Research on Bearing Life Trend Prediction Method Based on Principal Component Analysis and Grey Model

Abstract

To predict the bearing residual life by using the characteristic index of vibration signal, the principal component analysis (PCA), the method is proposed to remove the redundancy and correlation between many characteristic indexes, to achieve the purpose of dimensionality reduction. The first principal component is used as the degenerate characteristic quantity to describe the bearing residual life, and the degenerate characteristic sequence is formed. The grey model is trained with the degenerate feature sequence, and the changing trend of the degraded feature series is predicted by the trained grey model. It is verified by the actual life cycle vibration signal of the measured bearing. The results show that the method based on principal component analysis and the grey model can effectively predict the life of bearing.