Effective Wind Power Density Prediction Based on Neural Networks

As a green renewable resource, wind energy has been emphasized to solve the problem of world energy crisis and environmental pollution. Prediction of the effective wind power density is a significant work for wind power evaluation. In this paper, the effective wind power density values are regarded as a time series, furthermore, Back Propagation Neural Networks (BPNN) and Generalized Regression Neural Networks (GRNN) are employed to forecast the effective wind power density in the future. Both the two models predict the effective wind power density only based on historical data and statistical calculation, combining the long term prediction and the short term prediction. Forecasting and calculation results show that the neural networks have strong learning ability that can well capture the random changes of time series researched. Compared the forecasting results of the two neural networks mentioned above, the Back Propagation Neural Networks performs better than the Generalized Regression Neural Networks for effective wind power density prediction in Hexi Corridor.