High-Reynolds-number turbulent boundary layers under adverse pressure gradients. Part 1. Decoupling local and upstream pressure gradient effects

This study utilizes high-Reynolds-number wind tunnel experiments with controlled pressure gradients to demonstrate that while the von Karman coefficient in the logarithmic law remains invariant, the additive coefficient varies systematically with both local adverse pressure gradients and upstream pressure-gradient history, which energize motions in the wake region without penetrating the inner layer.

The Gist

Imagine you are driving a car down a long, straight highway. The air rushing past your car is like a Turbulent Boundary Layer (TBL)—a chaotic, swirling layer of air that clings to the surface of the car.

Usually, scientists study what happens when the road is perfectly flat and the wind is steady. This is called a "Zero Pressure Gradient" (ZPG) flow. It's the "textbook" scenario, and we have a very simple, famous rule (the Log-Law) that predicts exactly how fast the air is moving at different heights above the road.

However, real life isn't flat. Sometimes you drive up a hill (an Adverse Pressure Gradient or APG). The air has to work harder to climb, and the turbulence changes. For decades, scientists have been arguing about how to update that simple rule to account for these hills.

The Big Problem: The "Memory" Effect
The tricky part is that the air doesn't just react to the hill you are on right now. It also remembers the hills you drove over yesterday.

• Local Effect: The hill you are currently climbing.
• History Effect: The bumps and dips you encountered before you got here.

In previous experiments, it was impossible to separate these two. If you changed the hill, you also accidentally changed the path you took to get there. It was like trying to figure out if a runner is tired because of the current steep slope or because they ran a marathon the day before. You couldn't tell which was which.

The Experiment: The "Wind Tunnel Time Machine"
The researchers at the University of Melbourne built a special wind tunnel that acts like a time machine for air. They used a series of adjustable "bleed slots" (like tiny windows in the ceiling) to control the pressure.

They designed a clever experiment with two groups of air:

1. The "Clean" Group: These air particles traveled on a perfectly flat road (Zero Pressure Gradient) for a long time, then suddenly started climbing a moderate hill.
2. The "Perturbed" Group: These air particles took a bumpy road first (a small, controlled hill), then went back to a flat road for a while, and then started climbing the exact same moderate hill as the first group.

Crucially, by the time they reached the measurement point, both groups were in the exact same spot, facing the exact same hill, with the exact same speed. The only difference was their history.

The Findings: What the Air Told Us

1. The "Golden Rule" (Von Kármán Coefficient) Stays the Same
The researchers checked the "Golden Rule" of the air flow (the slope of the log-law). They found that it didn't change, no matter the hill or the history. The fundamental shape of the flow remained universal. It's like the engine of the car still runs the same way, regardless of the road.

2. The "Offset" (Additive Coefficient) Changes
However, the position of the flow shifted.

• The Hill Effect: The steeper the hill, the more the flow shifted down.
• The Memory Effect: The air that took the bumpy road earlier ended up in a slightly different position than the air that took the smooth road, even though they were on the same hill now.
• Analogy: Imagine two people walking up a ramp. One walked up a ramp earlier today, and the other didn't. Even if they are now on the same ramp, the one who walked earlier is slightly more tired (shifted position) than the fresh one.

3. Big Swirls vs. Small Swirls
The researchers looked at the "swirls" (eddies) in the air.

• Small Swirls (Near the ground): These are like tiny dust motes. They react instantly to the current hill and forget the past immediately. They don't care about the history.
• Big Swirls (Higher up): These are like massive weather systems. They react slowly. If you drove over a bumpy road earlier, these big swirls carry that "memory" for a long time, affecting the air flow even after you've been on a smooth road for a while.

Why Does This Matter?
For a long time, scientists thought the "Log-Law" was broken or inconsistent when applied to real-world flows (like on airplane wings or wind turbines). This paper proves that the law isn't broken; we just didn't account for the memory of the air.

The Takeaway
To predict how air flows over complex shapes (like a whole airplane wing), we can't just look at the shape at one specific point. We have to understand the entire journey the air took to get there.

The researchers have now provided a "high-fidelity map" that separates the immediate hill from the road history. This allows engineers to build better models for designing more efficient airplanes, wind turbines, and cars, ensuring they handle the wind exactly as nature intended.

In a nutshell: The air has a memory. If you want to predict how it behaves, you can't just look at where it is now; you have to know where it's been.

Technical Summary

1. Problem Statement

Turbulent Boundary Layers (TBLs) subjected to Adverse Pressure Gradients (APGs) are critical in engineering applications (e.g., airfoils, diffusers). Their behavior is governed by three factors:

1. Local Friction Reynolds number ($Re_\tau$): Describes the scale of the flow.
2. Local Pressure Gradient ($\beta$): The immediate streamwise pressure gradient.
3. Pressure Gradient (PG) History: The upstream flow conditions experienced by the TBL before reaching the measurement location.

The Core Issue: In previous studies, these factors were often coupled. It was difficult to determine whether deviations from classical scaling laws (specifically the logarithmic law for mean velocity) were caused by the local APG strength or the upstream PG history. This ambiguity has led to conflicting reports in the literature regarding the universality of the von Kármán coefficient ($\kappa$) and the additive coefficient ($B$) in the log-law equation:
$$U^+ = \frac{1}{\kappa} \ln(z^+) + B$$
Furthermore, many previous studies were conducted at low Reynolds numbers or under strong APGs, which compress the "overlap region" where the log-law applies, making rigorous evaluation difficult.

