A shell-model approach to fractal-induced turbulence

We present a shell-model of fractal induced turbulence which predicts that structure function scaling exponents decrease in absolute value as the fractal dimension of the turbulence-inducing fractal object increases. This qualitative prediction is in agreement with laboratory measurements. Finer details of the fractal induced turbulence statistics and dynamics depend on the fractal force's phases, i.e. on the detailed construction of the fractal stirrer. In a case of deterministic forcing phases, a critical fractal dimension exists below which the average rate of inter-scale energy transfer <T _n> is a decreasing function of the wavenumber k_n and the structure function scaling exponents take close to Kolmogorov values. Above this critical fractal dimension, <T _n> is an increasing function of k_n and the structure function scaling exponents deviate significantly from Kolmogorov values. Received 25 June 2001 / Received in final form 5 April 2002 Published online 19 July 2002     

Local multifractal thermodynamics of 3D turbulence

Using results of a direct numerical simulation (DNS) of 3D turbulence we show that the observed generalized scaling (i.e. scaling moments versus moments of different orders) is consistent with a lognormal-like distribution of turbulent energy dissipation fluctuations with moderate amplitudes for all space scales available in this DNS (beginning from the molecular viscosity scale up to largest ones). Local multifractal thermodynamics has been developed to interpret the data obtained using the generalized scaling, and a new interval of space scales with inverse cascade of generalized energy has been found between dissipative and inertial intervals of scales for sufficiently large values of the Reynolds number. Received 21 July 2000     

Turbulence over arrays of obstacles in low temperature helium gas

Turbulence produced in low temperature helium gas flowing over arrays of rectangular- and triangular-shaped blunt obstacles is investigated experimentally. The set-up allows both low fluctuation rates (down to 8%), and high microscale Reynolds numbers, (up to 1 150). The forced Kolmogorov equation is found to apply accurately. Similar to another flow configuration (counter rotating flow case [1]), the analysis of the flatness factor evolution with the Reynolds number reveals a transitional behavior around 650. Received 26 August 1999 and Received in final form 28 August 2000     

Lagrangian and Eulerian velocity intermittency

Usual turbulence experiments, based on the Taylor hypothesis, differ from true Eulerian measurements. This is the origin of the apparent discrepancy between a recent two point correlation analysis and the multiplicative cascade picture. Indeed, both Eulerian and Lagrangian observations perfectly agree with this picture. Received 19 June 2002 / Received in final form 29 July 2002 Published online 14 October 2002 RID="a" ID="a"e-mail: bcastain@ens-lyon.fr     

Nonperturbative renormalization group approach to turbulence

We suggest a new, renormalization group (RG) based, nonperturbative method for treating the intermittency problem of fully developed turbulence which also includes the effects of a finite boundary of the turbulent flow. The key idea is not to try to construct an elimination procedure based on some assumed statistical distribution, but to make an ansatz for possible RG transformations and to pose constraints upon those, which guarantee the invariance of the nonlinear term in the Navier-Stokes equation, the invariance of the energy dissipation, and other basic properties of the velocity field. The role of length scales is taken to be inverse to that in the theory of critical phenomena; thus possible intermittency corrections are connected with the outer length scale. Depending on the specific type of flow, we find different sets of admissible transformations with distinct scaling behaviour: for the often considered infinite, isotropic, and homogeneous system K41 scaling is enforced, but for the more realistic plane Couette geometry no restrictions on intermittency exponents were obtained so far. Received: 28 December 1997 / Accepted: 6 August 1998     

A statistical estimator of turbulence intermittency in physical and numerical experiments

The velocity increments statistic in various turbulent flows is analysed through the hypothesis that different scales are linked by a multiplicative process, of which multiplier is infinitely divisible. This generalisation of the Kolmogorov-Obukhov theory is compatible with the finite Reynolds number value of real flows, thus ensuring safe extrapolation to the infinite Reynolds limit. It exhibits a estimator universally depending on the Reynolds number of the flow, with the same law either for Direct Numerical Simulations or experiments, both for transverse and longitudinal increments. As an application of this result, the inverse dependence is used to define an unbiased value for a Large Eddy Simulation from the resolved scales velocity statistics. However, the exact shape of the multiplicative process, though independent of the Reynolds number for a given experimental setup, is found to depend significantly on this setup and on the nature of the increment, longitudinal or transverse. The asymmetry of longitudinal velocity increments probability density functions exhibits similarly a dependence with the experimental setup, but also systematically depends on the Reynolds number. Received 7 January 2000 and Received in final form 17 March 2000     

