Towards PR-DNS of scour around a wall-mounted cylinder in turbulent open channel flow

This study utilizes particle-resolved direct numerical simulation to demonstrate that a wall-mounted cylinder in turbulent open channel flow generates intense vortical structures and enhanced turbulence that significantly alter local wall shear stress, leading to preferential particle accumulation or depletion in the wake and increased wall-normal transport, effects that are further amplified by the addition of wall roughness.

The Gist

Imagine a river flowing smoothly over a flat, sandy bottom. Now, imagine dropping a large, vertical pole (like a bridge pillar) into that river. What happens to the sand on the riverbed? Does it just sit there, or does the water start digging a hole around the pole? This is the question of scour, and it's a big deal because if the sand washes away too much, the bridge can collapse.

This paper is a high-tech "virtual experiment" run by scientists to understand exactly how water moves sand around such a pole. Instead of building a physical model in a lab, they used a supercomputer to simulate every single water molecule and every single grain of sand.

Here is the breakdown of their study using some everyday analogies:

1. The Setup: The "Digital Sandbox"

The scientists created a virtual river.

• The Water: It's a turbulent, fast-flowing open channel (like a river, not a closed pipe).
• The Pole: A cylinder (a round pole) stands straight up from the bottom, reaching all the way to the surface.
• The Sand: Instead of a thick bed of sand, they started with just a few heavy, round marbles floating near the bottom.
• The Twist: They ran two versions of the experiment:
  1. Smooth Floor: A perfectly flat riverbed.
  2. Rough Floor: A riverbed covered in a layer of fixed pebbles (to mimic a real, rocky riverbed).

2. The Main Character: The "Whirlpool Dance"

When water hits a pole, it doesn't just stop; it gets angry. It creates a chaotic dance of swirling water called vortices.

• The Horseshoe: As the water hits the front of the pole, it splits and wraps around the sides, creating a "horseshoe" shape of swirling water.
• The Wake: Behind the pole, the water is a mess of spinning eddies, like the wake behind a boat.

The computer showed that these swirls are incredibly strong right behind the pole. They act like tiny, invisible vacuum cleaners and fans mixed together.

3. The Sand's Reaction: "The Party and the Empty Zone"

The most interesting part is how the "marbles" (the sand) react to this swirling water.

• The "No-Go" Zone: Right next to the pole, the water is moving so fast and swirling so violently that the sand grains are actually pushed away. It's like a dance floor where the DJ is playing music so loud that no one dares to stand near the speakers. The area immediately around the pole is kept surprisingly clean of sand.
• The "Stripes" of Sand: Behind the pole, the sand doesn't spread out evenly. Instead, it forms long, thin stripes.
  • Some stripes are packed tight with sand (accumulation).
  • Some stripes are empty (depletion).
  • Why? Think of the water swirling behind the pole like a giant, invisible carousel. The spinning motion pushes the sand into specific lanes, creating these striped patterns.

4. The Rough Floor Effect: "The Trampoline"

When the scientists added the layer of fixed pebbles (the rough floor), the story changed slightly.

• On the smooth floor, the sand mostly stays low, just skimming the bottom.
• On the rough floor, the fixed pebbles act like trampolines. When the swirling water hits the roughness, it kicks the mobile sand grains much higher into the water column.
• The Result: The combination of the pole and the rough floor created the most "suspended" sand. Even far away from the bottom, there were more floating grains than in any other scenario. It's as if the rough floor gave the sand a boost, and the pole gave it a lift.

5. The Big Picture: Why This Matters

The scientists found a direct link between how hard the water pushes on the bottom (wall shear stress) and where the sand goes.

• Where the water pushes hardest, the sand is often swept away.
• Where the water swirls in specific ways, the sand gets trapped in those stripes.

The Takeaway:
This study is a first step toward predicting exactly how and where bridges might get undermined by water. By understanding that a pole creates specific "sand-free zones" and "sand-trapping stripes," and that a rocky riverbed makes the sand fly higher, engineers can design better foundations.

In short: The pole creates a whirlwind that clears a circle around itself, paints stripes of sand behind it, and if the riverbed is rocky, it launches the sand into the air like popcorn in a hot pan.

Technical Summary

1. Problem Statement and Motivation

The study addresses the phenomenon of scour, which is the erosion of sediment around hydraulic structures (like bridge piers or cylinders) due to local flow conditions. Scour weakens foundations and can lead to structural failure.

• Current Limitations: Existing numerical modeling approaches (Eulerian two-fluid, Eulerian-Lagrangian with point-particles) rely on closure models for fluid-particle exchange terms derived from averaging, introducing significant uncertainties.
• Objective: To perform Particle-Resolved Direct Numerical Simulation (PR-DNS) of scour around a wall-mounted cylinder. This approach resolves the flow around every individual particle without requiring closure models.
• Specific Goals:
  1. Investigate how a vertical cylinder influences the boundary layer and the motion of heavy, mobile particles in a turbulent open channel.
  2. Determine if particles avoid the cylinder wake and how the cylinder affects vertical particle transport (against gravity).
  3. Assess the impact of wall roughness (simulated by a fixed layer of particles) on vertical transport and entrainment.

