Proposal and evaluation of a technique based on ultrasound for composite insulator monitoring

Summary

In the operation of a high voltage electrical system unexpected interruptions can cause many damages to electricity companies and consumers. Among various causes of interruptions, this paper highlights those related with failures in polymeric insulators in transmission lines. In fact, composite insulators have emerged as the solution for several problems regarding outdoor insulation. However, as any relatively new technology, polymeric insulator also brought with them some new challenges. The most intriguing among these challenges are, probably, monitoring and diagnosis of those insulators in the field. In this work, a technique for polymeric insulator monitoring is proposed. Furthermore, an analysis of sensitivity of the parameters related with the technique is performed. The technique is based on the acoustic interpretation of the ultrasonic noise coming from the electrical discharges close to the insulators. For this purpose, tests were carried out in a high voltage laboratory, employing several samples of 230 kV polymeric insulators, extracted from transmission lines, with different pollution and age conditions. The Spectral Subband Centroid Energy Vectors algorithm was used to extract attributes and process the ultrasonic noise signals. This algorithm splits the frequency spectrum into a number of overlapping subbands, locates the centroids of each subband and calculates the energy in the proximity of each centroid. The output of the algorithm is an energy vector with size equal to the number of subbands chosen. Aiming to support the decision-making, an artificial neural network was employed. Thus, the artificial neural network made possible to identify and distinguish accurately the energy vectors from cleaned insulators, polluted insulators and damaged insulators with success rates above 80%. Based in these results, and aiming to define which parameters and topologies would give the best results, an analysis of sensitivity was executed. The conclusion gives technical basis for implementation of the method with a larger amount of samples insulators, making it possible to statistically evaluate its applicability in the field.

Additional informations

Authors

MOULAI, Da Costa, COLE, ZHU, PAN, ZHAN, LEHOUIDJ

Keywords

Artificial neural network, Polymeric insulators, Spectral sub-band centroid, Sensitivity analysis, Ultrasonic noise