Skip to main content
Hemyar Alharazi

Technical note · Aerospace

CFD for Small UAV Wings at Low Reynolds Numbers: A Verification and Validation Checklist

Abstract

Small UAV wings often operate at chord Reynolds numbers where laminar separation bubbles and transition dominate the aerodynamics — exactly the regime where default CFD settings mislead. This note gives a practical verification and validation checklist: choosing a turbulence and transition approach, meshing for the boundary layer, demonstrating mesh independence, and comparing against trustworthy reference data before any lift or drag figure is quoted.

Why low Reynolds numbers are different

The chord Reynolds number

Re=ρVcμRe = \frac{\rho V c}{\mu}

for a small UAV wing is frequently in the range where the boundary layer stays laminar over much of the surface, separates, transitions in the free shear layer and may reattach as a laminar separation bubble. The bubble changes the effective shape of the aerofoil, so lift, drag and especially the drag polar near stall depend on predicting it correctly.

A fully turbulent RANS model assumes transition at the leading edge. At these Reynolds numbers that can overpredict drag and miss hysteresis entirely — while still producing a smooth, convincing-looking result.

Verification: solving the equations right

  1. Near-wall resolution. Resolve the viscous sublayer (first-cell y+ around 1) when using a transition model; wall functions discard exactly the physics that matters here.
  2. Growth rate. Keep boundary-layer cell growth gentle so the separation and reattachment region is captured.
  3. Mesh independence. Run at least three systematically refined meshes and show that the quantities you intend to report settle. Report the refinement ratio and the change between the last two levels.
  4. Iterative convergence. Monitor the forces themselves, not only residuals. Unsteady bubbles may need a time-accurate solution.
  5. Domain size. Place far-field boundaries far enough away that they do not constrain the flow.

Validation: solving the right equations

Compare against experimental data for the same aerofoil at a similar Reynolds number and turbulence level, from a published wind-tunnel campaign. Match the case, not just the shape: Reynolds number, angle-of-attack range and free-stream turbulence all move transition.

If no comparable data exist, say so and treat the results as trends rather than numbers.

Choosing the model

  • A transition-sensitive RANS model (for example a γ–Reθ family model) is a reasonable default for steady polars.
  • Where the bubble is unsteady or the angle of attack approaches stall, consider time-accurate simulation and, for key cases, higher-fidelity methods.
  • Panel-method tools with boundary-layer coupling remain useful for fast, early design sweeps — and as a sanity check on CFD.

What to report

State the model, mesh levels, y+, convergence criteria, reference data and the difference from it. A lift coefficient quoted without these is a number; with them it is evidence.

Share