The Behaviour of Shallow Foundations on Silica Sand Subjected to Inclined Load

Abstract

In this thesis, the behaviour of shallow circular foundations on silica sand under inclined loads is investigated by performing experimental model-scale tests and numerical and semianalytical analyses. A series of inclined loading tests on circular footings on dense, medium and loose Sydney sand, under three different surcharge pressures (OkPa, 25kPa, and 50kPa), is conducted using a displacement controlled apparatus. The design of the loading system enables the model footing to rotate and move freely in vertical and horizontal directions, and avoids any kinematic restraints from the loading rod, which is used to apply load on the footing. The testing system designed in this research is intended to apply axially inclined loads on the footing. However, as the footing rotates about the tip of the loading rod, load eccentricity occurs and this results in moments being applied to the footing, in addition to the applied inclined load. All the classical bearing capacity theories assume the soil to obey the Mohr-Coulomb failure criterion, using a single value of friction angle. However, the peak friction angle of most granular soils decreases with increasing mean normal stress, owning to the stress dependent nature of dilatancy. In addition, because of the variation of stress level, the soil beneath the footing possesses a significant variation of mobilised frictional strength. This leads to uncertainty as to what value is to be used in the conventional bearing capacity solutions, and hence to the uncertainty in reliability of the estimated bearing capacity. In this research, a modified conventional method proposed by Perkins and Madson (2000), which accounts indirectly for the non-linear strength behaviour of sand through the use of Bolton’s (1986) strength-dilatancy equations, is used to predict the vertical test results for maximum load. The method eliminates the need for triaxial data to define peak strength friction angles and requires knowledge of soil parameters consisting only of the relative density, unit weight and constant volume friction angle. Overall the approach of Perkins and Madson (2000) is shown to give good predictions for the vertical test results on dense sand. However, this approach is shown to have some uncertainties when applied to medium and loose sands due to the compressible nature of these sands. A semi-analytical finite element model incorporating the Mohr-Coulomb failure criterion is used to simulate the behaviour of the experimental model-scale circular footings on sand subjected to inclined load. It is demonstrated that this model, together with a non-associated flow rule for the sand, can provide a reasonable match between the numerical predictions and the experimental observations in most cases. The degree to which the experimental results can be matched depends to a significant degree on the choice of input parameters for the Mohr-Coulomb model. The choice of a non-associated flow rule is particularly important when assessing the ultimate bearing resistance. In order to achieve good agreement, a non-associated flow rule is essential, but the high non-associativity required to match experimental data is often accompanied by instability in the numerical solution. The model predictions indicate significant dependence of the maximum bearing resistance on the dilation angle of the soil, and this becomes much more significant as the applied load becomes inclined. One clear advantage of the finite element method over conventional bearing capacity theory is that it provides a good indication of the dominant deformation mechanisms. Numerical analyses corresponding to the model tests are also performed using the strain hardening system-level plasticity model, known as “Model C”, developed by Houlsby and Cassidy (2002). In general, Model C gives realistic predictions for the experimental data of the footing tests. The experimental validation of Model C from this research confirms the suitability of the model for predicting the behaviour of circular footings on sands which are subjected to both vertical and inclined loading.