Energy Cascades in Driven Granular Liquids : A new Universality Class? I : Model and Symmetries

This article deals with the existence and scaling of an energy cascade in steady granular liquid flows between the scale at which the system is forced and the scale at which it dissipates energy. In particular, we examine the possible origins of a breaking of the Kolmogorov Universality class that applies to Newtonian liquids under similar conditions. In order to answer these questions, we build a generic field theory of granular liquid flows and, through a study of its symmetries, show that indeed the Kolmogorov scaling can be broken, although most of the symmetries of the Newtonian flows are preserved.

💡 Research Summary

The paper “Energy Cascades in Driven Granular Liquids: A new Universality Class? I: Model and Symmetries” investigates whether the well‑known Kolmogorov‑41 (K41) scaling of energy cascades in Newtonian turbulence survives in dense granular liquids that are driven into a steady shear flow. Granular liquids consist of macroscopic particles that lose kinetic energy through inelastic collisions, so they are intrinsically out of equilibrium and cannot be described by equilibrium statistical mechanics. The authors focus on the regime where the packing fraction φ lies between roughly 0.4 and 0.6, a dense “liquid” state that still exhibits collective eddy‑like structures similar to those in ordinary turbulence.

The study proceeds in several stages. First, a qualitative argument based on von Weizsäcker and Heisenberg’s derivation of the –5/3 energy spectrum for Newtonian fluids is reproduced. The argument relies on a hierarchy of length scales L_n, a mixing length proportional to L_n, and a viscous dissipation term η_n ∝ ρ l_n w_n. By assuming scale‑invariant velocity gradients, one recovers the classic K41 result v_n ∝ L_n^{1/3} and consequently an energy spectrum F(k) ∝ k^{−5/3}.

The authors then adapt this reasoning to granular media. The crucial difference is that dissipation is not viscous but originates from binary collisions. They replace the viscous dissipation S_n with a collision‑based term S_G n = ρ Γ_d ω_n T_n, where Γ_d is a restitution‑dependent damping rate, ω_n is a scale‑dependent collision frequency, and T_n is the granular temperature (the kinetic variance of velocity fluctuations). By relating ω_n to the inverse of the eddy turnover time at scale L_n, they obtain a new scaling relation v_n ∝ L_n^{1/2}. This leads to an energy spectrum F(k) ∝ k^{−3/2}, a clear deviation from K41. The authors point out that this exponent matches several three‑dimensional numerical studies (e.g., Saitoh’s group) and explains the spread of reported exponents (−1, −5/4, −6/5, −7/5) in two‑dimensional simulations.

To place the argument on a firm theoretical footing, the paper constructs a field‑theoretic description of granular liquids. The starting point is the Granular Integration Through Transients (GITT) formalism, which yields a general, non‑Newtonian viscosity tensor η_{ijkl}(φ, γ̇, e) that is valid for all flow regimes. This tensor is coupled to the incompressible Navier–Stokes equations, producing a set of nonlinear stochastic partial differential equations for the velocity field v(x,t) and the stress tensor σ_{ij}(x,t). From these equations an effective action S_eff