Reduction of continuous and pulse-shape noise by digital signal processing for reliable PD detection in HV substations

This paper presents newly developed noise reduction methods for partial discharge detection and measurements of electric power apparatus using digital signal processing techniques. We have examined noise diminution by two algorithms in both laboratory and on-site environments based on measurements of PD signals in SF₆ gas and real noise as well as interference in practical HV and EHV substations. As a result, by the two techniques utilizing digital signal processing, we are able to reduce both sinusoidal continuous and pulse-shape noise signals considerably. The techniques are promising for future application in actual substations