Summary
Since the insulation defects in electric equipment usually lead to partial discharge (PD), PD detection can be used to evaluate the insulation condition. PD pattern recognition is of considerable practical interest to power system utilities. In this paper, several kinds of artificial oil-paper defect models are made. PD tests on them under AC voltage are performed, and a comprehensive classifier is proposed to recognize different defects. The phase resolved partial discharge (PRPD) patterns are recorded. Several statistical operators representing PD patterns are extracted and used for PD recognition features. A comprehensive classifier using three kinds of recognition methods including minimum distance, neural network and support vector machine are adopted. The recognition results are indicated as defect type and confidence level, showing good classification accuracy.
Additional informations
| Publication type | ISH Collection |
|---|---|
| Reference | ISH2015_536 |
| Publication year | 2015 |
| Publisher | ISH |
| File size | 399 KB |
| Price for non member | Free |
| Price for member | Free |
Authors
Izawa Yasuji, Koo Ja-Yoon, Tamakoshi, Nishijima Kiyoto, Porkar Babak, Sack Martin, Tsutsumi Ryota, Huang Bo, Tran Ngoc Thach
