A car works as one system. So should its model. Watch the tyres, upper vents, steering and stored energy respond together — then explore the cold-pressure idea at its source.
Calculated trajectories and forces. Illustrative vehicle styling. The vehicle has not been built.
Start with the complete air system enabled and disabled. Then replay unequal grip, a crosswind, a lost upper port or unavailable motor. Every trajectory comes from the recorded numerical runs.
Same assumed mass, tyres and finite stores. Air systems enabled and disabled.
Matched mass and stores
Dry straight stop
Car shape and air paths are illustrations, not CFD. Wheel markers show braking force relative to normal load; the separate plot uses actual model distances. The shaded band marks a ±0.8 m centre-path guide. The pass check includes the rotated vehicle footprint within a 3.6 m corridor and peak heading below 3°. These are research thresholds.
Grip multipliers are model sensitivities, not measured wet, snow or ice coefficients. The steering and braking model does not know an invisible road boundary: it follows a specified straight reference and reacts to delayed vehicle-state feedback.
The matched air-off case takes 543.3 m. Both start at 400 km/h with the same 2,464.4 kg mass, effective tyre model, brake limits and stored energy. That is a 18.7% conditional reduction in this dry, straight-line scenario.
Removing an assumed 270 kg air-hardware allowance from the air-off model gives 522.2 m. This sensitivity separates some mass cost from the aerodynamic benefit; it is not a measured production-car reference.
The published Chiron reference is 491 m from 400 km/h. The earlier effective tyre parameter was fitted to that point, so the CryoPulse result is a design projection built on a calibrated baseline. It is not independent proof that a real car beats the Chiron.
Bugatti’s published technical data ↗ · P3.4 assumptions and comparison ↗
| Condition | Wheel ESC Fixed upper route | Airflow-aware coordination | Coordinated peak lateral position | Coordinated peak heading | Both controllers |
|---|---|---|---|---|---|
| Dry straight stop | 441.8 m | 441.8 m | 0.000 m | 0.000° | Within stated limits |
| Unequal left/right grip | 872.5 m | 881.1 m | 0.431 m | 1.248° | Within stated limits |
| Crosswind pulse | 620.1 m | 620.1 m | 0.184 m | 0.892° | Within stated limits |
| Unequal grip and wind | 863.7 m | 872.8 m | 0.452 m | 1.252° | Within stated limits |
| One upper port lost | 444.7 m | 444.7 m | 0.009 m | 0.044° | Within stated limits |
| Unequal grip and port loss | 874.7 m | 882.2 m | 0.435 m | 1.267° | Within stated limits |
| Low-grip straight stop | 1266.9 m | 1266.9 m | 0.000 m | 0.000° | Within stated limits |
| Low grip and port loss | 1284.8 m | 1284.8 m | 0.017 m | 0.048° | Within stated limits |
| Upper motor unavailable | 678.0 m | 678.0 m | 0.000 m | 0.000° | Within stated limits |
Both virtual controllers already have the CryoPulse fan, underfloor suction and upper-route hardware. Air trim adds roughly 8–9 m in the three unequal-grip cases; its direction metrics change only slightly. It earns no claim of uniformly shorter stops. The stronger current result is the revised architecture meeting all nine stated conditions.
The retained first prototype failed the unequal-grip corridor test: it reached about 15 m lateral deviation and 17.3° heading, with no completed stop. It is not a production-car comparator.
The revised design sets rear cornering stiffness to 90 kN/rad per tyre instead of 70, adds dynamic steering feedback, keeps grip estimates from drifting upward without informative measurements, and reserves tyre capacity for direction control. The airflow planner predicts how each command changes wheel loads. Those stiffness, sensing and actuator targets still need hardware qualification.
This revision uses the prior finite-energy acceleration model to initialise the braking event. With the stated engine envelope, electrical traction boost, launch suction, an 8 MJ initial pulse store and the new mass, it calculates 2.07 s to 100 km/h and 24.39 s to 400 km/h. About 3.06 MJ remains before braking. These acceleration figures inherit the earlier longitudinal assumptions; four-wheel yaw dynamics are applied to the braking phase.
P3.5 does not establish a new top-speed result. Earlier unrestricted speed screens remain separate model cases with their own power, drag and pulse assumptions. Explore the earlier performance comparison ↗
The useful mechanism is a pressure difference. Cooling a sealed gas lowers its pressure. Open a path to warmer ambient air and air can flow in — until that difference is used up.
This gives us a simple demonstration of the original idea. Start with 10 litres of dry air at 20°C and atmospheric pressure. Keep the volume fixed and cool it to 0°C. The ideal-gas pressure falls to 94.41 kPa absolute.
