Comparison of inviscid and viscous vortex shedding from translating and rotating plates

This study compares an inviscid vortex sheet model with Navier-Stokes simulations across approximately 70 unsteady plate motions at moderate Reynolds numbers, demonstrating that the inviscid approach accurately predicts forces and flow structures in body-dominated regimes while showing reduced accuracy at low angles of attack in flow-dominated configurations.

The Gist

Imagine you are trying to predict how a flat piece of cardboard moves through the air when you wave it, spin it, or flick it. To do this perfectly, you need to understand the invisible "swirls" of air (vortices) that form around the edges of the cardboard.

This paper is a giant experiment comparing two different ways of calculating these air swirls:

1. The "Real World" Simulator (Viscous Model): This is like a high-definition, slow-motion video camera that captures every tiny detail, including the friction of the air rubbing against the cardboard. It's incredibly accurate but takes a massive amount of computer power to run.
2. The "Magic Sketch" (Inviscid Model): This is a simplified, super-fast drawing. It ignores the friction of the air and treats the air as if it were perfectly slippery. It assumes that when air hits the sharp edge of the cardboard, it peels off smoothly and instantly, creating a swirl.

The Big Question:
Can this fast, friction-free "Magic Sketch" actually predict the forces on the cardboard as well as the slow, detailed "Real World" simulator?

The Main Discovery: It Depends on Who is Driving the Car

The researchers tested about 70 different ways of moving the plate (waving it up and down, spinning it, flipping it). They found that the answer depends entirely on what is causing the motion.

1. When the Plate is the Boss (Body-Dominated)

Imagine you are holding the cardboard and you suddenly jerk your hand forward or spin it rapidly. The air doesn't have time to decide what to do; it just reacts to your sudden movement.

• The Result: The "Magic Sketch" works amazingly well. It predicts the force on the plate almost perfectly.
• The Analogy: Think of a swimmer doing a sudden, powerful dive. The water splashes exactly where the swimmer's body pushes it. The friction of the water against the skin matters less than the sheer force of the dive. In these cases, the fast model is a reliable shortcut.

2. When the Air is the Boss (Flow-Dominated)

Imagine you hold the cardboard steady and let the wind blow past it, or you move it very slowly. Now, the air itself starts to get chaotic. It forms complex, swirling patterns that detach from the plate and float away on their own.

• The Result: The "Magic Sketch" gets a bit messy. It gets the general idea right, but it starts to drift off course, especially if the plate is moving at a shallow angle.
• The Analogy: Think of a leaf floating in a stream. If you push the leaf (body-dominated), it goes where you push. But if you just let the current carry it (flow-dominated), the leaf starts to spin and wobble in ways that are hard to predict without looking at the tiny eddies in the water. The "Magic Sketch" misses some of these tiny, chaotic details because it ignores the "stickiness" (viscosity) that helps stabilize the real air.

The Secret Sauce: Continuous Shedding

A major hurdle in these "Magic Sketch" models has always been the leading edge (the front tip of the plate).

• The Old Problem: In the past, these models would get confused when air hit the front tip. They would either stop making swirls or create messy, unstable ones, causing the math to crash.
• The New Fix: The authors developed a new rule that allows the air to peel off the front tip smoothly and continuously, just like it does in the real world. They call this "continuous leading-edge shedding."
• Why it matters: This new rule acts like a stabilizer. It stops the math from breaking and allows the "Magic Sketch" to handle complex movements (like spinning plates) much better than before.

The Takeaway

The paper concludes that if you want to quickly test thousands of different ways to move a plate (like designing a robot wing or a drone), the "Magic Sketch" is a fantastic tool—as long as the movement is driven by the object itself.

However, if you are studying how a plate behaves in a steady, chaotic wind where the air is doing the heavy lifting, you still need the slow, detailed "Real World" simulator to get the exact numbers right.

In short: The fast model is a great map for driving on a highway (controlled motion), but for navigating a bumpy, chaotic off-road trail (chaotic airflow), you still need the detailed satellite view.

Technical Summary

Problem Statement
The aerodynamic forces generated by unsteady plate motions, such as those found in insect and bird flight or bio-inspired robotics, are typically governed by complex vortex shedding patterns at moderate Reynolds numbers ($Re \approx 10^2$–$10^4$). While direct Navier-Stokes (NS) simulations and experiments provide accurate data, exploring large kinematic parameter spaces with these methods is computationally prohibitive. Consequently, researchers often turn to inviscid vortex sheet models as efficient alternatives. However, a persistent challenge in these models is the accurate and stable treatment of leading-edge vortex shedding. Traditional inviscid models often halt shedding at the leading edge when flow is directed onto it or rely on empirical parameters like the "leading-edge suction parameter" (LESP) to determine shedding onset. These approaches can introduce numerical instabilities, particularly near-singular integrals, and limit the ability to perform uniform comparisons across diverse unsteady motions. This study addresses the question: To what extent can an inviscid vortex sheet model, specifically one allowing for stable, continuous leading-edge shedding, quantitatively reproduce forces and vortex dynamics from direct Navier-Stokes simulations across a wide range of unsteady plate maneuvers?

