落石潛在危險度之不確定性評估初探

Abstract

臺灣是個多山島嶼，擁有複雜且破碎之地質條件，北中南三條橫貫公路邊坡每逢大雨過後必傳出落石與土石流災害。在大地工程中，落石坡研究相較於其他範疇，明顯有高不確定性之特徵。在以往對不穩定邊坡評估中，各項評估因子一般都採用絕對確定性之分級劃分原則，然而幾乎所有之邊坡相關參數卻多具有隨機性，因此落石危險度評估實有必要以機率與統計之觀念與方法來探討。本論文之研究，對評估落石相關調查資料之取得，除了探討現地調查外之照片影像處理技術，並嘗試利用蒙地卡羅模擬方法以評估落石尺寸之機率分佈，及利用統計方法探討邊坡危險度相關參數與其分級之可能範圍，期望能對落石潛在險度提供較合理之評估方法。

Rock-fall often occurs along many highway slopes, especially during or after a large rainfall. The consequence of rock-fall may result in traffic interrupt, property damage, and serious personal injury. The hazard potential of rock-fall depends on several factors including geology, topography, and environmental conditions. Compared to other aspects of geotechnical engineering, the rock-fall problem has a higher degree of uncertainty. In the conventional approach, however, the factors determining the hazard potential of rock-fall are usually evaluated in a deterministic manner. It seems reasonable to take the probabilistic nature of the roch-fall problem into account. This thesis aims to investigate the aspeccts for determining the probability and statistics of rock-fall hazard. In addition to a discussion on the site investigation for evaluating rock-fall hazard, this study proposes an approach for using photo-analysis to collect data of discontinuities. The collected data is then used for obtaining the probability and statistics of discontinuities. Using the information, the probability and distribution of rock-fall size is studied by using the Monte-Carlo simulation. Moreover, the uncertainty and reliability of the hazard classification of rock-fall potential is elaborated.