Study On Characteristics Of Electomagnetic Spectrum During The Discharge Of Contaminated Insulator

Pollution flashover is an frequent and pernicious accident which threatens thereliability of power supply and results in interrupt power supply and productiondamages. Consequently, investigation of effective approach to online montitoringinsulator contamination state is of great significance.This paper attempts to study from the ponit of electromagnetic field, analyze thespectrum’s characteristics of the different contaminatied insulator discharge’selectromagnetic radiation, and eventually determine the contamination status.First of all, a model for effectively analyzing how contaminated insulator dischargesand radiates electromagnetic wave. In the model, the spectrum’s of electromagneticradiation of insulator discharge is obtained via simulations based on FDTD software andcomputation according to fast Fourier transform. Through analyzing E, H spectrums ofdifferent level contaminated insulators, it could be found that they have similar shape,and the amplitude size is proportional to the contamination’s degree, which can initiallydetermine the contamination status of the insulator. By observing the analog signal ofthe insulator and waveforms and spectrum of the measured discharge current signal, itcould be found that the real discharge current signal could be regarded as superpositionof multiple analog signals, and it proves the analog signal is reasonable. Simulation ofdischarge radiation electromagnetic radiation by different waveforms (amplitude of thesame) of the analog current signal, in consequence, its shape is different, but theamplitude proportional to contamination degree.Then, considering the fact that the UHF method can effectively shield the majorityof environment electromagnetic interference and get effective electromagnetic data, thestudy measures partial discharge electromagnetic radiation of the contaminated insulatorby applying the UHF method. In addition, the isolation transformer, no PD transformerand blank contrast item are used to eliminate interference. According to the comparisonof spectrum, average and effective values of its magnitudes between different tests,spectrums are different from each other, the number of spikes of spectrums isproportional to contamination degree, so do the mean and effective values of spectrum.Via statistical analysis and classifying the different contamination degreeelectromagnetic radiation spectrum regarding frequency band and amplitude, it could befound thar in the same band and in the some range of amplitude the the number of data is proportional to contamination degree, which prove that discharge intensity isproportional to contamination degree. Investigating on voltage influence on thedischarge, spectrum of voltage in10kV and20kV is very different, the number of spikes,mean and rms of spectrum in voltage of20kV is much larger.Further to analyze the electromagnetic radiation spectrum’s data of contaminationinsulators and extract characteristic from the discharge electromagnetic radiationspectrum. In this paper, count the number of spectrum’s data by sub-band and sub-rangeof amplitude, and establish the three-dimensional map of the sub-band, sub-range ofamplitude and the amount of data, as the fingerprint features of discharge spectrum. Atthe same time, by using the fractal theory and Katz algorithm calculate the fractaldimension of discharge spectrum, as its another characteristic quantity, and the fractaldimension curve changes with different voltage and contamination. Establishingfingerprint database, will provide data for insulator contamination degree’s judgmentand recognition.