Remaining Useful Life Prediction of Rolling Bearings Based on PCA and GSACO-SVR Model

Abstract

Abstract The prediction of the remaining useful life (RUL) of rolling bearings facilitates the better development of maintenance programs. It is very important to improve prediction accuracy. We proposed an improved optimized support vector regression (GSACO-SVR) model to accurately predict the RUL of bearings, which is based on a new golden sine ant colony algorithm (GSACO) aiming to optimize the support vector regression (SVR) parameters. Compared with SVR, fruit fly algorithm, and ant colony algorithm under different working conditions by experiments, the GSACO-SVR model has more accurate prediction results and better bearing life degradation trend.