Spatiotemporal statistics of the dissipation rate at the boundary of a turbulent flow using Diffusing-Wave Spectroscopy

Basic Information

• Paper ID: 2511.03462
• Title: Spatiotemporal statistics of the dissipation rate at the boundary of a turbulent flow using Diffusing-Wave Spectroscopy
• Authors: Enzo Francisco, Julien Lambret, Sébastien Aumâıtre
• Affiliation: Service de Physique de l'État Condensé, Université Paris-Saclay, CNRS, CEA, France
• Classification: physics.flu-dyn (Fluid Dynamics)
• Submission Date: November 5, 2025
• Paper Link: https://arxiv.org/abs/2511.03462

Abstract

This study employs Diffusing-Wave Spectroscopy (DWS) to achieve the first direct spatiotemporal measurement of the dissipation rate ε at the boundary of a turbulent flow. The research team generated turbulence through impeller stirring in a square container and conducted measurements near the boundary. Despite the small measurement region, results demonstrate that the dissipation rate is proportional to the injected power and follows the turbulent scaling relation ε∝Re³ (Reynolds number range 1.5×10⁴ to 6×10⁵). Unlike pipe flow, this flow exhibits no logarithmic correction near the flat boundary. Furthermore, the spatiotemporal fluctuations of dissipation near the boundary are substantial (relative fluctuations ~50%), following a lognormal distribution, though the power spectral density deviates from expectations for homogeneous isotropic turbulence (HIT).

Research Background and Motivation

Research Problem

The turbulent dissipation rate is a key physical quantity in fluid mechanics, yet its direct measurement is extremely challenging, particularly near boundaries. This research aims to develop an experimental technique capable of directly measuring the spatiotemporal distribution of the dissipation rate at turbulent flow boundaries.

Problem Significance

1. Engineering Application Value: Dissipation rate determines energy consumption in vehicles and pressure losses in pipelines
2. Geophysical Significance: Affects energy balance and transport efficiency in climate models
3. Fundamental Scientific Significance: Dissipation structures may drive turbulent intermittency
4. Boundary Issues: Many practical applications focus on dissipation and shear rates near surfaces, such as drag on immersed objects and erosion efficiency in geophysical flows

Limitations of Existing Methods

1. Hot-wire Technique: Requires assumption of Taylor's frozen turbulence hypothesis; multi-wire probes cause flow disturbance
2. Particle Image Velocimetry (PIV):
   • Accuracy limited by interrogation window grid size
   • Applicable only to moderate Reynolds number ranges
   • Dual-plane stereoscopic PIV is technically complex
3. Direct Numerical Simulation (DNS): High computational cost, difficulty with long-time simulations, potential underestimation of rare events

Research Motivation

The dissipation rate in Newtonian fluids is defined as: $\epsilon = \sum_{i,j} \frac{\nu}{2}(\partial_j u_i + \partial_i u_j)^2$

This requires measurement of spatial derivatives of the velocity field in all directions, i.e., the norm of the strain rate tensor. DWS technology provides a novel approach to addressing this challenge.

Core Contributions

1. First Implementation: First direct spatiotemporal measurement of dissipation rate at turbulent flow boundaries
2. Scaling Relation Verification: Demonstrates that boundary dissipation rate is proportional to injected power, following the ε∝Re³ scaling law without logarithmic correction
3. Fluctuation Characteristics Revealed: Discovers that boundary dissipation fluctuations reach 50%, following a lognormal distribution
4. HIT Comparative Analysis: Finds that boundary dissipation statistics partially conform to HIT theory (PDF) but differ in power spectral characteristics
5. Technique Extension: Successfully extends DWS technology from Taylor-Couette flow to fully turbulent flow
6. High Reynolds Number Measurement: Achieves measurements at Re up to 6×10⁵, far exceeding traditional methods

Detailed Methodology

Task Definition

Input: Time series of scattered light intensity at turbulent flow boundary
Output: Spatiotemporal distribution map of dissipation rate ε near the boundary
Constraints:

• Particle radius r and light wavelength λ are much smaller than the light transport mean free path l*
• Gradient characteristic length exceeds the light path in the fluid
• Temporal resolution must be smaller than flow evolution timescale

