Effect of Expansion Geometry on Turbulence in Axisymmetric Pipe Flows

Using refractive index-matched stereo PIV, this study reveals that gradual ($45^\circ$) pipe expansions generate higher turbulence levels and stronger Reynolds stress anisotropy than abrupt ($90^\circ$) expansions due to geometry-induced modulation of the return flow, which sustains shear layer interaction and turbulence production in the former while confinement limits it in the latter.

The Gist

Imagine you are driving a car down a highway. Suddenly, the road widens. How the road widens matters a lot for how the traffic behaves.

This paper is a scientific investigation into exactly that scenario, but instead of cars, it's about water flowing through a pipe, and instead of traffic, it's about turbulence (the chaotic, swirling mess that happens when fluid speeds up or slows down).

The researchers wanted to answer a simple question: Does it matter if the pipe gets wider all at once (like a sharp step) or gradually (like a gentle ramp)?

Here is the breakdown of their findings using everyday analogies:

The Setup: Two Different Roads

The team built a special water tunnel where they could see inside the pipes clearly (using a trick where the water and the plastic pipe have the same "optical density," like looking through clear glass). They tested two scenarios with the same amount of water flowing:

1. The "Back-Facing Step" (90°): Imagine the road suddenly dropping down a curb. The water hits a sharp corner and has to make a hard 90-degree turn to fill the wider space.
2. The "Wedge" (45°): Imagine the road widening gradually on a gentle ramp. The water slides up the slope before entering the wider section.

The Surprise: The Gentle Ramp is "Messier"

You might think a sharp, sudden change would cause more chaos. But the researchers found the opposite. The gentle ramp (45°) created much more turbulence and energy loss than the sharp step.

Here is why, using a metaphor:

The "Return Flow" Dance

When water hits a sudden expansion, it doesn't just fill the new space; it creates a "recirculation zone." Think of this like a swirl in a bathtub after you pull the plug. The water spins in a big circle, and some of it tries to flow backward toward the source.

• In the Sharp Step (90°): The backward-flowing water hits a vertical wall (the step). It's like a dancer trying to spin but hitting a wall. It gets stuck, creates a small, messy secondary spin, and then bounces off. This interaction is "confined." The chaos stays in a small box.
• In the Gentle Ramp (45°): The backward-flowing water slides smoothly up the ramp. It doesn't hit a wall; it keeps moving with momentum. When it finally reaches the main stream of water coming forward, it slams into it at a sharp angle.

The "Head-On Collision"

This is the key discovery. In the Gentle Ramp case, the backward water (return flow) and the forward water (free-stream) collide with much more force and over a wider area.

• Analogy: Imagine two people running toward each other.
  • Sharp Step: They are running on different floors of a building and barely glance off each other. Not much energy is exchanged.
  • Gentle Ramp: They are running on the same floor and collide head-on. BOOM. That collision creates a huge amount of energy, shaking everything around them.

The Results: What Happens to the Water?

1. More Swirling (Turbulence): Because the "head-on collision" in the ramp case is so violent, it creates a much larger, more chaotic swirl. The water mixes much more intensely.
2. 3D Chaos: In the sharp step, the chaos is mostly flat (2D). In the ramp, the collision is so strong it pushes the water sideways and out of the plane, creating complex 3D swirls.
3. Energy Loss: All that extra mixing and violent collision requires energy. This is why the ramp causes a bigger drop in pressure (energy loss) than the sharp step. It's like the difference between a gentle bump in the road (sharp step) and a massive pothole that shakes your whole car (ramp).

Why Does This Matter?

You might wonder, "If the ramp causes more chaos and energy loss, why do we ever use ramps?"

• The Trade-off: In engineering, sometimes you want that chaos. If you are mixing fuel and air in an engine, you want that violent collision to mix them perfectly so they burn efficiently.
• The Cost: If you just want water to flow through a pipe without losing pressure, the sharp step is actually more efficient (less energy lost).

The Bottom Line

The paper reveals that geometry dictates the dance.

• A sharp step forces the water to separate and reattach in a contained, less energetic way.
• A gentle ramp allows the water to maintain its momentum, leading to a violent, high-energy collision with the main stream, creating a much larger, more turbulent, and "messier" flow field.

The researchers used high-speed cameras and laser lights to film this invisible dance, proving that the shape of the pipe expansion is the "conductor" that decides how wild the water's music will be.

Technical Summary

1. Problem Statement

The study investigates the influence of expansion geometry on flow separation, shear layer development, and turbulence structure in axisymmetric pipe flows. While the flow over a sudden (90°) backward-facing step (BFS) is well-documented, the specific mechanisms governing flow through gradual (sloped) expansions remain poorly understood, particularly regarding turbulent flow at high Reynolds numbers.

A known paradox in fluid mechanics is that gradual expansions (e.g., 45°) often result in higher pressure losses than abrupt (90°) expansions, despite the latter causing more severe flow separation. Previous literature has largely focused on 2D geometries or laminar flows, with limited experimental data on turbulent axisymmetric expansions due to optical challenges (refraction and reflection) in curved pipe walls. This study aims to resolve the fundamental mechanisms dictating why gradual slopes induce higher turbulence and energy losses compared to abrupt steps.

