Modified Duncan-Chang Constitutive Model for Modeling Supported Excavations in Granular Soils

Abstract

Numerical analyses have been widely used to investigate the stability of deep excavations. The results of this type of analyses rely on the constitutive model that is adopted for the analyses. Although many advanced constitutive models have been developed for the simulation of soil behavior, they are often complex, expensive to use, and not well understood by practicing engineers. Hence, simple constitutive models, such as the Mohr–Coulomb model, are widely used in geotechnical engineering practice. However, these simple models, fail to represent some key features of soil behavior. Of particular importance in granular soils is the effect of changes in soil density on the soil behavior, an aspect that is often overlooked when simple constitutive models are used. In this study, a series of triaxial tests will be conducted on reconstructed gravelly sand to investigate the effect of the relative density and confining pressure on strength and stiffness characteristics of the soil. According to the test results, variations of the soil stiffness and peak deviatoric stress with relative density and confining pressure will be established. A novel disturbance function will then be proposed based on the disturbance state concept to capture the effect of soil density on its behavior. Subsequently, a constitutive model that is based on the original Duncan and Chang model will be developed that considers the effects of the confining pressure and relative density of the soil using the proposed disturbance function. The model is implemented in ABAQUS software and is verified by comparing the numerical results with experimental data. The application of the model is presented through the simulation of a deep supported excavation. The numerical results show that the model successfully simulates horizontal displacements of the support piles, with errors of <5% for the monitoring points. In particular, it is highlighted that the model is superior to the widely used Mohr–Coulomb model in the simulation of excavation problems, and yielded a safer design than that obtained using the Mohr–Coulomb model.