Abstract: This study presents a new manifold learning framework for machinery fault diagnosis, in order to further improve fault diagnosis accuracy. The new manifold learning framework contains two stages: unsupervised manifold learning for nonlinear denoising and supervised manifold learning for feature extraction. Firstly, the nonlinear denoising method with unsupervised manifold learning was introduced, which combined advantages of manifold learning in revealing nonlinear manifold structure as well as advantages of phase space reconstruction in representing spatial distribution of signal and noise. Then, fault feature extraction was carried out according to the frequency spectrum of vibration signals after denoising. In order to reduce…the high dimension and remove redundant information of frequency spectrum, an improved supervised local tangent space alignment (ISLTSA) was proposed to further enlarge diversity of the fault samples and thus increase separability. Finally, the extracted low-dimensional fault features were inputted into a pattern recognition method for fault identification. The effectiveness of the proposed method was verified by studying the fault diagnosis of bearings.
Show more
Abstract: To overcome the low diagnosis accuracy caused by the scarcity of labeled training samples, a fault diagnosis method was proposed using orthogonal Semi-supervised linear local tangent space alignment (OSSLLTSA) for feature extraction and transductive support vector machine (TSVM) for fault identification. Through extracting the statistical features were extracted from the sub-bands of vibration signals decomposed by wavelet packet decomposition (WPD), the high-dimensional feature set could be obtained. Following that, the improved kernel space distance evaluation method was applied to remove non-sensitive fault features. Then, a semi-supervised manifold learning method (OSSLLTSA) was proposed to reduce the dimensionality of the fault feature…set, and thus to extract fused fault features with high clustering performance. OSSLLTSA overcomes the over-learning of supervised manifold learning and projection aimlessness of unsupervised manifold learning. Finally, the low-dimensional feature set after dimension reduction was inputted into TSVM for fault diagnosis. TSVM was able to completely utilize the fault information contained in unlabelled samples to modify the model, and the trained fault diagnosis model has better generalization ability. The effectiveness of the proposed method was verified based on the case of gearbox fault. Experimental results showed that the proposed method is able to achieve very high fault diagnosis accuracy even when labeled samples were insufficient.
Show more