改良樁對黏土層開挖穩定之數值分析

Abstract

地盤改良技術已廣泛應用於提高土壤的強度與勁度，其中柱狀型式改良樁常用於軟弱黏土層之深開挖工程中，以抑制擋土牆的變形並提供阻抗開挖底部的隆起，進而提升開挖過程之穩定性。然而，近期發生建築基地下方黏土層設置柱狀型式改良樁仍發生隆起破壞，引發對改良樁佈設位置是否能有效提供垂直向與側向承載，對隆起破壞機制之影響等問題應進一步討論。本研究透過 PLAXIS 有限元素法分析，探討改良樁排列方式與壁體變位、地表沉陷及開挖穩定性之關係。另外，在模擬軟弱黏土於開挖過程中之力學行為，常用組成律模式為Mohr-Coulomb (ϕ= 0) model (MCUC) 不排水總應力分析與 Hardening soil model with undrained A (HSUA) 不排水有效應力分析，故本研究探討兩者對於變形量分析與穩定性分析結果之異同。首先，以 Song-San 深開挖分析案例，利用三維四分之一模型以壁體監測資料校正土壤組成律參數，結果顯示縱使兩個組成律能良好的預測壁體變形，但 HSUA 之沉陷量分析結果有大於 MCUC 之趨勢。此外，鑑於三維分析會耗費較長的計算時間與模型建置不容易，現階段常用於分析柱狀改良樁的單一等值化材料法過於簡化真實的幾何情形，目前尚未有文獻針對深開挖中改良樁之行為提出恰當等值化平面應變之方法。因此，本研究引入前人對於群樁基礎之各種簡化模型方法，將各種方法對應成柱－牆簡化之平面應變模型，並與三維模型進行比對。結果顯示，在不同改良樁的間距與直徑之條件下，Equivalent shear strength (ESS) 與三維強度折減分析之結果相似。以變形分析而言，ESS 與 Equivalent area (EA) 分別能預測到 HSUA 與 MCUC 三維分析之變形量最大值。於強度折減分析結果顯示，柱狀改良樁只影響節點位移曲線之破壞初期，且當前勁度參數 (current stiffness parameter, CSP) 與節點位移曲線之發展有相關，基於此論點，本研究以 CSP 下降斜率最大位置定義為初始破壞點，用以評估柱狀改良樁對於穩定性之影響。

The ground improvement technology is widely used to enhance the strength and stiffness of soil. Particularly in soft clay layers, it can restrain the displacement of retaining walls and provide additional resistance against basal heave. However, the recent excavation projects that adopted the column-type ground improvement method in soft clay layers have still encountered basal heave failure. Therefore, the effectiveness of column-type improvement in contributing to vertical and lateral capacity, as well as its impact on basal heave failure mechanisms, needs further investigation. This study uses the PLAXIS finite element method to examine the relationship between the geometric arrangement of improvement piles and wall deflection, ground settlement, and excavation stability. Two constitutive models commonly adopted for analysis of deep excavation behavior in soft clay layers are compared: the Mohr-Coulomb (ϕ=0) model (MCUC) for undrained total stress analysis and the Hardening soil model with undrained A (HSUA) for undrained effective stress analysis. Evaluate the performances of these two models on the deformation and stability analysis results. First, a three-dimensional quarter model was applied to simulate the Song-San deep excavation project, calibrating the soil parameters through wall monitoring data. Both constitutive models demonstrated an ability to predict wall deformation; however, the HSUA model yielded greater settlement values than the MCUC model. This study also aims to reduce the computational cost of three-dimensional analysis and modeling, so simplified plane strain approaches for column-type improvement piles are needed. The single equivalent material method is commonly used but oversimplifies the actual geometric conditions. No literature currently provides an appropriate equivalent plane strain method for predicting the behavior of improvement piles in deep excavations. Therefore, this study aims to compare various simplified plane strain approaches proposed in the previous research, which were formulated for pile group foundations and then applied to ground improvement piles in deep excavations. The performance of various simplified column-wall plane strain models is evaluated and compared with those of the three-dimensional model. Under varying pile spacing and diameter, the Equivalent Shear Strength (ESS) method yielded results comparable to those obtained from three-dimensional strength reduction analysis. In terms of deformation analysis, the ESS and Equivalent Area (EA) methods could predict the maximum values of deformations from the HSUA and MCUC three-dimensional analyses, respectively. The results of strength reduction analysis indicated that column-type improvement piles primarily affect the initial failure behavior. The current stiffness parameters (CSP) are related to the development of the nodal displacement curve. Based on these observations, this study defines the initial failure point as the position where the CSP greatest decays and uses this definition to evaluate the impact of column-type improvement piles on excavation stability.