Methodology
The authors conduct a systematic comparison between two numerical models for a zero-thickness plate of unit length at $Re = 1000$:

1. Viscous Model: A direct Navier-Stokes solver formulated in the body frame of the plate. It employs a finite-difference method on a non-uniform, staggered MAC grid clustered near the plate tips to resolve inverse-square-root singularities in pressure and vorticity. The solver uses second-order backward differentiation (BDF2) for time stepping and a modified SIMPLE scheme for pressure, validated against benchmark results for impulsively started plates.
2. Inviscid Model: An improved vortex sheet formulation based on previous work [LA25]. This model treats the plate as a "bound" vortex sheet and allows for continuous shedding of "free" vortex sheets from both the leading and trailing edges. Key technical features include:
   • Continuous Leading-Edge Shedding: The model permits shedding even when flow is directed onto the leading edge, avoiding the need for ad-hoc stopping criteria.
   • Regularization: To handle the ill-posed nature of the Birkhoff-Rott equations and logarithmic singularities at the plate edges, the authors employ a "blob-regularized" approach. They use mismatched smoothing parameters and kernels: a smoothed kernel ($K_\delta$) for free sheet advection and a singular kernel ($K$) for the no-penetration condition on the bound sheet.
   • Numerical Stability: A "fencing" technique is used to clip displacements of free sheet points normal to the plate, preventing penetration errors caused by velocity mismatches.
   • Kutta Condition: Enforced numerically by ensuring the vortex strength is continuous across the plate edges, solved simultaneously with the circulation constraint via Kelvin's theorem.

The study examines approximately 70 distinct kinematic configurations, categorized into body-dominated motions (driven by prescribed accelerations, e.g., oscillation, pitch-up) and flow-dominated motions (driven by natural vortex dynamics, e.g., steady translation, uniform rotation).

Key Results
The comparison reveals distinct regimes of agreement between the inviscid and viscous models:

• Body-Dominated Regimes: For motions where fluid dynamics are primarily driven by plate acceleration (e.g., normal oscillation, pitch-up, heaving/pitching), the inviscid model accurately reproduces the time histories of the net normal force and the qualitative structure of the induced vorticity field near the plate. The average relative $L_1$ error for these cases ranges from 11% to 19%. The model captures the formation and shedding of leading-edge vortices effectively, even though it lacks viscous diffusion.
• Flow-Dominated Regimes: For motions with quasi-periodic vortex shedding and little body acceleration (e.g., steady translation at low angles of attack, uniform rotation without background flow), agreement is reduced. The inviscid model often exhibits phase shifts in vortex shedding or differences in the timing of vortex roll-up compared to the viscous case. The maximum $L_1$ error in these regimes can exceed 50%.
• Sensitivity to Parameters: In flow-dominated cases, the inviscid results are sensitive to the smoothing parameter $\delta$ and the angle of attack. At low angles of attack, viscous effects stabilize the flow and delay vortex shedding, whereas the inviscid model may shed vorticity earlier or with different structures.
• Reynolds Number Sensitivity: A limited sensitivity analysis at $Re = 500, 1000, 2000$ for pitch-up motions shows that the error between the models decreases slightly as $Re$ increases, suggesting the inviscid approximation remains robust across this moderate range.

Significance and Claims
The paper claims that the ability of the present formulation to accommodate stable, continuous leading-edge vortex shedding is the critical factor enabling uniform comparisons across diverse motions. By removing the need for model-specific assumptions or empirical parameters to trigger shedding, the authors demonstrate that inviscid vortex sheet models can be reliably used for:

1. Force Prediction: Providing accurate force histories in body-dominated regimes where acceleration drives the flow.
2. Physical Interpretation: Clarifying the vortex-dynamical features of unsteady high-$Re$ flows that can be approximated by inviscid theory.
3. Efficient Exploration: Serving as a computationally efficient tool (orders of magnitude faster than viscous simulations) for exploring large kinematic parameter spaces or control strategies, provided the motion falls within the body-dominated regime.

The authors modestly conclude that while these models are powerful for interpreting force generation mechanisms, their limitations become apparent in flow-dominated dynamics where viscous dissipation and Reynolds-number-dependent effects play a dominant role in determining the precise timing and structure of vortex shedding.