Engineering work
I built AutoCFD, a pipeline that accepts a vehicle CAD file, sets up a Luminary Cloud simulation, monitors the run, and extracts drag, side force, lift, center of pressure, and convergence information. The application also logs results to Google Sheets so geometry and wind cases can be compared. The code includes farfield and mesh setup, optional rotating-wheel treatment, and a job dashboard. These are implemented workflow features, not claims that every simulation has been validated experimentally.
The same workflow can run Shellpower, a solar-array layout and simulation tool, on an exported car-body mesh. Shellpower places cells under a 6.0 m² target-area constraint, evaluates exposure and occlusion across sampled headings and race hours, and estimates peak power and daily energy. The current code supports a symmetric layout and an optional shadow-aware comparison run. I worked on bringing aerodynamic and solar estimates into one run because a sail can affect both drag and array shading.
Adaptive retractable sail
This study tests a retractable, auto-cambering sail for crosswind energy harvesting. The proposed mechanism changes camber and trim when apparent wind is favorable and stows when the combined drag, rolling resistance, shading, and control costs exceed the aerodynamic benefit. The mechanism remains a design study. Deployment logic, durability, control power, and race integration still need validation.
Scroll sideways to read the four study slides.
The second rendering above the page shows the sail installed in the rear bodywork. The concept uses a paired skin around a mast with a retraction and trim mechanism. The CAD images show packaging intent; they do not establish structural adequacy or a tested actuation cycle.
CFD and force comparison
The Luminary visualization below shows surface results and flow paths for one car geometry. The color scales represent skin-friction coefficient and turbulence intermittency. Read alone, the visualization is not a validated prediction of race energy savings.
The following comparison comes from a single preliminary working table. “Net resisting force” is the supplied drag estimate plus the supplied increase in rolling resistance. Negative drag means the aerodynamic force acts forward in that simulation. Cases and boundary conditions require a matched rerun before these differences can support a design decision.
| Configuration | Drag (N) | Side force (N) | Added rolling resistance (N) | Net resisting force (N) | Change from base (N) |
|---|---|---|---|---|---|
| Base, 90° | 32.10 | 196.00 | 3.35 | 35.45 | 0.00 |
| Fin mast | 14.89 | 299.00 | 7.80 | 22.69 | 12.76 lower |
| Fat fin | -1.97 | 402.79 | 14.16 | 12.19 | 23.27 lower |
| AC75 profile | 2.91 | 375.86 | 12.33 | 15.24 | 20.21 lower |
| AC75 profile + jib | -6.65 | 493.39 | 21.24 | 14.59 | 20.87 lower |
A separate set of notes lists different drag and side-force values for similarly named cases. I have kept those results out of this comparison until the run settings and reference forces are reconciled.
Working assumptions and decision checks
One working route scenario assumes a 3 m/s southeast wind during the first 40% of the race to Alice Springs. The noted relative-wind vector is (-26.71, -2.12, 0) m/s, giving a magnitude of about 26.79 m/s and an angle of about 4.54°. This is a scenario for analysis, not a verified race forecast. The current energy estimate of 0.88 kWh/day gained and 0.20 kWh/day lost to shading and drag is also preliminary; the underlying wind frequency, electrical conversion, and deployment assumptions need a documented calculation.
The sail also changes side force, tire load, and roll moment. The notes include a 260 kg vehicle mass, a 550 N maximum side-force case, and a reported center of pressure for one AC75 case at (-0.9, 0.41, 0.31) m relative to the center of the front wheels at ground level. Those inputs need a common coordinate frame and a tire-load model before a rollover or steering conclusion. The initial actuator estimate is 16.25 N·m of boom torque under a 130 N side load; gearing and holding-power claims still require a mechanical test.
The next study needs a control map over apparent-wind speed and direction: when to deploy, what sail angle of attack to command, and how to adjust camber. Lift-to-drag ratio is one aerodynamic measure, but the useful objective for the car is net race energy after side-force rolling losses, shading, and actuator power. Lead-car anemometer data could supply a wind distribution for that calculation. Before production, the mechanism also needs repeatable trim and retraction tests, structural and fatigue checks, mass and packaging limits, and a failure-safe stowed state.