[关键词]
[摘要]
为进一步提高风电功率预测计算效率及准确性,建立基于熵关联数据挖掘的MPSO-Elman风电功率预测模型:在分析信息熵与互信息的熵相关系数(ECC)后,对各个历史日数据样本和待测时段参考样本间的复杂非线性映射关系进行量化评估,经过高关联度样本筛选,Elman模型隐含层结构优化以及权值初值选取改进,最后采用改进粒子群算法(MPSO)对网络参数进一步优化,并以某风电场实测数据为依据进行实例分析。结果表明,该模型使得功率预测准确度达到91.24%,预测效果要优于RBF-BP模型,证明了该模型的先进性与有效性。
[Key word]
[Abstract]
In order to improve the prediction accuracy and the computational efficiency further,this paper establishes the MPSO-Elman wind power prediction model based on entropy association data mining.It adopts an index of entropy correlation coefficients(ECC) based on information entropy and mutual information to quantitatively evaluate the complex non-linear relationship between historical data samples and the data to be measured.After intimate-samples selection,Elman model′s hidden layer structure improvement and network weights choice,the network parameters are further optimized with MPSO.By analyzing the data of a wind farm in Jiangsu,the presented method has the power prediction accuracy of 91.24%,more accurate than the RBF-BP model,indicating its advancements and effectiveness.
[中图分类号]
TM614;TP18
[基金项目]