THERMODYNAMICALLY CONSISTENT GRADIENT PLASTICITY THEORY FOR POROUS MEDIA

GUILLERMO ETSE; JAVIER L. MROGINSKI; SONIA M. VRECH

Columbus, Ohio

Congreso; 10th US National Congress on Computational Mechanics (USNCCM-X); 2009

United States Association for Computational Mechanics

The simulation of deformation behavior of porous media like soils requires appropriate constitutive models to objectively describe their complex strength degradation processes under monotonic type of loading. This complexity is mostly related to the strong dependence of the softening behavior in the post-peak regime on the boundary conditions in terms of the govern stress and hydraulic conditions. Precisely, the govern stress and hydraulic conditions represented by the suction play both a fundamental role in the failure mode, i.e. diffuse or localized, that develop in porous media like soils when subjected to monotonic loading. With other words, the failure behavior of porous media may vary from brittle to ductile depending on the acting stress and suction. In the framework of FE evaluations it is further required that failure predictions of porous media under monotonic loading are objective with respect to the size and orientation of the considered special discretization. In this work a theormodynamically consistent gradient plasticity theory for porous media like cohesive-frictional soils is proposed. The non-local softening formulation of the proposed constitutive model incorporates the dependence of the gradient characteristic length on the governing stress and hydraulic conditions to realistically predict the size of the maximum energy dissipation zone. Firstly the basis of the thermodynamically consistent gradient plasticity theory is discussed. Then this theory is extended for the case of porous media like soils and particularly for the formulation of the non-local softening law. The localization tensor of the proposed non-local theory for porous media is obtained and the conditions for localized type of failure in uniaxial and triaxial compression tests are numerically evaluated.