[Paper Review] Macroscopic effects of the spectral structure in turbulent flows

Tuan Tran, Pinaki Chakraborty+6 more|Sep 15, 2009

Fluid Dynamics and Turbulent Flows25 references3 citations

TL;DR

This paper establishes a direct link between the macroscopic frictional drag in turbulent flows and the spectral structure of turbulent fluctuations, demonstrating that the drag scaling depends on the spectral exponent 𝛼. Using soap-film flows (𝛼 = 3), it finds drag scales as Re⁻¹/², contrasting with pipe flows (𝛼 = 5/3) where drag scales as Re⁻¹/⁴, validating a new theory that connects drag to the turbulent spectrum via eddy velocity scaling.

ABSTRACT

Two aspects of turbulent flows have been the subject of extensive, split research efforts: macroscopic properties, such as the frictional drag experienced by a flow past a wall, and the turbulent spectrum. The turbulent spectrum may be said to represent the fabric of a turbulent state; in practice it is a power law of exponent α(the "spectral exponent") that gives the revolving velocity of a turbulent fluctuation (or "eddy") of size s as a function of s. The link, if any, between macroscopic properties and the turbulent spectrum remains missing. Might it be found by contrasting the frictional drag in flows with differing types of spectra? Here we perform unprecedented measurements of the frictional drag in soap-film flows, where the spectral exponent α= 3 and compare the results with the frictional drag in pipe flows, where the spectral exponent α= 5/3. For moderate values of the Reynolds number Re (a measure of the strength of the turbulence), we find that in soap-film flows the frictional drag scales as Re^{-1/2}, whereas in pipe flows the frictional drag scales as Re^{-1/4} . Each of these scalings may be predicted from the attendant value of αby using a new theory, in which the frictional drag is explicitly linked to the turbulent spectrum. Our work indicates that in turbulence, as in continuous phase transitions, macroscopic properties are governed by the spectral structure of the fluctuations.

Motivation & Objective

• To investigate the relationship between macroscopic drag in turbulent flows and the spectral structure of turbulent fluctuations.
• To resolve the long-standing gap between classical theories of drag and the underlying turbulent spectrum.
• To test whether different spectral exponents (𝛼) in 2D vs. 3D turbulence lead to distinct drag scaling laws.
• To validate a new theoretical framework that explicitly links frictional drag to the turbulent energy spectrum.

Proposed method

• Measured frictional drag f = τ / (ρU²) in turbulent soap-film flows using a Laser Doppler Velocimeter at 5 kHz sampling rate.
• Applied Taylor’s frozen-turbulence hypothesis to convert time-series velocity data into spatial velocity fluctuations for spectral analysis.
• Computed the longitudinal turbulent spectrum E(k) from Fourier transforms of velocity fluctuations, assuming E(k) ∝ k⁻ᵅ.
• Used the spectral exponent 𝛼 to predict eddy velocity uₛ ∝ U(s/L)⁽ᵅ⁻¹⁾/² and smallest-scale eddy velocity uₙ ∝ ν/η.
• Derived theoretical drag scaling f ∝ Re⁽¹⁻ᵅ⁾/⁽¹⁺ᵅ⁾ from momentum transfer by smallest eddies, linking drag to spectral structure.
• Compared experimental drag scaling in soap-film flows (𝛼=3) with pipe flow data (𝛼=5/3) and theoretical predictions.

Experimental results

Research questions

• RQ1 Does the spectral exponent 𝛼 of the turbulent energy spectrum determine the scaling of frictional drag in wall-bounded turbulent flows?
• RQ2 Can the discrepancy between Re⁻¹/⁴ scaling in pipe flows and Re⁻¹/² scaling in soap-film flows be explained by differences in spectral structure?
• RQ3 Is there a theoretical framework that unifies macroscopic drag with the spectral properties of turbulent fluctuations?
• RQ4 Can 2D soap-film flows serve as a testbed for probing the role of spectral structure in turbulence?
• RQ5 Why do classical theories of drag fail to distinguish between 2D and 3D turbulent flows?

Key findings

• In turbulent soap-film flows with spectral exponent 𝛼 = 3, the frictional drag scales as f ∝ Re⁻¹/², consistent with experimental data.
• In pipe flows with 𝛼 = 5/3, the drag scales as f ∝ Re⁻¹/⁴, matching the Blasius empirical scaling.
• The new theoretical model f ∝ Re⁽¹⁻ᵅ⁾/⁽¹⁺ᵅ⁾ successfully predicts both scaling laws using only the spectral exponent 𝛼.
• The theory links drag to the velocity of the smallest-scale turbulent eddies, uₙ ∝ ν/η, which are viscous (Reₙ ≈ 1).
• The experimental results show that classical drag theories are incomplete because they do not account for spectral structure or dimensionality.
• The findings suggest that macroscopic properties in turbulence are governed by the spectral structure of fluctuations, analogous to critical phenomena in statistical mechanics.