Signal Detection and Pattern Recognition of Switchgear Partial Discharge
Partial Discharge(PD)is caused by the deterioration of insulation materials,whose detection and recognition are of great significance in preventing insulation breakdown and catastrophic failure.In order to solve these problems,this paper presents a detection system of partial discharge of switchgear based on the method of transient earth voltage(TEV).Partial discharge test of switchgear is carried out by four partial discharge models,and corresponding partial discharge data which will be uploaded to the computer are obtained.Due to the electromagnetic interference generated by the electromagnetic wave and hardware in the surrounding space,the reasonable de-noising of the partial discharge signal can greatly improve the signal-to-noise ratio,and it will be more conducive to signal feature extraction and pattern recognition.Wavelet de-noising based on the experience usually takes a fixed decomposition of the number of layers,but in this paper,a Mattat algorithm combined with the optimal decomposition level adaptive algorithm for noise signal separation and reconstruction is proposed,the result shows that the algorithm can be very well to filter out noise and retain the original discharge signal greatly.The characteristic parameters of the PD signal that have been de-noised are extracted,and BP neural network is used to identify the type of partial discharge of switchgear.The accuracy of different training errors is slightly different.When equals 0.001,the recognition accuracy of discharge pattern reached 98.5%.
Yun ZOU Zhen-sheng WU Jun-feng GUI Xiao-nan JIANG
School of Electrical Engineering,Beijing Jiaotong University,100044,Beijing,China
国际会议
上海
英文
301-307
2017-12-30（万方平台首次上网日期，不代表论文的发表时间）