DWS Technology Principles

Basic Equations

Coherent light illuminating a moving fluid containing small particles produces a speckle pattern. The autocorrelation function of backscattered light intensity is:

$g_2(\tau) = \frac{\langle I(t+\tau)I(t)\rangle}{\langle I(t+\tau)\rangle\langle I(t)\rangle} = \beta \exp\left(-\gamma\sqrt{6(\tau/\tau_o + \tau^2/\tau_V^2)}\right) + 1$

Where:

• β: intensity contrast
• γ: parameter depending only on optical geometry and boundary conditions
• τₒ = 1/(Dk²): timescale induced by Brownian motion
• τᵥ: timescale induced by fluid motion

Dissipation Rate Extraction

The relationship between τᵥ and dissipation rate:

$\tau_V = \sqrt{30}/\left(l^*k\sqrt{\langle\epsilon\rangle/\nu}\right)$

Since τᵥ ∝ λ/l* ≪ 1/√(⟨ε⟩/ν), the temporal evolution of ⟨ε⟩ can be captured.

Experimental Apparatus

Flow System

• Container: Square water tank with 60 cm sides
• Impeller: Radius R=20 cm, consisting of 1.2 cm thick disk and four 3.6 cm high blades
• Rotation Rate Range: Ω = 0.02-0.8 rev/s
• Reynolds Number Definition: Re = 2πUL/ν, where U = RΩ
• Reynolds Number Range: 1.5×10⁴ ≤ Re ≤ 6×10⁵
• Fluid: 170 liters of deionized water + 1.7 kg TiO₂ particles (1% mass fraction)

Tracer Particle Characteristics

• Mean Hydrodynamic Radius: r = 200 nm
• Stokes Number: Sk ≤ 0.1 (ensures passive tracer behavior)
• Sedimentation Velocity: < 10⁻⁷ μm/s (negligible)
• Transport Mean Free Path: l* = 80 μm

Characteristic Length Scale Comparison (Table 1)

Optical System

1. Illumination System: 2W Nd:YAG laser (λ=532 nm), expanded through 20× microscope lens
2. Two Detection Methods:
   • PMT + Correlator: Detects average dissipation over 12 cm diameter disk area
   • High-speed Camera: Resolves fluctuations in 5.1×5.1 cm² region (128×128 pixels)
3. Temporal Resolution:
   • Correlator sampling time: 1.28 μs
   • Camera frame rate: 430,000 fps
   • Correlation function convergence requires: 25,000 frames (0.06 s)

Measurement Protocol

1. Measure g₂(τ) in stationary fluid; derive γ from τₒ using Stokes-Einstein formula
2. Run impeller at set rotation rate; determine τᵥ from equation (3)
3. Verify γ value through stationary measurement
4. Most measurements last approximately 2 seconds (camera RAM limitation)
5. At Re=2×10⁵, conduct 90-second long-duration measurement (45 segments of 2 seconds)

Technical Innovations

1. Spatial Resolution Advantage: Strain rate tensor estimation accuracy independent of camera resolution; enables exploration of large Reynolds number ranges with same configuration
2. Non-invasive: Optical measurement without flow field disturbance
3. Direct Measurement: No need for Taylor hypothesis or complex velocity field reconstruction
4. Spatiotemporal Resolution: Simultaneously obtains spatial distribution and temporal evolution information
5. High Reynolds Number Applicability: Overcomes Reynolds number limitations of traditional methods

Experimental Setup

Fluid System Parameters

• Working Fluid: Deionized water (kinematic viscosity ν ≈ 1×10⁻⁶ m²/s)
• Tracer Particles: TiO₂ nanoparticles
  • Density: ρₛ > ρw (but sedimentation negligible)
  • Concentration: 1% mass fraction
  • Suspension Stability: ensured through deionized water
• Power Measurement: Torque meter measures torque C; injected power calculated as P = C·Ω

Optical Parameter Verification

• In-situ l Measurement*: Spatial-resolved reflection technique
  • Uses another 250 mW Nd:YAG laser
  • CCD camera measures radial intensity decay of diffuse light spot
  • Follows ∝(r/l*)⁻³ law
  • l* remains stable over 48-hour experimental period (under continuous stirring)

