Numerical Analysis of Drained Direct and Simple Shear Tests

Abstract

Drained direct and simple shear tests are analyzed assuming an isotropic elastoplastic constitutive law and plane strain conditions. The stress and strain distributions within the sample are examined. The numerical analyses are interpreted as if they were actual tests, using the computed stresses and displacements on the top boundary of the sample. Stiffness and strength results for the direct shear are then compared with simple shear results based on computations over the whole top boundary (average) and over the middle third (core). With the ideal simple shear as a reference state, simple shear results based on average measurements underestimate initial stiffness and peak strength by as much as 20% while core measurements give a better but not perfect agreement. Direct shear tests overestimated the initial stiffness and in most cases the peak strength by as much as 7.5%. Stress and strain distribution in the simple shear test become highly nonuniform as peak conditions are approached. The findings agree in general with some reported experimental observations. Analyses where strain softening was modeled did not develop any significant progressive failure, despite the nonuniform stresses and strains.