Partial discharge location based on time difference of energy accumulation curve of multiple signals

Abstract

Acquired partial discharge (PD) signals are always affected by various noises in the process of propagation and detection. These errors have great impact on the accuracy of location of the PD source. In this study, a new PD location algorithm based on the time difference on an energy accumulation curve of multiple signals is proposed to solve this problem. First, errors from different propagation paths of signals are eliminated by correlation shift. Then, averaged signal is used to calculate accumulated energy curves. By searching the maximum correlation of those curves, the time differences of signals are calculated to locate the PD sources. This method has been used in a transformer model in the laboratory and it is proved to be a more effective way to locate the PD source in the transformer model compared to the single signal method. Based on this algorithm, the errors of time difference are in 0.1 ns range and errors of the PD source positions are in centimetre range.