人民长江

• 论文 • 上一篇    下一篇

基于偏相关法的SWMM模型参数全局敏感性分析

段明印,李传奇,熊剑智   

  • 出版日期:2017-09-18 发布日期:2017-09-18

Global sensitivity analysis for SWMM model parameters using partial correlation method

DUAN Mingyin,LI Chuanqi,XIONG Jianzhi   

  • Online:2017-09-18 Published:2017-09-18

摘要: 基于拉丁超立方采样和偏相关分析方法,对暴雨管理模型(SWMM)中的水文水力参数进行了全局敏感性分析。以某小区为例,构建SWMM模型进行径流模拟,研究参数变化对模型输出结果的影响。结果表明:在设计降雨情景下,对峰值流量最敏感的参数是不透水区曼宁系数和子汇水区宽度比例因子;对峰值时间敏感的参数是不透水区曼宁系数、子汇水区宽度比例因子和子汇水区坡度比例因子,其他参数敏感性较小;对总产流最敏感的参数是子汇水区面积比例因子。通过分析,量化评估了各个参数的敏感性,可为精确构建模型提供指导。

关键词: SWMM模型, 偏相关法, 全局敏感性, 降雨径流

Abstract: Based on the Latin hypercube sampling method and partial correlation analysis method, we analyzed the global sensitivity of input parameters of storm water management (SWMM) model. Taking a residential quarter as an example, the SWMM was established to simulate the runoff and the influence of input parameters variation on runoff was studied. The results show that the most sensitive parameter to peak flood flow is the ratio of Manning coefficient at impervious area and the sub-catchment width; the greatest impact parameters on flood peak time are Manning coefficient at impervious area, sub-catchment width ratio and sub-catchment gradient ratio; the most sensitive parameter to the total runoff is the sub-catchment area ratio.

Key words: SWMM model, partial correlation, global