Assessing the effects of power quality on partial discharge behaviour through machine learning

Catterson, V.M. and Rudd, S.E. and McArthur, S.D.J. and Bahadoorsingh, S. and Rowland, S.M. (2010) Assessing the effects of power quality on partial discharge behaviour through machine learning. In: 7th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies (CM 2010 and MFPT 2010), 2010-06-22 - 2010-06-24.

[thumbnail of strathprints026479.pdf]
Preview
Text. Filename: strathprints026479.pdf
Preprint

Download (1MB)| Preview

Abstract

Partial discharge (PD) is commonly used as an indicator of insulation health in high voltage equipment, but research has indicated that power quality, particularly harmonics, can strongly influence the discharge behaviour and the corresponding pattern observed. Unacknowledged variation in harmonics of the excitation voltage waveform can influence the insulation's degradation, leading to possible misinterpretation of diagnostic data and erroneous estimates of the insulation's ageing state, thus resulting in inappropriate asset management decisions. This paper reports on a suite of classifiers for identifying pertinent harmonic attributes from PD data, and presents results of techniques for improving their accuracy. Aspects of PD field monitoring are used to design a practical system for on-line monitoring of voltage harmonics. This system yields a report on the harmonics experienced during the monitoring period.