The use of PSA for fault detection and characterization in electrical apparatus

The monitoring of the actual condition of high voltage apparatus has become more and more important in the last years. One well established tool to characterize the actual condition of electric equipment is the measurement and evaluation of partial discharge data. Immense effort has been put into sophisticated statistic software-tools, to extract meaningful analyses out of data sets, without taking care of relevant correlations between consecutive discharge pulses. In contrast to these classical methods of partial discharge analysis the application of the Pulse Sequence Analysis allows a far better insight into the local defects. The detailed analysis of sequences of discharges in a voltage- and a current-transformer shows that the sequence of the partial discharge signals may change with time, because either different defects are active at different measuring times or a local defect may change with time as a consequence of the discharge activity. Hence for the evaluation of the state of degradation or the classification of the type of defect the analysis of short `homogeneous´ sequences or sequence correlated data is much more meaningful than just the evaluation of a set of independently accumulated discharge data. This is demonstrated by the evaluation of measurements performed on different commercial hv-apparatus