Papers by Keyword: Machinery Fault Diagnosis

Abstract: Wavelet analysis, being a popular time-frequency analysis method has been applied in various fields to analyze a wide range of signals covering biological signals, vibration signals, acoustic and ultrasonic signals, to name a few. With the capability to provide both time and frequency domains information, wavelet analysis is mainly for time-frequency analysis of signals, signal compression, signal denoising, singularity analysis and features extraction. The main challenge in using wavelet transform is to select the most optimum mother wavelet for the given tasks, as different mother wavelet applied on to the same signal may produces different results. This paper reviews on the mother wavelet selection methods with particular emphasis on the quantitative approaches. A brief description of the proposed new technique to determine the optimum mother wavelet specifically for machinery faults diagnosis is also presented in this paper.

Abstract: Aimed at the problem of low resolution and cross term interference of the traditional time-frequency analysis methods, a new time-frequency filtering method based on generalized S transform is proposed. The method is extended under the premise of the linearity, lossless invertibility, high time-frequency resolution of S transform. On the basis, a coefficient which is direct to the signal energy distribution is introduced. In this way, the resolution of the S transform can be adjust adaptively. Eventually, this method is applied to the time-frequency filtering. The results of simulation and faulty bearing show that the proposed methodology can achieve good effect of noise reduction, and be more suitable for the non-stationary characteristics of vibration signals.

Abstract: Aiming at problem on limiting development of machinery fault intelligent diagnosis due to needing many fault data samples, this paper improves a multi-classification algorithm of support vector machine, and a multi-fault classifier based on the algorithm is constructed. Training the multi-fault classifier only needs a small quantity of fault data samples in time domain, and does not need signal preprocessing of extracting signal features. The multi-fault classifier has been applied to fault diagnosis of steam turbine generator, and the results show that it has such simple algorithm, online fault classification and excellent capability of fault classification as advantages.