Measurement Region

• PMT Detection Area: 12 cm diameter disk, depth ~5l* ≈ 400 μm
• Camera Detection Area: 5.1×5.1 cm² square, depth ~5l* ≈ 400 μm
• Pixel Size: ~400 μm (larger than l*, satisfying requirements)
• Measurement Location: Container wall at impeller height

Experimental Condition Control

• Stirring Strategy: Alternate high and low rotation rates to maintain uniform particle mixing
• Measurement Duration:
  • Routine measurements: ~2 seconds
  • Long-duration measurements: 90 seconds (for PDF statistics)
• Macroscopic Timescale: 1/(4Ω) = 0.3-12.5 seconds

Experimental Results

Main Results

1. Dissipation Rate Scaling Relation (Figure 2)

Core Findings:

• Boundary dissipation rate ⟨ε⟩ is proportional to injected power P
• Follows turbulent zeroth law: ⟨ε⟩ ∝ P ∝ (ν³/L⁴)·Re³
• No Logarithmic Correction Required: Unlike pipe/channel flows

Quantitative Results:

• Measured boundary dissipation rate approximately 2.4 times the average injected power
• This excess dissipation conforms to expectations for excess dissipation at boundaries near impellers

Physical Interpretation:

• Single impeller in square container produces highly non-uniform flow
• Dissipation structures spatially non-uniformly distributed
• Excess dissipation exists at boundaries near impeller

2. Spatial Fluctuation Characteristics (Figure 3)

Observations:

• Relative Fluctuations: σε/⟨ε⟩ ≈ 50% (independent of Re)
• Standard Deviation Scaling: σε ∝ (ν³/L⁴)·Re³
• Dissipation structure scale significantly larger than pixel size

Structure Characteristics:

• Dissipation structures advected by large-scale flow
• Structures stretched and deformed
• Primary flow direction: azimuthal flow (y-direction) + rising polar component

Temporal Evolution:

• PMT measurement (12 cm diameter average): ~30% fluctuations
• No characteristic frequency correlated with blade passage frequency observed

3. Probability Density Function (Figure 4)

Statistical Analysis:

• 90-second long-duration measurement (Re=2×10⁵)
• Variable: (log(ε) - ⟨log(ε)⟩)/σlog(ε)

Key Findings:

• PDF Conforms to Lognormal Distribution
• Consistent with HIT theory expectations
• Yet measurement location is at turbulent boundary layer, not HIT region

Theoretical Comparison:

• Lognormal distribution in HIT based on cascade processes
• Complex hierarchical structures exist in boundary layer
• Experimental results indicate lognormal property persists in boundary layer

4. Power Spectral Density (Figure 5)

Experimental Spectral Characteristics:

• Small-scale fluctuations enhanced with increasing Re
• Pixel resolution sufficient to capture all relevant structures
• Reaches noise level before kηₖ = 1/ηₖ (assuming inertial range)

Comparison with HIT Theory:

• Refined Kolmogorov Theory Prediction: $\langle\epsilon(x)\epsilon(x+r)\rangle \propto \langle\epsilon^2\rangle(L/r)^\mu$ Corresponding power spectrum: k⁻⁽¹⁻μ⁾, μ ≈ 0.2-0.3
• Experimental Observation: Spectral decay steeper than k⁻⁰·⁸

Anisotropy Analysis:

• Difficult to identify preferred direction
• Uses radial wavenumber k = √(k²ₓ + k²y)

DNS Verification (Johns Hopkins Database)

Simulation Parameters

• Flow Type: Turbulent channel flow
• Reynolds Number: Re ≈ 4×10⁴
• Grid: 2048×1536×512 nodes
• Computational Domain: 8π × 3π × 2
• Boundary Conditions: Periodic in x, y directions; no-slip in z direction
• Observation Window: 20×40 nodes (corresponding to experimental 5.1×5.1 cm²)

DNS Results Comparison

Dissipation Structures (Figures 3c, 3d):

• Experiment and DNS show similar structures
• DNS exhibits more obvious anisotropy (clear mean flow direction)

PDF Comparison (Figure 4):

• DNS fluctuations also conform to lognormal distribution
• Consistent with experimental results

Power Spectrum Comparison (Figure 5 inset):

• Spanwise and streamwise power spectra
• Both directions show decay steeper than HIT prediction k⁻⁰·⁸
• Consistent with experimental observations

Important Conclusion: DNS results confirm experimental findings are not measurement artifacts but genuine characteristics of boundary turbulence.

