水壩安全之綜合風險分析

Integrated Risk Analysis for Dam Safety

許永佳(Yung-Chia HSU)

指導教授 ： 郭振泰

共同指導教授 ： 湯有光(Yeou-Koung Tung)

國立臺灣大學/工學院/土木工程學研究所/博士(2007年)

https://doi.org/10.6342/NTU.2007.00683

摘要

台灣位於東南亞的太平洋板塊與歐亞板塊交界上，因此經常遭遇颱風的襲擊以及遭受地震所造成之災害。由於台灣的工業及經濟發展、人口稠密，國人對於水資源的需求增加的同時，對水壩的安全也愈來愈重視。過去的許多統計資料顯示水壩溢頂、水壩滲流是土石壩潰壩之最主要因素；另外，台灣位於西太平洋的地震帶上，地震可能造成壩坡永久位移，甚至潰壩。因此本研究探討三種主要潛在的造成水壩失敗事件的型式，包括：因洪水及風造成之水壩溢頂、因滲流引起的壩坡不穩定，以及因地震造成之壩坡位移。
水壩溢頂模式考慮水文的觀測量之不確定性以及洪水頻率分佈曲線之不確定。在水文頻率模式中，本研究考慮年最大序列以及月最大序列等兩種模式；兩種模式均考慮三種機率分佈作分析。本研究並提出一個取樣法IS-LHS method，其結合重點取樣法(importance sampling method)以及拉丁高次取樣法(Latin hypercube sampling method)，從洪水與風速之頻率模式的分佈中取樣。取樣之樣本作為洪水演算之輸入，用來評估水壩溢頂風險；其中，洪水演算模式結合洪水時期之操作規則(operation rules)、風引起之湧浪及爬坡等模式。
在壩坡穩定分析中，首先須決定滲流面(phreatic surface)的位置。本研究考慮穩定及非穩定的水庫水位所造成之滲流面。其中，非穩定水位之滲流面由洪水時期之水庫操作所造成水庫水位急上升以及水位急卸降等兩種情況。滲流面的分析結果將作為壩坡穩定性分析之初始假設。考慮土壤強度參數之變異性，本研究以蒙地卡羅法將土壤強度參數作取樣，即可分析壩坡穩定之可靠性。
在地震造成之壩坡位移模式中，本研究採用林晉祥教授及惠曼教授之模式(Lin and Whitman’s model, 1986)，評估地震造成之壩坡永久位移(permanent displacement)。本研究蒐集75筆地震資料，考慮地表運動參數的變異性，土壤強度對於臨界加速度(critical acceleration)的影響，以及地震危害度(seismic hazard)，以此模式估計水壩可能遭遇之地震壩坡位移之風險。
最後，本研究以系統可靠度分析將上述三種可能之失敗型式整合並計算石門大壩之失敗風險。本研究探討之三種可能失敗型式，不代表水壩一定會遭受到結構性的崩潰(例如，發生水壩溢頂不必然會發生潰壩事件)，僅代表石門大壩因這三種失敗型式發生的機率。

關鍵字

風險分析 ； 水壩溢頂 ； 滲流 ； 地震 ； 壩坡穩定性分析 ； 壩坡位移 ； 可靠度分析

並列摘要

Taiwan, located in Southeast Asia, frequently encounters typhoons and earthquakes. By receiving increasing attention from the public, dam safety has become an important issue. Many researchers have reported that overtopping, seepage, and piping are the main causes of dam failures. This study takes Shihmen Dam as a case study of three potential failure modes: overtopping induced by flood and wind events, slope instability due to seepage under steady and transient states, and permanent displacement due to earthquakes.
The overtopping model accounts for the uncertainties involved in hydrological observations and flood frequency models. The annual maximum (AM) and monthly maximum (MM) series models are considered; each model considers the Gumbel, Lognormal, and LogPearson type III distributions. A proposed sampling method replicates the sample set of peak flow rate and wind speed. Reservoir routing incorporates operation rules, wind setup, and run-up models and is used to evaluate dam overtopping risk with the input of the peak flow rate and wind speed sampled by the proposed sampling method.
Determining the locations of the phreatic water table under steady reservoir water levels is the first step of slope stability analysis. Results of seepage analysis are then treated as initial conditions in the stability analysis of earth dams. The slope stability model for seepage under steady and transient states is used to assess the reliability of slopes of Shihmen Dam. Monte Carlo sampling (MCS) is used to reproduce the parameters associated with soil strengths, and to evaluate the reliability of the slope stability of Shihmen Dam.
A probabilistic model is used to assess the permanent displacement induced by earthquakes. This model is verified by the simulated results through time domain double integration on 26 real earthquake motions on rock sites. The ground motion is modeled as a portion of a stationary random process defined by a Kanai-Tajimi spectral density function. In addition, 38 sets of ground motion parameters on rock sites and 11 time history records observed at Shihmen are also used to portray the variability of ground motion parameters. Moreover, the uncertainty of the critical acceleration is also accounted in this study by considering the variability of soil strengths. This model has been proved to yield the best estimate of the distributional properties of permanent displacement.
The probabilities of the three failure modes are then integrated through the system reliability algorithm, which yields the overall risk of Shihmen Dam. The occurrence of three failure modes investigated in this study does not necessarily lead to structural failure of Shihmen Dam. For example, the occurrence of dam overtopping does not necessarily result in dam collapse. Thus, the probabilities of the three failure modes should not be treated as the probability of dam collapse.

並列關鍵字

Dam safety ； overtopping ； flood frequency ； slope stability ； seepage ； permanent displacement ； uncertainty ； earthquakes

參考文獻

117. Yeh, Y.T., Loh, C.H., Yeh, Y.H., Hsin, J.C., Tsai, C.C., Jian, W.Y. Cheng, S.N. and Huang, W.J. (1993). The seismic assessment of the reserved highway corroder. The Chinese Geophysical Society. (in Chinese)

1. Afshar, A. and Marino, M.A. (1990). “Optimizing spillway capacity with uncertainty in flood estimator.” J. of Water Resources Planning and Management, 116(1), 74-81.

5. Askew, J.A., Yeh, W.G., and Hall, A.H. (1971). “Use of Monte Carlo Techniques in the Design and Operation of a Multipurpose Reservoir System.” Water Resources Research, 7(4), 819-826.

8. Bishop, A.W. (1955). “The use of the slip circle in the stability analysis of earth slopes.” Géotechnique, 5(1), 7–17.

9. Casagli, N., Rinaldi, M., Gargini, A., and Curini, A. (1999). “Pore water pressure and streambank stability: results from a monitoring site on the Sieve River, Italy.” Earth Surface Processes and Landforms, 24(12), 1095-1114.