Intelligent Mechanical Fault Diagnosis Of High Voltage Circuit Breaker Based On Weighted Evidence Theory

As an important part of the field of electrical equipment monitoring and fault diagnosis,the on-line monitoring and mechanical fault diagnosis of high-voltage circuit breaker has become the focus of current research.However,after considering the present resarch situation,the current related researchs focus on the fault diagnosis system for monitoring single electrical or mechanical signals.Based on the reasonable analysis of present resarch combination,in consideration of the exigence and development trend of making efficacious use of high-voltage circuit breaker on-line monitoring,taking into account the extent defects of the single sensor fault diagnosis system,the thesis propose amechanical fault diagnosis method of high voltage circuit breaker based on multi-sensor information fusion technology.In this thesis,the vibration signal and the displacement-travel characteristic of the circuit breaker are measured,and compares with the characteristic information of the typical fault signal stored in the database,and the state information of the circuit breaker is obtained.Finally,the weighted D-S evidence theory is used to complete the fault diagnosis.This thesis introduces the basic principle of data fusion,and analyzes and classifies the common data fusion methods.It points out the uncertainty sources of single sensor fault diagnosis system in various fault diagnosis links,which proves the necessity of multi-sensor data fusion applied to circuit breaker online monitoring and mechanical fault diagnosis system.Based on the theory of wavelet analysis,the influence factors of wavelet basis function,the choice of decomposition level,the threshold size and the threshold function in wavelet denoising are analyzed,and a new method is proposed to evaluate the wavelet basis function.Based on the SNR formula and the LBQ correlation test method,an adaptive algorithm for the optimal decomposition of the number of layers in the wavelet threshold noise reduction algorithm is proposed.The vibration signal and the travel signal are collected by the vibration sensor and the rope displacement sensor respectively,and the eigenvector is extracted.On the basis of the traditional D-S evidence theory,the concept of evidence credibility is introduced to generate weighted evidence,which can solve the data conflict and combinatorial explosion phenomenon in data fusion.The hardware and software platform of on-line monitoring and mechanical fault diagnosis system of high voltage circuit breaker is set up.Through the simulation fault experiment of sulfur hexafluoride circuit breaker,the measured data is obtained as the fault knowledge base and the rest of the data is used as experimental test sample.The experimental results vertify the proposed method.