On?load tap?changer fault mode recognition based on the singular value of Hilbert energy spectrum time?frequency matrix and spectrum entropy

To solve the difficulty of eigenvalue extraction on-load tap-changer (OLTC) vibration signals, an method time-frequency matrix based on Hilbert energy spectrum and spectral entropy is proposed in this study. By optimizing Hilbert-Huang transform (HHT) algorithm, signal distribution obtained. The was extracted to characterize chaotic degree distribution, avoided problem inaccurate feature quantity such as similarity due minute time difference during acquisition. A new constructed by frequency band division singular value decomposition (SVD) carried out. vector obtained contains time, information, plane which represent essential features original signal. mechanical states OLTC were characterized value. measured signals under three typical faults driving mechanism are input into intelligent multi-classification support machine (DAG-SVM) for pattern classification. results show that can accurately identify state OLTC, has practical application significance.