Vibration Monitoring of On-load Tap-changer by Resonance-based Sparse Signal Decomposition

On-load Tap-changer (OLTC) is one of the significant components in power transformer. The breakdown of frequently switched OLTC may cause the cutoff of the corresponding in-service transformer. Hence, suitable indicators are needed for effectively monitoring the operation condition of OLTC. Although the vibration analysis method has been proven to be effective for the mechanical condition assessment of the OLTC, the diagnostic parameters are mostly submerged in the noises during the switch-over process of OLTC. To solve this problem, this paper presents a new methodology - resonance-based sparse signal decomposition (RSSD) - to extract the determining vibration signals' features for the monitoring of OLTC. The vibration signal is decomposed into high-resonance components, low-resonance components, and redundant components. The obtained fault characteristics are mainly contained in the high-resonance components, which can be used for the fault detection.