With the rapid development of deep underground infrastructure, such as deep tunnels, high-rise building foundations, and underground energy storage facilities, soils are increasingly subjected to high confining pressures.High confining pressure not only significantly alters the strength, stiffness, and volumetric deformation characteristics of granular soils, but may also induce particle breakage, gradation reconfiguration, and pore-structure evolution, causing their macroscopic mechanical response to differ from that under conventional confining pressure conditions [1].Consequently, in deep engineering applications, reliable characterization of the constitutive behavior of granular soils is critical to stability analyses and deformation predictions.
However, most existing constitutive models for granular soils have been developed primarily for conventional confining pressures [2-5]. This often leads to prediction errors when these models are applied under high confining pressure conditions, because they may not adequately reflect the pressure dependence of the hardening law. In recent years, various enhancements have been proposed to better represent the mechanical behavior of granular soils under high confinement. For example, Russell and Khalili [6] formulated a critical-state relation composed of three linear segments and, on this basis, developed a corresponding bounding-surface constitutive model; Liu et al. [7] found that particle breakage would cause CSL to move according to experiments, so that they developed a CSL expression considering particle breakage and formulated an elastoplastic constitutive model of coral sand. Xiao et al. [8] incorporated a particle breakage parameter into the plastic modulus and developed a constitutive model for sands applicable to high-pressure conditions. Although the above models can predict the stressstrain behavior of sands over a wide range of stresses, the large number of parameters and the difficulty in determining them limit their practical application.
To this end, this study proposes a simplified S-shaped normal compression line (NCL) to avoid predicting a negative void ratio. Subsequently, within the framework of the Modified Cam Clay (MCC) model, two parameters are introduced to correct the geometry of the yield surface. On this basis, an elastoplastic constitutive model for granular soils that is applicable across different pressure levels is established and validated against experimental data.
International Conference on Advances and Innovations in Soft Soil Engineering (ICAISSE2026)
Constitutive, Numerical, and Machine Learning Modelling