Wind speed forecasting based on variable weight combination forecasting model of neural network and grey model

Accurate forecast of wind speed can reduce or avoid the adverse effect of wind farms to power system effectively, which improves the competitiveness of wind farms in the power market. Although sets of models, including artificial neural network(ANN) and grey model(GM) are applied to wind speed forecasting, the forecasting accuracy is poorly satisfied. Here we propose a new variable weight combined (VWC) model which is based on the ANN and GM. The forecasting accuracy of VWC is higher than either of the two single models. We also found that the accuracy of VWC is higher than the unchanged weight combination(UWC) model. Our data show a new method for wind speed forecasting and the reduction of auxiliary service costs of wind farms.