Research On State Monitoring And Signal Joint Analysis System Of Circuit Breaker

High voltage circuit breaker is increasing in number with the expansion of the power grid capacity and scale.Therefore,the problem which State Grid is in urgent need is how to monitor and judge the operation status of the circuit breaker,study new methods for state monitoring of circuit breaker,and further improve the diagnostic accuracy of defects according to a variety of signals emitted during the operation.Based on the analysis of the structural features and the motion process of the circuit breaker operating mechanism,this paper proposes to apply the FMECA method to the determination of the state monitoring signal of the circuit breaker.Through the failure mode,influence and hazard analysis of breaker spring operating mechanism,the current,sound wave and vibration signal are identified as the monitoring parameters in system based on qualitative and quantitative results in order to master the mechanical condition of circuit breaker better.A circuit breaker status monitoring system based on multi-source signal joint analysis is established according to the existing online monitoring experience and the analysis of structure,function and realization scheme of circuit breaker status monitoring system.Signals of current,sound wave and vibration are recorded in the process.And the functions of signal acquisition,waveform display,data storage and fault diagnosis are realized by software.It meets the monitoring requirements of accurate data,comprehensive information and simple operation.The deficiencies of existing monitoring methods are overcome.Then,the general average empirical mode decomposition method is applied to the construction of the characteristic vector of the acoustic and vibration signals of the circuit breaker.After the EEMD decomposition of the acoustic and vibration signals of the circuit breaker,the energy distribution of the IMF component is selected to constitute the eigenvector.The classification accuracy of LibSVM for various states is used as the weight of the basic reliability.On the basis of studying the principle of multi-signal joint analysis,a new evidence fusion algorithm based on adaptive weight is proposed.It can avoid the possible conflict of evidence.Finally,the fault condition such as loose and unsmooth of circuit breaker is simulated in the laboratory.The accurate identification results are obtained and the feasibility of this method is verified according to the test.