Abstract
It is difficult to measure the wind speed accurately in short term. This reveals challenges for wind turbine control, especially for maximum power point tracking with adaptive control strategies. In this paper, a genetic algorithm based support vector machine model is adopted to estimate the wind speed, using physically measurable signals, such as the electrical power, pitch angle, and rotor speed, while the desired rotor speed can be obtained accordingly. Further, by combining the radial basis function neural networks with adaptive algorithms, a novel virtual parameter based neuroadaptive controller is developed to accommodate the system uncertain and external disturbances. The effectiveness and performances of the proposed method are validated and demonstrated with FAST (Fatigue, Aerodynamics, Structures, and Turbulence) and Simulink.
| Original language | English |
|---|---|
| Pages (from-to) | 7754-7764 |
| Number of pages | 11 |
| Journal | IEEE Transactions on Industrial Electronics |
| Volume | 63 |
| Issue number | 12 |
| Early online date | 19 Jul 2016 |
| DOIs | |
| Publication status | Published - Dec 2016 |
| Externally published | Yes |
Funding
This work was supported in part by the Talent Foundation of Beijing Jiaotong University under Grant W15RC00110, China Postdoctoral Science Foundation under Grant 2016M590040, the Major State Basic Research Program of China under Grant 2012CB215202, the National Natural Science Foundation of China under Grant 51207007, and the Fundamental Research Funds for the Central Universities under Grant 2015JBM003.
Keywords
- FAST (Fatigue Aerodynamics Structures and Turbulence)
- neuroadaptive control
- support vector machine (SVM)
- virtual parameter
- wind turbine (WT)