| Autor: |
Niya Chen, Rongrong Yu, Yao Chen |
| Předmět: |
|
| Zdroj: |
International Workshop on Large-Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Plants; 2015, p639-643, 5p |
| Abstrakt: |
Rapidly developing wind energy leads to growing interest in turbine condition monitoring which can detect early fault and reduce the maintenance cost. In this paper, several approaches for wind turbine condition monitoring using only SCADA (Supervisory Control and Data Acquisition) data are applied and analyzed. The methods are: Gaussian Processes (GP), Artificial Neural Network (ANN), and Nonlinear State Estimation Technique (NSET). In each method, the model is trained to reflect the operational function of wind turbine when it is normal. The residual of test set can be calculated by estimated output and actually measured data, and this residual can be used for wind turbine condition determination - larger residual means turbine further away from normal status. Real-world dataset from a commercial wind farm in China is used to train and validate our models. The results show that all the methods can reflect wind turbine condition, but with different detection accuracy and false alarm times. Detailed comparison on GP, ANN and NSET are made on experimental results of 20 wind turbines. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
