[Paper Review] The Breakdown of Kinetic Theory in Granular Shear Flows

Gregg Lois, Anaël Lemaı̂tre+1 more|arXiv (Cornell University)|Jul 13, 2005

Granular flow and fluidized beds18 citations

TL;DR

This paper demonstrates that kinetic theory breaks down in dense, inelastic granular shear flows due to the emergence of spatially correlated clusters of grains, invalidating the binary collision and molecular chaos assumptions. Using Contact Dynamics simulations, the authors quantify cluster size via force-force correlations, showing that cluster formation—driven by low restitution coefficients and high packing fractions—renders standard kinetic theory inadequate for dense granular flows.

ABSTRACT

We examine two basic assumptions of kinetic theory-- binary collisions and molecular chaos-- using numerical simulations of sheared granular materials. We investigate a wide range of densities and restitution coefficients and demonstrate that kinetic theory breaks down at large density and small restitution coefficients. In the regimes where kinetic theory fails, there is an associated emergence of clusters of spatially correlated grains.

Motivation & Objective

• To test the fundamental assumptions of kinetic theory—binary collisions and molecular chaos—in dense granular shear flows.
• To identify the regimes where kinetic theory breaks down in granular materials under varying density and inelasticity.
• To quantify the emergence of spatially correlated grain clusters that invalidate the binary interaction assumption.
• To provide a quantitative measure of cluster size using force-force correlation functions to assess the breakdown of kinetic theory.
• To determine the phase space regions where kinetic theory remains valid versus where it fails due to persistent contacts and collective behavior.

Proposed method

• Employed two-dimensional Contact Dynamics (CD) simulations with Lees-Edwards boundary conditions to model steady-state granular shear flow.
• Used spatial force-force correlation functions $ C(\ell) = \langle \vec{F}(0) \cdot \vec{F}(\ell) \rangle $ to detect spatial correlations in grain forces.
• Defined average cluster size $ N_c $ via the correlation length $ \langle \ell \rangle $ from $ C(\ell) $, normalized so $ N_c = 2 $ for dilute, nearly elastic conditions.
• Excluded pre-collisional pairs from force correlation averages to ensure pre-collisional correlations were measured.
• Mapped $ N_c $ as a function of packing fraction $ \nu $ and restitution coefficient $ e $, producing contour plots to visualize breakdown regions.
• Compared collisional stress from kinetic theory with static stress to assess validity of the binary collision assumption.

Experimental results

Research questions

• RQ1 At what combinations of packing fraction and restitution coefficient does the binary collision assumption in kinetic theory break down in granular shear flows?
• RQ2 How do spatial correlations in grain forces indicate the emergence of persistent clusters in dense granular systems?
• RQ3 To what extent do pre-collisional velocity correlations invalidate the molecular chaos assumption in inelastic granular materials?
• RQ4 How does the cluster size $ N_c $, derived from force-force correlations, quantify the failure of kinetic theory in dense regimes?
• RQ5 In which phase space regions (defined by $ \nu $ and $ e $) is kinetic theory still applicable for granular shear flows?

Key findings

• The binary collision assumption fails at high packing fractions and low restitution coefficients, where force-force correlations reveal the formation of spatially correlated clusters.
• Cluster size $ N_c $ increases with decreasing restitution coefficient and increasing packing fraction, with $ N_c \to \infty $ near the jamming transition.
• The average cluster size $ N_c $, derived from the correlation length $ \langle \ell \rangle $, shows exponential decay of force correlations at large distances.
• For $ e = 0.92 $ and low density, $ N_c \approx 2 $, confirming the binary collision regime; for $ e = 0 $ and $ \nu = 0.81 $, $ N_c \gg 2 $, indicating strong clustering.
• The breakdown of the binary collision assumption correlates with the divergence of pressure and shear stress ratios near jamming, confirming its physical significance.
• The molecular chaos assumption also fails for small $ e $, due to pre-collisional velocity correlations, but the binary collision failure is more fundamental and harder to correct within standard kinetic theory.