2. Methodology

The authors utilized a unique experimental facility at the University of Melbourne to systematically decouple these effects.

• Facility: A high-$Re_\tau$ wind tunnel with a 27m test section and adjustable air-bleed slots on the ceiling.
• Control Strategy: By selectively opening and closing air-bleed slots and adjusting outlet blockage, the researchers could prescribe specific pressure coefficient ($C_P$) profiles along the test section.
• Experimental Cases:
  • Reference Cases: Zero-Pressure-Gradient (ZPG) and APG cases with minimal PG history (strictly ZPG upstream of the APG development region).
  • Perturbed Cases: Cases with controlled upstream PG perturbations (a weak upstream APG followed by a long relaxation region) but matched local conditions ($Re_\tau$ and $\beta$) downstream.
  • Comparison: The study compared cases with identical local $Re_\tau$ and $\beta$ but distinct upstream histories.
• Measurements:
  • Hot-wire anemometry: High-frequency sampling to resolve turbulence statistics and energy spectra.
  • Oil-Film Interferometry (OFI): Independent, direct measurement of friction velocity ($U_\tau$) to ensure accurate normalization of data.
  • Reynolds Numbers: The study focused on high $Re_\tau$ ($\approx 4,000 - 10,000$) and low-to-moderate $\beta$ ($\approx 0 - 1.5$), ensuring a well-defined overlap region.

3. Key Contributions

1. Decoupling Mechanism: The first systematic experimental demonstration of how to isolate PG history effects from local APG effects in high-$Re_\tau$ TBLs.
2. High-Fidelity Dataset: Provided a dataset with sufficient scale separation to distinguish between inner, overlap, and wake regions, and between large-scale (LSM) and small-scale motions.
3. Resolution of Log-Law Coefficients: Rigorously tested the universality of $\kappa$ and $B$ under varying APG and history conditions.
4. Scale-Dependent Response: Revealed how different scales of turbulence (inner vs. outer, small vs. large) respond differently to local vs. historical pressure gradients.

4. Key Results

A. Log-Law Coefficients ($\kappa$ and $B$)

• Von Kármán Coefficient ($\kappa$): Found to be invariant across all cases (ZPG, APG, and varying histories) within experimental uncertainty. The value remained consistent with the canonical $\kappa \approx 0.39$.
• Additive Coefficient ($B$): Found to vary systematically.
  • $B$ decreases as the local APG strength ($\beta$) increases.
  • $B$ is also sensitive to PG history. Cases with different upstream histories but identical local $\beta$ showed different $B$ values.
  • Significance: This explains the variability in reported $B$ values in previous literature, attributing it to unaccounted PG history effects and low Reynolds numbers.

B. Mean Velocity Profiles

• Relaxation Region: Upstream APG perturbations caused the mean velocity profile in the wake region to lie above the ZPG profile, while the overlap region was slightly below. However, after a sufficiently long relaxation length ($\approx 26\delta$), the mean velocity profile recovered the classical ZPG scaling, though the boundary layer thickness and growth rate remained elevated.
• Development Region: Even with matched local $Re_\tau$ and $\beta$, the "Perturbed" case (with upstream history) did not collapse with the "Reference APG" case (minimal history). The Perturbed case showed a "lag" in developing the full APG signature.

C. Turbulence Statistics and Energy Spectra

• Inner Region ($z^+ < 20$): Small-scale motions were largely unaffected by PG history. They responded rapidly to local conditions.
• Wake Region ($z \approx 0.4\delta$): Both local APG and PG history energized large- and small-scale motions.
• Overlap Region ($z \approx 0.25\delta$): PG history primarily influenced large-scale motions (LSMs).
  • Local APGs energized motions across all scales in the wake.
  • PG history left a "memory" in the large-scale motions extending into the overlap region.
  • The Perturbed case showed a "lag" in the energization of APG-related scales compared to the Reference APG case, with the peak energy growth shifted closer to the wall.

D. Scale Separation

The high $Re_\tau$ allowed the authors to confirm that PG history effects are confined to the overlap and wake regions. Unlike low-$Re_\tau$ studies where history effects appeared to penetrate the inner region, this study showed that near-wall small-scale motions remain robust against upstream history variations.

5. Significance and Implications

• Fundamental Physics: The study establishes that the classical log-law universality holds for $\kappa$ but not for $B$ in non-canonical flows. The additive constant $B$ is a function of both local $\beta$ and PG history.
• Modeling: Current models relying on a single parameter to describe PG history effects are insufficient. The authors propose that at least two independent parameters are needed: one for the inner/overlap region (affecting $B$) and one for the wake region (affecting wake strength $\Pi$ and profile shape).
• Experimental Design: Highlights the critical need for long relaxation regions and controlled upstream histories in TBL experiments to ensure that observed statistics reflect local conditions rather than facility-specific history artifacts.
• Future Work: This paper (Part 1) provides the physical basis for Part 2, which will introduce a new composite mean velocity profile formulation that explicitly incorporates local pressure gradients and PG history effects.

In summary, this research provides a high-fidelity, controlled experimental validation that local pressure gradients and upstream history act as distinct, decoupled mechanisms influencing TBL statistics, with history effects primarily manifesting as a shift in the additive log-law constant and a modification of large-scale motions in the outer layer.