Research On Partial Discharge Detection And Pattern Recognition In High-voltage Switchgear Based On VMD-PSOChOA-ELM Algorithm

High voltage switchgear plays an important role in power supply and distribution network.As a power distribution equipment,the stable operation of high voltage switchgear is an important basis to ensure the reliability of power supply and distribution and power quality.However,the phenomenon of aging or deterioration of insulation will gradually occur in the operation and use of high voltage switchgear for a long time,resulting in partial discharge in the switchgear,which will cause problems in the power supply and distribution system and power equipment damage,serious and even lead to important safety accidents.Therefore,it is very necessary to study the running state of high voltage switchgear online monitoring,which can improve the reliability and safety of power supply and distribution system,and effectively avoid safety accidents caused by high voltage switchgear.In order to understand the electromagnetic wave signal propagation law and electromagnetic wave position distribution law in the switchgear cabinet more directly,this paper takes KYN28-12 type high voltage switchgear as the research reference object,based on the idea of finite difference Time domain(FDTD)on the switchgear for modeling,specific analysis of the electromagnetic wave signal transmission characteristics and partial discharge signal propagation process in the switchgear cabinet.The optimal installation position of UHF antenna sensor for detecting partial discharge of switchgear is determined.Based on the propagation law and characteristics of UHF signal in the switchgear cabinet,according to the performance requirements of UHF sensor,the basic principle of Archimedes spiral antenna is cited in this paper,and the parameters of UHF antenna sensor are calculated and set through modeling and simulation of HFSS software,and a UHF antenna sensor is designed and produced.The sensor will be used to detect partial discharge in the switchgear in subsequent tests in this paper.In this paper,three kinds of partial discharge defect test models common in high voltage switchgear are designed,and a test platform for partial discharge test of switchgear is built.The UHF sensor is used to effectively detect three kinds of partial discharge test models in switchgear,and then signals are collected and data are extracted.Finally,the discharge characteristics of three kinds of discharge signals are analyzed.Finally,an optimized pattern recognition algorithm is adopted in this paper to realize accurate identification of three discharge signals: surface discharge,air gap discharge and needle plate discharge.Firstly,with the local minimum envelope entropy as the objective function,particle swarm optimization algorithm(PSO)was used to optimize the two key parameters of the variational mode decomposition algorithm(VMD),namely the mode decomposition number and penalty factor.Then,the signal feature vector was constructed by calculating the IMF component permutation entropy.Then the extreme learning machine(ELM)algorithm is used to identify the three signal states.The innovation of this paper lies in that,aiming at the problem of poor recognition ability of the standard ELM model,an optimization algorithm combining particle swarm optimization and chimpanzee optimization(PSOCh OA)is proposed in this paper.The PSOCh OA algorithm proposed in this paper is used to optimize the weight threshold of the ELM model.The final results show that VMD-PSOCh OA-ELM algorithm can greatly improve the recognition rate,increase recognition rate to 98%.