由傾度管資料推求基樁反應之研究

Abstract

摘要 欲實際了解基樁之側向承載能力，工程實務上通常會進行基樁側向載重試驗並埋置傾度管以量測樁身傾角變化，進而求得樁身變位。許多學者利用彈性梁理論之基樁變形與受力間的微分關係，將現地量測的樁身變位透過微分關係，推求樁身反應及土壤反力，並進而建立現地p-y曲線。 事實上，將現地量測的樁身變形直接進行微分，求得樁身反應之分析結果並不理想，尤其是須透過四次微分才可得到的土壤反力分佈，因為現地傾度管中量測的樁身變位曲線常含有誤差，微分四次將會造成誤差的放大，另外，彈性梁理論是適用於材料皆為彈性階段，因此當樁材受力變形進入非線性狀態時，彈性梁理論之微分關係便不適用。 基於上述的原因，邱俊翔及陳正興 ( 2000 ) 發展一套以彎矩函數曲線為出發點的回歸分析模式，其僅需要微分一至二次，便可得到樁身剪力及土壤反力，而其微分關係與材料性質無關，因此當材料進入非線性後依然可以使用，有效解決上述問題。 從現地試驗中發現基樁承受側向力後，其反應較大的地方皆集中在地表處，而樁的下半部仍然在彈性範圍內，因此彎矩函數之選取採用兩種不同的函數型式進行模擬，在反應較大的地方(上半部)，採用七次多項式函數，反之，在反應較小的地方，採用半無限長樁彈性解的基本特徵函數。 本研究進一步簡化此分析模式，將樁身上半部反應由七次多項式模擬降為五次多項式，並探討合成函數連續條件及樁材性質不同之影響。研究結果發現，無論在數值實驗或現地試驗資料分析上，使用五次多項式即可模擬不錯的結果。此外，由樁材性質的敏感度分析發現，樁材性質會影響彎矩、剪力及土壤反力分佈。因此樁材性質的正確與否，將影響本分析模式所得之樁身反應的正確性。 關鍵詞：側向樁、基樁側向載重試驗、傾度管資料、p-y曲線、回歸分析模式。

Abstract In a laterally loaded pile test, the pile deflection is usually measured by using a buried slope inclinometer. The deflection of the pile can be obtained by integrating the data of measured slopes. To deduce the internal forces of the pile and associated soil reactions on the pile, the elastic beam theory is usually applied in engineering practices. The procedure firstly adopts a smooth curve function to fit the slope or deflection data of the test pile. Then, on the basis of the fitted function, the bending moments, shear forces, and soil reactions along the pile can theoretically be deduced from the beam theory by successive differentiations. Since many differentiations may amplify measurement errors and when a pile moves into nonlinear response the pile internal forces and soil reactions obtained by differentiating the slope or deflection of the pile are incorrect, the above procedure sometimes leads poor results. To solve this problem, Chiou and Chen (2000) developed a model that can effectively reduce orders of differentiation and consider the nonlinear behavior of the pile by using a composite moment function to do regression analysis with inclinometer slope data, instead of directly using a slope function to fit the inclinometer slope data. The composite function consists of upper and lower parts: the upper part of the pile where the response exhibits large variations is modeled by a seventh-degree polynomial and the lower part of the pile where response is small is modeled by the characteristic function of an elastic solution of a semi-infinite pile In order to increase effectiveness of Chiou and Chen’s model, this study further simplifies their model by using a fifth-degree polynomial to model the moment distribution of the upper part of the pile. For verification, both the solution of finite element analyses and the results of in-situ load tests have been adopted and the results of regressive analyses can satisfactorily obtain the pile responses. In addition, on the basis of curves of pile deflection and soil reaction deduced from the results of the pile test, a set of site-specific p-y curves can thus be established. Finally, a sensitivity study has shown that accurate estimate of moment-curvature relationship of a pile is essential to the accuracy of the deduced pile internal forces, soil reactions, and p-y curves. Key Words: lateral pile, lateral pile test, inclinometer data, p-y curve, regressive model.