Modern power spectral density estimation of discharge pulses of polluted insulators

Although the insulation level of polluted insulators and the number of ultra-violet discharge pulses have close relationship, it is difficult to determine the insulation level of polluted insulators only by the number of discharge pulses. This paper illustrates the method of time-series analysis to study power spectral density (PSD) of ultra-violet discharge pulses from one string of 110 kV polluted insulators, by applying Burg algorithm to establish the autoregressive model based on field data. Results indicate that different polluted level has its own corresponding curve, so it is possible to classify the insulation level of insulators through PSD method.