Searcharxiv⌕ Search

arXiv subjects

Gregory L. Eyink

Publications and source records attributed to Gregory L. Eyink.

At least 73 records · Page 4Linked to original sources

Multi-Scale Gradient Expansion of the Turbulent Stress Tensor

We develop an expansion of the turbulent stress tensor into a double series of contributions from different scales of motion and different orders of space-derivatives of velocity, a Multi-Scale Gradient (MSG) expansion. The expansion is proved to converge to the exact stress, as a consequence of the locality of cascade both in scale and in space. Simple estimates show, however, that the convergence rate may be slow for the expansion in spatial gradients of very small scales. Therefore, we develop an approximate expansion, based upon an assumption that similar or `coherent' contributions to turbulent stress are obtained from disjoint subgrid regions. This Coherent-Subregions Approximation (CSA) yields an MSG expansion that can be proved to converge rapidly at all scales and is hopefully still reasonably accurate. As an application, we consider the cascades of energy and helicity in three-dimensional turbulence. To first order in velocity-gradients, the stress has three contributions: a tensile stress along principal directions of strain, a contractile stress along vortex lines, and a shear stress proportional to `skew-strain.' While vortex-stretching plays the major role in energy cascade, there is a second, less scale-local contribution from `skew-strain'. For helicity cascade the situation is reversed, and it arises scale-locally from `skew-strain' while the stress along vortex-lines gives a secondary, less scale-local contribution. These conclusions are illustrated with simple exact solutions of 3D Euler equations. In the first, energy cascade occurs by Taylor's mechanism of stretching and spin-up of small-scale vortices due to large-scale strain. In the second, helicity cascade occurs by `twisting' of small-scale vortex filaments due to a large-scale screw.

nlin.CD↗

Recovering Isotropic Statistics in Turbulence Simulations: The Kolmogorov 4/5th-Law

One of the main benchmarks in direct numerical simulations of three-dimensional turbulence is the Kolmogorov 1941 prediction for third-order structure functions with homogeneous and isotropic statistics in the infinite-Reynolds number limit. Previous DNS techniques to obtain isotropic statistics have relied on time-averaging structure functions in a few directions over many eddy turnover times, using forcing schemes carefully constructed to generate isotropic data. Motivated by recent theoretical work which removes isotropy requirements by spherically averaging structure functions over all directions, we will present results which supplement long-time averaging by angle-averaging over up to 73 directions from a single flow snapshot. The directions are among those natural to a square computational grid, and are weighted to approximate the spherical average. The averaging process is cheap, and for the Kolmogorov 1941 4/5ths law, reasonable results can be obtained from a single snapshot of data. This procedure may be used to investigate the isotropic statistics of any quantity of interest.

nlin.CD↗

Local 4/5-Law and Energy Dissipation Anomaly in Turbulence

A strong local form of the ``4/3-law'' in turbulent flow has been proved recently by Duchon and Robert for a triple moment of velocity increments averaged over both a bounded spacetime region and separation vector directions, and for energy dissipation averaged over the same spacetime region. Under precisely stated hypotheses, the two are proved to be proportional, by a constant 4/3, and to appear as a nonnegative defect measure in the local energy balance of singular (distributional) solutions of the incompressible Euler equations. Here we prove that the energy defect measure can be represented also by a triple moment of purely longitudinal velocity increments and by a mixed moment with one longitudinal and two tranverse velocity increments. Thus, we prove that the traditional 4/5- and 4/15-laws of Kolmogorov hold in the same local sense as demonstrated for the 4/3-law by Duchon-Robert.

nlin.CD↗

The Joint Cascade of Energy and Helicity in Three-Dimensional Turbulence

Three-dimensional (3D) turbulence has both energy and helicity as inviscid constants of motion. In contrast to two-dimensional (2D) turbulence, where a second inviscid invariant--the enstrophy--blocks the energy cascade to small scales, in 3D there is a joint cascade of both energy and helicity simultaneously to small scales. The basic cancellation mechanism which permits a joint cascade of energy and helicity is illuminated by means of the helical decomposition of the velocity into positively and negatively polarized waves. This decomposition is employed in the present study both theoretically and also in a numerical simulation of homogeneous and isotropic 3D turbulence. It is shown that the transfer of energy to small scales produces a tremendous growth of helicity separately in the + and - helical modes at high wavenumbers, diverging in the limit of infinite Reynolds number. However, because of a tendency to restore reflection invariance at small scales, the net helicity from both modes remains finite in that limit. The net helicity flux is shown to be constant all the way up to the Kolmogorov wavenumber: there is no shorter inertial-range for helicity cascade than for energy cascade. The transfer of energy and helicity between + and - modes, which permits the joint cascade, is shown to be due to two distinct physical processes, advection and vortex stretching.

