Case study · Automotive
FSAE AERO
External aerodynamics CFD with Argus & ANSYS Fluent
Tools
Argus · ANSYS Fluent 2026 R1 · SpaceClaim
Outcome
50% engineering time saved
Case study · Automotive
What’s inside
A walk-through of how Argus took an FSAE external-aero CFD study from ~12 hours of manual work down to ~6: mesh, setup, results and design takeaways.
Summary
A full external-aero CFD workflow, delivered in half the time.
A Formula Style (FSAE) open-wheel race car was simulated at 20 m/s in ANSYS Fluent, driven through Argus. End-to-end engineering time dropped from ~12 hours to ~6 hours, and the agent reduced the mesh to the minimum number of nodes needed to find convergence, from 5.44M cells down to 799,538, without losing physical fidelity. Final: Cd = 1.145, Cl = −1.794 (downforce ~439 N).
50%
Engineering time saved
85%
Reduction in mesh size achieved
1.57
L/D ratio predicted (downforce/drag)
Problem
Open-wheel aero is a high-cell-count problem, and these constraints compound for any team running on a constrained workstation.
01
Best-practice sizing
Hitting both the cell budget and a log-law y+ band usually takes a senior CFD engineer several iterations.
02
Post-processing overhead
Building velocity, pressure, wall-shear, y+ and wake contours manually is 4+ hours of GUI work per case.
+ Benefit to engineer
Minimum-fidelity mesh without losing k-ω SST log-law physics · auto-generated contours with embedded flow-physics insight · design recommendations quantified by ΔCd / ΔCl impact.
Methodology
Half-car FSAE geometry imported live from SpaceClaim. Argus proposed mesh settings, predicted cell counts before running the mesher, iterated through mesh convergence and proposed targeted reductions, turning an open-ended blocker into a 5-step problem.
FINAL SURFACE MESH · 2 PRISM LAYERS · MAX SKEW 0.69
Iter | Surf min | BOI near / far | BL | Cells |
|---|---|---|---|---|
1 | 0.5 | 16 / 64 | 10 | ~5.44 M |
2 | 0.5 | 24 / 64 | 4 | 3.68 M |
3 | 2 | 32 / 128 | 3 | 1.46 M |
4 | 2 | 48 / 132 | 2 | 1.35 M |
5 | 4 | 64 / 256 | 2 | 799 K |
+
Argus proposed setup with best-practices
Turbulence model (k-ω SST), inlet turbulence quantities (TI 5%, VR 2), reference frontal area (0.4997 m²) and rotating-wheel BCs all proposed by Argus.
+
Quantified design guidance
Prioritized L/D wins with ΔCd / ΔCl estimates: aggressive diffuser ramp (ΔCd −0.05, ΔCl −0.30), wheel turning vanes (ΔCd −0.05, low effort), targeting Cd ~ 0.95 next iteration.
Results
Phase | Manual | With Argus | Savings |
|---|---|---|---|
Mesh generation | ~6 h | ~3 h | −50% |
Case setup (BCs, turbulence, references) | ~2 h | ~1 h | −50% |
Post-processing + report preparation | ~4 h | ~2 h | −50% |
Total | ~12 h | ~6 h | −50% |
Every contour below was generated with Argus, no manual scripting or clicking in Fluent. After constructing results, the engineer asked the agent for design recommendations, and it paired each visualization with flow-physics commentary that translated directly into design actions.
Streamlines · front-wing inlet
Ground-effect speed-up at 24–32 m/s, strong upwash over chassis, coherent bundle through the wing channel, consistent with the negative Cl.
Wall shear stress
Highest WSS on rotating tire shoulders (doubled effective slip velocity), confirms rotating-wheel BCs active. Wheels dominate viscous drag.
Velocity, symmetry plane
Bow-wave acceleration above the helmet (~30–35 m/s), wake deficit drops to ~10–18 m/s, ~3× car length downstream.
Wake envelope · total pressure
Iso-surface at >60% P0 loss. Rear plume from wings, chassis and wheels is the dominant Cd driver. Underbody stays clean, attached, high-energy.
Agent Design Input
After the run, the engineer asked Argus to suggest design changes to reduce drag and increase downforce. The agent paired each recommendation with a ΔCd / ΔCl estimate and an effort tag, and led with a mesh-fidelity caveat so the suggestions are read as directional, not absolute.
!
Mesh-physics caveat, flagged by the agent first
Current mesh has 2 prism layers, y+ mostly 30–250, no wake refinement behind wheels/rear-wing. Skin friction is captured cleanly; wake and separation regions carry ±10–15% uncertainty on pressure drag. Treat suggestions as directional, and confirm winners with A/B comparisons on the same mesh.
Quick-win priority list · best L/D gain per effort
# | Change | Est. ΔCd | Est. ΔCl | Effort |
|---|---|---|---|---|
1 | Front-wing 2nd element + endplate curl | −0.02 | −0.20 | Medium |
2 | Aggressive diffuser + floor strakes | −0.05 | −0.30 | High |
3 | Front-wheel turning vanes | −0.05 | 0 | Low |
4 | Rear-wing Gurney + slight incidence reduction | −0.03 | −0.10 | Low |
5 | Roll-hoop fairing + boat-tail engine cover | −0.04 | 0 | Medium |
A
Drag: wake is the dominant Cd source
Wheels are the single biggest contributor: add fairings / partial shrouds and cover unused brake ducts. Reduce rear-wing AoA by 1–2° or swap incidence for a Gurney flap. Energize the base wake with a beam wing or boat-tailed engine cover.
B
Downforce: load the underbody
Add a 2nd-element flap to the front wing and curl the endplate footplate to strengthen the leading-edge vortex. Steepen the diffuser ramp (5–10°), add floor strakes to seal the underbody, drop ride height 5–10 mm if mechanically feasible.
→
Before re-running CFD on any redesign, fix the mesh
Increase prism layers to 5–8 · refine the wake region behind car and front-wing endplates · target y+ in the 30–80 band globally. Stays inside the 1M cell budget while removing the ±10–15% pressure-drag uncertainty.
Think through > click through.
Let Argus handle the rest.
Your AI agent for engineering workflows
Want to try Argus on your own CFD workflow?
Visit www.cosmon.com to start a trial.
www.cosmon.com
