Gear faults diagnosis based on wavelet-AR model and PCA

Abstract

Gear mechanisms are an important element in a variety of industrial applications and about 80% of the breakdowns of the transmission machinery are caused by the gear failure. Efficient incipient faults detection and accurate faults diagnosis are therefore critical to machinery normal operation. The use of mechanical vibration signals for fault diagnosis is significant and effective due to advances in the progress of digital signal processing techniques. Through virtual prototype simulation analysis and experimental study, a novel method for gear multi-faults diagnosis was presented in this paper based on the wavelet-Autoregressive (AR) model and Principal Component Analysis (PCA) method. The virtual prototype simulation and the experimental test were firstly carried out and the comparison results prove that the traditional Fast Fourier Transform Algorithm (FFT) analysis is not appropriate for the gear fault detection and identification. Then the wavelet-AR model was applied to extract the feature sets of the gear fault vibration data. In this procedure, the wavelet transform was used to decompose and de-noise the original signal to obtain fault signals, and the fault type information was extracted by the AR parameters. In order to eliminate the redundant fault features, the PCA was furthermore adopted to fuse the AR parameters into one characteristic to enhance the fault defection and identification. The experimental results indicate that the proposed method based on the wavelet-AR model and PCA is feasible and reliable in the gear multi-faults signal diagnosis, and the isolation of different gear conditions, including normal, single crack, single wear, compound fault of wear and spalling etc., has been effectively accomplished.