| Autor: |
Takamitsu Araki, Ryosaku Ikeda, Doan, Van Q., Ishizaki, Noriko N., Hiroyuki Kusaka |
| Předmět: |
|
| Zdroj: |
International Workshop on Large-Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Plants; 2018, p1-5, 5p |
| Abstrakt: |
To further spread the wind generation, an accurate wind forecast is an essential process. To improve the accuracy of a short term wind forecasts given by a numerical weather prediction (NWP) model, nonlinear regression methods using observed wind speed data and the NWP outputs have been applied. In this study, we introduce an additive model based on B-spline basis expansion into the NWP wind forecast correction, which consists of nonlinear and flexible coordinate regression functions of input variables.We also estimate parameters of the functions using the maximum penalized likelihood method to obtain smooth and flexible regression functions. We compared the proposed method with two nonlinear regression methods that have been widely used, an artificial neural network (ANN) and analog ensemble (AnEn) method through the correction of the wind forecasts of the Weather Research and Forecasting (WRF) model. As the results, our method corrected the WRF wind speed forecasts more accurately than the ANN and AnEn method and revealed the relationships between the WRF wind speed forecasts and the observed wind speed. The relationship was not able to be obtained by the ANN and AnEn method. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
