Inspection in overhead insulators through the analysis of the irradiated RF spectrum

Abstract Failures in high voltage insulators can cause significant injuries for the consumers, since these piece of equipment play a fundamental role for the transmission lines operation. Insulator inspection and diagnosis techniques are widely researched, varying mainly regarding to the data collection method, as example: ultrasound, thermal imaging, leakage current, corona, electric field and RF signals. In this article, the RF signals radiated by the insulators were used to classify them into three levels of pollution: “light”, “medium” and “heavy”. For this, we used the standard IEC 60815-1 as reference, which takes into account the ESDD and NSDD indices. The proposed classification method was based on the wavelet transform, for signal processing, and artificial neural network for the classification. The success rate was approximately 95% in the 1.2–1.4 GHz frequency range.