Summary
Frequency response analysis is a sensitive technique for detecting damages in power transformers. The standardization of the measurement procedure has been already established, but the interpretation of the results is still far from reaching a widely-accepted and definitive methodology. Frequency response analysis is normally performed by comparing only the magnitude of frequency response data and neglecting the phase information. The approach proposed in this paper is based on the time-frequency transformation of the frequency response measurements so that they can be analyzed in a more descriptive domain for detecting faults. The time-frequency transformation of a frequency response measurement uses magnitude and phase information to generate a single two-dimensional map where the data is explained in time and frequency domains simultaneously. The Complex Morlet wavelet is used as a transformation function and four indicators using the transformation coefficients are additionally computed to compare the measurements and establish the condition of the power transformer. The results show that the application of time-frequency transformation to the frequency response allows an improved detection of failures and emphasizes the time-frequency ranges where differences appear. Time-frequency representation of the frequency response measurements displays more details than the typical magnitude plots so it is a promising tool for the development of a pattern recognition application to improve the classification of faults and the automation of the FRA interpretation process.
Additional informations
| Publication type | ISH Collection |
|---|---|
| Reference | ISH2017_567 |
| Publication year | |
| Publisher | ISH |
| File size | 4 MB |
| Pages number | 6 |
| Price for non member | Free |
| Price for member | Free |
Authors
YAO
