Lagrangian Markovianized field approximation for turbulence

In a previous communication (W.J.T. Bos and J.-P. Bertoglio 2006, Phys. Fluids, 18, 031706), a self-consistent Markovian triadic closure was presented. The detailed derivation of this closure is given here, relating it to the Direct Interaction Approximation and Quasi-Normal types of closure. The time-scale needed to obtain a self-consistent closure for both the energy spectrum and the scalar variance spectrum is determined by evaluating the correlation between the velocity and an advected displacement vector-field. The relation between this latter correlation and the velocity–scalar correlation is stressed, suggesting a simplified model of the latter. The resulting closed equations are numerically integrated and results for the energy spectrum, scalar fluctuation spectrum and velocity–displacement correlation spectrum are presented for low, unity and high values of the Schmidt number.     

Corrections to the scaling of the second-order structure function in isotropic turbulence

The approach of Obukhov assuming a constant skewness was used to obtain analytical corrections to the scaling of the second order structure function, starting from Kolmogorov＇s 4/5 law. These corrections can be used in model applications in which explicit expressions, rather than numerical solutions are needed. The comparison with an interpolation formula proposed by Batchelor, showed that the latter gives surprisingly precise results. The modification of the same method to obtain analytical corrections to the scaling law, taking into account the possible corrections induced by intermittency, is also proposed.     

Mixing in Isotropic Turbulence with Scalar Injection and Applications to Subgrid Modeling

A new technique for injecting scalar fluctuations in a DNS of isotropic turbulence is presented. It is used to study statistically steady states associated with different levels of mixing. The results are analysed in terms of spectra and PDF, and they are used as a database to investigate the effect of the filtering operation that is performed in LES. It is shown that the PDF of the scalar is substantially affected by the filtering operation. It is also shown that the Cook and Riley [1] subgrid model allows reconstruction of a PDF which is in fairly good agreement with the unfiltered DNS results. The consequences of estimating the scalar subgrid variance by scale similarity assumptions are investigated. It is found that the results are improved by a local determination of the model constant. This revised version was published online in July 2006 with corrections to the Cover Date.     

Transportation of passive scalar in inhomogeneous turbulence

The paper studies the transportation of passive scalar in the inhomogeneous turbulence by means of large eddy simulation. The prediction accuracy is verified by the well-known Comte-Bellot spectrum of the homogeneous turbulence. The mean transportation properties are predicted with satisfaction and the underestimation of the thermal flux by the phenomenological models has been disclosed. The high intermittence of the temperature fluctuations has been found in the inhomogeneous turbulence and it is the reason for the underestimation of turbulent thermal flux by the phenomenological model. The project supported by the National Natural Science Foundation of China (19732005), National Climbing Project and LIAMA     

An elementary geometric nonstandard proof of the Jordan curve theorem

We give an elementary proof, using nonstandard analysis, of the Jordan curve theorem. We also give a nonstandard generalization of the theorem. The proof is purely geometrical in character, without any use of topological concepts and is based on a discrete finite form of the Jordan theorem, whose proof is purely combinatorial.Some familiarity with nonstandard analysis is assumed. The rest of the paper is self-contained except for the proof a discrete standard form of the Jordan theorem. The proof is based on hyperfinite approximations to regions on the plane.Research of the first author partially supported by FONDECYT Grant # 91-1208 and of the second author, by FONDECYT Grant # 90-0647.     

Recent understanding on the subgrid-scale modeling of large-eddy simulation in physical space

In the last 50 years, the methodology of large-eddy simulation (LES) has been greatly developed, while lots of different subgridscale (SGS) models have appeared. However, the understanding of the procedure of SGS modeling is still not clear. The present contribution aims at reviewing the recent SGS models and, more importantly, expressing our recent understanding on the SGS modeling of LES in physical space. Taking the Kolmogorov equation for filtered quantities (KEF) as an example, it is argued that the KEF alone is not enough to be a closure method. Three physical laws are then introduced to complete this closure procedure and are expected to inspire the future researches of SGS modeling.     

Effects of plane strain on inhomogeneous turbulence

The influence of the mean plane strain on the turbulence transportation is investigated by large eddy simulation (LES) in the shearless turbulence mixing layer. It is found that the mean strains enhance the turbulent fluctuations in the mixing region. Compression in the inhomogeneous direction can greatly increase the transport of turbulent kinetic energy by triple correlation terms, while stretching in the inhomogeneous direction decreases the turbulence transportation. The gradient diffusion models for turbulent transportation are evaluated and it is found that the intermittency consideration can improve the prediction ability of the gradient-type models for the triple correlation terms. Project supported by the Sino-French Laboratory in Beijing and the National Natural Science Foundation of China (Grant No. 19572041).     

A pointwise spectrum and representation of operators

For a self-adjoint operator commuting with an increasing family of projections we study the multifunction an open set of the topology containing , where is the spectrum of on . Let be the measure of maximal spectral type. We study the condition that is essentially a singleton, is not a singleton. We show that if is the density topology and if satisfies the density theorem, in particular if it is absolutely continuous with respect to the Lebesgue measure, then this condition is equivalent to the fact that is a Borel function of . If is the usual topology then the condition is equivalent to the fact that is approched in norm by step functions , where the set of intervals covers the set where is a singleton.