Research On Simulation And Diagnosis Method Of Mechanical Fault Of High Voltage Disconnect Switch

As one of the largest electrical devices used in the power system,the complex working environment of the high-voltage disconnect switch makes its mechanical failure rate high,while the current inspection methods in engineering practice are mainly based on backward overhaul methods such as operating feel and dismantling inspection.A new fault diagnosis method is therefore necessary to maintain grid security.This paper addresses the shortcomings in the current research on disconnecting switch fault diagnosis methods,proposes a method of applying dynamics simulation to expand data samples,verifies the accuracy of the calculation results of dynamics simulation from theoretical analysis and simulation experiments,analyzes the characteristics of the change of operating torque during the action of the disconnecting switch,combines the Bayesian optimization algorithm and the random forest algorithm,establishes a disconnecting switch mechanical state classification model,and carries out the following work:(1)A hydrostatic analysis of the action process of the disconnector switch was carried out in stages,and expressions were derived for the relationship between the friction force,the contact finger contact force and the operating torque at different stages,and it was found that the operating torque was not only related to the friction force and the contact finger contact force,but also to the position of each linkage during the action process.(2)A simulation model of the rigid-flexible coupling dynamics of the disconnecting switch was established,and flexible modelling was carried out for components such as contact fingers and spring plates that undergo elastic deformation during the operation of the disconnecting switch,and simulation calculations were carried out for the model.By modifying the simulation duration,friction coefficient and linkage length,simulation calculations were carried out for defects and faults such as jamming,and the variation pattern of the operating torque of the disconnector switch in different states was analyzed and summarized,laying the foundation for subsequent research on diagnostic methods.Simulation experiments were carried out on the isolating switch experimental platform and the operating torque was measured using a stepper motor-based operating torque measurement device,which was compared with the simulation data to verify the accuracy of the simulation results.(3)From the operating torque data,seven characteristic quantities were extracted to characterize the mechanical state of the disconnector and used as data samples,and a sample library of the mechanical state of the disconnector was constructed using a combination of simulation data and actual data.The random forest algorithm is used to classify the mechanical state of the disconnector and to achieve fault diagnosis.To improve the diagnostic efficiency,the importance of the feature quantities was calculated,and the seven-dimensional features were downscaled into four-dimensional features according to the importance ranking,and the diagnostic model was optimized,which greatly improved the efficiency of the diagnostic method.The accuracy of the final diagnostic model for the measured samples was 90%,and it was able to effectively diagnose the disconnecting switch fault types.