Multiple Health Phases Based Remaining Useful Lifetime Prediction on Bearings

摘要

Bearings are key components for all industrial machinery systems. The health status of bearings has great impact on the performance of rotating machineries. Remaining useful lifetime (RUL) estimation on bearings can effectively improve the reliability and availability of industrial machineries. In this paper, a multiple health phases based method is proposed for RUL estimation with application to bearings. Bags of word is brought into the method to model the time-frequency domain features of bearing vibration signals. Besides that, a gaussian mixture model is utilized to model the lifetime of various bearings to build accurate lifetime prediction model. Finally, the experiments demonstrate that the proposed method achieves a good performance comparing with other existing methods.