基于经验模式分解和神经网络的短期风速组合预测  

Short-term wind speed forecasting based on EMD and ANN

作　　者：王韶[1] 杨江平[1] 李逢兵[1] 刘庭磊 

机构地区：[1]输配电装备及系统安全与新技术国家重点实验室(重庆大学电气工程学院),重庆400044 [2]重庆南岸供电局,重庆400060

出　　处：《电力系统保护与控制》2012年第10期6-11,18,共7页Power System Protection and Control

基　　金：国家'111'计划项目(B08036);输配电装备及系统安全与新技术国家重点实验室自主研究项目(2007DA10512709212)

摘　　要：风速时间序列具有很强的间歇性和随机性,属于非平稳时间序列。为提高预测精度,提出了经验模式分解法(EMD)和神经网络相结合的短期风速组合预测模型。该方法运用EMD将风速序列分解为一系列不同频率的相对平稳的分量,减少了不同特征信息之间的干扰;根据各个分量的变化规律,选择合适的神经网络模型来分别预测,对高频分量采用神经网络组合预测模型,低频分量采用合适的预测模型直接进行预测;将各分量预测值叠加得到最终预测值。算例结果表明,所提方法与单一的径向基神经网络模型(RBF)和支持向量机模型(SVM)相比,预测精度得到了大幅度的提高。Wind speed time series, due to its strong intermittency and randomness, belongs to non-stationary time series.In order to improve the forecasting accuracy, a new forecasting model based on Empirical Mode Decomposition （EMD） and ANN is presented.By means of the EMD technique, the original wind speed series are decomposed into a series of components of different frequencies with relatively stationary variation. Thus the interferences among the characteristic information embedded in the wind speed can be weakened. According to the change regulation of each intrinsic mode function, the appropriate ANN model is chosen to forecast each intrinsic mode function. For high frequency components, we can use combination of ANN model, while use one appropriate model for low frequency components, and then add up each forecasting result to get the final forecasting value. The simulation results indicate that the accuracy of the forecasting model discussed in the paper is higher than that of RBF model and SVM model.

关 键 词：短期预测 经验模式分解 径向基神经网络 支持向量机 广义回归神经网络 组合预测 

分 类 号：TM614[电气工程—电力系统及自动化]