Abstract Geological materials have inherent uncertainties from variability in particle properties, such as inter‐particle friction. Traditional deterministic models overlook this variability, leading to inaccurate predictions of geophysical flow behavior. We developed a novel 3D discrete element method framework incorporating stochastic field theory, capturing granular materials' random flow behavior. Combined with ring shear tests and PFC, we also propose a virtual ring shear test. The effect of coefficient of variation (COV) and correlation distance of inter‐particle friction coefficient μ on particle mechanical properties is discussed. Monte Carlo simulations show ignoring μ uncertainty does not affect its strain softening and strengthening characteristics, but may overestimate shear strength. Under quasi‐static shear, shear stress‐shear displacement curves of heterogeneous samples with nearly identical particle arrangement but varied properties formed curve clusters of a certain width. The cluster width positively correlates with both and COV. Additionally, μ heterogeneity does not affect microscopic force chain's overall evolution trend, but significantly impacts its characteristic peak value. Specifically, as the average sliding friction coefficient decreases, the mean contact force reduces while the mean coordination number increases. Furthermore, we propose shear strength reduction coefficients for different confidence levels. These findings provide valuable insights into granular soil mechanical behavior during landslide evolution and hazard assessment.
Huang et al. (2026) studied this question.