Roadster Aerodynamics

The Brief

Roadsters lose aerodynamic efficiency because the open cockpit disrupts airflow and enlarges the wake. The aim was to improve high-speed stability by adding downforce without a large drag penalty, while keeping a realistic, attractive form.

Goals

Generate downforce for high-speed stability
Keep the drag penalty small
Maintain a realistic, attractive roadster form

Research

Benchmarking

Features that improve the aerodynamics
Features that could be improved

Porsche Boxster

The Porsche Boxster is benchmarked as a modern two-seat roadster, representing a higher-performance reference for open-top sports-car aerodynamics.

0.43Roof down C_d

Annotated side, front and rear line drawings of a Porsche Boxster highlighting aerodynamic features: the raked windshield, recessed intake, front air inlets, rear lip spoiler and rear diffuser as strengths, and the sudden roof termination, short rear overhang, exposed wheels, large frontal area and rounded nose as areas to improve

Mazda MX-5

The Mazda MX-5 is benchmarked as a widely used, lightweight roadster, providing a representative baseline for typical open-top road-car aerodynamics.

0.53Roof down C_d

Annotated side, front and rear line drawings of a Mazda MX-5 highlighting aerodynamic features: the wheel deflector, smooth side surfaces, rear lip spoiler, compact frontal area, rounded bumper corners and rear bumper aero outlets as strengths, and the steeper windshield transition, sudden roof termination, antenna, short rear deck and cooling intake as areas to improve

Comparison

Benchmarking the Porsche Boxster and Mazda MX-5 highlights two contrasting aerodynamic approaches. The Boxster uses active rear lift reduction and underbody shaping to improve high-speed stability, while the MX-5 relies on compact proportions and clean body surfaces to minimise drag through lightweight simplicity. Across both vehicles, drag is mainly driven by rear wake behaviour rather than cabin flow alone. Aerodynamic performance depends on controlling rear separation, with front underbody entry and wake management offering the greatest design impact.

Rear Spoiler Optimisation

Why use a rear lip spoiler?

Open-top roadsters tend to generate rear lift. A rear lip spoiler fixes where the flow separates at the rear edge, reducing lift and improving high-speed stability. Kept small, it adds downforce with very little extra drag.

Audi TT Benchmark

The first-generation Audi TT was fitted with a rear spoiler after reports of high-speed instability. The added lip reduced rear lift and improved stability at motorway speeds.

Three sketches comparing airflow behind a car with no spoiler, a small lip spoiler and a large spoiler — No spoiler Small lip spoiler: less drag Large spoiler

What affects performance

Height: how strongly the flow is redirected
Edge sharpness: where separation occurs
Kick-up angle: the balance between downforce and drag

Small to moderate spoilers can stabilise the wake and improve stability with minimal drag penalty, which is why a lip spoiler was chosen over a larger wing.

Splitter

Why use a splitter?

A front splitter controls how air enters the underbody. By restricting high-pressure air from flowing beneath the car, it raises pressure on top of the splitter and lowers pressure underneath. This creates front downforce and reduces front lift.

Sketch of airflow around a car, showing slower high-pressure air above and faster low-pressure air beneath the splitter

Porsche 911 Splitter Benchmark

Front splitter for a Porsche 911, shown alone and fitted to the car

Underbody flow restriction: limits high-pressure air entering beneath the car, lowering front lift.
Clear separation edge: defines where airflow detaches at the bumper, improving pressure control.
Low-profile integrated design: keeps ground clearance while generating efficient front downforce.

Front-End Pressure Management

Front-end shape controls how stagnation pressure forms at the nose and how smoothly air accelerates over the bonnet. Abrupt geometry creates strong pressure gradients and early separation, while gradual curvature spreads pressure more evenly and delays separation.

Flat front

Creates a concentrated stagnation region and abrupt separation, producing a large wake and high pressure drag.

Circular front

Smooths the stagnation region and delays separation compared with a flat face, but still produces a relatively large wake.

Teardrop

Spreads pressure gradually and keeps flow attached, minimising wake size and reducing drag.

Rear Deck Pressure Recovery

Rear-end geometry determines how smoothly pressure recovers before the flow separates. A gradual taper reduces wake size and pressure drag, while an abrupt cut-off causes early separation and a large low-pressure wake.

Volkswagen XL1 in a wind tunnel, with smoke streamlines following its long tapered rear

Volkswagen XL1 Benchmark

The XL1's long, gradual taper is near-ideal for minimising pressure drag.

Airflow slows down gradually before leaving the car
Flow stays attached longer, so separation happens further back
A smaller wake and higher rear pressure reduce pressure drag

Sketch Development

Concepts explored nose shape, windscreen angle, splitter size and rear length, with each iteration responding to the limitations of the last.

Roadster concept sketch 1 of 5 — 3-wheel teardrop with a rounded nose. Rejected, as it's less stable than a 4-wheel car.

Roadster concept sketch 2 of 5 — Larger splitter and prominent spoiler. The windscreen angle didn't follow the nose, risking pressure build-up.

Roadster concept sketch 3 of 5 — Rear hump and small spoiler with more realistic proportions. A larger hump risked adding frontal area.

Roadster concept sketch 4 of 5 — Single smooth bonnet-to-windscreen line, but the blunter nose and short rear limited pressure recovery.

Roadster concept sketch 5 of 5 — Longer rear and more angled front, for gentler pressure recovery and a smaller wake.

Final Design

The final iteration best balances stable downforce with low pressure drag for an open-top roadster. The smooth bonnet-to-windscreen line reduces front pressure, while the splitter and spoiler give front and rear stability. A longer rear allows more gradual pressure recovery, helping produce a smaller, more organised wake.

