泵站水电转换系数的BP神经网络模型  

BP neural network model of hydroelectricity conversion coefficient of pumping station

作　　者：周璐 张健 孟凡继 黄国情[1] 金秋[1] 赵广举 ZHOU Lu;ZHANG Jian;MENG Fanji;HUANG Guoqing;JIN Qiu;ZHAO Guangju(Rural Water Management Department,Nanjing Hydraulic Research Institute,Nanjing 210029,China;Rural Water Conservancy Technology Development Center,Department of Water Resources of Jiangsu Province,Nanjing 210029,China;Jiangsu Province Water Engineering Sci-tech Consulting Co.,Ltd.,Nanjing 210029,China)

机构地区：[1]南京水利科学研究院农村水利研究所,江苏南京210029 [2]江苏省水利厅农村水利科技发展中心,江苏南京210029 [3]江苏省水利工程科技咨询股份有限公司,江苏南京210029

出　　处：《水利经济》2024年第6期71-75,共5页Journal of Economics of Water Resources

基　　金：南京水利科学研究院中央级公益性科研院所人才项目(Rc923003)。

摘　　要：为实现区域性水电转换系数的智能预测和动态修正,以江苏省连云港市为研究区,采用贝叶斯正则化算法进行含噪数据分析,构建了大中型灌区典型提水泵站水电转换系数的BP神经网络模型,并通过泵站实测数据对模型进行验证。结果表明,连云港市泵站水电转换系数区域特性影响因素主要包括配套功率、流量、转速和效率,模型拟合优度为0.961,且泵站预测值与实测值误差均在允许范围内。由该模型及56个建模数据得到连云港市泵站水电转换系数分布区间为[11.03,69.30];通过参数优化所建立的BP神经网络模型能够实现区域性泵站水电转换系数的智能预测和动态修正,为多元水泵承包主体实行用电定额节水管控提供了新思路。In order to realize the intelligent prediction and dynamic correction of regional hydroelectricity conversion coefficients,the Lianyungang City in Jiangsu Province is taken as study area,and the Bayesian regularization algorithm was used for noisy data analysis.A BP(back propagation)neural network model was established for the conversion coefficient of typical pumping stations in large and medium-sized irrigation areas.The model was validated through measured data from the pumping stations.The results show that the factors affecting the regional characteristics of the hydroelectricity conversion coefficient of the second level pumping station in Lianyungang City,Jiangsu Province mainly include supporting power,flow rate,speed,and efficiency.The goodness of fit of the model is 0.961,and the error between the predicted and measured values of the pumping station is within the allowable range.The distribution range of hydroelectricity conversion coefficient in Lianyungang City,Jiangsu Province is[11.03,69.30].By optimizing parameters and establishing a BP neural network model,intelligent dynamic correction of regional hydroelectricity conversion coefficient can be achieved.This research also provides a new idea for multi-water pump contractors to implement water-saving quota management.

关 键 词：灌区 提水泵站 水电转换系数 相关性分析 BP神经网络模型 

分 类 号：S275[农业科学—农业水土工程]