Research On Mechanical Fault Pattern Recognition Of Circuit Breaker Based On Vibration Signal Analysis

Circuit breakers are a vital part of power system equipment.If a fault occurs in the operation of the circuit breaker,it directly threatens the stable operation of the power grid,which brings hidden dangers to the economic development of the society.Therefore,the reliability of the circuit breaker has very strict requirements.The vibration signal of the circuit breaker contains a rich feature of the mechanical operating state.Traditional Time-frequency analysis methods cannot obtain the characteristic parameters in the vibration signal accurately due to the defects of them which a faster and more accurate fault diagnosis method is needed instead.Firstly,this paper takes the Z65 type circuit breaker as the object and builds the hardware platform.It mainly collects the vibration signals under the condition of fault of on-off and off-on,mechanism stuck,loose pedestal bolt and normal operating conditions,which is used for subsequent feature extraction and fault identification in the next steps.Secondly,this paper takes the Empirical Mode Decompositon(EMD)as an example.this paper Features of vibration signal can reflect the running state of the circuit breaker.For the purpose of In order to solve the inherent problems such as modal aliasing,improve the high-frequency extraction property and fault identification speed,a method based on improved variational mode decomposition(VMD)is proposed to get better fault diagnosis.Utilize VMD to decompose the vibration signal under different conditions,several modes with finite bandwidth and center frequency can be obtained.Fréchet distance is capable of detecting the duplicate mode.Then calculating the singular value of each mode as the input feature vector of the relevance vector machine(RVM)to identify the fault.Finally,compare with the diagnosis method based on empirical mode decomposition(EMD)and support vector machine(SVM),the method proposed performs better feature extraction property,increasing the success-rate effectively and identification speed as well and has better performance in the detection of weak faults.