Friction in unconforming grain contacts as a mechanism for tensorial stress–strain hysteresis

Materials composed of consolidated grains and/or containing internal contacts are widespread in everyday life (e.g. rocks, geomaterials, concretes, slates, ceramics, composites, etc.). For any simulation of the elastic behavior of this class of solids, be it in seismology, in NDT, or in the modeling of building constructions, the stress–strain constitutive equations are indispensable. Since the most common loading patterns in nature considerably deviate from simple uniaxial compression, the problem of tensorial stress–strain representation arises. In simple loading cases it may be sufficient to use a phenomenological constitutive model. However, in a more general case, phenomenological approaches encounter serious difficulties due to the high number of unknown parameters and the complexity of the model itself. Simplification of the phenomenology can help only partly, since it may require artificial assumptions. For instance, is it enough just to link the volumetric stress to the volumetric strain, or do we have to include shear components as well, and if yes, in what form? We therefore propose a physical tensorial stress–strain model, based on the consideration of plane cracks with friction. To do this, we combine known relations for normal displacements of crack faces given by contact mechanics, the classical Amonton's law of dry friction for lateral displacements, and the equations of elasticity theory for a collection of non-interacting cracks with given orientation. The major advantages of this model consist in the full tensorial representation, the realistic stress–strain curves for uniaxial stress compression and quantitative comparison with experimental data, and a profound account for hysteretic memory effects.