云南南盘江大桥为7跨预应力钢构连续桥,1、2号墩为桩基,采用人工挖孔灌注桩。No.2嵌岩桩底下部12 m处为强风化白云岩。对拟定的端承区域勘探发现在桩底的岩石-混凝土接触区域有一溶洞,溶洞高约2 0 m,垂直发育,洞底有粘土、碎石土充填,结构松散,粘土为软塑状。东南大学采用与O- cell试验类似的平衡内部千斤顶系统进行了桩的荷载实验。O- cell试验安装在距桩底12 m处,即靠近桩下部的弱风化与强风化岩石之间的接触区域。本文阐述了O- cell试验数值分析中用来反算桩体/岩石交接区域特性的部分。采用的数值分析软件为BAQUS,版本6.4(ABAQUS,2 0 0 4 )。同时采用二维轴对称模型来对桩径2 .5 m、桩长49m的钻孔桩进行O- cell试验数值模拟。假设加载的钻孔桩的变形与屈服发生在混凝土桩体与岩石的交接区域,混凝土桩为线弹性,强风化白云岩与弱风化白云岩为弹-塑性,并可用Drucker- Prager模型(ABAQUS,2 0 0 4 )表示,对岩石和混凝土-岩体接触面的特性进行反算,然后用反算结果模拟桩顶荷载特征。数值模拟计算和现场载荷试验检测的载荷-位移曲线对比表明,除了非弹性和永久变形外,整个向上与向下的变形与试验检测到的值基本相同。这主要是由于接触面模型的弹性性质所决定的,即它在卸载过程中会将滑动的接触面复原;而永久变形是岩体-混凝土接触面屈服的结果。必须指出的是,由于桩的承载力相对较高(2 .5 m桩径,嵌岩长度37m) ,试验不能达到最终荷载,所以检测到的变形也较小(<10 mm)。分别在桩端、O-单元荷载盒上37m处检测向上弯沉,并将结果与O-单元荷载盒顶部的弯沉作比较。结果发现,由于桩是由O-单元荷载盒底部向上加载,所以桩端的弯沉预计比O-单元荷载盒的弯沉小约5 mm。这一差值几乎等于施加荷载状态下无限制(Δ=PL/AE)桩体的理论弹性缩短量。通过O- cell现场试验可以有效地观测加载状态下荷载弯沉反应情况和调整或校准数值分析中的材料模型,并可将之推广应用到相同地质条件下类似项目施工的其它桩的反应过程中。此外,还可以通过溶洞区域中的高承载力桩的O- cell试验结果来构建桩体摩擦反应的模型;在模型摩擦反应基础上,可以估计桩端预计的变形反应值;在交接面反应中,可以采用正弦接触面有效地引入人工粘结和表面粘结 Yunnan Nanpanjiang Bridge is a 7-span prestressed steel continuous bridge, pier No. 1 and 2 are piles, and manual digging and pouring pile is adopted. No.2 rock-socketed pile at the lower part of 12 m strong weathered dolomite. For the proposed end bearing area exploration, there is a cave in the rock-concrete contact area at the bottom of the pile. The cave is about 20 m high and vertically developed. Clay and gravel soil are filled in the cave bottom. The structure is loose and the clay is in soft plastic form. Southeast University with O-cell test similar to the balance within the jack system for pile load test. The O-cell test was installed at a contact area between weak weakened and strongly weathered rocks at a distance of 12 m from the bottom of the pile, ie near the lower part of the pile. This paper describes the part of the numerical analysis of the O-cell test that is used to reverse the characteristics of the pile / rock interface. The numerical analysis software used was BAQUS, version 6.4 (ABAQUS, 2004). At the same time, a two-dimensional axisymmetric model was used to simulate the O-cell test of a pile with a pile diameter of 2. 5 m and a pile length of 49 m. It is assumed that the deformation and yielding of the loaded bored piles occur in the interface area between the concrete piles and the rock. The concrete piles are of linear elasticity. The strong-weathered dolomite and the weakly weathered dolomite are elastic-plastic and can be identified by the Drucker-Prager model (ABAQUS, 2 0 0 4) shows that the characteristics of rock and concrete-rock contact surface are inversely calculated, and then the back-calculation results are used to simulate the load-bearing characteristics of the pile top. Comparisons of load-displacement curves measured by numerical simulations and on-site load tests show that, except for inelastic and permanent deformations, the whole upward and downward deformations are basically the same as those measured by the test. This is mainly due to the elastic nature of the contact surface model, which means that it will restore the sliding contact surface during unloading; and the permanent deformation is the result of rock-concrete contact surface yielding. It must be pointed out that due to the relatively high carrying capacity of the pile (2.5 m pile diameter, sock length 37 m), the test can not reach the final load, so the detected deformation is also smaller (<10 mm). The upward deflection was measured at 37 m at the pile end, O-unit load box and the results were compared with the deflection at the top of the O-unit load box. As a result, it was found that due to upward loading of the pile from the bottom of the O-cell load box, the deflection of the pile tip is expected to be about 5 mm less than the deflection of the O-cell load box. This difference is almost equal to the theoretical elastic shortening of the pile without load (Δ = PL / AE) under applied load. The O-cell field test can effectively observe the load deflection response under load and adjust or calibrate the material model in numerical analysis, and can be extended to the reaction process of other piles under similar geological conditions under the same geological conditions . In addition, the model of pile-body friction can be constructed by the O-cell test results of high-bearing piles in the cave area. Based on the model frictional response, the predicted deformation response of the pile-end can be estimated. In the interface reaction , You can use the sinusoidal contact surface effectively introduce artificial bonding and surface bonding