基于CEEMD-CNN-BiGRU-RF模型的短期风电功率预测  

Short-term wind power prediction based on CEEMD-CNN-BiGRU-RF model

作　　者：曾亮 狄飞超 兰欣 王珊珊 Zeng Liang;Di Feichao;Lan Xin;Wang Shanshan(School of Electrical and Electronic Engineering,Hubei University of Technology,Wuhan 430068,China;Hubei Key Laboratory of Solar Energy Efficient Utilization and Energy Storage Operation Control,Hubei University of Technology,Wuhan 430068,China)

机构地区：[1]湖北工业大学电气与电子工程学院,湖北武汉430068 [2]湖北工业大学,太阳能高效利用及储能运行控制湖北省重点实验室,湖北武汉430068

出　　处：《可再生能源》2022年第2期190-195,共6页Renewable Energy Resources

基　　金：国家自然科学基金项目(51977061;61903129);湖北省重点研发计划项目(2020BAB114)。

摘　　要：风电功率的准确预测对电网的安全运行和经济调度起着重要作用,为进一步提高风电功率的预测精度,文章提出了一种基于CEEMD-CNN-BiGRU-RF模型的短期风电功率预测模型。首先,利用完全集成经验模态分解(CEEMD)对风电功率时间序列进行模态分解;其次,对分解的各个风电功率时间序列利用卷积神经网络(CNN)进行特征提取;再次,建立双向门控循环单元(Bi GRU)模型对各个风电功率时间序列进行预测,叠加各个分量的预测值;最后,对误差进行进一步分析与预测,利用随机森林(RF)进行误差修正,得到最终的风电功率预测值。实验仿真表明,该模型的预测效果明显优于传统模型,模型的平均绝对百分比误差(MAPE)仅为2.09%。Accurate prediction of wind power plays an important role in the safe operation and economic dispatch of power grids.In order to further improve the prediction accuracy of wind power,this paper proposes a short-term wind power prediction model based on CEEMD-CNNBiGRU-RF model.Firstly,complete ensemble empirical mode decomposition(CEEMD)is used to decompose the wind power time series.Secondly,the decomposed wind power time series are extracted by convolutional neural networks(CNN)for features.Thirdly,establishing a bidirectional gated recurrent unit(BiGRU)model to predict each wind power time series and superimpose the predicted values of each component.Finally,to further analyze and predict the error,use the Random Forest(RF)to perform error correction to obtain the final wind power forecast value.Experimental simulation shows that the prediction effect of the model is significantly better than the traditional model,and the mean absolute percentage error(MAPE)of the model is reduced to 2.09%.

关 键 词：风电功率预测 完全集成经验模态分解 卷积神经网络 双向门控循环单元 随机森林 

分 类 号：TK81[动力工程及工程热物理—流体机械及工程]