One day ahead wind speed forecasting using wavelets

Electricity markets are evolving throughout the world, accommodating the penetration of alternative energy sources, in particular wind. As the wind energy penetration is increasing, there is a need for accurate and efficient wind forecasting technique. If the wind speed can be reliably forecasted up to several hours, the generating schedule can efficiently accommodate wind power generation, leading to higher profits for wind plant developers and lower integration costs for utilities. This paper proposes a new application of wavelets in the field of wind speed forecasting. A new wavelet coefficient predictor technique is developed for forecasting wind speed up to 24 hours. This technique is successfully applied on the wind speed data obtained from Colorado public utility sites. The results obtained are compared with the common benchmark persistence method.