Fractal dimensions and spectra of interfaces with application to turbulence

This paper is concerned with the analysis of any convoluted surface in two or three dimensions which has a self-similar structure, and which may simply be defined as a mathematical surface or as an interface where there is a sharp change in the value of a scalar field F(x) (say from 0 to 1). The different methods of analysis that are related to each other here are based on spectra of the scalar which have the form Γ(k) ∝ k-p, the Kolmogorov capacity DK of the interface or D'K of the intersections of the interface with a plane or a line (both being defined by algorithms for counting the minimum number of boxes of sizes ∊ either covering the surface or its intersection), and the Hausdorff dimensions DH, D'H (which are defined differently). It is demonstrated that interfaces with a localized self-similar structure around accumulation points, such as spirals, may have non-integer capacities DK and D'K even though their Hausdorff dimension is integer and equal to the topological dimension of the surface. It is explained how the same surface can have different values of DK and D'K over different asymptotic ranges of ∊. There are other (fractal) surfaces where both DH and DK are non-integer which are convoluted on a wide range of scales with the same form of self-similarity everywhere on the surface. Distinctions are drawn between these two kinds of interface which have local and global self- similarity respectively. If the intersections of the interface with any line form a set of points that is statistically homogeneous and independent of the location and orientation of the line and also that is self-similar over a sufficiently wide range of spacing that a capacity D'K can be defined, it is shown that the scalar F has a power spectrum of the form of (1) and that the exponent p is related to D'K by p + D'K = 2. This quite general result for interfaces is verified analytically and computationally for spirals. In experiments with scalar interfaces in different turbulent flows at high Reynolds and Prandtl number, K. R. Sreenivasan, R. Ramshankar and C. Meneveau’s measurements showed that D'K = 0.33 for values of ∊ within the inertial range and D'K = 0 for smaller values (in the microscale range). The values of p derived from (2) are consistent with the theory of G. K. Batchelor and many measurements of scalar spectra. For fractal interfaces with global self-similarity, the values of DH of an interface and D'H of the intersections of the interface with a line, have been shown previously to be related simply to each other by the topological dimension E of the interface so that DH = D'H + E. No such theorem exists in general for the Kolmogorov capacities DK and D'K. But it is shown analytically and computationally that for the case of spirals, over a certain range of resolutions ∊, DK = D'K + E. This corresponds in practice to the measurable range of typical experimental spirals with fewer than about five turns. Over a range of smaller length scales ∊, where a larger number of turns is resolved and which is experimentally difficult to measure, (3) is not correct. The result (3) has been previously suggested based on experimental results. Finally we demonstrate that interfaces for which there is a well-defined value of capacity (which is indeed the case for spirals of three turns) are only found to have self-similar spectra if there is a much wider range of length scales (e. g. more than 50 turns of the spiral) than is needed for the capacity DK to be measurable. As well as demonstrating this computationally, this is proved mathematically for interfaces having a particular class of accumulation points whose intersections with straight lines form a self-similar sequence of points xn ∝ n-α; the power spectrum of F only tends to the self-similar form (1) if (∊min/∊max)1-D'K≪ 1, whereas the capacity measure simply requires that ∊min/∊max ≪ 1. So when D'K > 0, the criteria for the spectra requires a wider range of scales in the convolutions of the interface. This is consistent with the finding that reliable measurements of DK can be computed from measurements of interfaces in laboratory experiments, but in the same experiments computations of spectra are often not of the form (1). Therefore, despite this apparent discrepancy with the general result (2) (which implies a value of p given a value of D'K), the above theoretical argument supports the deduction from such experiments that if a non-integer value of DK is measured, the interface does indeed have a self-similar structure; but it is not self-similar over a wide enough range of length scales to satisfy (2). Consequently, for turbulent flows, stating that the capacity should have its asymptotic value rather than (as is usual) the spectrum should be equal to its asymptotic form (e. g. in (1) p = 5/3) may be the correct necessary condition for deciding whether, in a given flow, the interfaces have the characteristic structure found at very high Reynolds number. Many of the results here may be of value in other scientific fields, where convoluted interfaces are also studied.

[1]  D. A. Santavicca,et al.  The Fractal Nature of Premixed Turbulent Flames , 1990 .

[2]  Charles Meneveau,et al.  The fractal facets of turbulence , 1986, Journal of Fluid Mechanics.

[3]  David Montgomery,et al.  Two-Dimensional Turbulence , 2012 .

[4]  A. Townsend,et al.  Small-scale variation of convected quantities like temperature in turbulent fluid Part 2. The case of large conductivity , 1959, Journal of Fluid Mechanics.

