Cyclic freezing-thawing induced rock strength degradation, crack evolution, heave and settlement accounted for by a DEM model

A general discrete element method-based model is proposed to investigate the cyclic freezing-thawing induced damage of fractured rocks. A set of cracks is defined as an assembly of interparticle pores that are interconnected due to self-definition (pre-existing) or contact breaking (external force induced). During the freezing-thawing simulation, only the rock grains surrounding the cracks are subjected to frost heave pressure, which are expected to cause contact breaking and crack propagation. In this sense, most of the pores that do not form cracks in the numerical model will not participate in the calculation, so the computational cost and time are substantially reduced. The cracks connected to the atmosphere, i.e. the model boundaries, are classified as inactive cracks and their frost heave pressures are removed accordingly. Therefore, model convergence is recognized when all cracks are converted to inactive. Parametric analysis is further performed to show the influence of some important input parameters, including the unfrozen water content, calibration factor of frost heave pressure, crack density and dip. Finally, the freezing-thawing damage of rocks is quantified by the reduction in uniaxial compressive strength after a prescribed number of freezing-thawing cycles. The results are validated with an analytical solution. It is shown that the rock strength degradation, crack evolution, heave and settlement can be well simulated by the proposed model. This methodology can be extended to large-scale problems such as glacier slope instability or landslides.