东篱科研大数据发现系统（DRDS）

位置：成果数据库 > 期刊 > 期刊详情页

两种不同类型的水文模型在贵州典型岩溶地区的应用

ISSN号：0469-5097
期刊名称：南京大学学报(自然科学版)
时间：0
页码：409-415
语言：中文
分类：P334[天文地球—水文科学;水利工程—水文学及水资源;天文地球—地球物理学]
作者机构：[1]南京大学地球科学与工程学院水科学系,南京210093
相关基金：国家自然科学基金（40725010,40671030）
相关项目：我国西南喀斯特山区雨洪资源有效利用与拦沙减淤研究

作者：王栋|尚晓三|

关键词：岩溶地区, 后寨河流域, 遗传算法, BP神经网络, 岩溶水文模型, karst area, Houzhaihe basin, genetic algorithm, BP neural network, karst hydrological model

中文摘要：

本文建立了两种不同类型的水文模型以研究岩溶地区特殊的产汇流过程．一种是将人工神经网络应用到岩溶地区，采用成因分析法和自相关分析法确定模型的输入变量，进而建立了Back—propagation（BP）神经网络岩溶水文模型．另一种是根据系统理论方法，建立起概念性岩溶水文模型，并采用遗传算法率定模型参数，进而建立了基于遗传算法率定参数的概念性岩溶水文模型．以贵州普定后寨河流为例，将后寨测站不同时段水文资料对这两种模型进行了检验，并分析比较了两种模型在岩溶地区的适用性．研究结果表明，这两种模型均能模拟及预报岩溶地区产汇流过程；按照相对误差RE、互相关系数R和确定性系数QR这3个指标，所建的BP神经网络岩溶水文模型优于基于遗传算法率定参数的概念性岩溶水文模型．

英文摘要：

In this paper, two different hydrological models are established to study special rainoff-runoff process in the karst area. For one thing, artificial neural network is applicated to simulate rainfall-runoff processes of the karst area. The karst hydrological model based on BP neural network is established to forecast runoff processes, whose input variables are determined based on genesis analysis methods and autocorrelation analysis. For another, acording to the system theory approach, the karst hydrological model is established, whose parameters are calibrated and determined by genetic algorithm. In order to analyze their application in karst area the two different models are tested by using hydrological data in two different period＇s of Houzhai station which locates in Houzhaihe catchment. The prediction shows that the two different models could simulate and forecast the special runoff yield and concentration process of the karst area. Acorrding to three criteria （relative error, correlation and correlation coefficient and qualified rate）, the proposed karst hydrological model based on BP neural network is better than the conceptual karst hydrological model whose parameters are optimized by genetic algorithm in the karst area.

同期刊论文项目