闫孟婷,黄炜斌,张天遥,马光文,赵丽伟

• 1:四川大学水利水电学院

摘要(Abstract)：

【目的】梯级水电站间存在的水力联系导致下游电站的运营模式往往受制于上游电站，以往的梯级水电站优化调度中通常选择忽略流量滞时或认为其是常数，少有考虑流量滞时与上游水电站出库流量及区间降雨等因素的动态关系。为提高下游电站预判流量精确度，将神经网络应用到梯级电站间流量动态滞时研究中。【方法】首先采用联合互信息理论选取下游电站入库流量的主要影响因素作为模型输入因子，其次根据卷积神经网络和长短时记忆神经网络的互补特性，建立上游出库流量与下游入库流量的JMI-CNN-LSTM深度学习网络模型，最后结合实际算例，将所建立模型的拟合结果与随机森林回归模型、固定滞时模型进行对比。【结果】结果显示：本文所建立的模型较相同条件下其他方法各类误差均存在不同程度的减少，其中MAE至少减少了14.6%。【结论】结果表明：相较其他方法，JMI-CNN-LSTM耦合模型预测精度更佳，能够更准确的体现梯级电站间流量滞时的动态关系。

关键词(KeyWords)： 梯级水电站;动态滞时;联合互信息;卷积神经网络;长短期记忆网络

Abstract:

Keywords:

基金项目(Foundation): 国家重点研发计划(2016YFC0402208);国家重点研发计划(2018YFB0905204)

