Simulating Analysis Of Mechanical Properties On Soil-Pile Interface

The bearing capacity of pile foundation consists of two parts: lateral frictionresistance and end bearing capacity. As for the lateral friction resistance, mostof the design works are based on the assumption of a linear relationship betweenmaximum shear stress and normal stress at the interface of pile and soil, whichworks well in the field. But the intrinsic transfer mechanism of stress and strainnear the interface of soil and pile has been seldom discussed. The shear strengthproperty of the soil-concrete interface should largely depend on the stress andstrain distribution. By using the field materials from Zhaolong SonghuajiangBridge Project, medium-size laboratory shear tests were conducted, and a numericalmodel has been built to achieve the full-scale data collection. This research willhelp to have a better understanding of the shear behaviors of the soil-concreteinterface.Firstly, medium-size shear tests are performed using the materials fromconstruction site. The soil samples are classified into three groups: clay, coarsesand and fine sand. The normal stress values are determined from the designdocuments. The experimental results show that, the peak shear resistance ofsoil-concrete interface has a good linear relationship with the normal stress.As for clay, the results also show an existence of cohesion. So it is suitableto utilize the Mohr-Coulomb Theory of soil to describe the shear behavior ofsoil-concrete interface. It is difficult to collect enough data from the sheartest sample due to its size and cost. Numerical method was introduced here. Inverseanalysis was conducted to build a similar model as the shear test samples, andthe stress and strain were discussed later.Particle Flow Code in3-Dimension were used in this study. As a kind of discreteelement methods, the particle flow model represents the normal, shear and slidinginteraction among sand particles well. The constitutive relations are describedin the particle level, which means it represents materials in a new version comparedwith finite element method.In a discrete element model, the mechanical properties are determined by the microparameters of particles. So the selection of particle parameters plays an important role in the model built process. In Chapter Four, orthogonal table were used todesign the four-factor four-level triaxial tests. The four considering factorsare porosity, friction coefficient, normal stiffness and shear-normal stiffnessratio. The target parameters are internal friction angle of sand, Poisson ratioand elastic modulus. The results indicate that: the internal friction of sand isdetermined by the friction coefficient and shear-normal stiffness ratio; theelastic modulus of sand is largely influenced by the normal stiffness andshear-normal stiffness ratio; the Poisson ratio of sand is determined by theshear-normal stiffness ratio and porosity.Then, by implementing the micro parameters fixed from Chapter Four, a medium-sizeshear test model was built in PFC3D, and shear tests were conducted under differentnormal stresses. The results of numerical models match with experimental data well,indicating the rationality of this discrete element method. By comparing with thestress, displacement and porosity distribution among particles, the transferringmechanical regularities of sand model under shear test are discussed. Supportvector machine method was used to draw the boundary of shear band, which has apretty good result. Finally the maximum thicknesses of shear band were comparedwith different normal stresses and particle sizes. Sand samples consisted of largeparticles, or bearing small normal stress, likely have a larger shear bandthickness.This research will contribute to the understanding of soil-concrete shear failuremechanism.