以振動台試驗探討地表不透水層厚度對砂性地盤液化噴砂與沉陷之影響

Abstract

臺灣位於環太平洋地震帶上，地震頻繁發生，當疏鬆飽和的砂性地盤在強烈地震作用下之時，土壤孔隙水壓上升，導致有效應力降低，進而引發土壤液化現象。土壤液化常伴隨著噴砂，尤其地表存在不透水層，超額孔隙水壓在消散的過程中會沿著弱面將懸浮之砂土噴出地表。本研究使用剛性盒於1g振動台進行液化噴砂縮尺模型試驗，探討輸入振動之最大加速度、地表不透水層厚度和振動歷時對土層總沉陷量、噴砂量、液化時間以及激發水壓段歷時的影響。試驗中使用飽和石英砂作為液化層，而不透水層則使用石英粉與高嶺土混合而成，試體高度為20公分，模型縮尺率為百分之一。透過量測設備記錄不同深度的加速度和水壓隨時間變化，並進行動態分析。試驗前後拍攝了砂箱之模型地表面，以利於後續進行三維數值建模，並量化液化後之噴砂量與沉陷量。試驗結果顯示，輸入振動之最大加速度越大，地表沉陷量、總噴砂量和平均每孔噴砂量也越多。隨著不透水層厚度增加，其覆土壓力也相應增加，地表沉陷量、總噴砂量和平均每孔噴砂量也呈現增加的趨勢。振動歷時的增加，地表沉陷量、總噴砂量和平均每孔噴砂量會增加，與振動歷時呈正相關。較大的愛氏强度與地表沉陷量、總噴砂量和平均每孔噴砂量呈正相關，並且不透水層覆土壓力對沉陷量和噴砂量的影響較振動歷時顯著。在相同愛氏强度下，振動歷時對噴砂量的影響比最大加速度顯著。較高的最大加速度及較厚的不透水層將提早液化發生的時間，較大的最大加速度及較厚的不透水層會延長相同深度下激發水壓段的持續時間。觀察加速度歷時圖，在液化過程中，地表下17.5 m的加速度未降至趨近於零，因覆土應力較大，較不容易發生液化現象。本研究顯示空中三角測量技術可應用於土壤液化噴砂縮尺試驗，並且液化後引致的噴砂量與地表沉陷量分別和輸入振動之最大加速度、地表不透水層厚度與振動歷時呈正相關。

Taiwan is located on the Circum-Pacific seismic belt and experiences frequent earthquakes. The phenomenon of soil liquefaction occurs when loose saturated sand is subjected to intense seismic loading, increasing pore water pressure and decreasing effective stress. Soil liquefaction is often associated with sand boils, particularly when an impermeable layer deposits on the ground surface, causing the excess pore water pressure to dissipate along weak planes and eject suspended soil particles onto the ground surface. This study aims to investigate, under cyclic loading, the effects of maximum acceleration, the thickness of an impermeable layer, the shaking duration on ground settlement, the volume of the sand ejection, liquefaction duration, and the duration of the excess pore pressure. Hence, a series of 1g shaking table model tests were conducted, and a rigid box was utilized. The model used saturated silica sand as the liquefiable layer, while the impermeable layer was composed of a mixture of silica powder and kaolin clay. The model specimen has a height of 20 cm, and the scaling factor between the model and prototype is 1/100. Acceleration and pore water pressure at different depths were measured using sensors and analyzed automatically. Pre- and post-test images of the model ground surface were captured to facilitate subsequent three-dimensional modeling (3D modeling) and quantification of sand boils and ground settlements for post-liquefaction. The experimental results show that ground settlement, total ejection volume, and average ejection volume are increased with higher maximum accelerations. As the thickness of the impermeable layer increases, the overburden pressure also increases, resulting in a corresponding trend of increased ground settlement, the total volume of the ejection, and the average volume of the ejection. The increase in shaking duration is positively correlated with the increase in ground settlement, the total volume of the ejection, and the average volume of the ejection. The ground settlement, total ejection volume, and average ejection volume positively related to Arias intensity were observed. Additionally, the impermeable layer, representing the overburden pressure, has a greater impact on settlement and sand ejection than shaking duration. Under the same Arias intensity, the effect of shaking duration on sand ejection is more significant than the effect of maximum acceleration. Higher maximum acceleration and thicker impermeable layers caused liquefaction to occur earlier and extended the duration of excess pore water pressure at the same depth. The acceleration-time history reveals that acceleration at a depth of 17.5 m beneath the ground surface did not approach zero during the liquefaction process due to greater overburden pressure, making liquefaction less likely to occur. This study demonstrates the applicability of aerial triangulation in scaled model tests of soil liquefaction-induced sand boiling. The results indicate a positive correlation between the volume of the sand ejection and ground settlement with the maximum acceleration, thickness of the impermeable layer and shaking duration.