Prediction of Bearing Raceways Superfinishing Based on Least Squares Support Vector Machines

Electrochemical belt superfinish (ECBS) technology, which has the advantage of high surface quality and efficiency, has been applied in bearing raceway finishing. However, the finishing effect of ECBS is dominated by many factors. Due to the relationship between the finishing effect and the factors is very complicated, there is a complex and dynamic behavior. Therefore, it is difficult to predict the finishing results and select the suitable processing parameters in ECBS. In this paper, Least Squares Support Vector Machines (LS-SVM) is proposed to solve this problem. An intelligent predictive multiple regression model of the non-linear relationship between machining parameters and the machining effect was established based on the experimental data. The comparison between the calculated results of SVMs and the experimental results under the same conditions was carried out, and the result indicates that it is feasible to apply LS-SVM in determining the processing parameters and forecasting the surface quality effects in ECBS.