Open AccessRF‐based location of partial discharge sources using received signal features
Author(s) -
Iorkyase Ephraim T.,
Tachtatzis Christos,
Glover Ian A.,
Atkinson Robert C.
Publication year - 2019
Publication title -
high voltage
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.732
H-Index - 20
ISSN - 2397-7264
DOI - 10.1049/hve.2018.5027
Subject(s) - radio frequency , partial discharge , signal (programming language) , feature selection , feature (linguistics) , computer science , dimensionality reduction , curse of dimensionality , radio signal , frequency band , radio spectrum , radio propagation , pattern recognition (psychology) , electronic engineering , artificial intelligence , acoustics , physics , telecommunications , electrical engineering , engineering , voltage , antenna (radio) , linguistics , philosophy , programming language
Partial discharges (PDs) are symptomatic of some localised defects in the insulation system of electrical equipment. PD activity emits electrical pulses in the form of radio frequency (RF) signals which can be captured using appropriate sensors. The analysis of the measured RF signals facilitates localisation of PD. This study investigates the plausibility of using purely RF received signal features of PD pulses to locate PD at low cost. A localisation approach based on the analysis of these features has been developed, with the assumption that PDs generate unique RF spatial patterns due to the complexities and nonlinearities of RF propagation. In this approach, two distinct frequency bands which hold different PD information are exploited. PD location features are extracted from the main PD signal and the two sub‐band signals. Correlation‐based feature selection (CFS) is employed for feature selection and dimensionality reduction. Experimental results show that PD location can be inferred from the features of the PD pulses. The application of CFS to PD data reduces the memory/computational demand and improves localisation accuracy.