SWMM─RUNOFF參數優選模式之開發

Development of the SWMM─RUNOFF Parameters Optimization Model

江政育(Cheng-Yu Chiang)

指導教授 ： 林旭信

中原大學/工學院/土木工程研究所/碩士(2010年)

https://doi.org/10.6840/cycu201000980

摘要

本研究結合暴雨經理模式(SWMM)和遺傳演算法程式庫(GAlib)，開發可自動優選SWMM-RUNOFF參數之模式(SWMM-RUNOFF Parameters Optimization Model，簡稱SRPOM)，進行SWMM-RUNOFF之水文地文參數自動化校驗，減少人工方式憑經驗調整參數所需花費之時間，提升參數優選之效率，達到自動化優選SWMM-RUNOFF水文地文參數之目的。並以台北市第四排水分區中港抽水站系統作為本研究實際案例，選取2004至2006年間實際九場颱風事件，各別作為SRPOM模式之訓練、驗證及測試資料。

由其中七場颱風事件當成SRPOM模式之訓練資料，以推算案例區域之最佳參數(optimal_parameters)；以兩場驗證之颱風資料(桑美、寶發颱風)，比較最佳參數與預設參數(default_parameters)對於中港抽水站系統實際中港1人孔水位記錄之影響，經模擬結果顯示最佳參數設定下得到之人孔水位歷線較符合實際人孔水位歷線。最後測試之颱風資料(寶發颱風)，不考慮PCTERR(質量誤差百分比)之下，其目標函數收斂值相對於限制質量誤差百分比下之目標函數收斂值會來得較低；此外，進行SRPOM模式參數優選過程中，可依不同質量誤差百分比限制，設定不同GA operator，以提升目標函數之收斂效率。

關鍵字

自動化優選參數 ； SWMM ； GAlib ； 模式開發

並列摘要

This research integrates Storm Water Management Model(SWMM) and Genetic Algorithm Library(GAlib), into a model that optimizes SWMM-RUNOFF parameters automatically(SWMM-RUNOFF Parameters Optimization Model, named SRPOM), the SRPOM model calibrate the hydrological and geographical parameters of SWMM-RUNOFF automatically to reduce the cost of time in calibrating parameters, and it will promote the efficiency of calibrating parameters in order to achieve the hydrological and geographical parameters of SWMM-RUNOFF optimization automatically. The 4th drainage system, Zhung-Gung drainage system, in Taipei city is employed as a case study. Furthermore, we selected nine of the typhoon events during 2004 to 2006 which were considered to be the training, verification, and examination data for SRPOM model.

Seven typhoon events were considered to be the training data for SRPOM model in order to compute the optimal parameters; two typhoon events(typhoon Saomai, typhoon Bopha) were considered to be the verification data for the influence on the optimal parameters and the default parameters with the manhole water level of Zhung-Gung_1 where in the Zhung-Gung drainage system, the simulation indicated that manhole water level hydrograph in the optimal parameter sets were much more conformed to the actual manhole water level hydrograph. In the examination data(typhoon Bopha), if we were not considered the percentage error that objective function will got the lower convergence than restrained the percentage error; Furthermore, when we were in the optimizes parameters process for SRPOM model, depended on different percentage error, we could set up different GA operator to promote the convergence efficiency of objective function.

並列關鍵字

SWMM ； GAlib ； model integration ； automatically parameters optimization

參考文獻

17.邱煌升, 2008,“以OpenMI架構整合都市降雨逕流模式與雨水下水道模式,”中原大學土木工程學研究所碩士論文.

22.蔡文雄, 2009,“整合雨水下水道即時水理模擬與Web GIS展示技術於都市內水預警之應用─以台北市第四排水分區為例,”中原大學土木工程學研究所碩士論文.

2.Elliott, A.H. and Trowsdale, S.A., 2007, “A Review of Models for Low Impact Urban Stormwater Drainage,”Environmental Modelling & Software,Vol.22, pp.394-405.

3.Goldberg, D.E., 1997, “Genetic Algorithms in Search, Optimization, and Machine Learning,”Reading,Mass., Addison-Wesley.

4.Gironás J., Roesner L.A., Rossman L.A., and Davis J., 2010, “A New Applications Manual for the Storm Water Management Model (SWMM), ” Environmental Modelling & Software,Vol.25, pp.813–814.