An improved artificial bee colony-BP neural network algorithm in the short-term wind speed prediction

The accuracy about short-term wind speed prediction is helpful to the economy and safety in the wind power grid. In view that the nonlinear and complexity characteristics related to the wind speed change, a new approach for short-term wind speed forecasting is put forward. The proposed method belongs to the back propagation (BP) neural network based on improved artificial bee colony algorithm (ABC-BP). Some factors such as historical moment of wind speed, current moment of air pressure and temperature are all the model input. The weight and threshold values of BP neural network are optimized by using the improved artificial bee colony algorithm. Furthermore, the optimized BP neural network is used to predict future moment of wind speed. Results of analyzing calculation example show that the algorithm has the characteristics of high precision, fast convergence rate compared with traditional BP neural network and genetic BP neural network.