Experimental Findings Summary

1. Scaling Law Verification: Boundary dissipation rate follows ε∝Re³ without logarithmic correction
2. Local Excess Dissipation: Dissipation rate at impeller height is 2.4 times the average value
3. Large Fluctuations: Relative fluctuations of 50%, independent of Re
4. Lognormal Property: PDF conforms to HIT theory expectations
5. Spectral Characteristic Differences: PDS decay faster than HIT predictions
6. DNS Consistency: Channel flow DNS confirms boundary characteristics

Related Work

DWS Technology Development History

Early Development (1980s)

• Pine et al. (1988)20: PMT measurement of coherent scattered light in simple flows
• Wu et al. (1990)21: Theoretical foundation established
• Bicout & Maynard (1993)22,23: Theory refined

Limitations: PMT sensitive and fast but unable to perform spatiotemporal measurements

Application Extensions

• Complex Fluid Rheology: Durian et al. (1991)24
• Particulate Flows: Menon & Durian (1997)26
• Solid Deformation: Palmer et al. (1999)28
• Foam Research: Cohen-Addad & Höhler (2001)29
• Particle Physics: Le Bouil et al. (2014)30

Prior Work by This Team

Francisco et al. (2023)1:

• First quantitative spatiotemporal measurement with high-speed imaging
• Applied to Taylor-Couette flow
• Verified quantitative accuracy of technique

Turbulent Dissipation Measurement Methods

Traditional Method Comparison

Boundary Layer Turbulence Research

Theoretical Framework

Monin & Yaglom (1981)9:

• Logarithmic correction theory for wall friction
• Applicable to channel and pipe flows

Innovation in This Work:

• Impeller-driven flow shows no logarithmic correction
• Boundary layer characteristics differ from classical wall turbulence

Dissipation Statistics Theory

Kolmogorov Refined Theory (1962)4:

• Intermittency coefficient μ = 0.2-0.3
• Dissipation correlation function scaling law
• Power spectrum k⁻⁽¹⁻μ⁾

Boundary Layer Structure: Robinson (1991)35: Complex hierarchical structures

Turbulent Statistics Research

Lognormal Hypothesis

• Based on cascade processes in HIT
• First verification in boundary layer in this work

Power Spectrum Research

• Sreenivasan & Kailasnath (1993)36: Atmospheric boundary layer spectrum estimation, consistent with theory
• Brachet (1991)38: Taylor-Green vortex DNS, verifies theoretical spectrum
• This Work Finding: Boundary layer spectrum differs from HIT

Conclusions and Discussion

Main Conclusions

1. Technical Breakthrough: DWS achieves first direct spatiotemporal measurement of turbulent boundary dissipation rate with good spatial and temporal resolution even at Re up to 6×10⁵
2. Scaling Relation: Boundary dissipation rate proportional to injected power, following pure power-law scaling ε∝Re³ without logarithmic correction required in pipe flow
3. Local Characteristics: Excess dissipation easily estimated at impeller height (2.4 times)
4. Fluctuation Characteristics:
   • Strong dissipation structure fluctuations (50% relative fluctuations)
   • Conform to lognormal distribution (consistent with HIT)
   • Power spectrum decay faster than HIT predictions
5. Universality: Channel flow DNS confirms measurement characteristics represent universal behavior of turbulent boundaries

Physical Significance

Boundary Dissipation Mechanism

• Boundary dissipation in impeller-driven flow differs from classical wall turbulence
• Bulk dissipation structures dominate boundary layer dissipation
• Statistical weight of viscous sublayer varies with Re (93.6%→2.9%)

Dual Nature of Statistical Properties

• Lognormal Property Preserved: Cascade process effects extend to boundary
• Spectral Characteristics Changed: Complex boundary layer structures alter spatial correlations

