination of post-processed cloud images from the EDEM software and wear nephograms, along with the impact wear surface of the sample, led to the conclusion that the primary wear mechanisms of austenitic steels were abrasive wear and adhesive wear. Using a random forest approach, the critical parameters influencing wear depth were identified. A particle swarm optimization support vector machine was employed to accurately predict wear depth. The mean values of the determination coefficient (R 2 ) before and after feature selection were 0.96294 and 0.98074, respectively, further validating the accuracy of the feature selection process and providing a ranking of importance for identifying key factors to enhance wear resistance.