The colder gas can admit about 0.882 g more air before equalising, if held at 0°C. A pressure difference can do work, but the cooling and reset require energy. No new energy appears when the space fills.
The 69.1 N value is a pressure force on a fixed 100 cm² sensing surface. It is not vehicle thrust. The work figure is an ideal upper bound for a reversible fill expander, not a measured tiny turbine.
Animation timing is illustrative. No flow rate is predicted.
The separate nitrogen calculation pumps cold liquid, warms it with an available heat source, expands the resulting dry gas through a generator, then uses the remaining cold for another cooling duty.
Liquid nitrogen at atmospheric pressure. A finite offboard-prepared cold store.
Electrical pump input is deducted. Pump efficiency target: 65%.
Heat is supplied before expansion. It is an explicit input.
Expansion to atmospheric pressure. Expander efficiency target: 65%.
Further heat uptake after expansion. The same nitrogen is used once.
The state calculation closes its first-law and exergy balances. The net result subtracts the specified liquid-pump motor input and generator loss. The selected case has dry gas at the expander and a positive heat-exchanger temperature approach.
Circulation power, valve losses, real exchanger area, tank and equipment mass, boil-off and replenishment have not been closed. Nitrogen manufacture and re-liquefaction need an offboard energy account before any whole-cycle efficiency claim.
This is a thermodynamic branch to develop, with zero driving-energy credit in the results above. It is not self-charging propulsion.
All 12 pressure/temperature states and balances ↗5 solids · 240 mm housing length
A shared plenum, four outlets and four gate discs. The interactive view uses the mesh of the delivered STEP geometry.
Three discrete poses were checked for solid overlap. Independent STEP imports verify five solids and a 240 mm housing length. The main inlet bore is 84.4 mm; each outlet is 27.64 mm.
The next detail design must add shafts, seats, seals, mounts and actuators. The nominal gate clearance is not an airtight seal, and the three-pose check does not establish continuous motion or structural strength. This is a low-pressure bench layout, not the car’s installed duct geometry.
For the assumed axle geometry and front stiffness, the linear unbraked bicycle model requires rear cornering stiffness of at least 76.4 kN/rad per tyre for nonnegative understeer. The 90 kN/rad target satisfies that condition. The original 70 kN/rad target does not; its linear model has a critical speed near 210 km/h.
This is a mathematical result within a specified model. It does not establish nonlinear braking stability, measured tyre characteristics or a certified speed limit. The numerical scenarios add saturation, finite actuator response, sensing delay and energy limits.
Read the linear checks and eigenvalues ↗The wheel model uses force actuators and an ideal inner slip-limiting loop. It does not solve wheel rotational speeds, a measured tyre-slip curve or actual hydraulic pressure. Pitch and roll transfer are quasistatic. Tyre stiffness and grip estimation, brake heat fade, sealing, steering authority, fan maps, valve dynamics and the equipment mass all need measured data.
Local outlet reaction and inlet momentum are included, but changed pressure forces on the surrounding vehicle body are withheld. Installed full-car CFD and force measurements remain open; the earlier wake and mesh-quality gates are not declared passed here. These deterministic scenarios do not establish universal stability.
The braking event follows a separately calculated acceleration run from an initial 8 MJ pulse store. Approximately 3.06 MJ remains for braking. The shared pulse node is limited to about 217.4 kW. Fan startup, valve/vector movement and DC/DC input share that node. Wheel regeneration can replenish the store within its capacity; ground suction, steering and control loads draw from the protected auxiliary supply. Neither cold-pressure work nor nitrogen-generation potential is added to that account.
NASA: momentum and pressure thrust · NASA: ideal-gas equation of state · MathWorks: combined-slip wheel modelling · Bosch: electronic stability programme · CoolProp: nitrogen equation of state.
P3.3: airflow, energy and thermal branches · P3.4: outlet backpressure, axle loads and shared energy. None of these sources validates this vehicle; they support the equations and established mechanisms used in the research model.
The next gains have to come from closing the hardware requirements: an installed body and duct shape, measured wheel and fan maps, a practical cold-recovery assembly, and control that keeps the car composed as conditions change.
Measure the shared manifold’s forces, leakage and power. Fit measured tyre and actuator data. Add suspension transients and wheel-speed dynamics. Rework air trim where it increases distance without a useful direction benefit. Then repeat the combined driving cases with those measured inputs.
Acceleration, unrestricted speed, thermal management, materials, accessible doors and the complete luxury package remain part of the programme. This revision closes a numerical coordination stage; it does not finish the vehicle or claim a percentage of engineering completion.