nlin.CD↗

Intermittency in the Joint Cascade of Energy and Helicity

The statistics of the energy and helicity fluxes in isotropic turbulence are studied using high resolution direct numerical simulation. The scaling exponents of the energy flux agree with those of the transverse velocity structure functions through refined similarity hypothesis, consistent with Kraichnan's prediction \cite{Kr74}. The helicity flux is even more intermittent than the energy flux and its scaling exponents are closer to those of the passive scalar. Using Waleffe's helical decomposition, we demonstrate that the existence of positive mean helicity flux inhibits the energy transfer in the negative helical modes, a non-passive effect.

nlin.CD↗

Athermodynamic Alignment in the Two-Dimensional Enstrophy Cascade

We study inertial-range statistics in the direct enstrophy cascade of two-dimensional turbulence via a numerical simulation of the forced Navier-Stokes equation. In particular, we obtain the distribution of the enstrophy flux and of the angle of alignment of the large-scale vorticity gradient with the small-scale vorticity transport vector. These distributions are surprisingly symmetrical and ``athermodynamic'', not explainable by a local eddy-viscosity approximation with coefficient of either positive or negative sign. By a systematic evaluation of the role of various triadic interactions, we can trace the origin of the strong symmetry to the dominance of infrared non-local triads in the enstrophy cascade.

nlin.CD↗

Dissipation in Turbulent Solutions of 2-D Euler

We establish local balance equations for smooth functions of the vorticity in the DiPerna-Majda weak solutions of 2D incompressible Euler, analogous to the balance proved by Duchon and Robert for kinetic energy in 3D. The anomalous term or defect distribution therein corresponds to the ``enstrophy cascade'' of 2D turbulence. It is used to define a rather natural notion of ``dissipative Euler solution'' in 2D. However, we show that the DiPerna-Majda solutions with vorticity in $L^p$ for $p>2$ are conservative and have zero defect. Instead, we must seek an alternative approach to dissipative solutions in 2D. If we assume an upper bound on the energy spectrum of 2D incompressible Navier-Stokes solutions by the Kraichnan-Batchelor $k^{-3}$ spectrum, uniformly for high Reynolds number, then we show that the zero viscosity limits of the Navier-Stokes solutions exist, with vorticities in the zero-index Besov space $B^{0,\infty}_2$, and that these give a weak solution of the 2D incompressible Euler equations. We conjecture that for this class of weak solutions enstrophy dissipation may indeed occur, in a sense which is made precise.

math.AP↗

A Variational Formulation of Optimal Nonlinear Estimation

We propose a variational method to solve all three estimation problems for nonlinear stochastic dynamical systems: prediction, filtering, and smoothing. Our new approach is based upon a proper choice of cost function, termed the {\it effective action}. We show that this functional of time-histories is the unique statistically well-founded cost function to determine most probable histories within empirical ensembles. The ensemble dispersion about the sample mean history can also be obtained from the Hessian of the cost function. We show that the effective action can be calculated by a variational prescription, which generalizes the ``sweep method'' used in optimal linear estimation. An iterative numerical scheme results which converges globally to the variational estimator. This scheme involves integrating forward in time a ``perturbed'' Fokker-Planck equation, very closely related to the Kushner-Stratonovich equation for optimal filtering, and an adjoint equation backward in time, similarly related to the Pardoux-Kushner equation for optimal smoothing. The variational estimator enjoys a somewhat weaker property, which we call ``mean optimality''. However, the variational scheme has the principal advantage---crucial for practical applications---that it admits a wide variety of finite-dimensional moment-closure approximations. The moment approximations are derived reductively from the Euler-Lagrange variational formulation and preserve the good structural properties of the optimal estimator.

physics.data-an↗

Free Decay of Turbulence and Breakdown of Self-Similarity

It has been generally assumed, since the work of von Karman and Howarth in 1938, that free decay of fully-developed turbulence is self-similar. We present here a simple phenomenological model of the decay of 3D incompressible turbulence, which predicts breakdown of self-similarity for low-wavenumber spectral exponents $n$ in the range $n_c<n<4$, where $n_c$ is some threshold wavenumber. Calculations with the eddy-damped quasi-normal Markovian approximation give the value as $n_c\approx 3.45$. The energy spectrum for this range of exponents develops two length-scales, separating three distinct wavenumber ranges.

