以數值分析探討不同降雨重現期下光華地區地下水變化及邊坡變形行為

Abstract

邊坡潛在大規模崩塌之穩定性及其變形行為，為坡地災防研究的重點項目之一。然而，許多潛在大規模崩塌地變形速率緩慢，甚至存有僅局部位置活躍運動之現象。為此，本研究旨在釐清光華大規模崩塌潛勢區在不同重現期降雨下之地下水位變化與邊坡變形行為。首先以現地調查確認地表露頭位態、地表滲水處位置等，探討光華崩塌地的地質模型。以烟花颱風事件分析降雨與地下水位變化關係，建立地下水位直接步推歷線(Direct Step Method of Groundwater Hydrograph, DSM-GH)。藉由二維、三維有限元素法數值分析擬合烟花颱風事件之地表位移監測數據和DIC分析成果，獲得地盤模型參數。透過DSM-GH評估不同降雨重現期所對應之地下水位歷線，探討不同重現期地下水位變化及其對邊坡變形行為之影響，並以強度折減法(Shear Strength Reduction, SSR)探討光華崩塌地之崩塌潛勢發展趨勢。最後，以橫向集水管建立地工模型，進行排水穩定工法性能評估。 研究結果顯示DSM-GH擬合烟花颱風事件R2約為0.73，在驗證梅花颱風事件時R2約為0.85，因此認為DSM-GH可適用於監測資料量不足時，評估不同降雨重現期時可能的地下水位變化。利用DIC分析成果與二維數值模型比較之R2約為0.62，而比較三維數值模型之R2約為0.74，兩者差異來自於二維數值模型在分區2下邊坡處因不受束口狀地型效應影響，下邊坡處位移量不減反增，而三維數值模型則可以較為合理的模擬現地狀況。當重現期為50年之地下水位時，M1塊體中的分區1在二維數值分析時會發生破壞，而在三維數值分析時，受束口狀地形影響則不會發生破壞，且兩者之最大總位移量分別為16.7 m和4.4 m。在利用三維數值模擬進行SSR分析時，光華崩塌地在M1塊體失去強度後，首先會自M4塊體產生張力裂縫(SRF=1.15)，M3塊體會在後續發展出張力裂縫(SRF=1.25)，最後兩處張力裂縫會聯通並形成潛在滑動面(SRF=1.55)，因此認為除持續監測M1塊體外，針對M3、M4塊體的整治與監測工作亦需要被重視，並建議可在FST02-G2站上邊坡處再增設GPS/GNSS站及設置伸張計。而橫向集水管可以防止M1塊體分區1發生破壞，並減緩M3、M4塊體的剪裂帶發展。

The stability and deformation of large-scale landslides poses a long-term threat to the protected objects in mountainous areas. However, assessing the potential failure of large-scale landslides remains challenging due to their extremely low deformation velocity. This research aims to investigates the relationship between groundwater variation and deformation kinematics of Guanghua large-scale landslide. Outcrop investigations were first conducted to measure the attitudes of weak plane, in order to establish a geological model of the Guanghua landslide. To correlate the rainfall pattern and groundwater change, the groundwater well data and the rainfall record from 2021 Typhoon In-Fa were used to assess the Direct step method of groundwater hydrograph (DSM-GH). A ground models in 2D and 3D FEM models were established based on calibrate the performance of the surficial displacement data from GNSS observation and Digital image correlation (DIC). The DSM-GH was used to estimate groundwater variations under different rainfall return periods and evaluate their influence on deformation Kinematics. The failure potential of the Guanghua landslide was then assessed using the Shear Strength Reduction (SSR) method. A geotechnical model incorporating horizontal drainage pipes was subsequently developed to evaluate the performance of the drainage stabilization method. The results indicated that DSM-GH yielded an R² of approximately 0.73 for the Typhoon In-Fa event and around 0.85 for the Typhoon Muifa. This demonstrates its applicability in estimating potential groundwater variations under different rainfall return periods, especially when only short-term monitoring data are available. The R² value between the DIC analysis and the 2D numerical model is approximately 0.62, compared to 0.74 for the 3D model. The difference arises from the 2D model’s inability to capture topographical constraints at the lower slope of Zone 2, leading to overestimated displacements, whereas the 3D model offers a more realistic simulation of the site conditions. Under the 50-year return period groundwater table, Zone 1 of the M1 block fails in the 2D analysis but remains stable in the 3D analysis due to topographical constraints. The maximum total displacements are 16.7 m (2D) and 4.4 m (3D). In the 3D numerical simulation using the SSR method, failure of the key block M1 induces tensile cracking in M4, followed by M3. These cracks eventually connect to form a potential sliding surface. Therefore, in addition to continued monitoring of M1, stabilization and monitoring of M3 and M4 should also be emphasized. It is recommended to install additional GPS/GNSS stations and extensometers on the upper slope near station FST02-G2.