Four Lenses Toward Principle-Based Understanding of Granular Segregation in Dense Flows: Particle, Continuum, Statistical, and Conceptual Approaches

Abstract

This review examines the development of concepts and models describing the spontaneous spatial separation—segregation—of dense flowing granular mixtures. Although particle-scale interactions are largely understood and modern discrete element methods can resolve system dynamics in detail, a central challenge remains: prediction has advanced more rapidly than understanding. Simulation can reveal what happens; it does not, by itself, explain why. We argue that progress requires integrating four complementary approaches: particle-based simulations, continuum descriptions based on continuity, statistical methods extending kinetic theory to dissipative systems, and conceptual models that isolate dominant mechanisms. These are not competing paradigms but distinct lenses on a multi-scale phenomenon. Across them, common mathematical structures recur, including Newtonian dynamics at the particle level and flux--divergence formulations governing transport. A central thesis is that conceptual models remain indispensable. By distilling mechanisms such as kinetic sieving, squeeze expulsion, and force balance, they provide explanatory structure that complements both simulation and formal theory. Recent advances applying particle-level force models of segregation illustrate the value of combining these perspectives. At the same time, each framework encounters limitations. The central challenge is therefore to identify organizing principles that connect mechanisms across scales. Granular segregation serves as a testbed for a broader question: how understanding emerges in systems where the microscopic governing laws are known, yet collective behavior remains difficult to explain.