BP神经网络在密云水库入库水量预测中的应用  

Application of BP neural network in water inflow prediction of Miyun Reservoir

作　　者：李婵娟 焦有权[2] 温江丽 荀志成 李光涛 吴迪 LI Chanjuan;JIAO Youquan;WEN Jiangli;XUN Zhicheng;LI Guangtao;WU Di(Beijing Miyun Reservoir Management Office,Beijing 101512,China;Beijing Vocational College of Agriculture,Beijing 102442,China)

机构地区：[1]北京市密云水库管理处,北京101512 [2]北京农业职业学院,处北京102442

出　　处：《北京水务》2022年第4期14-20,共7页Beijing Water

摘　　要：水库入库水量预测是水库调度和水资源优化配置的重要依据。密云水库是华北地区最大的水库,是确保首都供水安全和社会经济发展的重要水源地。反向传播BP神经网络由于具有高维性、自适应、自组织及自学习等优点,被广泛应用于水文预报中。基于BP神经网络模型对密云水库入库水量进行模拟预测,将入库流量分为汛期(6—9月)和非汛期,非汛期又分为1—5月和10—12月,不同时段分别建立BP神经网络模型,共建立3种BP神经网络模型,利用1960—2004年共45年系列样本数据分别训练3个BP网络模型,用于预报入库水量,并利用2005—2021年共17 a资料检验样本,分别代入3个BP网络模型,对预报结果进行验证,最后得到模型预报及检验结果。结果表明,利用BP神经网络进行汛期和非汛期入库水量预报,误差小,精度高,应用BP神经网络对2022和2023年入库水量进行预报,可为水资源合理优化配置提供重要依据。Reservoir inflow prediction is an important basis for reservoir scheduling and optimal allocation of water resources. Miyun Reservoir is the largest reservoir in North China and an important water source to ensure water supply security and socio-economic development in the capital. BP neural network is widely used in hydrological forecasting due to its advantages of high dimensionality, self-adaptation,self-organization and self-learning. Based on the BP neural network model, this paper simulated and predicted the inflow of Miyun Reservoir, and divided the inflow into the flood season(June-September) and the non-flood season(January-May and October-December). The BP neural network models were established in different seasons respectively and a total of three BP neural network models were established.Then, three BP network models were trained to predict the inflow of water using the 45 a series of sample data from 1960 to 2004, and the 17 a data test samples from 2005 to 2021 were substituted into three BP network models to verify the prediction results, Finally, the model prediction and test results were obtained. The results show that the BP neural network is used to predict the inflow in flood and non-flood seasons with small error and high precision. The BP neural network is used to predict the inflow of 2022 and 2023, which provides an important basis for the reasonable and optimal allocation of water resources.

关 键 词：密云水库 BP神经网络模型 入库水量 降雨量 蒸发量 

分 类 号：TV213.4[水利工程—水文学及水资源]