Study of coal mill vibration faults based on SVD-VMD and LSSVM

In order to deeply analyze the fault types of coal mill under different vibration anomalies, based on the collected vibration signals, a coal mill vibration fault feature extraction method based on SVD decomposition fused with VMD-HHT marginal spectra is proposed and combined with the LSSVM model to realize the vibration fault diagnosis of coal mill. Firstly, VMD decomposition is performed on the vibration signals of different parts of the coal mill and the IMF component matrix is reconstructed using SVD decomposition; then, HHT marginal spectrum analysis is performed on the reconstructed IMF components to extract the feature vectors; finally, the LSSVM model is used to classify each fault feature to complete the fault diagnosis, and the validity and accuracy of the proposed method are demonstrated by examples. The results show that the SVD-VMD-HHT marginal spectrum can express the fault information more accurately, and the fault diagnosis accuracy of this method can be up to 98%, which is higher than that of the VMD-HHT marginal spectrum with a single sensor; this method can accurately identify the fault type of vibration abnormality 2.5 minutes after the fault occurs.