基于EEMD-LSTM的小水电站发电量预测  

Power generation prediction of small hydropower station based on EEMD-LSTM

作　　者：杨锋 何青松 詹毅 何庭辉 冯磊华[2] YANG Feng;HE Qingsong;ZHAN Yi;FENG Leihua(HNAC Technology Co.,Ltd.,Changsha 410006,China;School of Energy and Power Engineering,Changsha University of Science and Technology,Changsha 410114,China;Hunan huarun electric power repair Co.,Ltd.,Chenzhou 423000,China)

机构地区：[1]华自科技股份有限公司,湖南长沙410006 [2]长沙理工大学能源与动力工程学院,湖南长沙410114 [3]湖南华润电力检修有限公司,湖南郴州423000

出　　处：《能源工程》2024年第5期23-30,共8页Energy Engineering

摘　　要：准确而有效的发电量预测对于智能电网的经济运行意义重大。提出一种结合集合经验模态分解方法构建长短期记忆网络预测小水电站发电量的模型,利用集合经验模态分解对水电站历史发电量时序进行特征分解,得到包含不同特征的发电量分量。将分解的历史发电量时序与相关影响因素(河道径流量、导叶开度、月份)作为模型的输入,并通过训练调整模型中各参数,然后叠加预测结果得到最终的预测发电量。经过实验对比分析后发现,结合集合经验模态分解方法构建的长短期记忆网络预测模型较传统长短期记忆网络预测模型有更高的精度,并验证了分解-预测-重构的方法在水电站发电量预测中的可行性,同时对新能源大规模并网和电网整体经济运行提供科学依据和决策指导。Accurate and effective power generation prediction is of great significance to the economic operation of smart grids.In this paper,a model that combines the ensemble empirical mode decomposition method to construct a long short-term memory network was proposed to predict the power generation of small hydropower stations.The ensemble empirical mode decomposition is used to perform feature decomposition on the historical power generation time series of the hydropower station to obtain power generation components containing different characteristics.The decomposed historical power generation time series and related influencing factors(river runoff,guide vane opening,month)are used as the input of the model,and each parameter in the model is adjusted through training,and then the prediction results are superimposed to obtain the final predicted power generation.After experimental comparative analysis,it was found that the long-short-term memory network prediction model constructed by combining the ensemble empirical mode decomposition method has higher accuracy than the traditional long-short-term memory network prediction model,and verified that the decomposition-prediction-reconstruction method can generate electricity in hydropower stations.It also provides scientific basis and decision-making guidance for large-scale grid connection of new energy and overall economic operation of the power grid.

关 键 词：集合经验模态分解 长短期记忆神经网络 小水电 发电量预测 

分 类 号：TM612[电气工程—电力系统及自动化]