风速预测可以有效提升风电场输入电网的能量。采用小波包分析方法对原始风速数据作分解,用Shannon熵选取最佳小波树,对不同的分解部分用神经网络作预测,将各部分预测结果叠加得到风速预测值。实例表明本方法的有效性。
Wind speed prediction can effectively enhance the power of wind energy input. Original wind speed data was decomposed into different frequency by means of wavelet packet analysis. Shannon entropy was used to select the best wavelet tree, and different decomposed parts were predicted by BPNN (back-propagation neural network ), whose prediction results were added together as the final wind prediction data. The examples verify the effectiveness of the proposed method.