Bachelor Thesis - Aerodynamic Evaluation of an FSAE Car

This project encompassed my Bachelor’s Thesis at K.J Somaiya. It covers the end-to-end design, simulation, and manufacture of a complete aerodynamic package for Orion Racing India’s Formula Student Electric car, competing at Formula Student East 2023. The package — front wing, rear wing, bullhorn wing, side diffusers, and rear diffuser — was developed by me and my colleagues Ritwik Deo and Aditya Shah, as our final-year thesis at K.J. Somaiya College of Engineering, Mumbai, under the guidance of Prof. (Dr.) Shailesh R. Nikam.

The primary goal was to maximise downforce while staying within the drag budget, and to achieve an aerodynamic balance of 54% front — a slightly oversteering setup that lap-time simulations on the Hockenheim FSG track confirmed to be optimal.

The complete aerodynamic package: front wing with flow-conditioning devices, bullhorn wing, side and rear diffusers, and a rear wing with slat–main–auxiliary configuration and louvred endplates.

Motivation & System-Level Targets

Aerodynamic downforce improves cornering speed without adding the inertia penalty of extra mass — a well-designed package can generate 4–5× its own weight in downforce at racing speeds. Lap-time simulations using IPG CarMaker on the Hockenheim FSG track showed a 5.2-second improvement (80.75 s vs 85.92 s) when the aerodynamic package was fitted.

Three top-level targets were set before any geometry work began:

Maximum drag coefficient: C_d ≤ 1.62 (0.75× the top-speed-limiting C_d of 2.165, derived from the 55 kW motor and a target top speed of 125 km/h)
Aerodynamic balance: 54% front, tuned to complement the vehicle dynamics team’s slightly oversteering setup
Structural compliance: front wing must fail at ≤ 120 kN crush load per Formula Student rules; carbon fibre tubes verified by 3-point bend test (fracture at 520 N, well above the 200 N maximum in-service load)

CFD Setup & Verification

All aerodynamic simulations were run in OpenFOAM using the k-ω SST turbulence model, chosen for its accuracy in adverse pressure gradients, separated flows, and near-wall regions without requiring additional damping functions.

Simulation domain (42 m × 18 m × 12 m bounding box). Boundary conditions: velocity inlet −16 m/s, moving floor −16 m/s, rotating wheels at 70 rad/s, pressure outlet at 0 Pa gauge.

A systematic mesh independence study fixed the bounding box dimensions and cell edge length at 0.2 m (where residuals plateau). The mesh was then validated against published Ahmed Body experimental data — C_d error within 1.3% and C_L error within 3.2% at 25° slant angle.

Front Wing

Custom airfoils were designed from scratch for both the main element and the outboard section, using the S1223 camber line as a baseline. XFLR5 comparisons at 430 mm chord showed the custom airfoil outperforming all standard candidates (C_L = −1.6, C_d = 0.022 vs S1223’s −1.2 / 0.017).

Final front wing CFD pressure coefficient

Left: final front wing geometry after five design iterations. Right: total pressure coefficient (C_pt) side view showing clean attached flow over the main and auxiliary elements, with an outwash vortex at the endplate keeping tyre wake away from the car body.

The final design features a custom main element, an S1223 inboard ladder (keeps the wing inboard-loaded, reducing overhang bending), and two flow-conditioning vortex generators that produce the outwash effect. The five-iteration CFD-driven process progressively refined endplate geometry, auxiliary element chord, and the dihedral of the outboard section.

Final result: C_L = −1.62, C_d = 0.28, downforce = 248.8 N @ 16 m/s.

Rear Wing

For the rear element a modified Benzing 153-175 airfoil was adopted — camber reduced to 80% of its original value to raise the stall angle and ease manufacturing (fewer MDF stacking layers per mould). This yielded C_L = −1.42 with moderate stall characteristics, compared to the abrupt stall of the S1223.

Final rear wing CFD pressure coefficient

Left: final rear wing with slat–main–auxiliary configuration and louvred endplates. Right: C_pt side view showing how the slat re-energises the boundary layer, allowing the main element to run at a higher angle of attack without separation.

Velocity streamlines showing the counter-rotating vortex pair generated by the endplate louvres, which reduce induced drag at the tip.

The three-element Slat–Main–Auxiliary configuration allows an aggressive main-element angle of attack by re-energising the boundary layer, while the endplate louvres generate counter-rotating vortices that cut induced drag.

Final result: C_L = −1.278, C_d = 0.553, downforce = 196.3 N @ 16 m/s.

Drag Reduction System (DRS)

A servo-actuated DRS was implemented on the rear wing auxiliary element. CFD pivot-position analysis identified the 60% chord point as the optimum — minimising both opening torque (0.095 Nm) and holding torque (−0.175 Nm) to reduce strain on the servo. An AGFRC SA81BVMW smart servo was selected and mounted directly inside the element.

