Experimental and finite element analysis of nonlinear moment-rotation behavior of rocking shallow foundations on stiff clay

Abstract

This thesis presents a better understanding of nonlinear moment-rotation behavior of shallow foundations resting on stiff clay based on the results of 3D finite element analysis and large scale field experiments. In past decades several researchers have carried out both numerical modelling and experimental studies to understand the rocking behavior and earthquake response of shallow foundations. Algie (2011) performed a series of snap-back tests to study the moment-rotation response of shallow foundations on stiff residual clay. The results showed that snap-back testing is an effective tool for obtaining insight into the nonlinear behavior and earthquake response of shallow foundations. Tests also gave a good assessment of magnitude of damping of rocking shallow foundations resting on stiff clay. Another advantage was that the static moment-rotation curves are obtained during the initial pull-back phase of the test and it turns out that the static response provides considerable insight into the dynamic behavior of the foundation. This first half of the thesis comprises of development of a 3D finite element model in PLAXIS 3D, its validation with available experimental data and the analysis results. The moment-rotation behavior predicted from PLAXIS 3D converged with the experimental data and showed considerable nonlinearity once the footing uplift and partial loss of contact occurs. The contact pressure distribution under the footing showed that the footing uplift and nonlinearity is initiated when the applied moment exceeds around 0.45Mult. The analysis results also showed that direction of the applied moment is important as footing exhibited considerably different rotational stiffness about short and long axis. The analyses also showed that effect of footing L/B ratio, static initial factor of safety and the undrained shear strength of the soil considerably affect the non-linear moment-rotation behavior. The evaluation of actual footing contact area during rocking is critical for assessment of seismic performance of the rocking shallow foundations and it depends on the ultimate moment capacity of the footing. Based on numerical modelling results, a new modified hyperbolic equation was proposed to evaluate nonlinear moment-rotation response of footings on stiff clay incorporating the effects of L/B ratio, direction of the applied moment and undrained shear strength. The remaining half of the thesis deals with field experiments on surface shallow foundations resting on stiff residual clay. The experiments were performed on a site located in Silverdale, Auckland. The field experiments included slow-cyclic and snap-back tests on two different sets of footing configurations to simulate static non-linear moment-rotation response and footing rocking. For this study snap-backs were carried out in alternating directions, thus simulating the back and forth motion experienced during an earthquake. The test rig had two parallel foundation strips; in one case the longitudinal axis of the footings was in the plane of the rocking, in the other the footings were oriented perpendicular to the plane of the rocking and placed at the outer edges of the rig. This means that results are obtained for rocking in which there is a gradual uplifting of the foundation strip (foundation longitudinal axis in the plane of rocking), and when the entire footing lifts clear of the underlying soil (foundation strips at the outer edges of the rig). The purpose here was to investigate the validity of the damping mechanism proposed by Housner in his paper on inverted pendulum structures gave the best representation of the foundation damping; so the tests duplicate the partial uplifting of foundations and also foundations that lift completely clear of the underlying soil. The moment-rotation curves from the tests showed that the small strain stiffness is only around 25% of the theoretical stiffness. Also, it was observed the for the case of snap-back tests at higher vertical loads, the damping ratio was generally higher around 20% due to large footing rotation and partial detachment with underlying soil at the time of snap. The direction of rocking and footing L/B ratio also affect the magnitude of damping. Lastly the conclusions from both the finite element analysis and experimental studies are presented in detail. The need of further research and future scope of work is outlined.