LES against the wind tunnel DrivAer hp-F rear wing
How closely AOX LES predicts the downforce a rear wing adds, compared with Cranfield wind tunnel data.
- Lift change from the 15° rear wing in AOX LES
- −0.262
- The same change measured in the Cranfield wind tunnel
- −0.256
- Bare-body drag in the tunnel, LES vs experiment (uncorrected)
- +1.1%
- One full-size run on a single GPU
- 72 min
Downforce changes from parts are where LES and the wind tunnel agree best, so AOX uses LES to compare part and wing settings.
DrivAer hp-F
A public research car, tested in a wind tunnel and in AOX LES.
DrivAer is a public research car body for aerodynamic studies. The hp-F is its high-performance fastback variant, with rear wing and spoiler options.
The reference is Cranfield University’s open wind tunnel data: a 35% scale model in an 8 × 6 m closed tunnel at 40 m/s, reported without blockage correction.
AOX scaled the model to full size and ran standard-grade LES at 120 km/h and 0° yaw. Every configuration ran on the same fixed grid, so any difference comes from the geometry alone.
Wind tunnel · Cranfield University
- Data
- CERES open repository
- Model
- 35% scale
- Tunnel
- 8 × 6 m, closed
- Speed
- 40 m/s
- Correction
- No blockage correction
AOX LES
- Model
- Scaled to full size
- Speed
- 120 km/h
- Yaw
- 0°
- Grade
- Standard
The downforce the wing adds
Set at the same attitude as in the tunnel: 15° about the bolt axis.
Adding the 15° rear wing changes lift by ΔCl −0.262 in AOX LES. The wind tunnel measured −0.256. A negative change means more downforce.
Without the wing, the bare body’s drag coefficient is 0.3151 in AOX LES and 0.3114 in the tunnel, a difference of +1.1%. Both are tunnel values before blockage correction.
- ΔCl from the 15° wing · AOX LES
- −0.262
- ΔCl from the 15° wing · Cranfield wind tunnel
- −0.256
- Bare-body Cd · AOX LES, uncorrected tunnel value
- 0.3151
- Bare-body Cd · Cranfield wind tunnel, uncorrected
- 0.3114
Colour shows speed relative to the oncoming flow, from 0 (dark blue) to 1.4 (dark red). Both images come from AOX LES on the same grid.
Five wing angles
One fixed grid. Only the wing angle changes.
From 5° to 25°, the downforce the wing adds grows steadily with every step in angle.
| Wing angle | ΔCl · AOX LES |
|---|---|
| 5° | −0.160 |
| 10° | −0.201 |
| 15° | −0.262 |
| 20° | −0.316 |
| 25° | −0.374 |
ΔCl is the lift change from adding the wing, on the same grid as the bare body. The tunnel’s own angle series was measured on a different rig, with a moving belt and 9.6% blockage, so it is not mixed into this table.
How we ran it
Full-size LES on a single GPU.
- Simulation grade used for every run
- Standard
- Per run on one NVIDIA RTX PRO 6000
- 72 min
- Bare-body Cd when the same setup was run again
- 0.3151 = 0.3151
- Wing settings on one fixed grid
- 7
References
- Cranfield University. DrivAer hp-F wind tunnel dataset. CERES open data, 35% scale model, 8 × 6 m tunnel.
- Heft, A., Indinger, T. & Adams, N. (2012). Introduction of a new realistic generic car model for aerodynamic investigations. SAE 2012-01-0168 (DrivAer).