THREE-DIMENSIONAL FINITE ELEMENT MODELLING ON PILESUPPORTED EMBANKMENT WITH GEO-GRID REINFORCEMENT by Lao Jun-Yuan Thesis Supervisor: Prof. Wan-Huan Zhou Geotechnical and Structural Engineering ABSTRACT Pile-supported geogrid-reinforced (PSGR) embankment has been widely used for highway and high-speed railway construction on soft soil. This thesis presents a comprehensive numerical study using the finite element (FE) method in order to investigate the performance of PSGR embankments and the development of soil arching. A three-dimensional (3D) FE model was established using a software ABAQUS, and on the basis of a case study of a full-scale high-speed railway embankment experiment constructed in Zhejiang University of China. The simulation was carried out according to the field conditions in three experimental stages: (a) the construction of embankment fill in steps, (b) static loading procedure on the surface of embankment, and (c) the drainage of water bags under control. The birth-to-death element technique and the modulus variation controlled by the filed variables were used, to achieve the modelling on embankment constructing and draining process of water bags, respectively. The parameters of all materials used in the FE model were carefully selected and calibrated from the laboratory test results. Then, the numerical model verification was carried out by comparing the variations in earth pressure on both the IV pile caps and subsoil with the measured data. A good agreement can be obtained between the computed and measured data, in terms of the vertical stresses on the pile caps and in the soils at different experimental stages. Based on the present 3D FE model, some characteristics of soil arching were investigated, including the evolution of the arching foot position, arching height and shape at different experimental stages. After applying the static loading on the surface of embankment, the growth rate of the center of the pile cap in vertical stress was obviously higher than that at the edge and corner of the pile cap. On the other hand, it is found that when the effect of soil arching reaches its full potential, the subsoil settlements has little effect on the arching height. The shape and size of the soil arching at the shoulder of the embankment is similar as that in the middle. The height of arching increases with an increase in loading and that variations occurs only in the middle part of the embankment. Finally, group behavior of the soil arching within the PSGR embankment system was investigated. During the construction stage of embankment fill, the results of soil arching ratio (SAR) show that the development of soil arching in the middle of embankment is the same as that in the embankment shoulder. However, after the completion of the construction, the load transfer efficiency in the middle region of embankment is superior to that in the shoulder from the results of stress concentration ratio (SCR) and the surface distribution of vertical stress. During the static loading test on the top surface of the embankment, the stresses in the embankment fill will be redistributed, mostly in the middle of the embankment. It was also found that the effect of the static loading on the stress distribution and arching in the embankment fill are irreversible, even after the unloading stage. With the simulation of different stages of water drainage testing, the development of soil arching at different locations can be clearly shown in FEM modeling. The subsoil settlement within the adjacent pile caps have great influence on the above soil arching, and little influence on the adjacent arching. The soil arching height is further developed and reaches limit during the water drainage test.