Self-learning digital filter for the field location of partial discharge in cables

The resolution and robustness of a cable PD locating instrument, based on time-domain reflectrometry (TDR), are critically affected by noise present in a field environment. The permanent background noise encountered during field tests has been shown to be mainly due to amplitude modulation (AM) radio signals. This paper describes a self-learning digital filter aimed at mitigating this noise. The filter, in conjunction with a signal averaging routine, has been successfully used to reduce noise and enhance the PD signal obtained from a cable in a field environment. Examples are provided to demonstrate the ability of the filter to significantly improve the signal-to-noise ratio (SNR) and effectively extract PD signals from field recordings