基于数据特征提取和SSA-BiLSTM的短期风电功率预测  

Short-term wind power prediction based on data feature extraction and SSA-BiLSTM

作　　者：文博 陈芳芳[1] 王华玉 WEN Bo;CHEN Fangfang;WANG Huayu(School of Electrical and Information Engineering,Yunnan Minzu University,Kunming 650031,China;School of Law,Yunnan Minzu University,Kunming 650031,China)

机构地区：[1]云南民族大学电气信息工程学院,云南昆明650031 [2]云南民族大学法学院,云南昆明650031

出　　处：《应用科技》2023年第4期71-78,共8页Applied Science and Technology

摘　　要：为提高风电功率预测的准确性,提出了一种基于数据特征提取和麻雀算法优化双向长短期记忆网络(sparrow search algorithm optimised bi-directional long and short-term memory network,SSA-BiLSTM)短期风电功率预测模型。首先根据皮尔逊相关系数(Pearson correlation coefficient,PCC)分析风电数据中各影响因素与风电功率之间的相关性,根据计算结果将功率无关的因素去除。然后,采用自适应噪声完全集成经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)将原始风电功率序列进行分解,得到一系列子序列分量。再将所有子序列输入麻雀算法(sparrow search algorithm,SSA)优化的双向长短期记忆(bi-directional long short-term memory,BiLSTM)模型中进行预测,根据所得预测值对风速序列进行修正。将修正所得的风速序列与风电功率序列作为输入,送入SSA-BiLSTM模型中进行预测。最后,由实验结果分析并对比得出,该模型具有更好的风电功率预测精度。In order to improve the accuracy of wind power prediction,we propose a short-term wind power prediction model based on data feature extraction and sparrow search algorithm optimised bi-directional long and short-term memory network(SSA-BiLSTM).Firstly,the correlation between each influencing factor in wind power data and wind power is analyzed according to Pearson correlation coefficient(PCC),and the power-independent factors are removed according to the calculation results.Then,the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)is used to decompose the original wind power series,obtaining a series of sub-series components.Then,all subsequences are input into the bi-directional long short-term memory(BiLSTM)model optimized by the sparrow search algorithm(SSA)for prediction,and the wind speed series are corrected according to the obtained predicted values.The corrected wind speed series and wind power series are used as inputs,which are fed into the SSA-BiLSTM model for prediction.Finally,the experimental results are analyzed and compared,drawing a conclusion that the model has better wind power prediction accuracy.

关 键 词：双向长短期记忆模型 麻雀优化算法 皮尔逊相关系数 风速修正 短期风电功率预测 数据特征提取 

分 类 号：TM614[电气工程—电力系统及自动化]