Research And Implementation Of On-line Monitoring And Fault Diagnosis For Intelligent Vacuum Circuit Breaker

Vacuum circuit breaker is an important equipment in power system.With the development of digital technology,the State Grid puts forward the requirements of on-line monitoring and fault diagnosis for vacuum circuit breakers.At present,the on-line monitoring and fault diagnosis of vacuum circuit breaker is mainly based on a single signal,which is difficult to achieve a comprehensive diagnosis.Aiming at this defect,this paper takes 12kV iVD4 intelligent vacuum circuit breaker of ABB company as the research object to carry out the research of on-line monitoring and fault diagnosis.Firstly,the mechanism of vacuum circuit breaker is studied and is divided into auxiliary module,transmission module and major loop module.Referring to the literature,it is determined that in addition to the insulation fault,the most common fault of vacuum circuit breaker is refused operation fault,which often occurs in auxiliary module and transmission moduleSecondly,the type and installation location of sensor of the auxiliary module and the transmission module are determined.A variety of simulation methods of refused operation fault are designed.Under no-load condition of vacuum circuit breaker,fault implantation test is carried out to obtain current signal,vibration signal,angular displacement signal and pressure signal.After the signal samples are smoothed by wavelet denoising,they are used for feature extraction and fault diagnosis.Then,aiming at the transmission module,the mean square root value,skewness index and kurtosis index of the signal are extracted by using the time-domain waveform of vibration;A algorithm of the complete empirical mode decomposition with adaptive noise and autocorrelation is studied,which decomposes the vibration signal into IMF components with strong correlation with the original vibration signal and extracts Hilbert marginal spectrum energy from the high frequency IMF1-IMF3 components;The wavelet packet is used todecompose the vibration signal and extract the energy spectral entropy of each scale.According to the angular displacement and pressure signals,a combined method of travel-pressure is studied to extract the important mechanical parameters of vacuum circuit breaker.All the features make up the feature space,and the sensitive features of the transmission module are selected by the feature evaluation technology as the fault diagnosis features of the module.For the auxiliary module,the method of current contour is used to extract the features of the opening/closing release and the energy storage motor;The sample entropy method is used to extract the current sample entropy of the release;The integration value of the three stages of the energy storage motor,such as starting,idling and working,is extracted by the subsection integration method.Similarly,the feature evaluation technology is used to select the sensitive features of the auxiliary module to be the fault diagnosis features of the module.Finally,non-linear state estimation technology,support vector machine and probabilistic neural network are applied to diagnose the faults of vacuum circuit breakers by using sensitive features of transmission module and auxiliary module respectively,and the diagnostic results are compared.Under limited samples,the optimal fault diagnosis algorithm of the auxiliary module is the non-linear state estimation and the detection rate is 100%;The optimal fault diagnosis algorithm of the transmission module is the probabilistic neural network and the detection rate is 97.59%.Both of them cooperate with each other to realize on-line monitoring and fault diagnosis of intelligent vacuum circuit breaker comprehensively and efficiently.