[5]  S. Taylor The α-dimensional measure of the graph and set of zeros of a Brownian path , 1955, Mathematical Proceedings of the Cambridge Philosophical Society.

[6]  J. Jiménez,et al.  Fractal interfaces and product generation in the two‐dimensional mixing layer , 1991 .

[7]  S. Lovejoy Area-Perimeter Relation for Rain and Cloud Areas , 1982, Science.

[8]  K. W. Cattermole The Fourier Transform and its Applications , 1965 .

[9]  K. Falconer The geometry of fractal sets: Contents , 1985 .

[10]  C. Sparrow The Fractal Geometry of Nature , 1984 .

[11]  G. Batchelor Computation of the Energy Spectrum in Homogeneous Two‐Dimensional Turbulence , 1969 .

[12]  A. Gilbert Spiral structures and spectra in two-dimensional turbulence , 1988, Journal of Fluid Mechanics.

[13]  A. Grossmann,et al.  DECOMPOSITION OF HARDY FUNCTIONS INTO SQUARE INTEGRABLE WAVELETS OF CONSTANT SHAPE , 1984 .

[14]  Claude Tricot,et al.  Dimensions des spirales , 1983 .

[15]  G. Batchelor,et al.  The theory of homogeneous turbulence , 1954 .

[16]  Shigeo Kida,et al.  Numerical simulation of two-dimensional turbulence with high-symmetry , 1985 .

[17]  G. Batchelor Small-scale variation of convected quantities like temperature in turbulent fluid Part 1. General discussion and the case of small conductivity , 1959, Journal of Fluid Mechanics.

[18]  F. Gouldin An application of fractals to modeling premixed turbulent flames , 1987 .

[19]  N. Peters Laminar flamelet concepts in turbulent combustion , 1988 .

[20]  P. Sulem,et al.  Small-scale dynamics of high-Reynolds-number two-dimensional turbulence. , 1986, Physical review letters.

[21]  G. Batchelor,et al.  The effect of homogeneous turbulence on material lines and surfaces , 1952, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences.

[22]  H. K. Moffatt SIMPLE TOPOLOGICAL ASPECTS OF TURBULENT VORTICITY DYNAMICS 223 , 1984 .

[23]  J. Yorke,et al.  Dimension of chaotic attractors , 1982 .

[24]  H. Chaté A Minimal Model for Turbulent Flame Fronts , 1988 .

[25]  Steven A. Orszag,et al.  Turbulence: Challenges for Theory and Experiment , 1990 .

[26]  H. G. E. Hentschel,et al.  Relative diffusion in turbulent media: The fractal dimension of clouds , 1984 .

[27]  B. Mandelbrot On the geometry of homogeneous turbulence, with stress on the fractal dimension of the iso-surfaces of scalars , 1975, Journal of Fluid Mechanics.

[28]  J. C. Vassilicos On the Geometry of Lines in Two-Dimensional Turbulence , 1989 .

[29]  R. Kraichnan Inertial Ranges in Two‐Dimensional Turbulence , 1967 .

[30]  P. Saffman,et al.  On the Spectrum and Decay of Random Two-Dimensional Vorticity Distributions at Large Reynolds Number , 1971 .

[31]  M. M. Gibson,et al.  Spectra of turbulence in a round jet , 1963, Journal of Fluid Mechanics.

[32]  Charles Meneveau,et al.  Mixing, entrainment and fractal dimensions of surfaces in turbulent flows , 1989, Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences.

[33]  O. Phillips Spectral and statistical properties of the equilibrium range in wind-generated gravity waves , 1985, Journal of Fluid Mechanics.

[34]  Kenneth Falconer,et al.  The Hausdorff dimension of self-affine fractals , 1988, Mathematical Proceedings of the Cambridge Philosophical Society.

[35]  Richard F. Voss,et al.  Fractals in nature: from characterization to simulation , 1988 .

[36]  P. Saffman Lectures on Homogeneous Turbulence , 1968 .

[37]  S. Orey Gaussian sample functions and the Hausdorff dimension of level crossings , 1970 .

[38]  D. Ruelle Chaotic evolution and strange attractors , 1989 .

[39]  F. Hausdorff Dimension und äußeres Maß , 1918 .

[40]  Tim Bedford,et al.  The box and Hausdorff dimension of self-affine sets , 1990, Ergodic Theory and Dynamical Systems.

[41]  B. Mandelbrot How Long Is the Coast of Britain? Statistical Self-Similarity and Fractional Dimension , 1967, Science.

[42]  Katepalli R. Sreenivasan,et al.  Quantitative three-dimensional imaging and the structure of passive scalar fields in fully turbulent flows , 1990, Journal of Fluid Mechanics.