地表細胞自動機快速淹水－街道集流器－雨水下水道管網之整合模擬分析：臺北市強降雨即時淹水預報平台之研發

近年來發生極端降雨事件的頻率增加，短延時強降雨暴雨強度越來越大，嚴峻考驗排水設施的排水能力，使得都會區淹水災害事件頻傳。淹水模擬結果可提供防洪防災工作之參考，可有效降低災害帶來的生命財產損失。短延時強降雨事件造成排水系統無法負荷短時間內過量之逕流量，如道路進水口、連接管之排水能力，因此，要正確的進行都會區積淹水評估需考量道路進水口、連接管以及建築物之排水能力。但現今都市淹水模式因計算時間較長而導致其仍無法大量運用於即時積淹水預報，多為利用設計雨型以及雨量進行事前的淹水潛勢分析，缺乏時效性、難以做到即時預警應變之應用。本研究以臺大細胞自動機快速淹水模式（NTU-CAFIM）結合暴雨經理模式（SWMM）來建構預報平台，並考量道路進水口、連接管與建築物的排水能力來進行都會區快速淹水模擬，有利即時淹水評估以及大量淹水模擬。針對臺北市進行積淹水模擬評估，在歷史積淹水事件其偵測率約有60 %以上，而利用降雨預報之即時淹水評估可以提前30～60分鐘評估出可能積淹水的位置。

關鍵字

即時淹水模擬；細胞自動機；雨水下水道；道路進水口與連接管；淹水預報

並列摘要

In recent years, the frequencies of extreme weather events are increasing due to climate change. Severe urban flooding events occur when the capacity of the drainage networks is incapable to cope with the intense rainfall in such conditions, especially under short-duration rainfall events. To correctly reflect the aforementioned flow conditions in short-duration rainfall events, the urban 2D/1D inundation models used for providing flooding maps should consider the limited capacities of road inlets and connected pipes between the gullies and the sewer systems. Moreover, the runoffs from buildings should be appropriately handled. Although the numerical accuracies of current urban 2D/1D inundation models are acceptable by including the aforementioned key factors, these models often suffer from unacceptable long computational time due to the heavy operations they introduce. Such a condition causes urban 2D/1D inundation models to be inappropriate on real-time evaluations of flood inundation in urban areas. Consequently, these models are usually adopted on producing flood maps off-line. In the present study, the two-dimensional overland flow model based on Cellular Automata (National Taiwan University Cellular Automata Flood Inundation Model, NTU-CAFIM), which is remarkable to have high numerical efficiency, is coupled with SWMM to develop an urban 2D/1D inundation model. The limited capacities of road inlets and connected pipes are considered in the dynamic linkage between the overland and sewer flow models. The proposed model is used in Taipei City to construct a real-time flood inundation forecasting platform. Several historical flood events within Taipei City are selected to verify the proposed model. In terms of predicted flood extents, the values of POD (Probability of Detection) in these cases are beyond 60%, which demonstrates the satisfactory accuracy of the model. The real-time evaluation of flood inundation automatically provided by the platform can estimate the potential flooded regions 30 to 60 minutes ahead.

並列關鍵字

Real-time flood inundation modeling ； Cellular Automata ； SWMM ； Road inlets and connected pipes ； Flood inundation forecasting platform

國際替代計量

地表細胞自動機快速淹水－街道集流器－雨水下水道管網之整合模擬分析：臺北市強降雨即時淹水預報平台之研發

全文下載

主題瀏覽