Fault Condition Prognostic for Rotating Machinery based on New WEEMD and Adaptive Boosting Regression Algorithm

Pei Yao, Zhongsheng Wang, Hongkai Jiang, and Zhenbao Liu


Fault condition prognostic, adaptive boosting regression algorithm, WEEMD, high order spectrum slice


This paper addresses a fault condition prognostic for sudden failure of rotating machinery. The proposed method is based on the utilization of feature extraction by using signal processing technique, and adaptive boosting (adaboost) regression algorithm. In this paper, we decompose vibration signal using wavelet packet decomposition and ensemble empirical mode decomposition (WEEMD), and successively we utilize the high order spectrum slice to describe the process of a fault evolution, and finally adaptive boosting regression algorithm is adopted for predicting the fault conditions. Experimental results of rotating machinery show that adaboost regression is pronounced comparing with other regression methods for fault condition prognostics.

Important Links:

Go Back