Limitations

1. Measurement Location Constraints:
   • Only measures thin layer near boundary (depth ~400 μm)
   • Cannot directly measure bulk dissipation
2. Optical Condition Requirements:
   • Requires transparent or semi-transparent fluid
   • Requires suitable tracer particles (r, λ ≪ l*)
3. Temporal Resolution Constraints:
   • Correlation function convergence requires 0.06 seconds
   • Requires flow evolution timescale sufficiently long
   • Limits types of flows that can be studied
4. Spatial Resolution:
   • Pixel size ~400 μm, larger than minimum scale
   • l*=80 μm crosses Kolmogorov scale within Re range
   • May fail at high Re (l* ≤ ηₖ)
5. Flow Type:
   • This work's flow highly non-uniform
   • Differs from both HIT and classical wall turbulence
   • Universality requires further verification
6. Statistical Convergence:
   • Routine measurements only 2 seconds
   • PDF analysis based on single 90-second measurement
   • Statistical convergence needs improvement

Future Directions

1. Technical Improvements:
   • Increase camera RAM capacity for longer measurements
   • Improve temporal resolution
   • Optimize optical system
2. Flow Type Extension:
   • Apply to other boundary layer flows
   • Study different geometric boundaries
   • Explore higher Reynolds numbers
3. Physical Mechanism Exploration:
   • Dynamics of boundary dissipation structures
   • Relationship with bulk dissipation
   • Boundary layer cascade processes
4. Practical Applications:
   • Immersed object drag prediction
   • Geophysical flow erosion
   • Canopy dynamics
5. Theory Development:
   • Boundary layer dissipation statistics theory
   • Physical origin of lognormal property
   • Mechanism of power spectrum differences

In-Depth Evaluation

Strengths

1. Method Innovation (★★★★★)

• Originality: First direct spatiotemporal measurement of turbulent boundary dissipation rate
• Technical Breakthrough: Successfully extends DWS from simple to fully turbulent flows
• Non-invasiveness: Optical method without flow field disturbance
• High Reynolds Number Capability: Overcomes traditional method limitations, reaching Re=6×10⁵

2. Experimental Design (★★★★☆)

• Dual Measurement System: PMT (average) + high-speed camera (fluctuations) complementary
• In-situ Calibration: Continuous l* monitoring ensures data quality
• Reasonable Scale Design: Large apparatus achieves low characteristic time, satisfying DWS temporal resolution requirements
• Broad Parameter Range: Spans two orders of magnitude in Reynolds number

3. Result Reliability (★★★★★)

• Internal Consistency: PMT and camera measurements agree
• Theory Conformance: Scaling relation verifies zeroth law
• External Verification: DNS data confirms key findings
• Statistical Significance: Long-duration measurements support PDF analysis

4. Scientific Value (★★★★★)

• New Discoveries: Dual nature of boundary dissipation statistics (PDF vs PDS)
• Theory Challenge: Appearance of lognormal property in non-HIT region
• Application Prospects: Provides new tool for boundary layer research

5. Writing Quality (★★★★☆)

• Clear Structure: Logical, well-organized
• Excellent Figures: Effectively convey key information
• Thorough Discussion: In-depth comparison with theory and DNS

Weaknesses

1. Method Limitations

• Spatial Limitation: Only measures thin boundary layer, cannot obtain 3D information
• Fluid Restrictions: Requires specific optical properties
• Time Constraints: Correlation function convergence limits temporal resolution

2. Experimental Setup Defects

• Non-uniform Flow: Impeller-driven flow non-uniformity complicates result interpretation
• Short Measurement Duration: 2-second routine measurements may have insufficient statistical convergence
• Single Long Measurement: Only one 90-second measurement used for PDF

3. Insufficient Physical Interpretation

• Lognormal Origin: Lacks in-depth discussion of cascade processes in boundary layer
• Spectrum Difference Mechanism: No physical explanation for PDS deviation from HIT
• Excess Dissipation: 2.4× factor lacks theoretical explanation

