Abstract
This paper explores the possibility in using radar to automatically classify wind turbine faults. As a first step, a number of experiments were conducted in an anechoic chamber with a small wind turbine were different faults were artificially induced. Two basic clustering methods were used. One was based on using different statistical parameters of the corresponding time-domain signatures. The other used Principal Components Analysis (PCA) on the corresponding frequency-domain signatures. Subsequently, a K-NN algorithm was used as the classifier to investigate whether or not automatic classification is fundamentally possible and to provide an initial comparison between the two clustering methods which rely on different signal domains. The proof of concept results presented in the paper indicate that this may indeed be plausible, to encourage further development of this idea.
Original language | English |
---|---|
Article number | 7131011 |
Pages (from-to) | 286-291 |
Number of pages | 6 |
Journal | IEEE National Radar Conference - Proceedings |
Volume | 2015-June |
Issue number | June |
DOIs | |
Publication status | Published - 22 Jun 2015 |
Event | 2015 IEEE International Radar Conference, RadarCon 2015 - Arlington, United States Duration: 10 May 2015 → 15 May 2015 |
Keywords
- Doppler radar
- radar target classification
- structural health monitoring
- wind turbines