Short-term wind power forecasting using a double-stage hierarchical hybrid GA-ANN approach

Abstract

Power generation from wind generators is always associated with some intermittency due to wind speed and other weather parameters variation, and accurate short-term forecasts are essential for their efficient and effective operation. This can well support transmission and distribution system operators and schedulers to enhance the power network control and management in the smart grid context. This paper presents a double stage hierarchical genetic algorithm trained artificial neural network (double-stage hybrid GA-ANN) for short-term wind power forecast of a microgrid wind farm in Beijing, China. The approach has two hierarchical stages. The first GA-ANN stage employs numerical weather prediction (NWP) meteorological parameters to forecast wind speed at the wind farm exact site and turbine hub height. The second stage models the actual wind speed and power relationships. Then, the predicted next day's wind speed by the first stage is applied to the second stage to forecast next day's wind power. The presented approach has achieved considerable prediction accuracy improvements. The prediction performance of the proposed approach was also compared with another double-stage Back Propagation (BP) trained ANN prediction model and showed a better accuracy improvement.