2. Methodology

The researchers employed high-fidelity experimental techniques to overcome optical limitations and capture 3D turbulence statistics.

• Experimental Setup:

  • Facility: A refractive index-matched water tunnel at the Technion.
  • Working Fluid: A 62–63% aqueous solution of Sodium Iodide (NaI) with a refractive index of 1.488, matching the acrylic (PMMA) test section walls. This eliminated optical distortions and reflections, allowing clear imaging through curved surfaces.
  • Geometries: Two expansion configurations with an area ratio of 2.56:
    1. Abrupt (S): 90° step expansion (Backward-Facing Step).
    2. Gradual (W): 45° wedge expansion.
  • Flow Conditions: Experiments conducted at step-height Reynolds numbers ($Re_h$) of 25,000 and 35,000.

• Measurement Technique:

  • Stereo Particle Image Velocimetry (PIV): Utilized two 25-megapixel cameras and a double-pulse Nd:YAG laser to resolve three-component velocity fields ($U_z, U_r, U_\theta$).
  • Data Acquisition: 3,000 instantaneous flow field realizations were captured for ensemble averaging to ensure statistical convergence.
  • Resolution: High spatial resolution (vector resolution of 54 µm for mean fields) allowed for the calculation of Reynolds stresses and turbulence budgets.

• Analysis:

  • Calculation of Turbulent Kinetic Energy (TKE) and Reynolds stress tensors.
  • TKE Budget Analysis: Quantifying production, convection, diffusion, and dissipation terms.
  • Anisotropy Analysis: Using the Reynolds stress anisotropy tensor and Barycentric maps (Lumley triangle) to characterize turbulence structure and dimensionality.

3. Key Contributions

• First High-Fidelity Comparison: Provides the first comprehensive experimental comparison of turbulent flow in 90° vs. 45° axisymmetric expansions using refractive index matching to resolve 3D velocity fields.
• Mechanism Elucidation: Identifies the return flow dynamics as the governing mechanism for turbulence generation. It explains why gradual expansions cause higher losses: the slope allows the return flow to remain attached and impinge obliquely on the free stream, creating a distributed region of high shear.
• Anisotropy Insights: Reveals that gradual expansions induce significant out-of-plane fluctuations ($\langle u'_\theta u'_\theta \rangle$) and shift the turbulence state from axisymmetric contraction toward more isotropic or uniaxial states, unlike the confined turbulence in abrupt steps.

4. Key Results

A. Mean Flow Field and Recirculation

• Abrupt (90°) Case: Generates a secondary vortex near the step corner. This vortex disrupts the return flow, forcing it to separate from the wall and recirculate before reattaching. This limits the interaction between the return flow and the free stream.
• Gradual (45°) Case: The return flow remains attached to the sloped surface and travels upstream along the wall. It impinges obliquely on the free stream near the expansion corner without a secondary vortex disrupting the flow.
• Shear Layer: The shear layer in the gradual (wedge) case expands faster and wider into the free stream compared to the abrupt case.

B. Turbulence Statistics and TKE

• TKE Levels: The gradual expansion consistently yields higher Turbulent Kinetic Energy (TKE) levels, broader shear layers, and enhanced Reynolds stress anisotropy.
• Production ($P_k$): TKE production is significantly higher and more spatially distributed in the wedge case. In the abrupt case, production is confined near the step height due to the localized shear layer.
• Out-of-Plane Fluctuations: The wedge geometry induces strong out-of-plane velocity fluctuations ($u'_\theta$). Close to the corner, $\langle u'_\theta u'_\theta \rangle$ contributes up to 80% of the TKE, driven by the impingement of the return flow on the free stream.

C. Turbulence Budget and Anisotropy

• Budget: The production term ($P_k$) is the dominant driver of the difference. The wedge geometry creates a larger interaction zone between the high-momentum free stream and low-momentum return flow, sustaining turbulence production over a longer distance.
• Anisotropy (AIM & Barycentric Maps):
  • Abrupt: Turbulence remains closer to the axisymmetric contraction limit.
  • Gradual: The Reynolds stress ellipsoid shifts toward a uniaxial (one-component) or more isotropic state due to compressive stresses from the impinging flow. This indicates a "buckling" of turbulence in the flow direction, leading to increased 3D mixing.

5. Significance and Conclusions

• Resolution of the Pressure Loss Paradox: The study confirms that the higher pressure losses in sloped expansions are a direct result of enhanced turbulence production caused by the organized, attached return flow interacting strongly with the free stream. The abrupt step, by contrast, suppresses this interaction via the secondary vortex.
• Generalizable Mechanism: The findings suggest a general mechanism where the expansion slope controls the alignment and momentum of the return flow. A slope that keeps the return flow attached (gradual) maximizes shear and turbulence, while a sharp step disrupts it.
• Engineering Implications: These results provide critical data for validating turbulence models (RANS/LES) in separated flows and offer design insights for systems where flow separation is either detrimental (piping systems, minimizing loss) or beneficial (mixing chambers, combustors, where enhanced turbulence is desired).

In summary, the paper demonstrates that expansion geometry dictates the topology of the return flow, which in turn governs the shear layer interaction, turbulence anisotropy, and overall energy dissipation in axisymmetric pipe flows.