Finite size scale invariance

We develop a new approach to scale symmetry, which takes into account the possible finite cut-offs of the fields or the parameters. This new symmetry, called finite size scale symmetry: i) includes the traditional self-similarity as a limiting case, when the cut-offs are set to infinity (infinite size-system); ii) is consistent with the traditional finite size scaling approach already used in critical phenomena; iii) enables the computation of some of the universal functions appearing in the finite size scaling formulation; iv) allows scale transformations leaving the cut-offs invariant, like in the traditional renormalization approach; v) can be formulated to allow for positive or negative fields and parameters; vi) leads to new predictions about the shape of some distributions in critical phenomena or turbulence which are in very good agreement with the experimental or numerical findings. Received 26 January 1999 and Received in final form 25 October 1999     

Intermittency in a turbulent low temperature gaseous helium jet

We analyse experimental velocity measurements on the axis of a low temperature gaseous helium jet. From independent increments arguments, we reproduce the behaviour of structure functions. We show where this approach fails and how the intermittency phenomenon is a small correction. The physical arguments under the multiplicative cascade models for this intermittency imply an acceleration of this cascade close to the dissipative range, which we are able to evidence. This acceleration could be responsible of the apparent Extended Self Similarity between structure functions of various orders. Received 13 October 1999 and Received in final form 19 May 2000     

Large scale correlations for energy injection mechanisms in swirling turbulent flows

We report an experimental study of large scale correlations in the power injected in turbulent swirling flows generated in the gap between two coaxial rotating disks. We measure the pressure fluctuations on the blades of one disk, as well as the pressure drop between the leading and the trailing edges of the rotating blades, i.e. the local drag force. Measurements at different positions on one blade and on two successive blades display a correlation length much larger than the ones usually expected in turbulent flows. The time lag for which the correlation between two points is maximum, strongly depends on the global flow configuration. These results help us to understand the statistical properties of the injected power fluctuations in turbulent swirling flows. Received 2 September 1999     

Affine turbulence

We present a generalization of the multiplicative model for velocity increments involving an affine process. The consequences on the shape of the probability distribution functions for the velocity increments are explored, and shown to be better compatible with the existence of a scale variation of the skewness. Received 29 March 1999 and Received in final form 14 September 1999     

Detecting vorticity filaments using wavelet analysis: About the statistical contribution of vorticity filaments to intermittency in swirling turbulent flows

Swirling turbulent flows display intermittent pressure drops associated with intense vorticity filaments. Using the wavelet transform modulus maxima representation of pressure fluctuations, we propose a method of characterizing these pressure drop events from their time-scale properties. This method allows us to discriminate fluctuations induced by just formed (young) as well as by burst (old) filaments from background pressure fluctuations. The statistical characteristics of these filaments (core size, waiting time) are analyzed in details and compared with previously reported experimental and numerical findings. Their intermittent occurrence is found to be governed by a pure Poisson's law, the hallmark of independent events. Then we apply the wavelet transform modulus maxima (WTMM) method to the background pressure fluctuations. This study reveals that, once removed all the filaments, the “multifractal” nature of pressure fluctuations still persists. This is a clear indication that the statistical contribution of the filaments is not important enough to account for the intermittency phenomenon in turbulents flows. Received 27 July 1998 and Received in final form 23 November 1998     

Mean magnetic field and noise cross-correlation in magnetohydrodynamic turbulence: results from a one-dimensional model

We show that a recently proposed [J. Fleischer, P.H. Diamond, Phys. Rev. E 58, R2709 (1998)] one-dimensional Burgers-like model for magnetohydrodynamics (MHD) is in effect identical to existing models for drifting lines and sedimenting lattices. We use the model to demonstrate, contrary to claims in the literature, that the energy spectrum of MHD turbulence should be independent of mean magnetic field and that cross-correlations between the noise sources for the velocity and magnetic fields cannot change the structure of the equations under renormalisation. We comment on the scaling and the multiscaling properties of the stochastically forced version of the model. Received 29 October 1998 and Received in final form 8 December 1998     

