Font Size: a A A

Study On Partial Discharge Feature Extraction And Pattern Recognition Of High Voltage Switchgear

Posted on:2021-05-23Degree:MasterType:Thesis
Country:ChinaCandidate:M C ZhangFull Text:PDF
GTID:2392330605959290Subject:Engineering
Abstract/Summary:PDF Full Text Request
High-voltage switchgear is a very important electrical equipment in power system.Its safety and stability determine the reliability of power supply system.However,insulation defects will inevitably occur during the manufacture and operation of switchgear.The insulation damage of switchgear is largely caused by partial discharge.Therefore,the detection of partial discharge is of great significance for insulation fault diagnosis of switchgear.Partial discharge is not only an important inducing factor for insulation failure of high voltage switchgear,but also a manifestation of latent insulation failure.Therefore,the research of partial discharge signal detection,feature extraction and pattern recognition has been paid much attention.Based on the partial discharge test and data collection,this paper studies the technology of partial discharge feature extraction and pattern recognition for switchgear.The main work is as follows:Firstly,the principle of Pulse Current Method and air gap discharge is elaborated.The air gap discharge is simulated by SIMULINK,and the effects of discharge characteristics and physical parameters on discharge characteristics are studied.Thus,the process of partial discharge is deeply understood.Three types of common insulation defects in switchgear are designed and fabricated.The Pulse Current Method is used to detect partial discharge,and the time-domain waveform data are collected.Based on Hilbert-Huang transform and energy entropy theory,the discharge characteristic parameters are extracted as input vectors of pattern recognition network.Discharge types are identified by Extreme Learning Machine(ELM),Probabilistic Neural Network Model(PNN)and Support Vector Machine(SVM).Optimal algorithms are used to optimize PNN and SVM networks.The recognition accuracy before and after optimization is compared,and the feasibility of feature extraction and pattern recognition scheme studied in this paper is verified.Based on the research on partial discharge of high-voltage switchgear,a partial discharge simulation platform based on PRPD spectrum was researched and designed,and the platform was debugged with simulation parameters under laboratory conditions.Spectrum analysis and other functions.
Keywords/Search Tags:Partial Discharge, Hilbert-Huang Transform, Pattern Recognition, PRPD Spectrum
PDF Full Text Request
Related items