chao-dyn↗

Fluctuation-Response Relations for Multi-Time Correlations

We show that time-correlation functions of arbitrary order for any random variable in a statistical dynamical system can be calculated as higher-order response functions of the mean history of the variable. The response is to a ``control term'' added as a modification to the master equation for statistical distributions. The proof of the relations is based upon a variational characterization of the generating functional of the time-correlations. The same fluctuation-response relations are preserved within moment-closures for the statistical dynamical system, when these are constructed via the variational Rayleigh-Ritz procedure. For the 2-time correlations of the moment-variables themselves, the fluctuation-response relation is equivalent to an ``Onsager regression hypothesis'' for the small fluctuations. For correlations of higher-order, there is a new effect in addition to such linear propagation of fluctuations present instantaneously: the dynamical generation of correlations by nonlinear interaction of fluctuations. In general, we discuss some physical and mathematical aspects of the {\it Ansätze} required for an accurate calculation of the time correlations. We also comment briefly upon the computational use of these relations, which is well-suited for automatic differentiation tools. An example will be given of a simple closure for turbulent energy decay, which illustrates the numerical application of the relations.

chao-dyn↗

Linear Stochastic Models of Nonlinear Dynamical Systems

We investigate in this work the validity of linear stochastic models for nonlinear dynamical systems. We exploit as our basic tool a previously proposed Rayleigh-Ritz approximation for the effective action of nonlinear dynamical systems started from random initial conditions. The present paper discusses only the case where the PDF-Ansatz employed in the variational calculation is ``Markovian'', i.e. is determined completely by the present values of the moment-averages. In this case we show that the Rayleigh-Ritz effective action of the complete set of moment-functions that are employed in the closure has a quadratic part which is always formally an Onsager-Machlup action. Thus, subject to satisfaction of the requisite realizability conditions on the noise covariance, a linear Langevin model will exist which reproduces exactly the joint 2-time correlations of the moment-functions. We compare our method with the closely related formalism of principal oscillation patterns (POP), which, in the approach of C. Penland, is a method to derive such a linear Langevin model empirically from time-series data for the moment-functions. The predictive capability of the POP analysis, compared with the Rayleigh-Ritz result, is limited to the regime of small fluctuations around the most probable future pattern. Finally, we shall discuss a thermodynamics of statistical moments which should hold for all dynamical systems with stable invariant probability measures and which follows within the Rayleigh-Ritz formalism.

chao-dyn↗

Fluctuations in the Irreversible Decay of Turbulent Energy

A fluctuation law of the energy in freely-decaying, homogeneous and isotropic turbulence is derived within standard closure hypotheses for 3D incompressible flow. In particular, a fluctuation-dissipation relation is derived which relates the strength of a stochastic backscatter term in the energy decay equation to the mean of the energy dissipation rate. The theory is based on the so-called ``effective action'' of the energy history and illustrates a Rayleigh-Ritz method recently developed to evaluate the effective action approximately within probability density-function (PDF) closures. These effective actions generalize the Onsager-Machlup action of nonequilibrium statistical mechanics to turbulent flow. They yield detailed, concrete predictions for fluctuations, such as multi-time correlation functions of arbitrary order, which cannot be obtained by direct PDF methods. They also characterize the mean histories by a variational principle.

chao-dyn↗

Shape-Dependent Thermodynamics and Non-Local Hydrodynamics in a Non-Gibbsian Steady-State of a Drift-Diffusion System

Shape-dependent thermodynamics and non-local hydrodynamics are argued to occur in dissipative steady states of driven diffusive systems. These predictions are confirmed by numerical simulations. Unlike power-law correlations, these phenomena cannot be explained by a hypothesis of ``criticality''. Instead, they require the effective Hamiltonian of the system to contain very long-range potentials, making the invariant probability measures formally ``non-Gibbsian''.

cond-mat.stat-mech↗

Turbulence Fluctuations and New Universal Realizability Conditions in Modelling

General turbulent mean statistics are shown to be characterized by a variational principle. The variational functionals, or ``effective actions'', have experimental consequences for turbulence fluctuations and are subject to realizability conditions of positivity and convexity. An efficient Rayleigh-Ritz algorithm is available to calculate approximate effective actions within PDF closures. Examples are given for Navier-Stokes and for a 3-mode system of Lorenz. The new realizability conditions succeed at detecting {\em a priori} the poor predictions of PDF closures even when the classical 2nd-order moment realizability conditions are satisfied.

