| dc.description.abstract | The strain-softening phenomenon has been extensively studied the last decades in context of numerical modelling and progressive failure mechanism. In this study, a vane is modelled in soft clay with a negative dilatancy. The vane is applied several deformation rates which results in different strain-softening patterns. It is of interest to investigate the interactive connection between shear band formation and local drainage mechanics
in FEM for different mesh qualities. A setup of four different mesh qualities are kept consistent for all cases. Coarse, medium, fine and very fine mesh refinements of 326, 520, 1400, 4186 elements are used. Consolidation analysis are carried out in Plaxis 2d with different time-steps. The shear band thickness is an important parameter as it explains
the strain-level and consolidation rate. Plaxis displays issues regarding objective shear band thicknesses and leaves these as undetermined values. Consequently, results are sensitive to the element size of the mesh. Hence, mesh dependency must be achieved to get objective results in terms of shear band formations and peak-strength. Using the conventional Hardening soil model in Plaxis 2d resulted in increasing peak-values for slower deformation
rates. Hardening soil model appeared rate-dependent and insufficient for the aim of this thesis. Afterwards, a ratedependent user designed soil model called Unified enhanced soft soil creep model were used to obtain adequate results. Results indicated rate-dependency as faster rates resulted in higher peak-values with steeper decrease to a lower residual strength. It was observed zones of suction behind each vane blade with maximum values of 60-80 kPa indicating a lack of tension cut-off criterion for the user-designed soil model.
It is concluded in this study that both Hardening soil and Unified enhanced soft soil creep model could not generate mesh-dependent results for a soft contractive clay. | |
