透過您的圖書館登入
IP:3.239.57.87
  • 學位論文

大規模崩塌運移模擬-連體及非連體模型案例擬合

Run-out Simulation of Large-scale Landslides - Back Analyses with Continuum and Dis-continuum Models

指導教授 : 潘以文

摘要


大規模崩塌之災害評估中不可或缺的要素為致災性參數,泛指運移距 離、影響範圍、衝擊速度、動量…等,不少過往研究提出推估致災性參數之方法,其中數值模擬法是求解致災性參數相對可靠之方法,且能對崩塌區作較精準的描述。本研究蒐集臺灣十處已發生崩塌之大規模崩塌案例,以連續體模型與非連續體模型數值軟體反演崩塌運移,率定最佳輸入參數以令模擬結果接近真實崩塌情形,連續體模型以 RAMMS (RApid Mass Movements Simulation)為代表;非連續體模型則以PFC3D (Particle Flow Code)為代表。 為率定輸入參數,本研究以真實崩塌案例之堆積範圍與堆積厚度作為 率定比較對象,以量化的方式調整 RAMMS 與PFC3D之輸入參數以令模擬 結果能最接近真實崩塌行為。本研究依據真實崩塌案例之堆積範圍及厚度率定 RAMMS 主控運移距離之摩擦係數與影響流速變化之紊流阻抗係數;PFC3D方面則以模擬之單壓試驗擬合現地室內試驗之巨觀力學參數藉以標定微觀參數,運移模擬之反演係經由調整摩擦係數及黏滯阻尼比以達最佳模擬結果。 就所蒐集之崩塌案例,本研究繼而根據不同因子分類,如破壞機制、 量體體積、平均坡度等,探討 RAMMS 反演所得最佳擬合參數與崩塌案例因子間關係,歸納得以下趨勢: 量體體積越大、崩塌區平均坡度越大之崩塌案例,擬合最佳之摩擦係數值越大;摩擦係數固定之情況下,紊流係數對崩塌量體之運移距離影響甚小,但對崩塌量體之運移速度有較大影響。本研究並就幾個不同破壞機制之崩塌案例分採 RAMMS 與PFC3D進行模擬反演,比較兩個軟體之適用性與效能。 RAMMS 由於建模與模擬速度快,有利於初步風險評估,盡速提供即時防災管理所需;PFC3D則因建模與運移所費時間較長,但卻能對微觀物理量與破壞模式能有較精確之描述,更適合作為詳細研究之用。 整體而言,兩軟體於本研究比較之幾個案例模擬表現均可接受。

並列摘要


Hazad parameters of landslides, such as run-out distance, affected zone, inpact velocity and momentum, are important factors for the risk assessment of landslides. Many previous studies have proposed a variety of methods for estimating these hazard parameters. Among those methods, numerical simulation is relatively more capable and reliable for obtaining the comprehensive hazard parameters. This study collected ten cases of large-scale landslide failure in Taiwan and adopted both the continuum model RAMMS (RApid Mass Movements) and the discontinuum model PFC3D (Paticle Flow Code) to conduct back analyses of landslide run-out simution. The back analyses enabled the calibration of the input parameters of RAMMS and PFC3D so that the simulated run-out deposition would best fit the actual deposition after landslide. For RAMMS, the friction coefficient is the dominant parametrer controlling the run-out distance, while the turbulence coefficient affects the runout velocity. For PFC3D, most of the input micro-parameters (except frictional coefficient and damping ratio) were at first calibrated through the stress-strain curve of unconfined compressive test. Subsequently, the friction coefficient and the viscous damping ratio were adjusted to let the simulated run-out results as close to the actual deposition as possible. All the simulated cases by RAMMS were classified by various factors, including their failure mechanism, source volume and average slope in order to examine the relations between the calibrated parameters and these factors. The calibrated friction coefficient tends to be larger as the source volume or the average slope increases. For a fixed friction coefficient, the turbulence coefficient has little effect on the run-out distance, but notably affects the runout velocity. With the simulation of several common landslide failure cases, this study also compared the capability and performance of RAMMS and PFC3D. RAMMS is advantageous for its easiness of use and low computation demand, hence it may be more suitable for the prompt risk assessment of a potential landslide. To the contrary, modeling by PFC3D is usually more timeconsuming and its computation demand is generally higher. However, PFC3D has the advantage for providing more detail microscopic variables in the course of landslide run-out process. For all the common cases for comparison, it appears the capability of landslide run-out simulation by RAMMS and PFC3D are both acceptable for risk-assessment purpose.

參考文獻


1. 中央地質調查所全球資訊網,https://www.moeacgs.gov.tw/
2. 林昊勳 「大規模崩塌的邊坡破壞類型 -以莫拉克風災高屏溪、曾文溪流域為例」,國立交通大學,碩士論文,民國 107 年。
3. 林務局,「國有林大規模崩塌潛勢區減災策略研擬及安全監測成果報告」,民國 106 年。
4. 施虹如、吳亭燁、蘇元風、劉哲欣、李欣輯、陳永明與張志新, 「極端降雨事件下淺層崩塌潛勢衝擊評估分析」, Journal of Engineering,
(34), pp. 77-96, 民國 104 年。

延伸閱讀