A New Feature Extraction Technique for Early Degeneration Detection of Rolling Bearings

Abstract

Feature extraction technology is an important part of bearing diagnosis, especially for early degradation detection. However, the traditional feature extraction technology can not effectively remove noise or is not sensitive to periodic weak faults, which leads to be inclined to raise false alarms and prediction delay for early degradation detection. In order to solve these two issues, a new feature extraction technique is presented based on Envelope Harmonic-to-noise Ratio (EHNR) and Adaptive Variational Mode Decomposition (AVMD). First of all, the minimum average envelope entropy is used as the objective function to search the optimal parameters of the Variational Modal Decomposition (VMD) adaptively by the Grey Wolf Optimization (GWO) algorithm. The problem of under-decomposition or over-decomposition caused by improper parameter setting is avoided. Then, a new index called Effective Weighted Sparseness Kurtosis (EWSK) is proposed. This index can separate the effective modal components and noise modal components only by the positive and negative results, so as to achieve the purpose of removing noise interference and retaining a large amount of fault information. Finally, the EHNR of the reconstructed signal is calculated, and its sensitivity to periodic fault shock is utilized to detect the early degradation starting point of the rolling bearing. Experimental results show that the proposed method outperforms several state-of-the-art detection methods in terms of early degradation point detection, false alarm rate and computational complexity. The superior performances of the presented AVMD-EHNR method can provide the basis for early fault diagnosis and remaining useful life prediction of rolling bearings.