Bias-correction method for wind-speed forecasting

Abstract

The improvement in wind-speed-forecast accuracy by modifying the bias between simulations and observations can significantly improve the wind-power efficiency of wind farms, increase grid stability, relieve grid pressure, and reduce economic costs. Gansu (China) is rich in wind-energy resources and has a large number of wind farms, but there are also difficulties in the simulation of the wind speed. While the correlation coefficient between the simulated wind speed and that observed is high, the bias is large, making it necessary for a fast cyclic-bias correction method based on historical data. This new bias-correction method uses three parameters (average, variance trend) to correct the simulated wind speed. The results show that the bias is reduced to 1-2 m/s from 2-3 m/s, which meets the forecasting demands of wind-farm operators. The percentage wind-speed biases of the different wind farms are also reduced by 17 % to 23 %.