Application of frequency band energy decomposition using wavelet packet in condition monitoring of circuit breakers

Based on the algorithm of wavelet packets, the vibration signals during the operation of circuit breakers were decomposed into the individual frequency bands. The variations of the signal energy in these bands reflect the different mechanical conditions of circuit breakers. Main energy ingredients were extracted to form the eigenvector for condition monitoring and the tolerance range of the eigenvector was determined by the test results under each condition. These tolerance ranges supplied the quantified standard for condition monitoring of the circuit breaker, and a simple classification of the mechanical conditions was implemented consequently. The vibration signals of a ZN12-35 type vacuum circuit breaker under specific closing simultaneity were recorded from a single vibration sensor mounted on the mechanism box. After the wavelet packet decomposition, different closing simultaneities distributed in the individual areas of a complex plane can be apparently distinguished, which verifies the feasibility of the proposed method applied to the condition monitoring of circuit breakers. The repeated tests indicate that the wavelet packet decomposition has a high monitoring accuracy based on a stable measurement system.