A multi-yield surface model in reference state soil mechanics for cohesionless soils and liquefaction problems.

摘要

Liquefaction of cohesionless soils is a phenomenon that has caused a number of catastrophic events in the past. The term "liquefaction" has been used by different researchers to define both flow liquefaction in strain-softening soils and cyclic liquefaction caused during cyclic loading with shear reversal. During the last decades, different approaches have been proposed to analyze these catastrophic phenomena. Empirical criteria based on field observations, total stress analysis and effective stress analysis are the three main approaches in this area. It is believed that effective stress analysis, using constitutive models, is the most powerful tool to determine permanent deformations during and after liquefaction. None of the elements in the effective stress approach is as important as the constitutive model.; The state of the soil has a significant effect on its behavior. In most existing constitutive models for cohesionless soils, this important factor has been overlooked. As a result, these models can be used only for a small range of stresses and void ratios for which the model has been calibrated. In this study, after showing the limitation of models with no reference, a framework, termed "Reference State Soil Mechanics," is proposed. In this framework the multi-yield surface theory is extended to embrace "Reference State Soil Mechanics". The modified model is capable of predicting drained and undrained responses of loose and dense sands under both monotonic and cyclic loading with a unique set of parameters. At large strains, this model shows the same ultimate state condition for both loose and dense sands. The model can also represent the effects of initial cross-anisotropy and induced anisotropy. Two different hardening rules have been used and compared with each other for specimens at different states. Validation of the model has been achieved using comprehensive comparisons of the model predictions with experimental observations on different sands with various physical properties under both monotonic and cyclic loading. It is shown that this model is capable of handling both flow liquefaction and cyclic liquefaction in a unique framework.