Research On Fault Diagnosis Of High Voltage Disconnector Based On The Fusion Of Vibration Signal And Torque Signal

High voltage disconnector is used in a large amount in the power supply system.Its operation state directly affects the safety and reliability of the power system.Due to the lack of perfect condition monitoring means,its hidden dangers and defects can not be found in time.The statistics of defect data on site for many years show that the failure rate of high-voltage disconnector remains high,and mechanical failure is one of the main failure types of disconnector.However,at this stage,the monitoring level of disconnector is low,the fault samples are scarce,there are many types of mechanical faults,and the self-awareness is weak.With the continuous improvement of the requirements of intelligent operation and maintenance of power equipment,the research on disconnector condition monitoring and fault diagnosis is of great significance to maintain the stable operation of power system.Under this background,this paper studies GW4 Disconnector.Firstly,based on the geometric basic structure and opening and closing action process of the disconnector,the dynamic model of the disconnector is established with the help of Solid Works and Adams,and the parametric linkage is established with the help of the parametric modeling function of Adams;The contact parameters and simulation steps of the disconnector simulation model are determined through the Hertz theoretical calculation results and ANSYS modal analysis results.The accuracy of the model is verified by comparing the force curve of the contact finger of the model with the theoretical curve and the operating torque curve of the model with the measured curve,so as to provide theoretical basis and simulation data for the subsequent research on the fault diagnosis of the disconnector.The fault characteristics of vibration signal and torque signal of disconnector are analyzed.The time-frequency energy matrix of vibration signal is extracted by wavelet packet transform as the fault characteristics of vibration signal.The time-domain waveform characteristic matrix is constructed as the fault characteristics of torque signal by selecting 8 time-domain indexes of operating torque signal,such as peak value,average value,variance,root mean square,kurtosis,skewness,waveform factor and pulse factor,A convolutional neural network(CNN)model is constructed by using the method of deep learning to fully mine the corresponding relationship between the extracted fault features and the fault types of disconnectors,and the final fault discrimination result is obtained by decision-making level fusion through Dempster Shafer(D-S)evidence theory.The effectiveness of this method is verified by the analysis of examples and the test of the diagnosis effect under different faults.Finally,the fault simulation and data acquisition scheme of disconnector are studied,and the fault simulation test of GW4-126 disconnector is carried out on site;The vibration signals and operating torque signals of the closing process of the disconnector under different mechanical states are obtained;The characteristics of measured operating torque signal and vibration signal in the closing process of disconnector under different faults are analyzed respectively,and the rationality of the simulation model is verified.Finally,the parameters of the diagnosis model are adaptively adjusted through the measured data,and its diagnosis effect is tested to verify the effectiveness and adaptability of the fault diagnosis method.