数学模型在漫湾水电站中长期预报中的应用  

Application of mathematical model in medium-long term forecast for Manwan Hydropower Station

作　　者：郑孝庆[1] 金鑫[1] 

机构地区：[1]桂林理工大学,广西桂林541006

出　　处：《广西水利水电》2014年第2期51-55,共5页Guangxi Water Resources & Hydropower Engineering

摘　　要：水文中长期预报作为防灾减灾以及进行水资源优化调度、水电站运行管理的重要依据,一直是水文工作中的重点和难点。为了探究不同数学模型在漫湾水库的适用性,选定人工神经网络、最近邻抽样回归、线性回归3个模型对漫湾水库进行月径流模拟,并将各个模型的模拟径流进行对比和精度评价。最终选定模拟精度较高的最近邻抽样回归方法建立漫湾水库月径流来水预报方案。As the important basis for disaster prevention and alleviation, water resources optimization and hydropow-er station operation, medium-long term forecast is the focus and difficulty of hydrological work. In order to study theapplicability of different mathematical models, artificial neural network, nearest neighbor bootstrapping regressionand linear regression were selected to simulate the monthly runoff of Manwan reservoir. The simulated runoff out ofthree models were compared, based on which the NNBR model with relatively higher simulation accuracy was adopt-ed to constitute the monthly runoff forecasting program of Manwan Reservoir.

关 键 词：径流中长期预报 线性回归 人工神经网络 最近邻抽样回归 漫湾水电站 

分 类 号：P338[天文地球—水文科学]