作者(Author): 闫孟婷,黄炜斌,张天遥,马光文,赵丽伟

DOI: 10.13928/j.cnki.wrahe.2023.03.014

参考文献(References)：

• [1] 肖贵友，陈仕军，李基栋，等.考虑水流滞时影响的梯级水电站群短期优化调度研究[J].水力发电，2018,44(4):37-41.XIAO Guiyou,CHEN Shijun,LI Jidong,et al.Short-term optimal operation of cascade hydropower stations considering the delay of flow propagation[J].Water Power,2018,44(4):37-41.
• [2] 张培，纪昌明，张验科.考虑滞时的水库群短期优化调度风险分析模型[J].人民长江，2017,48(11):107-111.ZHANG Pei,JI Changming,ZHANG Yanke.Risk analysis model of short-term optimal operation for reservoir group considering flow time-lag[J].Yangtze River,2017,48(11):107-111.
• [3] 李景刚，黄诗峰，任亚鹏，等.平移搜索最大相关系数法在渠道水流滞时分析中的运用[J].水电能源科学，2020,38(3):87-89.LI Jinggang,HUANG Shifeng,REN Yaping,et al.Application of translational search for maximum correlation coefficient method in time delay analysis of channel water flow[J].Water Resources and Power,2020,38(3):87-89.
• [4] 刘欢，廖胜利，刘战伟，等.耦合水流滞时的梯级库群中期调峰调度方法[J].水力发电学报，2022,41(4):37-46.LIU Huan,LIAO Shengli,LIU Zhanwei,et al.Medium-term peak-shaving scheduling method of cascade reservoirs coupled with flow time delay[J].Journal of Hydroelectric Engineering,2022,41(4):37-46.
• [5] 李晓晨.马斯京根法的文献综述[J].科技风，2020(7):187.LI Xiaocheng.Literature review of the muskingum method[J].Vent Technologique,2020(7):187.
• [6] BOZORG-HADDAD O,ABDI-DEHKORDI M,HAMEDI,et al.Generalized storage equations for flood routing with nonlinear muskingum models[J].Water Resources Management,2019,33(8):2677-2691.
• [7] REICHSTEIN M,CAMPS-VALLS G,STEVENS,B,et al.Deep learning and process understanding for data-driven earth system science[J].Nature,2019,566(7743):195-204.
• [8] 纪昌明，张培，吴月秋，等.基于空间映射原理的水库群短期优化调度模型[J].水力发电学报，2017,36(5):58-67.JI Changming,ZHANG Pei,WU Yueqiu,et al.Short-term optimal operation models based on space mapping principle[J].Journal of Hydroelectric Engineering,2017,36(5):58-67.
• [9] 魏勤，陈仕军，谭政宇，等.基于SA-LSTM的水电站站间流量动态滞时关系[J].水电能源科学，2021,39(6):16-19.WEI Qin,CHEN Shijun,TAN Zhengyu,et al.Dynamic flow lag time relationship between hydropower stations based on SA-LSTM[J].Water Resources and Power,2021,39(6):16-19.
• [10] 纪昌明，俞洪杰，阎晓冉，等.考虑后效性影响的梯级水库短期优化调度耦合模型研究[J].水利学报，2018,49(11):1346-1356.JI Changming,YU Hongjie,YAN Xiaoran,et al.Study on the coupling model of cascade reservoirs′ short-term optimal operation considering the influence of aftereffect[J].Shuili Xuebao,2018,49(11):1346-1356.
• [11] SHARMA A,LUK K C,CORDERY I,et al.Seasonal to interannual rainfall probabilistic forecasts for improved water supply management:Part 2-Predictor identification of quarterly rainfall using ocean-atmosphere information[J].Journal of Hydrology,2000,239(1):240-248.
• [12] 赵铜铁钢，杨大文.神经网络径流预报模型中基于互信息的预报因子选择方法[J].水力发电学报，2011,30(1):24-30.ZHAO Tongtiegang,YANG Dawen.Mutual information-based input variable selection method for runoff-forecasting neural network model[J].Journal of Hydroelectric Engineering,2011,30(1):24-30.
• [13] 陈璐，叶磊，卢韦伟，等.基于Copula熵的神经网络径流预报模型预报因子选择[J].水力发电学报，2014,33(6):25-29.CHEN Lu,YE Lei,LU Weiwei,et al.Determination of input variabes for artificial neural networks for flood forecasting using Copula entropy method[J].Journal of Hydroelectric Engineering,2014,33(6):25-29.
• [14] 纪昌明，俞洪杰，阎晓冉，等.基于联合互信息的水文预报因子集选取研究[J].水力发电学报，2017,36(8):12-21.JI Changming,YU Hongjie,YAN Xiaoran,et al.Selecting hydrological forecast factor sets based on joint mutual Information[J].Journal of Hydroelectric Engineering,2017,36(8):12-21.
• [15] HOCHREITER S,SCHMIDHUBER J J N C.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
• [16] 贺邓超，张宏军，郝文宁，等.基于Parzen窗条件互信息计算的特征选择方法[J].计算机应用研究，2015,32(5):1387-1389.HE Dengchao,ZHANG Hongjun,HAO Wenning,et al.Feature selection based on conditional mutual information computation with Parzen window[J].Application Research of Computers,2015,32(5):1387-1389.
• [17] 胡义明，陈腾，罗序义，等.基于机器学习模型的淮河流域中长期径流预报研究[J].地学前缘，2022,29(3):284-291.HU Yiming,CHEN Teng,LUO Xuyi,et al.Medium to long term runoff forecast for the Huai River Basin based on machine Learning algorithm[J].Earth Science Frontiers,2022,29(3):284-291.
• [18] 姬战生，章国稳，黄薇.基于在线SVM的平原河网河道水位预报方法[J].安徽农业科学，2021,49(14):191-195.JI Zhansheng,ZHANG Guowen,HUANG Wei.Water level forecast method of plain river network based on online SVM[J],Anhui Agric Sci,2021,49(14):191-195.
• [19] 李伶杰，王银堂，胡庆芳，等.基于随机森林与支持向量机的水库长期径流预报[J].水利水运工程学报，2020(4):33-40.LI Lingjie,WANG Yintang,HU Qingfang,Long-term inflow forcast of reservoir based on Random Forest and support vector machine[J].Hydro-Science and Engineering,2020(4):33-40.
• [20] JIUXIANG G,ZHENHUA W,JASON K,et al.Recent advances in convolutional neural networks[J].Pattern Recognition,2018,77:354-377.
• [21] ASIFULLAH K,ANABIA S,UMME Z,et al.A survey of the recent architectures of deep convolutional neural networks[J].Artificial Intelligence Review,2020,53(8):5455-5516.
• [22] 杨丽，吴雨茜，王俊丽，等.循环神经网络研究综述[J].计算机应用，2018,38(S2):1-6.YANG Li,WU Yuxi,WANG Junli,et al.Research of recurrent neural network[J].Journal of Computer Application,2018,38(S2):1-6.
• [23] ALEX K,ILYA S,GEOFFREY E H.ImageNet classification with deep convolutional neural networks[J].Communications of the ACM,2017,60(6):84-90.