4. Data Analysis

• Anisotropy: Insufficient analysis of x, y direction differences
• Frequency Characteristics: Unexpected absence of blade frequency, lacks discussion
• Re Dependence: Re-independence of some properties (e.g., relative fluctuations) lacks in-depth analysis

5. Universality Verification

• Single Flow: Only one flow configuration studied
• Limited DNS Comparison: Only compared with one Re channel flow
• Lack of Systematic Comparison: No direct comparison with other boundary layer measurement methods

Impact Assessment

Contribution to Field (★★★★★)

1. Opens New Direction: Provides entirely new tool for boundary layer turbulence research
2. Challenges Existing Theory: New discoveries about boundary layer statistical properties
3. Methodological Value: Successful DWS application demonstration

Practical Value (★★★★☆)

1. Engineering Applications:
   • Drag prediction and optimization
   • Mixing and heat transfer processes
   • Erosion and wear assessment
2. Limitations:
   • Requires transparent fluid
   • Laboratory scale
   • High cost

Reproducibility (★★★☆☆)

1. Advantages:
   • Detailed experimental parameters
   • Clear method description
   • Prior work provides additional details
2. Challenges:
   • Expensive equipment (high-speed camera, laser)
   • Large experimental apparatus
   • Requires specialized optical knowledge
   • Complex data processing

Applicable Scenarios

Ideal Application Scenarios

1. Laboratory Basic Research:
   • Boundary layer turbulence statistics
   • Dissipation structure dynamics
   • Theory verification
2. Transparent Fluid Systems:
   • Hydrodynamics
   • Certain chemical processes
   • Biological fluids (with adjustments)
3. Moderate to High Reynolds Number Flows:
   • Re > 10⁴
   • Characteristic time > 0.1 s

Inapplicable Scenarios

1. Opaque Fluids: Liquid metals, concentrated suspensions
2. Extremely High Reynolds Numbers: When l* < ηₖ
3. Rapid Transient Processes: Timescale < 0.06 s
4. Internal Flow Fields: Cannot measure regions far from boundaries

Recommendations for Future Research

Technical Improvements

1. Increase temporal resolution: Faster cameras or higher tracer particle concentration
2. 3D Measurement: Simultaneous multi-angle measurement
3. Automated Data Processing: Real-time feedback

Scientific Questions

1. Systematically study Re effects, particularly l* crossing ηₖ behavior
2. Explore physical origin of lognormal property
3. Establish quantitative relationship between boundary and bulk dissipation
4. Study universality across different flow types

Application Extensions

1. Biological flows (blood, cell suspensions)
2. Microfluidic systems
3. Industrial mixer optimization
4. Environmental flows (requires external field measurement techniques)

Key References

Methodological Foundation

1 Francisco et al., Exp. Fluids 64.9 (2023): DWS application in Taylor-Couette flow
20 Pine et al., Phys. Rev. Lett. 60, 1134 (1988): DWS technology foundation
22,23 Bicout & Maynard, Physica A (1993-1994): DWS theory

Turbulence Theory

4 Kolmogorov, J. Fluid Mech. 13, 82 (1962): Refined theory
9 Monin & Yaglom (1981): Classical statistical fluid mechanics
10 Frisch (1995): Turbulence monograph

DNS Data

2 Li et al., J. Turbulence 9 (2008): Johns Hopkins database
3 Graham et al., J. Turbulence 17(2), 181-215 (2016): Channel flow DNS
39 Turbulent Channel Flow dataset: DOI 10.7281/T10K26QW

Boundary Layer Research

35 Robinson, Annu. Rev. Fluid Mech. 23:601-39 (1991): Boundary layer structure
36 Sreenivasan & Kailasnath, Phys. Fluids A 5(2):512-514 (1993): Atmospheric boundary layer spectrum

Overall Assessment: This is a high-quality experimental physics paper with significant contributions in both methodological innovation and scientific discovery. The successful application of DWS technology opens new pathways for boundary layer turbulence research, and the discovered dual nature of statistical properties (lognormal PDF but non-HIT spectrum) raises interesting physical questions. Despite certain limitations, the paper's scientific value and potential impact are substantial, warranting attention from the fluid mechanics and experimental physics communities.