物質點法分析降雨引致不飽和土壤邊坡破壞之研究

Abstract

臺灣具有複雜的地理環境與氣候條件，山坡地時常受到高強度或強延時之降雨事件影響，使臺灣經常發生坡地破壞等災害，嚴重影響居住人民的生命財產安全。然而因為邊坡破壞機制的複雜性，目前仍難以找到全面且通用的方式進行評估與防災安全設計。就數值分析方面，常見之有限元素法分析適合模擬小變形問題，如邊坡破壞前的力學機制，當分析大變形問題，如邊坡破壞後之運移距離與堆積範圍，會發生網格扭曲導致運算無法繼續進行。因此本研究採用物質點法(Material Point Method)，藉由物質點法使用自由移動的物質點及背景網格的節點計算，可以模擬大地工程中大變形的問題。 本研究以Moriwaki et al.(2004)之降雨引致邊坡破壞之大型物理試驗進行數值分析，模擬不飽和土壤邊坡受降雨入滲之過程、邊坡破壞時之孔隙水壓變化、剪裂帶發展、邊坡地表位移、速度隨時間之關係，並與試驗結果進行比較驗證。在本研究中以此大型物理試驗進行全飽和與不飽和土壤兩種材料之模擬，全飽和模型為預設整體土壤已達飽和，即將發生破壞之狀態，而不飽和模型則能夠探討整體降雨入滲過程，分析降雨過程中土壤孔隙水壓、飽和度等變化對於邊坡破壞的影響。 結果顯示全飽和模型如前人之研究可以有效模擬此物理試驗之破壞過程；而不飽和模型若使用真實土壤滲透係數進行模擬將耗費過量計算效能及時長，因此本研究對於滲透係數及降雨強度進行尺度放大，並探討尺度放大後導致模擬結果不良之原因，並對其進行參數的調整，最終在現有計算效能下，針對不飽和土壤入滲分析提出一有效之模擬方式，並對此模擬之結果進行探討。

Taiwan's complex geography and climate often lead to heavy rainfall events that frequently cause slope failures, posing significant threats to the residents. The complexity of slope failure mechanisms makes it challenging to find a universal approach for assessment. This study used the Material Point Method (MPM) to simulate large deformation problems, overcoming the limitations in finite element method analysis for large deformation issues. This research conducts a material point method analysis based on a full-scale landslide flume test by Moriwaki et al. (2004), simulating the processes of rainfall-induced unsaturated soil slope failure, changes in pore water pressure, the development of shear zones, surface displacement, velocity of the slope, and comparing them with the flume test results. In this study, simulations for both fully saturated and unsaturated soil materials are performed; the fully saturated model assumes that the entire soil is saturated and is at the onset of failure, while the unsaturated model can explore the overall water infiltration process and analyze the effects of changes in pore water pressure and saturation during rainfall. The results show that the fully saturated model, as in previous studies, can effectively simulate the full-scale landslide flume test failure process; however, using the real soil permeability coefficient for the unsaturated model would consume excessive computational time. Therefore, this study scales up the permeability coefficient and rainfall rate and discusses the causes that lead to poor simulation results after scaling up, adjusts parameters accordingly, and finally proposes an effective simulation suggestion for unsaturated soil analysis of under existing computational capacities, and discusses the simulation results.