Research On Fault Diagnosis Method For Insulators

Insulator flashover is one of the major factors which affecting the safe and stablerunning of the power system for a long time, but also the main manifestations ofinsulator’s fault. Leakage current of insulator surface is an important parameter whichreacting the flashover. We can predict the extent of the contamination and the insulatorflashover levels by extracting leakage current information which reflect the status ofrunning insulators. Currently, the detection method based on the insulator surfaceleakage current is still the basic idea to diagnose fault insulators. The main contents areas follows:Firstly, this paper summarizes the current research status of the insulator pollutionflashover mechanism and the the prevention measures of pollution flashover. Inaddition, it also discusses the formation process and the mechanism of insulatorpollution flashover furtherly, which pointed out that the cause of the pollution flashoveroccurrence is the insulator surface pollution and damp. According to the plate model,this paper deducts the critical conditions of DC and AC pollution flashover, and thenstudy the change of leakage current characteristics in each stage of pollution flashover.Secondly, this paper designs an artificial pollution test based on the quantitativebrush, and then establishs a test system to get the leakage current signals of pollutedinsulators. According to the purpose of experiment, this paper chooses therepresentative insulators and satisfying test equipment to set up the artificial pollutiontest platform, in which using the high performance digital oscilloscope to gather theleakage current signals.Thirdly, the leakage current simulation signal with white noise is obtained byMATLAB, compare and analysis the effect of de-noising based on four commonwavelet thresholding and adaptive wavelet thresholding de-noising methods. Theconclusion as follow: under conditions of severe contamination, the effect of adaptivewavelet thresholding is significantly better than the other four threshold de-noisingmethods; under the conditions of mild contamination, the effect of wavelet adaptivethreshold de-noising is not as well as the level of severe contamination.Fourthly, the leakage current signal is decomposed into a series of intrinsic modefunctions (IMF) by using the overall average empirical mode decomposition (EEMD)method. Sort out the ingredients of discharge pulse and low-frequency harmonics. Thenextract the leakage current of the discharge after de-nosing each intrinsic mode functions by Adaptive threshold de-noising method. Analysis the extent of leakagecurrent discharge development by discharge fingerprints. Characterizing the degree ofthe pulse discharge of insulator by using amplitude, phase, and the number ofdischarges generated by discharge fingerprint. Thus reflecting the contaminated degreeof insulator.