Fault diagnosis of rotating machineries based on Laplacian eigenmaps

Aiming at the problem that fault signals of rotating machineries are complex and hard to extract, a novel fault diagnosis approach based on Laplacian Eigenmap (LE) for rotating machineries was proposed. The monitored signals of three typical faults of rotating machineries were extracted and converted, 26 time domain and frequency domain features were obtained. In the high-dimensional feature space constructed with those features, LE algorithm was used for feature fusion, and the fault essence and regularity hidden in the high dimensional feature space were extracted to identify incipient fault types. Using two-dimensional or three-dimensional images, the lower dimensional results extracted were expressed, and taking sample recognition rate, between-class scatter and within-class scatter of cluster analysis method as measuring indexes, they were analyzed from the perspective of pattern recognition. The results showed that compared with the principal component analysis (PCA) and kernel principal component analysis (KPCA), LE can be used to better extract effective features from the high-dimensional feature space to reveal the equipment running status, and realize classification and identification of rotating machineries' incipient faults.