基于偏秩相关-逐步回归法的SWMM模型全局敏感性分析
|
摘要
为了识别影响暴雨洪水管理模型(SWMM)的敏感性因素,给后续的模型率定和不确定性分析提供指导和参考,基于拉丁超立方抽样技术对各个参数进行抽样,分别以逐步回归法和偏秩相关法对SWMM模型进行全局敏感性分析,并以山东大学千佛山校区为例。结果表明:不透水区曼宁系数是对峰值时间和峰值流量最敏感的参数,并且随着不透水区曼宁系数的增大,峰值时间增大、峰值流量变小;汇水区面积比例因子是对总产流最敏感的参数,并且随着汇水区面积比例因子的增大,总产流量增大。逐步回归法和偏秩相关法都是线性分析方法,两种方法的对比使用,能够更加精确的识别SWMM模型的敏感性参数,对于精确建模和高效率定具有一定的借鉴意义。
In order to identify the sensitive factors that affect the rainstorm flood management model( SWMM) as well as to provide guidance and reference for the subsequent model rate determination and uncertainty analysis,this paper samples each parameter based on the Latin hypercube sampling technique. On the other hand,it also respectively analyzes the global sensitivity of SWMM model by using the stepwise regression method and partial rank correlation method as well as takes the Qianfoshan Campus of Shandong University as an example. The results show that: Firstly,the manning coefficient in impermeable region is the most sensitive parameter to peak time and peak flow. In addition,with the increase of Manning coefficient in impermeable region,the peak time will increase and the peak flow will decrease.Secondly,the ratio factor of catchment area is the most sensitive parameter to total production and flow. Therefore,with the increase of water catchment area ratio factor,the total production discharge will increase. The stepwise regression and partial rank correlation are the linear analysis methods. We also can identify the sensitive parameters of SWMM model more accurately through the comparison of the two methods,which is useful for the accurate modeling and high efficiency.
In order to identify the sensitive factors that affect the rainstorm flood management model( SWMM) as well as to provide guidance and reference for the subsequent model rate determination and uncertainty analysis,this paper samples each parameter based on the Latin hypercube sampling technique. On the other hand,it also respectively analyzes the global sensitivity of SWMM model by using the stepwise regression method and partial rank correlation method as well as takes the Qianfoshan Campus of Shandong University as an example. The results show that: Firstly,the manning coefficient in impermeable region is the most sensitive parameter to peak time and peak flow. In addition,with the increase of Manning coefficient in impermeable region,the peak time will increase and the peak flow will decrease.Secondly,the ratio factor of catchment area is the most sensitive parameter to total production and flow. Therefore,with the increase of water catchment area ratio factor,the total production discharge will increase. The stepwise regression and partial rank correlation are the linear analysis methods. We also can identify the sensitive parameters of SWMM model more accurately through the comparison of the two methods,which is useful for the accurate modeling and high efficiency.
引文
[1]张建云,王银堂,贺瑞敏,等.中国城市洪涝问题及成因分析[J].水科学进展,2016,27(4):485-491.
[2]黄爽.基于城镇水文模拟的雨洪系统规划对策研究[D].重庆:重庆大学,2017.
[3] HUBER W C,DICKINSON R E. Storm water management model user's manual, Version 4[M]. Georgia:Environmental Protection Agency,1992.
[4]蒋元勇,丰锴斌,刘学文,等.城市雨洪SWMM模型的敏感参数研究综述[J].生态科学,2015,34(2):194-200.
[5]吴积善,田连权,康志成,等.泥石流及其综合治理[M].北京:科学出版社,1993.
[6]宋明丹,冯浩,李正鹏,等.基于Morris和EFAST的CERESWheat模型敏感性分析[J].农业机械学报,2014,45(10):124-131,166.
[7]何亮,侯英雨,赵刚,等.基于全局敏感性分析和贝叶斯方法的WOFOST作物模型参数优化[J].农业工程学报,2016,32(2):169-179.
[8]高颖会,沙晓军,徐向阳,等.基于Morris的SWMM模型参数敏感性分析[J].水资源与水工程学报,2016,27(3):87-90.
