Research On Fault Diagnosis Of High Voltage Circuit Breaker Based On Mechanical Vibration Signal

As an import component of the electric system, HV circuit breakers play the key role of control and protection for power network. Therefore, the action reliability of HV circuit breaker is very import for the electric system. During the operation process of the high voltage circuit breaker, the changes of vibration signals reflect the machinery states of the circuit breaker. Therefore, the article main work concentrates the fault diagnosis research of HV circuit breakers based on mechanical vibration signal and mainly researched several aspects including the vibration signal acquisition, feature extraction of vibration signal and fault identification.Firstly, to build the vibration signal acquisition platform of the circuit breaker using the virtual instrument technology. In this paper, the type ZW32-12 of vacuum circuit breaker is as the research object. And collecting the four type vibration signals of the normal state?the lack of lubrication state?the foundation bolt looseness state and the energy storage spring shed state of circuit breaker.Then collecting a large number of signal samples for the subsequent feature extraction and fault identification.Secondly,this paper has researched several feature extraction methods of vibration signal. The feature extraction can be divided into two parts of features separation and characteristics calculation.In this paper, ensemble empirical mode decomposition and wavelet packet decomposition method were respectively used on the characteristics of vibration signals separation. Then,using the correlation dimension of the fractal theory and the energy entropy of the entropy theory for characteristics calculation respectively.So we can obtain the EEMD-correlation dimension,wavelet packet-correlation dimension and the EEMD-characteristic entropy of vibration signal characteristics.Finally,using the three kinds characteristics of vibration signals to identify the fault of circuit breaker. In this paper, the circuit breaker fault diagnosis respectively use the support vector machine(SVM) and BP(Back Propagation) neural network recognitionmethod.The recognition results show these methods can get more than 80% of the correct recognition rate.So, this research has great value of engineering practicability.