To deliver the mass flow of air that the water radiator and the oil-to-air heat exchanger need, at the lowest possible drag cost, and to keep that flow predictable and sealed so the cooling model can be trusted.
A diverging-converging duct formed by the sidepod inlet, the duct and the fan shroud creates a favourable pressure gradient through the entire system, which raises mass flow through the core beyond what a plain inlet would give. The sidepod inlet shape and position are chosen to avoid ingesting front tire wake and to avoid competing with the undertray inlet. The radiator is leaned forward 45 degrees, which trades some mass flow for much less drag and better packaging against the undertray. Ducts and shrouds are sealed to each other with foil tape so air cannot bypass the core.
The performance metric is mass flow rate through the core, not inlet area or duct volume. Targets are set from historical heat rejection measured on track, then the duct is designed to beat them with margin. Drag is a real cost and is measured. Coast-down testing with and without sidepods on CR26I showed CD 1.33 with, 1.37 without, so the ducted sidepods were a net drag reduction, not a penalty. 1. Powertrain sets the required heat rejection from previous-year endurance data; that becomes a target mass airflow (CR26I: 0.265 kg/s water, 0.060 kg/s oil). 2. Characterise the core on the flow bench to get pressure drop against velocity, and fit Darcy-Forchheimer coefficients (water radiator viscous 1,919,749.6, inertial 52.945; oil cooler viscous 7,387,951.3, inertial 132.358). 3. Model the core in ANSYS Fluent as a porous region and the fan as a pressure jump over a fan boundary, using the manufacturer fan curve. 4. Sweep radiator angle. CR26I found 0 deg gave 0.459 kg/s with 0.78 lbf drag, +45 deg forward lean gave 0.343 kg/s with 0.31 lbf drag. Forward lean was accepted because it still beat target and cut drag and lift. 5. Design the duct and shroud around the chosen core position, then validate the assembled system with tufting and coast-down testing.
- Beat the target mass airflow through both cores with margin at the average endurance vehicle speed (32 mph). - Reduce cooling system drag using sidepods rather than accepting it. - Create a predictable, sealed system so the cooling model input is real. - Keep the radiator serviceable without removing the undertray. - Support the Powertrain move to a single water radiator plus a separate oil-to-air heat exchanger.
- Ducts blowing off the radiator under pressure. Fixed with stronger metal fastening and by attaching the sidepod to hold the duct in place. - The radiator not using its full surface area to cool. Recognised in the Powertrain knowledge transfer notes; no solution recorded yet. - Porous media accuracy breaking down at high speed, and the fine mesh needed at the porous/free-flow interface. Accepted because the car operates at low speed. - Duct-to-sidepod sealing. Addressed with a lip on the duct that hooks the sidepod lip, plus foil tape at every joint.
ANSYS Fluent with a homogeneous porous media radiator model and a fan boundary condition. Flow bench testing for Darcy-Forchheimer coefficients. MATLAB NTU-epsilon sizing (owned by Powertrain, consumed here). Coast-down testing with a Bosch MM5.10 accelerometer and a digital anemometer for CD validation. Tufting for flow visualisation on the sidepod.