仙渡電廠基樁動力試驗與數值模擬分析

Abstract

摘 要 從施工振動引致鄰房損害、到晶圓設廠的微振控制，地表振動的問題常成為被關心的研究重點。由於現場全尺度振動量測資料稀少珍貴，而仙渡電廠又有一系列完整的基樁試驗量測資料，其中包含大尺寸基樁之靜動載重試驗與重錘落擊試驗。本研究於兩試驗時，於地表面上裝設一系列速度計量測地表振動歷時，並利用頻譜分析原理，將時間域訊號轉為頻率域以進行分析，並得到地盤振動隨距離衰減的關係。 在數值模擬分析部分，本研究選取ABAQUS 有限元素分析軟體，採二維軸對稱分析模式。所採用的土壤模式包含四種分析模式：1.無介面彈性土壤。2.無介面彈塑性土壤。3.有介面彈性土壤。4.有介面彈塑性土壤模式；再各別與試驗記錄作比較，並討論樁頭荷載-位移曲線、樁身應力、土壤塑性區、介面滑動情況與地表振動反應等內容。 兩試驗的數值模擬分析結果以有介面彈性土壤或彈塑性土壤模式與試驗結果最為接近。靜動載重試驗數值模擬於樁頭荷載-位移曲線部分，當樁頭達最大荷重前的沉陷量皆可與試驗值有相當程度之契合，但在後半段樁體因慣性力繼續沉陷的部分，此模式與試驗結果仍有一段差距；兩試驗在樁體應力歷時數值模擬部分已有不錯的分析結果；另外，檢視地盤振動的1/3倍頻頻譜，靜動載重試驗模式四的分析中，於低頻部分模擬結果較高頻部分佳，中間頻率5~16Hz部分數值分析會有高估的現象；而重錘落擊試驗反而是在低頻部分有高估的現象，但大體上來說，數值分析所得的趨勢與試驗結果類似。

Abstract This study aims at investigating the ground vibrations induced by dynamic pile tests. The cases selected for studying are the Statnamic and drop-hammer tests conducted at the Shian-Du Extreme-highly Voltage Substation located at the western tip of the Guan-Du Plain of Taipei Basin. The Statnamic test of maximum load 2000ton was conducted on a bored-pile of diameter 1.1m and length 52m and the drop-hammer tests of weight 16.7ton and falling-heights of 0.5m, 1.0m and 1.5m, respectively, were conducted on the same pile. During the tests, a series of velocity sensors were deployed along a line to a distance of 100m from the pile center to measure the ground vibrations induced. Data collected were analyzed to deduce the 1/3-octave-band vibration amplitude spectrum and the coefficients of ground vibration attenuation with distance were also deduced. In addition, numerical simulations for both tests are performed in this study by using the computer program ABAQUS. Axisymmetric analyses with 4-node soil and pile elements were adopted. Four types of model were selected for parameter studies. They are: (1) Linear soil elements without pile/soil interface elements; (2) Elasto-plastic soil elements without pile/soil interface elements; (3) Linear soil elements with pile/soil interface elements; and (4) Elasto-plastic soil elements with pile/soil interface elements. Both the piles responses and the ground responses obtained from numerical analyses are compared with the results in-situ of tests. Based on the comparison studies, it is found that both the models with pile/soil interface elements (i.e., model 3 and 4) are acceptable to simulate the behavior of dynamic pile tests. The responses of pile obtained are quite close to the test results. However, the responses of ground vibrations obtained are not so satisfactory due to the complexity and uncertainties in the in-situ conditions.