Acceleration statistics of inertial particles in turbulent flow

Turbulent transport of material inclusions plays an important role in many natural and industrial situations. Being able to accurately model and predict the dynamics of dispersed particles transported by a turbulent carrier flow, remains a challenge. One critical and difficult point is to develop models which correctly describe the dynamics of particles over a wide range of sizes and densities. Our measurements show that acceleration statistics of particles dispersed in a turbulent flow do exhibit specific, and so far unpredicted, size and density effects and that they preserve an extremely robust turbulent signature with lognormal fluctuations, regardless of particles size and density. This has important consequences in terms of modeling for the turbulent transport of dispersed inclusions.     

Dynamics of spatial Fourier modes in turbulence

We present the results of an experimental study of the spatial Fourier modes of the vorticity in a turbulent jet flow. By means of an acoustic scattering setup we have recorded the evolution in time of Fourier modes of the vorticity field, characterized by well defined wavevectors k. By computing the auto-correlation of the amplitude of the Fourier modes we evidence that, whatever the length scale (or equivalently k), the dynamic evolution of the vorticity field involves two well separated time scales. We have also performed the simultaneous acquisitions of pairs of Fourier modes with two wavevectors k and k'. Whatever the spectral gap k- k', any pair of Fourier modes exhibits a significant cross-correlation over long time delays, indicating a strong statistical dependence between scales.     

Convective heat transport in a fluid layer of infinite Prandtl number: upper bounds for the case of rigid lower boundary and stress-free upper boundary

We present the theory of the multi--solutions of the variational problem for the upper bounds on the convective heat transport in a heated from below horizontal fluid layer with rigid lower boundary and stress-free upper boundary. A sequence of upper bounds on the convective heat transport is obtained. The highest bound is between the bounds for the case of a fluid layer with two rigid boundaries and for the case of a fluid layer with two stress-free boundaries. As an additional result of the presented theory we obtain small corrections of the boundary layer thicknesses of the optimum fields for the case of fluid layer with two rigid boundaries. These corrections lead to systematically lower upper bounds on the convective heat transport in comparison to the bounds obtained in [5]. Received 29 September 1999     

Two-particle dispersion in model two-dimensional velocity fields

We consider two-particle dispersion in a velocity field, where the relative two-point velocity scales according to v ²(r) ∝r ^α and the corresponding correlation time scales as τ(r) ∝r ^β, and fix α = 2/3, as typical for turbulent flows. We show that two generic types of dispersion behavior arize: For α/2 + β < 1 the correlations in relative velocities decouple and the diffusion approximation holds. In the opposite case, α/2 + β > 1, the relative motion is strongly correlated. The case of Kolmogorov flows corresponds to a marginal, nongeneric situation. In this case, depending on the particular parameters of the flow, the dispersion behavior can be rather diffusive or rather ballistic. Received 13 March 2001     

Dependence of heat transport on the strength and shear rate of prescribed circulating flows

We study numerically the dependence of heat transport on the maximum velocity and shear rate of physical circulating flows, which are prescribed to have the key characteristics of the large-scale mean flow observed in turbulent convection. When the side-boundary thermal layer is thinner than the viscous boundary layer, the Nusselt number (Nu), which measures the heat transport, scales with the normalized shear rate to an exponent 1/3. On the other hand, when the side-boundary thermal layer is thicker, the dependence of Nu on the Peclet number, which measures the maximum velocity, or the normalized shear rate when the viscous boundary layer thickness is fixed, is generally not a power law. Scaling behavior is obtained only in an asymptotic regime. The relevance of our results to the problem of heat transport in turbulent convection is also discussed. Received 28 November 2001 Published online 25 June 2002     

Modeling an efficient Brownian heat engine

We discuss various properties of a homogeneous random multifractal process, which are related to the issue of scale correlations. By design, the process has no built-in scale correlations. However, when it comes to observables like breakdown coefficients, which are based on a coarse-graining of the multifractal field, scale correlations do appear. In the log-normal limit of the model process, the conditional distributions and moments of breakdown coefficients reproduce the observations made in fully developed small-scale turbulence. These findings help to understand several puzzling empirical details, which have been extracted from turbulent data already some time ago.