基于WPD-MAIWO-NN的短期风速多步预测方法

A Novel Multi-Step Prediction for Wind Speed Based on WPD-MAIWO-NN

提出了一种由小波包分解、多智能体入侵杂草算法和人工神经网络组成的混合预测方法,用于提高短期风速预测的准确性。利用小波包分解将风速时间序列分解成多个不同频率的子序列,然后利用多智能体入侵杂草算法优化后的神经网络对每个子序列进行预测,最后叠加各子序列的预测值,得出实际预测结果。以广东某风电场2014年1月的实测小时风速数据为例,使用提出的混合模型进行风速预测。仿真结果表明,与未经优化的神经网络相比,该文方法在进行风速直接多步预测时具有更好的整体误差指标。

This study proposes a hybrid forecasting approach that consists of the WPD （Wavelet Packet Decomposi- tion）, MAIWO （Multi Agent Invasive Weed Optimization） and ANN （Artificial Neural Network） for enhancing the accuracy of short-term wind speed forecasting. The WPD is employed to extract true information from a short-term wind speed series, and the ANN,which optimizes the parameters using a MAIWO algorithm,is used as the predictor to provide the final forecast. The proposed hybrid model is demonstrated to forecast a mean hour wind speed series obtained from a windmill farm located in Guangdong. The simulation results suggest that the developed forecasting method yields better predictions compared with those of other popular models, which indicates that the hybrid method exhibits stronger forecasting ability.