不同汇水面积下城市雨洪模型SWMM参数敏感性分析

Urban Flooding Model SWMM Parameter Sensitivity Analysis under Different Catchment Areas

参数敏感性分析是确定模型主要参数的重要步骤。以广州市2个不同汇水面积的区域为研究对象,基于蒙特卡罗的Sobol指数法分析不同汇水面积下模型不同结果(总径流量、流量峰值、峰值时间)的参数敏感性,探究不同汇水面积下影响总径流量、流量峰值、峰现时间的敏感参数。结果表明,不同汇水面积下,影响模型不同输出结果的主要敏感参数有所区别:①影响模型总径流量输出的主要敏感参数是汇水区基本特征参数,但是汇水面积不同,需加以考虑的敏感参数略有不同,较大汇水面积的研究区域需加以考虑渗透参数,较小汇水面积的研究区域则需加以考虑曼宁粗糙度参数;②影响模型流量峰值和峰值时间输出的主要敏感参数是汇水区基本特征参数和曼宁粗糙度参数,但是汇水面积不同,首要考虑的敏感参数略有不同,较大汇水面积的研究区域首要考虑汇水区基本特征参数,较小汇水面积的研究区域则首要考虑曼宁粗糙度参数。该研究可为将来在城市同类区域建模的主要参数选取提供参考,从而提高模拟精度和建模效率。

Parameter sensitivity analysis is an important step in determining the main parameters of a model.Taking two regions with different catchment areas in Guangzhou as the research objects,the Monte Carlo Sobol index method was used to analyze the parameter sensitivity of different results(total runoff,peak flow,peak time)with different catchment areas.The results show that with different catchment areas,the main sensitive parameters affecting different output results of the model are different.①The main sensitive parameters affecting the total runoff output of the model are the basic characteristic parameters of the catchment area.But when the catchment areas are different,the sensitive parameters to be considered are slightly different.The research area with a larger catchment area needs to consider infiltration parameter,and the research area with a smaller catchment area needs to consider Manning roughness parameters.②The main sensitive parameters affecting peak flow and peak time output are the characteristic parameters of the catchment area and the Manning roughness coefficient.But when the catchment areas are different,the sensitive parameters that are first considered are slightly different.The research area with a larger catchment area mainly considers the characteristic parameters of the catchment area,and the research area with a smaller catchment area mainly considers Manning roughness parameters.This study can provide reference significance for the main parameters determining in modeling similar regions in cities in the future and improve simulation accuracy and modeling efficiency.