Short-Term Wind Power Forecast Based on CNN&LSTM-GRU Model Integrated with CEEMD-SE Algorithm
杨国华 1祁鑫 2贾睿 1刘一峰 2蒙飞 2马鑫 1邢潇文1
扫码查看
点击上方二维码区域，可以放大扫码查看
作者信息
1. 宁夏大学电子与电气工程学院,宁夏银川 750021
2. 国网宁夏电力有限公司调度控制中心,宁夏银川 750001
折叠
摘要
为进一步提升短期风电功率的预测精度,提出了一种基于互补集合经验模态分解-样本熵(complementary ensemble empirical mode decomposition-sample entropy,CEEMD-SE)的卷积神经网络(convolutional neural network,CNN)和长短期记忆-门控循环单元(long short term memory-gated recurrent unit,LSTM-GRU)的短期风电功率预测模型.首先,利用互补集合经验模态分解将原始风电功率序列分解为若干本征模态函数(intrinsic mode function,IMF)分量和一个残差(residual,RES)分量,利用样本熵算法将相近的分量进行重构;其次,搭建卷积神经网络和长短期记忆网络的并行网络结构,提取数据的局部特征和时序特征,并将特征融合后输入门控循环单元网络中进行学习预测;最后,通过算例进行验证,结果表明采用该模型后预测精度得到了有效提升,其均方根误差降低了 15.06％、平均绝对误差降低了 15.22％、决定系数提高了 1.91％.
Abstract
In order to further improve the accuracy of short-term wind power forecast,a CNN & LSTM-GRU based short-term wind power prediction model using CEEMD-SE algorithm is proposed.First,the original wind power output series are decomposed into several intrinsic mode function components and one residual component by complementary set empirical mode decomposition,and those components of similar mode are reconstructed by sample entropy algorithm.Next,the parallel network structure of convolutional neural network and long short term memory network is set up,and the local and temporal features of the data are extracted.And then the features are fused and input into the gated cyclic unit network for learning and prediction.Finally,the feasibility of the model is verified through case studies.The results show that the forecast accuracy has been improved effectively.The root mean square error and average absolute error,of the proposed model are reduced by 15.06％and 15.22％respectively,while coefficient of determination is up by 1.91％.
关键词
短期风电功率预测/互补集合经验模态分解/样本熵/长短期记忆网络/门控循环单元
Key words
short-term wind power forecasting/complementary ensemble empirical mode decomposition/sample entropy/long short term memory network/gated recurrent unite
维普
万方数据