Research on the Optimized Algorithms for Support Vector Regression Model of Slewing Bearing's Residual life Prediction

Abstract

In this paper, with the difficulties of slewing bearing's residual life prediction, a new method of support vector regression (SVR) life prediction model is proposed. In order to improve the accuracy of SVR model, the algorithms for optimizing the SVR internal parameters are analyzed and researched. Binding assay, the effectiveness and stability of the SVR prediction model optimized respectively by Gridding Optimization Algorithm (GOA), Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) are analyzed according to the different input signal, which verifies the superiority of PSO. Finally, the SVR prediction model optimized by PSO is used to predict the remaining life of slewing ring, which is compared with its actual life to demonstrate the effectiveness of the SVR prediction model optimized by PSO on slewing bearing life prediction.