利用滲透試驗探討雙層土壤之滲流沖蝕行為

Abstract

滲流沖蝕(seepage erosion)係指土壤中的細顆粒因地下水流驅動通過較大顆粒間之孔隙，進而被帶離坡面，而滲流沖蝕一般會形成由坡面向坡體內部發展之漸進式破壞。針對台灣之林口台地蒐集相關案例及進行現地調查，該地區易發生滲流沖蝕破壞邊坡主要是因礫石層與其下方之阻水層間夾有一細砂層。一旦因滲流而產生之滲流力迫使細砂層之土粒產生移動情形，便可能造成邊坡開始發生破壞之現象。故為進一步減少因滲流沖蝕引發之邊坡穩定問題，對邊坡滲流沖蝕行為的瞭解有其必要性。 本研究就大南灣層之細砂層及室內混合不同粒徑之石英砂與高嶺土模擬透水係數較大之礫石層於不同水流條件下，進行滲透試驗並利用既有儀器觀察滲流沖蝕發展過程及情形。首先，針對單一土層試體分別求得細砂層及礫石層二種土壤之滲透係數，接著藉由改變試體裝填方式以獲得雙土層於水平滲流及垂直滲流狀態(水流方向與土層界面之關係)下之等值滲透係數。再者，為了解滲流沖蝕發生過程及破壞情形，將滲透室(permeability cell)水平擺放並移除頂蓋後，觀察單一細砂層、礫石層及雙土層(細砂層在下，礫石層在上)於水平滲流情形下，其滲流沖蝕發生過程及破壞情形。 由試驗結果發現，發現礫石層之試體於實驗過程中，因水流而造成礫石層內之基質材料(高嶺土)向下移動並流出土體，導致水力梯度與流速呈非線性關係。當試體為雙土層時，水平滲流之流速與水力梯度之關係符合達西定律；垂直滲流之流速與水力梯度之關係因受礫石層內之基質材料移動之影響而呈非線性關係。 接著，藉由觀察水平擺放後之試體發現，於單一礫石層內之滲流沖蝕行為始於基質材料流失，爾後底部土壤隨水流逐漸匯集而開始流動，進而導致上方土層崩落。於單一細砂層內之沖蝕行為，地下水存在使砂土呈流體狀，而造成蝕洞逐漸擴大並向土體內部發展，最終導致土體呈現滑動破壞。於雙土層中，其沖蝕行為亦始於細粒料流失，且隨水流持續發生而導致砂土持續向外流動並造成蝕洞之出現，最終將造成上方礫石層直接發生崩落之情形，而異於僅單一礫石層時之沖蝕行為。 最後，為了解數值模擬與實際情形間於滲流情形下之可行性及其模擬滲流沖蝕行為之適用性，本研究採用有限差分法軟體－FLAC5.0進行二階段模擬。第一階段為驗證數值軟體之可行性，進行定水頭滲透試驗之模擬；第二階段則進行滲流沖蝕實驗之模擬以了解數值程式於該領域之適用性。由模擬結果發現，於滲透試驗可獲得與室內試驗相近之結果；於滲流沖蝕實驗時，經過設置適當之臨界水力坡降與最大剪應變率門檻值，可獲得與試驗結果相似之滲流沖蝕行為，但因網格形狀、組成率及門檻值設定等限制，無法呈現出較微細之現象，如土壤流體化、蝕洞擴大等情形。

Seepage erosion is indicated that finer particles are dragging out from other soil particles by water and then causes progressive failure inward into the slopes and slope instabilities. When the groundwater begins to flow, the sand particles will move as soon as the seepage force is greater than the particle self-weights and inter particle friction forces. . Furthermore, according to site investigations and disaster histories on Lin-Kou tableland in Taiwan, we find out that there exists a particular sand layer between gravel layer and aquitard. This sand layer plays a crucial rule in affecting the seepage erosion of slopes. Therefore, it is necessary to explore the behavior of seepage erosion in order to prevent slopes from such failure mechanism. This study focuses on the particular sand layer in Tananwan formation (TN formation) and a gravel layer which has a greater saturated hydraulic conductivity by mixed up sands and kaolinite in the laboratory. And then, we begin to do permeability tests and seepage erosion experiments by using the apparatuses of permeability test. First, we do the permeability tests on the sand layer and the gravel layer separately to get their saturated hydraulic conductivities. Secondly, we do the tests to get the equivalents conductivities of two layers under parallel and perpendicular flows by changing the fill patterns of specimens. Furthermore, in order to know the seepage erosion behavior, we level the permeability cell without the cell cap to run the experiments on a one layer of sand, gravel and two layers (gravel above sand) separately under flows parallel to the layer. Based on the results of experiments, water forces the matrix of gravel (kaolinite) to move downward and flow out the specimens. Because of this phenomenon, the relation between hydraulic gradient and flow velocity become non-linear eventually. When the specimen is two layers, the relation was quite fit the Darcy’s law in parallel flow, but has a non-linear relation in perpendicular flow that affected by the movement of matrix particles. After that, we level the permeability cell without the cell cap to run the experiments. We can know that seepage erosion behavior started immediately with the dislocation of finer particles. Next, the particles close to the bottom of the soil specimen begin to flow with water and consequently the particles in the upper part of specimen collapsed as a block. However, in the sand specimen, water flows with soil particles and eventually the specimen fails as in sliding failure. However in the two layer specimen, the erosion behavior stated with finer particles dislocation just the same as sandy material. As seepage erosion takes place continuously, the two layer soils collapsed directly, unlike the seepage erosion process in the one layer or specimen. At the end, in order to know the suitability and feasibility of numerical analysis on simulating seepage erosion behavior, we choose FEM software – FLAC5.0 to do the work. In this part, we also started from the permeability tests and then the seepage erosion experiments as the same as the process of laboratory tests. According to the results from numerical analyses, the conductivities are similar to permeability tests. And the simulations of seepage erosion experiments, we can find the same states when specimens are failure, but not for the processes because of limitations in the software.