Cyclic loading tends to affect the strength and stiffness parameters of soils, degrade its structure and results in accumulation of the excess pore water pressure. Such behaviour often leads to premature failure of the soils. Over the past few decades, geotechnical research community has developed numerous constitutive models to predict the behaviour of soils with variable degrees of success. The constitutive models based on the concept of bounding surface plasticity have gained much attention owing to the simplicity in describing the development of stiffness. This thesis analyzes the performance of bounding surface SANICALY model in reproducing the stress path and stress-strain behaviour of kaolin clay under undrained cyclic loading conditions .
A driver was developed in MATLAB for the chosen constitutive model to simulate undrained triaxial loading conditions. The performance of the driver was verified against the data published from literature. Further, sensitivity analysis was carried out on chosen model parameters. This was followed by validating the model with the experimental data on kaolin clay. Particularly, model performance was examined with varying initial conditions such as change of over-consolidation ratio, change of initial anisotropy, variation of initial pressure and strain controlled loading. The obtained results from sensitivity analysis have shown to increase the strength and stiffness response of the model with increase in model parameters such as the rate of evolution of the anisotropy, bound for evolution of the anisotropy and change of initial stress-induced anisotropy. In the context of calibrating the model parameters against the experimental data, it was initially noticed that the experimental stress path in monotonic loading was not being reproduced by the model with various combinations of the initial parameters. Owing to such performance the model was subsequently assessed qualitatively. When the model is subjected to different initial loading conditions, certain aspects of the experimental behaviour were qualitatively captured by the model. These include faster rate of accumulation of pore water pressure with increase in the amplitude of cyclic loading, reduction in the rate of development of strains with increase in OCR values, increase in the hysteretic damping with increase in the amplitude of strains. However, with change of OCR there were differences in the development of stress path. Also contrasting results were observed with regard to the development of the stress-strain response with change of amplitude of cyclic loading and initial pressure. Analyzing the model formulations revealed that the chosen model did not take into consideration the fabric anisotropy and hence it explains the deviation of the stress path from the experimental stress path in monotonic conditions. In the chosen model, during the process of cyclic loading, stagnation in the evolution of stress path is observed whenever the plastic volumetric strains stop evolving. It is recommended to incorporate plastic deviatoric strains in the evolution of the bounding surface in order to stimulate the further development of stress path even when the plastic volumetric strains stop evolving. It is also suggested to validate the model against different clays since this thesis focused only on Kaolin clay.