CAD

Red render of the final roadster design from the front three-quarter view

1:17Scale
290 mmLength
110 mmWidth
55 mmHeight

Side, back, front and top view renders of the roadster

Fusion 360 Form model wireframe overlaid on the orthographic reference sketches

Modelling Process

The car was modelled in the Fusion 360 Form workspace to achieve smooth, continuous curvature. Orthographic sketches were drawn first and used as references to build the body around.

I kept a four-sided (quad) polygon topology wherever possible. This reduced surface irregularities and helped prevent meshing issues in the CFD.

The frontal area (6,129 mm² at model scale) was measured in Fusion 360 for use in the drag coefficient calculation.

CFD Setup

The aerodynamics were simulated in ANSYS Fluent at 30.5 m/s (68 mph) to find the drag and lift on the car.

Half of the roadster model cut along its centre line, with the symmetry plane shown as a transparent sheet

Side of the roadster model with two magnified circles highlighting short edges in red, near the windscreen and the front wheel arch

Six views of the box-shaped enclosure around the car, each highlighting one boundary: symmetry plane, walls, pressure outlet, stationary road, velocity inlet and the no-slip car surface

Mesh around the roadster on the symmetry plane, with polyhedral cells that get smaller close to the car

Triangular mesh around the roadster, very dense close to the car and coarser further away

01 Geometry

The car is symmetric, so only half was modelled, halving simulation time. The frontal area was halved to match.

02 Short edges

Six short edges were found and repaired before meshing. Left in, they create tiny mesh cells that increase cell count and can cause poor convergence.

03 Enclosure and boundary conditions

A box of air was built around the car, extending five car lengths behind it so the wake could fully develop. The blockage ratio was 3.37%, under the 5% limit, so the walls didn't affect the flow. I set a velocity inlet, pressure outlet, symmetry plane, stationary road and a no-slip wall on the car.

04 Surface mesh

A local sizing of 1 mm was applied around the car to capture sharp flow changes at the nose, over the roof curvature and at the rear, with cells growing coarser further away.

05 Volume mesh

The enclosure was filled with 3D cells, with a boundary layer on the car's surface for near-wall accuracy. The final mesh had 484,775 cells.

Solver Setup

Solver

Steady pressure-based solver, since the aim was one converged drag value at constant speed.

Turbulence model

k-epsilon, to capture the turbulent mixing that drives the wake and overall drag.

Numerics

SIMPLE algorithm with second-order upwind discretisation for accuracy in fast-changing regions like the wake.

Results

3D velocity streamlines flowing around the whole roadster, fastest over the nose and bonnet and slowing into the wake behind it

C_d 0.398In the typical range for open-top roadsters
C_l −0.135Moderate downforce, comparable to a Porsche 911 (997)
~136 NDownforce at 68 mph, scaled to full size
|C_l|/C_d 0.34Within the performance road car range

What worked

Flow stays attached over the bonnet and curved sides
The splitter creates a fast underbody jet, generating front downforce
The spoiler turns the flow upward, generating rear downforce

Limitations

Rear flow separation creates a large wake, adding drag
Air recirculates inside the open cockpit, wasting energy
Side inflow into the cabin feeds that recirculation

Flow Visualisations

2D pressure plot of the car's side, with high pressure at the nose and a smooth gradient over the bonnet and windscreen — A strong high-pressure region forms at the nose, with a smooth pressure gradient over the bonnet and windscreen.

Close-up 2D pressure plot at the front splitter, with high pressure ahead of it and low pressure beneath — Air stagnates at the front of the splitter, then speeds up beneath it, creating low pressure and front downforce.

Close-up 2D pressure plot inside the open cabin, with pressure rising at the rear cockpit wall — Pressure rises at the rear cockpit wall as recirculating air slows, showing energy lost inside the cabin.

Convergence

The drag and lift coefficients settled quickly and stayed stable, and all residuals levelled off at around 200 iterations, showing the solution had converged. The run was set to 500 iterations but stopped at around 200 once the convergence criteria were met.

Drag coefficient plot over 200 iterations: it swings between about 43 and minus 19 in the first few iterations, then settles flat at 0.398 by around iteration 20 — Drag coefficient settled at 0.398.

Lift coefficient plot over 200 iterations: it starts near minus 14, rises quickly, then levels off gradually at minus 0.135 — Lift coefficient settled at -0.135.

Fluent residuals plot over about 200 iterations, with the velocity residuals falling to around 10 to the minus 5 and the continuity and turbulence residuals levelling off around 10 to the minus 3 — Velocity residuals fell to around 10⁻⁵, with continuity and turbulence residuals at 10⁻³, acceptable for a steady k-epsilon simulation.

Proposed Refinements

Based on the CFD results, I refined the CAD model to tackle the two main issues: rear flow separation and air recirculating in the cockpit. These changes were not re‑simulated, so their effects are expected rather than tested.

Side view of the refined roadster with streamlines over the raised, extended rear — Expected streamlines over the refined rear. Move your cursor across to follow the flow.

Raised rear hump

A higher hump should guide the upper flow to follow the body for longer, delaying separation, shrinking the wake and improving rear pressure recovery.

Extended rear

A longer rear gives a more gradual reduction in area, lowering the pressure gradient at the back. This should delay separation and reduce the wake.

Cabin-edge lips

Small lips along the cabin edges act as flow deflectors, steering air down and rearward instead of into the cockpit. Less side inflow should weaken the vortex inside the cabin.

Blue render of the refined roadster from the front three-quarter view

Design Intent

The rear changes aim to delay separation and reduce the wake, while the cabin-edge lips limit side flow and calm the recirculation inside the cockpit. Together, they're expected to improve overall aerodynamic efficiency.