以可靠度為考量之颱洪期間區排護岸破壞預警架構研擬

Abstract

本研究旨在針對區排護岸破壞建立一預警架構。當河川流經都會區或者人口聚集之區域，護岸是洪氾來臨時的第一道防線，若能建立破壞預警機制，對於災害的防制及應變將有莫大的幫助。本研究以中央管區排德盛溪為對象，將過去有發生過破壞且臨近有住家的護岸列為優先目標。研究流程可分為四個部份：(1)統整護岸的設計與破壞機制，不論何種型式的護岸，破壞成因可大致歸為護岸因水流衝擊或護岸本體不穩定而破壞兩類；(2)採用拉丁超立方取樣法考慮曼寧n之不確定性，透過水理模式計算在各種設計流量與曼寧n之組合下，護岸是否會因水流剪力過大而破壞；以及不同基礎沖刷深度對於護岸傾倒安全係數的影響；(3)以水位做為預警之指標，並提出以型二錯誤(護岸破壞但未發布預警)發生風險為依據之護岸破壞警戒水位值訂定方法；(4)整理模式計算而得的資料，包括護岸發生破壞前1小時、前2小時之水理狀況，分析其統計特性，依據機率密度函數訂定警戒之指標，再加以驗證。本研究以混凝土、箱籠兩種型式之護岸為分析目標，結果顯示箱籠護岸在水流衝擊方面的預警不甚理想；但混凝土護岸不論因水流衝擊或本體不穩定而破壞兩方面都達到預期的預警效果。期望此研究成果能作為相關預警及規劃之參考。

Since revetment is the front line of defense against flood, the development of a reliable revetment failure early-warning framework is very important especially for metropolitan area. The purpose of this study is to develop a revetment failure early-warning framework considering the forecasting uncertainty. The Tesheng River is chosen as study site and the development of early-warning framework includes the following steps: (1) Analyze the mechanism and cause of revetment failure. In general, the major failure mechanism includes the hydraulic force resulted from shear stress and the geo-instability due to toe erosion. (2) Generate many realizations of Manning’s n and inflow hydrograph by Latin Hypercube Sampling technique (LHS). Shear stress and water level under different combinations of Manning’s n and hydrograph were simulated. The simulation results were then used to determine whether the revetment will fail or not. Two failure modes are considered in this study including the surface erosion due to excess shear stress and the overturning of revetment due to toe erosion. (3) The analysis results from the previous step show that the water level could be a major index to judge that the revetment will fail or not in the following hours. Therefore, a framework was proposed which enables the pre-determination of thresholds of water level for revetment failure early warning under various combinations of toe erosion, pore pressure behind the revetment, and stipulated probability of type II error (i.e., failure to warn). (4) Analyze the statistical properties of water levels at one and two hours prior to failure from which the warning water levels were determined. The verification results showed that the actual and stipulated probabilities of type II error are close for the concrete type revetment. However, the actual probability of type II error could be much less than the stipulated value for gabion.