Condition	C_L	C_d
DRS closed (full car)	−3.548	1.495
DRS open (full car)	−2.701	1.036

Opening the DRS reduces drag by ~31% on the full car, with a downforce cost of ~24%.

Diffusers

The diffuser system comprises two side diffusers and one rear diffuser, all operating under ground effect. Key design choices:

3D diffusing sections increase the effective expansion ratio beyond what a 2D profile would allow
Side diffuser inlet positioned to accept only the clean air conditioned by the front wing, keeping tyre wake out of the diffuser passage
Flat wall alongside the side diffuser prevents high-pressure air from striking the rear tyre, reducing tyre wake

Left: final diffuser assembly (side and rear). Right: pressure distribution showing the low-pressure region beneath the diffuser ramp generating downforce via ground effect; the blue suction peak at the rear is visible on the right side.

Final result: C_L = −1.13, C_d = 0.15, downforce = 173.6 N @ 16 m/s.

Full-Car CFD Results & Comparison

Full car pressure plot — top isometric view

Full car pressure plot — bottom isometric view

Full-car pressure distribution at 16 m/s. Top view (left) shows stagnation pressure on the front surfaces; bottom view (right) reveals the large suction regions under the diffusers generating ground-effect downforce.

Compared to the previous year’s design, the 2022–23 package achieved higher downforce with lower drag across all components:

Component	C_L (current)	C_d (current)	C_L (prev.)	C_d (prev.)
Front Wing	−1.62	0.28	−1.07	0.22
Rear Wing	−1.278	0.553	−1.06	0.47
Diffusers	−1.13	0.15	−1.12	0.19
Full Car	−3.548	1.49	−3.35	1.56

Total downforce improved from 514.6 N to 544.9 N while drag fell from 239.6 N to 228.9 N.

Aerodynamic Maps

To understand how the aerodynamics behave dynamically, four sets of CFD sweeps were performed, varying ride height (heave), roll angle, yaw angle, and the combination of front/rear ride height (pitch).

Left: C_L and C_d vs ride height — maximum downforce occurs near 30 mm, with a characteristic diffuser stall at lower ride heights. Right: C_L and C_d vs roll angle — the package is symmetric and well-behaved through ±2°.

C_L, C_d, and side force coefficient vs yaw angle (±5°). Downforce is broadly stable across the yaw sweep, confirming the package does not destabilise the car during cornering.

Aero balance map — front vs rear ride height

3D pitch maps (MATLAB-interpolated from CFD data). Left: total C_L as a function of front and rear ride height — maximum downforce is achieved at low front and intermediate rear ride height. Right: aerodynamic balance (%Front) across the same parameter space, showing how pitch attitude shifts the balance forward or rearward.

Manufacturing

All structural components were manufactured from 300 gsm Carbon Epoxy Prepreg over CNC-machined MDF moulds. Smaller intricate parts (flow conditioners, louvres, outwash vanes) were produced by FDM 3D printing in PLA.

Mould fabrication used a 3-axis CNC mill with Fusion 360 CAM toolpaths (15,000 mm/min feedrate, 10,000 rpm spindle, 0.8 mm stepover for finishing passes). MDF was chosen for its low coefficient of thermal expansion under autoclave conditions and cost-effectiveness over foam board or aluminium.

Carbon fibre prepreg layup on front wing mould

Left: carbon fibre prepreg layup for the front wing main element at Godrej Aerospace Plant 4B. Right: vacuum-bagged assemblies inside the autoclave — cure cycle: 10 min @ 50°C → 10 min @ 80°C → 60 min @ 120°C.

The sandwich layup sequence was: carbon prepreg skin → Nomex honeycomb core → carbon prepreg skin. CF tubes run spanwise inside the wings to resist bending, and aluminium inserts at all mounting points prevent the Nomex core from crushing under load.

Internal structure of wings and diffuser assemblies

Exploded view of the internal structure across the full aero package. CF tubes, foam inserts, and aluminium mounting inserts are visible inside both wing and diffuser assemblies.

Conclusions

The 2022–23 aerodynamics package met all primary objectives:

Total downforce of 544.9 N at 16 m/s — a 6% improvement over the previous year, with a simultaneous 4.5% drag reduction
Aerodynamic balance of ~50–54% front, consistent with the vehicle dynamics target
A functional servo-actuated DRS reducing full-car drag by ~31% on straights
Full compliance with Formula Student rules (geometry, TSAL visibility, crush load)
Validated CFD methodology (k-ω SST, Ahmed Body benchmark within 3.2%)

Identified areas for future improvement include the diffuser efficiency at low Reynolds numbers, and further integration of the bullhorn wing geometry with the rear wing slat to reduce upwash interaction losses.