A parameter prediction model of running-in based on surface topography

Abstract

Although surface topography is a directive and useful method to evaluate the quality of worn components, the common belief that the surface roughness after running-in is independent of the nature of initial roughness makes it difficult to investigate the surface topography of worn components after running-in based on unworn surface before running-in process. Trying to build the connection of surface topography before and after running-in, this article assumed running-in as a black box and adopted support vector machine as machine learning method to simulate the complex process. Through the training and validation of the simulation, the predictive model of surface topography after running-in based on unworn surface topography was established, which indicated the existence of a correlation between the surface topographies before and after running-in. The model analysis revealed that the hybrid properties of surface topographies before and after running-in have strong correlation. Especially, the correlation between input parameter Sk and output parameter Sq is a particular one among all relationships of parameter pairs.