Research on the prediction method of porcelain insulator pollution degree based on electric field monitoring

Abstract

The online monitoring of insulator surface pollution states is crucial for predicting/preventing pollution flashover accidents. This paper conducts artificial pollution tests on porcelain insulators at five pollution levels to obtain the electric field signal during the flashover process. Subsequently, the time-frequency performance of the electric field signal waveform at various discharge stages is discussed in detail. Finally, a method for predicting pollution degrees by combining back propagation (BP) neural networks and non-contact electric field signal monitoring is proposed and validated. Research results indicate that during the pollution flashover process of the porcelain insulator, the peak value of the electric field signal shows an upward trend in sections. As the equivalent salt deposit density (ESDD) increases, the curves of electric field values rise. The development of dry-band arcing clearly distorts the electric field signal waveform at its crest and trough parts, increasing the total harmonic distortion (THD) rate. The THD can reach 100–130 % when closing to flashover, with the third harmonic acting as the main component. By taking kmax (the rising rate of the E-field signal amplitude), g (the ratio of Emax to Erms), K3 (the third harmonic coefficient), and THD as inputs to the BP neural network model, an ESDD prediction with a relative error lower than 10.63 % is realized. This research can offer novel technical support for predicting pollution degrees.