Partial discharge classification in GIS using the narrow-band UHF method

This paper discusses the use of the narrow-band ultra high frequency (UHF) partial discharge (PD) measuring technique for the recognition and classification of PD in gas insulated substations (GIS). The principle of the identification of PD signals from external disturbances or UHF signals caused by switching operations is discussed. Additionally, the interpretation of the measured signals with the use of orthogonal functions, fuzzy logic and neural networks is reported. A genetic algorithm is presented which can be used to determine the optimum topology of the neural network used.