Tunable discrete fracture network for dynamic analyses of rock landslides by material point method

Abstract

The discrete fracture network (DFN) significantly influences the failure mechanisms of rock slopes. However, the integration of DFN within the hybrid mesh-particle material point method (MPM) remains ambiguous when juxtaposed with the frictional contact interfaces inserted in mesh-based methods and the degradation of contact bonds in particle-based approaches. This research introduces a tunable DFN adaptable to the MPM, employing hybrid congruence and normal probability algorithms to generate rock fractures with specific inclination angles and trace lengths. These fractures are then superimposed onto the computational domain of material points by image processing techniques, and the mechanical properties of fractures are assigned to the corresponding material points. The developed method effectively captures the critical features of rockslide, and the newly proposed parameter for intersection patterns of rock fractures allows for the examination of intricate slope failure modes, including slide-buckling-toppling, sliding-secondary toppling, and toppling-circular slope failure. This research further presents a comprehensive probability analysis of jointed slopes, where the mean sliding surface and deposit configuration can offer valuable insights for site characterization and risk assessment of rock slope engineering. This research contributes to a more nuanced understanding of complex interactions within rock slopes and enhances the predictive capabilities of slope stability models.