2. Methodology and Computational Setup

The authors utilized a custom PR-DNS code based on a fractional step formulation with a second-order finite difference solver and an immersed boundary method (Uhlmann, 2005) for fluid-solid coupling.

• Flow Configuration: Turbulent open channel flow driven by a constant pressure gradient.
  • Geometry: A rigid cylinder (diameter $D_{cyl}$) mounted flush on the wall, spanning the full channel depth.
  • Domain: Periodic in streamwise ($x$) and spanwise ($z$) directions; free-slip at the top surface. The domain length is $90 D_{cyl}$ to ensure wake decay.
• Particles:
  • Mobile Particles: Dilute set of rigid, heavy, spherical particles ($D_p \approx 8$ wall units, density ratio $\rho_p/\rho_f = 1.7$).
  • Fixed Particles (Roughness): In the rough-wall case, a fixed layer of particles simulates a rough bed.
• Cases Simulated:
  1. Reference: Smooth wall, no cylinder (matching Kidanemariam et al., 2013).
  2. Smooth Wall + Cylinder: Includes the cylinder and mobile particles.
  3. Rough Wall + Cylinder: Includes the cylinder, mobile particles, and a fixed layer of particles to simulate roughness.
• Key Parameters:
  • Bulk Reynolds number ($Re_b$) $\approx 3015$.
  • Cylinder Reynolds number ($Re_D$) $\approx 603$.
  • Friction Reynolds number ($Re_\tau$) $\approx 200-210$.
  • Particle Reynolds number ($D^+$) $\approx 7.8-8.5$.
  • Stokes number ($St^+$) $\approx 5.7-6.7$.

3. Key Results

A. Flow Structures and Vorticity

• Horseshoe Vortex System (HVS): A distinct HVS forms upstream and around the sides of the cylinder.
• Wake Dynamics: Intense vortical structures are generated downstream. The time-averaged wake features two quasi-streamwise, counter-rotating vortices near the wall, separated by the symmetry plane. These vortices drive high-speed fluid down toward the wall between them.
• Wall Shear Stress ($\tau_w$):
  • A horseshoe-shaped region of increased wall shear stress surrounds the cylinder, with a maximum amplification factor of 12.15 at an angle of $114^\circ$ and distance $r/D_{cyl} = 0.59$.
  • In the wake, a region of increased shear stress develops along the symmetry plane, flanked by regions of decreased shear stress.

B. Particle Distribution and Entrainment

• Vertical Distribution:
  • Smooth Wall (No Cylinder): 99.15% of particles reside in the second bin above the wall.
  • Smooth Wall (With Cylinder): 93.72% in the second bin, but a significant increase in particle concentration in higher bins. The cylinder enhances wall-normal transport (entrainment).
  • Rough Wall (With Cylinder): Only 33% and 36% of particles are in the 2nd and 3rd bins, respectively. The fraction of entrained particles (far from the wall) is the highest in this configuration. Roughness elements provide an additional lifting mechanism.
• Lateral Distribution (Wake):
  • Near Cylinder: Particles are largely avoided in the immediate vicinity of the cylinder, correlating with the high shear stress region.
  • Wake Region: Two elongated streaks of high particle concentration appear in the wake, separated by a central low-concentration zone. This pattern persists up to $x/D_{cyl} \approx 30$.
  • Mechanism: Near-wall counter-rotating vortices push particles away from the symmetry plane, while outer vortices push them toward it. Beyond $x/D_{cyl} \approx 30$, the near-wall vortices dissipate, causing the streaks to merge into a single high-concentration zone along the symmetry plane.
• Entrainment Locations:
  • Increased entrainment occurs downstream of the cylinder, driven by instantaneous vortical activity rather than mean flow shear stress alone.
  • The rough-wall case exhibits the highest fraction of entrained particles, even at significant heights from the wall.

4. Key Contributions

1. First PR-DNS of Scour with Obstacles: This work represents the initial steps in applying PR-DNS to scour problems involving complex geometries (cylinders), moving beyond continuum or point-particle approximations.
2. Quantification of Cylinder Effects: It demonstrates that a wall-mounted cylinder drastically alters the boundary layer, creating intense vortical structures that significantly enhance the vertical transport of heavy particles against gravity.
3. Roughness Interaction: The study reveals that combining wall roughness with a cylinder creates the most aggressive scour environment, maximizing particle entrainment.
4. Correlation of Shear and Accumulation: It establishes a direct link between the time-averaged wall shear stress distribution and the spatial patterns of particle accumulation/depletion (streaks) in the wake.

5. Significance and Conclusion

The study confirms that PR-DNS is a viable and powerful tool for investigating complex sediment transport phenomena where traditional closure models fail.

• Engineering Implication: The presence of a cylinder significantly increases the risk of scour by enhancing particle entrainment and altering the spatial distribution of sediment. The combination of rough beds and cylindrical obstacles poses the highest risk for foundation erosion.
• Physical Insight: The research highlights that particle entrainment in the near-wake is driven by instantaneous vortex dynamics rather than just mean shear stress, and that the interplay between near-wall and outer vortices dictates the lateral distribution of sediment in the wake.
• Future Work: The authors note that further statistical sampling is required to fully elucidate the detailed mechanisms of individual particle entrainment, particularly in the rough-wall case where statistics away from the wall are currently limited by computational runtime.