Design of Automatic Monitor for Cable Partial Discharge Anomaly Based on Generative Adversarial Network
During the operation of power equipment,cable insulation damage,internal defects,external environment and other factors are easy to cause local abnormal discharge of cable,and then cause power accidents and power failure.In order to timely and effectively mo-nitor local discharge,automatic monitoring of local discharge in the generated network environment is completed.According to the inter-nal structure of the cable,the local discharge reasons of the cable are analyzed,and the missing data are inserted to obtain the complete cable operation data.The random matrix is established to obtain the probability density function of the cable operation data,the feature vector is extracted,and the feature index matrix is constructed to implement the singular value decomposition of the feature vector,iden-tify the local discharge state of the cable,and realize the automatic monitoring of the cable local discharge anomaly.The experimental results show that the waveform vibration amplitude of the proposed method is small and the waveform is complete when extracting the pulse of cable partial discharge signal.The positioning of partial discharge position of cable is consistent with the actual position The relative error of partial discharge position of cable is less than 1%.
国家科技期刊平台
NSTL
万方数据
