Application of Huber-Similarity Measure on PD Detection

Abstract

Extraction of Partial Discharges (PD) is a key step in diagnosis and evaluation of the power system equipment condition. In field testing, besides high frequency noise and disturbance, Power Frequency (P.F.) harmonics also couple with PD measurement sensors. In order to deal with both types of noises and disturbances, in this paper, a new PD signal extraction algorithm is presented, which is based on a combination of Huber Function, Discrete Cosine Transform (DCT), l₁ and l₂ norms. This new method, which is introduced as Huber Similarity Measure for Partial Discharge (HSMPD), was evaluated through experimental laboratory constructed PD models. Results show this proposed algorithm successfully extracted PD signals in the presence of baseline and high frequency noises and disturbances. HSMPD can be employed as a backbone in intelligent diagnosis systems for improving the accuracy of PD condition monitoring equipment.