[9]王汉明,李传奇,熊剑智,等.基于互信息的SWMM模型参数全局敏感性分析[J].人民黄河,2017,39(10):22-26,93.
[10]段明印,李传奇,肖学,等.基于偏相关法的暴雨管理模型参数敏感性分析[J].人民黄河,2018,40(3):11-15.
[11]张小丽,彭勇,徐炜,等.基于Sobol方法的新安江模型参数敏感性分析[J].南水北调与水利科技,2014,12(2):20-24,33.
[12]王浩昌,杜鹏飞,赵冬泉,等.城市降雨径流模型参数全局灵敏度分析[J].中国环境科学,2008,(8):725-729.
[13]曹渊,王文科,王铁良,等.~3H在地下水中迁移的灵敏度分析[J].核技术,2012,35(10):771-774.
[14]王念秦,韩波,庞琦,等.泥石流防治工程效果后评价初探[J].工程地质学报,2015,23(2):219-226.
[15]廖爽,叶枫,武英.基于MATLAB的逐步回归法计算新疆伊犁地区煤的发热量[J].煤质技术,2011,(4):1-5.
[16]施小清,吴吉春,姜蓓蕾,等.基于LHS方法的地下水流模型不确定性分析[J].水文地质工程地质,2009,36(2):1-6.
[17]伍国军,陈卫忠,谭贤君,等.基于拉丁超立方抽样的有限元可靠度程序开发及应用[J].岩土力学,2015,36(2):550-554.
[2]黄爽.基于城镇水文模拟的雨洪系统规划对策研究[D].重庆:重庆大学,2017.
[3] HUBER W C,DICKINSON R E. Storm water management model user's manual, Version 4[M]. Georgia:Environmental Protection Agency,1992.
[4]蒋元勇,丰锴斌,刘学文,等.城市雨洪SWMM模型的敏感参数研究综述[J].生态科学,2015,34(2):194-200.
[5]吴积善,田连权,康志成,等.泥石流及其综合治理[M].北京:科学出版社,1993.
[6]宋明丹,冯浩,李正鹏,等.基于Morris和EFAST的CERESWheat模型敏感性分析[J].农业机械学报,2014,45(10):124-131,166.
[7]何亮,侯英雨,赵刚,等.基于全局敏感性分析和贝叶斯方法的WOFOST作物模型参数优化[J].农业工程学报,2016,32(2):169-179.
[8]高颖会,沙晓军,徐向阳,等.基于Morris的SWMM模型参数敏感性分析[J].水资源与水工程学报,2016,27(3):87-90.
[9]王汉明,李传奇,熊剑智,等.基于互信息的SWMM模型参数全局敏感性分析[J].人民黄河,2017,39(10):22-26,93.
[10]段明印,李传奇,肖学,等.基于偏相关法的暴雨管理模型参数敏感性分析[J].人民黄河,2018,40(3):11-15.
[11]张小丽,彭勇,徐炜,等.基于Sobol方法的新安江模型参数敏感性分析[J].南水北调与水利科技,2014,12(2):20-24,33.
[12]王浩昌,杜鹏飞,赵冬泉,等.城市降雨径流模型参数全局灵敏度分析[J].中国环境科学,2008,(8):725-729.
[13]曹渊,王文科,王铁良,等.~3H在地下水中迁移的灵敏度分析[J].核技术,2012,35(10):771-774.
[14]王念秦,韩波,庞琦,等.泥石流防治工程效果后评价初探[J].工程地质学报,2015,23(2):219-226.
[15]廖爽,叶枫,武英.基于MATLAB的逐步回归法计算新疆伊犁地区煤的发热量[J].煤质技术,2011,(4):1-5.
[16]施小清,吴吉春,姜蓓蕾,等.基于LHS方法的地下水流模型不确定性分析[J].水文地质工程地质,2009,36(2):1-6.
[17]伍国军,陈卫忠,谭贤君,等.基于拉丁超立方抽样的有限元可靠度程序开发及应用[J].岩土力学,2015,36(2):550-554.