Size-dependent finite strain analysis of cavity expansion in frictional materials

This paper presents unified solutions for elastic-plastic expansion analysis of a cylindrical or spherical cavity in an infinite medium, adopting a flow theory of strain gradient plasticity. Previous cavity expansion analyses incorporating strain gradient effects have mostly focused on explaining the strain localization phenomenon and/or size effects during infinitesimal expansions. This paper is however concerned with the size-dependent behaviour of a cavity during finite quasi-static expansions. To account for the non-local influence of underlying microstructures to the macroscopic behaviour of granular materials, the conventional Mohr-Coulomb yield criterion is modified by including a second-order strain gradient. Thus the quasi-static cavity expansion problem is converted into a second-order ordinary differential equation system. In the continuous cavity expansion analysis, the resulting governing equations are solved numerically with Cauchy boundary conditions by simple iterations. Furthermore, a simplified method without iterations is proposed for calculating the size-dependent limit pressure of a cavity expanding to a given final radius. By neglecting the elastic strain increments in the plastic zone, approximate analytical size dependent solutions are also derived. It is shown that the strain gradient effect mainly concentrates in a close vicinity of the inner cavity. Evident size-strengthening effects associated with the sand particle size and the cavity radius in the localized deformation zone is captured by the newly developed solutions presented in this paper. The strain gradient effect will vanish when the intrinsic material length is negligible compared to the instantaneous cavity size, and then the conventional elastic perfectly-plastic solutions can be recovered exactly. The present solutions can provide a theoretical method for modeling the size effect that is often observed in small-sized sand-structure interaction problems. (C) 2018 Elsevier Ltd. All rights reserved.