径向基函数在通榆河输水分析中的应用  

Application of Radial Basis Function in the Analysis of Water Transportation in Yancheng Section of Tongyu River

作　　者：潘道宏[1] 谭成宇 PAN Dao-hong;TAN Cheng-yu(Yancheng Branch of Jiangsu Province Hydrology and Water Resources lnvestigation Bureau,Yancheng 224015,Jiangsu,China;Dongtai City Water Conservancy Construction Co.Ltd.,Dongtai 224200,Jiangsu,China)

机构地区：[1]江苏省水文水资源勘测局盐城分局,江苏盐城224015 [2]东台市水利建设总公司,江苏东台224220

出　　处：《水利科技与经济》2020年第9期45-51,共7页Water Conservancy Science and Technology and Economy

摘　　要：通榆河是江苏省清水通道,是盐城市主要进水河道。为研究通榆河输水规律,为水源管理和防汛调度提供技术支撑,在介绍径向基函数网络基本原理的基础上,以通榆河盐城段为例,采用神经网络模拟河道水面曲线,对河道糙率数据进行分类,进而演算河道流量并与马斯京根法对比分析。结果表明,神经网络水面线拟合最大误差3 cm,河道糙率分类结果基本与实际相符,流量演算精度达到或优于马斯京根法。该方法为神经网络在水文分析中应用提供了新例证,所用代码可简易移植于防汛自动报汛系统。Tongyu river is a water-dilution channel in Jiangsu province,and it is the main water inflow channel in Yancheng city.In order to study the water delivery law of Tongyu river and provide technical support for water source management and flood control operation,based on the introduction of the basic principle of radial basis function network,taking Yancheng section of Tongyu river as an example,this paper uses neural network to simulate the river surface curve,classifies the river roughness data,and then calculates the river flow and compares it with masjingen method.The results show that the maximum fitting error of water surface profile of neural network is 3cm,the classification results of river roughness are basically in line with the actual situation,and the accuracy of flow calculation is up to or better than that of masjingen method.The method in this paper provides a new example for the application of neural network in hydrological analysis.The code used in this paper can be easily transplanted into the flood control automatic flood reporting system.

关 键 词：神经网络 水面线模拟 河道糙率分类 流量演算 通榆河 

分 类 号：TV672[水利工程—水利水电工程]