Abstract
Aimed at parameter optimization of fault classifier, the principle of kernel parameter optimization based on data distribution was discussed. Then a simplified algorithm was proposed in order to achieve parameter optimization of fault classifier for turbo-generator sets. Testing results show that the classification capability of the fault classifier can be improved by using the simplified algorithm with high efficiency. But the adaptability of the simplified algorithm needs to be improved.
| Original language | English |
|---|---|
| Pages (from-to) | 11-14 |
| Number of pages | 4 |
| Journal | Zhendong Ceshi Yu Zhenduan/Journal of Vibration, Measurement and Diagnosis |
| Volume | 26 |
| Issue number | 1 |
| State | Published - Mar 2006 |
Keywords
- Data distribution
- Fault classifier
- Kernel functions
- Parameter optimization
- Steam turbo-generator sets
- Support vector machine