chao-dyn↗

Rayleigh-Ritz Calculation of Effective Potential Far From Equilibrium

We demonstrate the utility of a Rayleigh-Ritz scheme recently proposed to compute the nonequilibrium effective potential nonperturbatively in a strong noise regime far from equilibrium. A simple Kramers model of an ionic conductor is used to illustrate the efficiency of the method.

chao-dyn↗

Existence and Uniqueness of $L^2$-Solutions at Zero-Diffusivity in the Kraichnan Model of a Passive Scalar

We study Kraichnan's model of a turbulent scalar, passively advected by a Gaussian random velocity field delta-correlated in time, for every space dimension $d\geq 2$ and eddy-diffusivity (Richardson) exponent $0<ζ<2$. We prove that at zero molecular diffusivity, or $κ= 0$, there exist unique weak solutions in $L^2\left(Ω^{\otimes N}\right)$ to the singular-elliptic, linear PDE's for the stationary $N$-point statistical correlation functions, when the scalar field is confined to a bounded domain $Ω$ with Dirichlet b.c. Under those conditions we prove that the $N$-body elliptic operators in the $L^2$ spaces have purely discrete, positive spectrum and a minimum eigenvalue of order $L^{-γ}$, with $γ=2-ζ$ and with $L$ the diameter of $Ω$. We also prove that the weak $L^2$-limits of the stationary solutions for positive, $p$th-order hyperdiffusivities $κ_p>0$, $p\geq 1$, exist when $κ_p \rightarrow 0$ and coincide with the unique zero-diffusivity solutions. These results follow from a lower estimate on the minimum eigenvalue of the $N$-particle eddy-diffusivity matrix, which is conjectured for general $N$ and proved in detail for $N=2,3,4$. Some additional issues are discussed: (1) Hölder regularity of the solutions; (2) the reconstruction of an invariant probability measure on scalar fields from the set of $N$-point correlation functions, and (3) time-dependent weak solutions to the PDE's for $N$-point correlation functions with $L^2$ initial data.

chao-dyn↗

Action Principle in Nonequilibrium Statistical Dynamics

We introduce a variational method for approximating distribution functions of dynamics with a ``Liouville operator'' $\hL,$ in terms of a {\em nonequilibrium action functional} for two independent (left and right) trial states. The method is valid for deterministic or stochastic Markov dynamics, and for stationary or time-dependent distributions. A practical Rayleigh-Ritz procedure is advanced, whose inputs are finitely-parametrized ansatz for the trial states, leading to a ``parametric action'' for their evolution. The Euler-Lagrange equations of the action principle are Hamiltonian in form (generally noncanonical). This permits a simple identification of fixed points as critical points of the parametric Hamiltonian. We also establish a variational principle for low-order statistics, such as mean values and correlation functions, by means of {\em least effective action.} The latter is a functional of the given variable, which is positive and convex as a consequence of Hölder realizability inequalities. Its value measures the ``cost'' for a fluctuation from the average to occur and in a weak-noise limit it reduces to the Onsager-Machlup action. In general, the effective action is shown to arise from the nonequilibrium action functional by a constrained variation. This result provides a Rayleigh-Ritz scheme for calculating just the desired low-order statistics, with internal consistency checks less demanding than for the full distribution.

chao-dyn↗

Intermittency and Anomalous Scaling of Passive Scalars in Any Space Dimension

We establish exact inequalities for the structure-function scaling exponents of a passively advected scalar in both the inertial-convective and viscous-convective ranges. These inequalities involve the scaling exponents of the velocity structure functions and, in a refined form, an intermittency exponent of the convective-range scalar flux. They are valid for 3D Navier-Stokes turbulence and satisfied within errors by present experimental data. The inequalities also hold for any ``synthetic'' turbulent velocity statistics with a finite correlation in time. We show that for time-correlation exponents of the velocity smaller than the ``local turnover'' exponent, the scalar spectral exponent is strictly less than that in Kraichnan's soluble ``rapid-change'' model with velocity delta-correlated in time. Our results include as a special case an exponent-inequality derived previously by Constantin \& Procaccia [Nonlinearity {\bf 7} 1045 (1994)], but with a more direct proof. The inequalities in their simplest form follow from a Kolmogorov-type relation for the turbulent passive scalar valid in each space dimension $d.$ Our improved inequalities are based upon a rigorous version of the refined similarity hypothesis for passive scalars. These are compared with the relations implied by ``fusion rules'' hypothesized for scalar gradients.

chao-dyn↗