To hold enough fuel for the endurance event, deliver it to the injectors at a stable pressure under every acceleration case, and do it without ever leaking.
A composite fuel cell with a CFRP sump and an aluminium tophat and filler neck. The sump geometry is carried over between years so cells stay cross-compatible. A Holley HydraMat covers the bottom of the tank as the pickup; it wicks fuel by surface tension so the pump keeps pressure even at high lateral acceleration with very little fuel remaining, and it doubles as a 15 micron pre-pump filter, which removes the need for a secondary filter. A DeatschWerks DW100 pump feeds the fuel rail; a DW1000c regulator, adjustable 30-120 psi, sets rail pressure at 80 psi and returns excess. The rail is milled from 6063-T5 extrusion with a 0.5 in centre bore, black anodised, carrying a Bosch PST-F 1 combined pressure and temperature sensor.
There can be absolutely no leaks from the fuel tank, ever. Every fuel cell revision has been driven by a sealing failure mode rather than by capacity or weight. Size the tank from a real failure. CR21 experienced fuel starvation during the 2022 endurance at 0.96 g lateral with 0.66 gal remaining, costing 10 seconds of lap time. That set the volume target and put the HydraMat in the car; with it, the system tolerates 1.66 g peak lateral with 10% (0.14 gal) remaining. Measure fuel temperature. Before the PST-F 1, the ECU assumed a default 30 C for density, which could run the engine lean when fuel got hot. Measured fuel temperature during a mock endurance was 51 C. 1. Set tank volume from endurance consumption plus margin, with the HydraMat area and the tank CG as constraints. CR26: 1.4 gal cell, 1.56 gal effective system. 2. Design the sump to the acceleration cases: 1.42 g at 45 deg from vertical under acceleration against a rear wall 30 deg from vertical, and 2.51 g at 66.5 deg from vertical under braking against a vertical front wall. 3. Lay up the sump: 1 layer of Kevlar over 5 layers of stitched carbon twill on HIPS dissolvable molds, bonded with 3M DP420, sealed with Caswell epoxy gas tank sealer. 4. Design the top plate: -4 AN weld bungs instead of fittings, stud bulkhead terminals so no fuel path runs through wire, an aluminium terminal block for proper bulkhead seating. 5. Leak test the assembled system before it goes in the car.
- No leaks. - Eliminate the risk of fuel starvation during endurance. - Reduce system mass. The CFRP cell cut 0.45 lb (11%), from 4.19 to 3.73 lb including pump and hardware; the fuel temperature sensor package cut 1.26 lb. - Integrate high-pressure PTFE fuel lines and leak-test the system. BMRS line raised supported pressure, added wear-resistant braiding, removed hose clamps and cut 0.256 lb. - Validate the fuel selection with data rather than assumption.
- Fuel cell sealing. Threads stripped before reaching the clamping pressure needed to seal, and fuel leaked through the threads unless nylon crush washers were used. Fixed by moving from 8-32 to 10-32 bolts, raising torque spec from 20 to 30 in-lb, raising average clamping pressure from 689.4 to 951.5 psi, and moving the bolts from the interior to the exterior of the cell, which eliminated thread sealant entirely. - Fuel travelling through the wires of the previously selected bulkhead. Fixed with stud bulkhead terminals. - Occasional failure of the nylon crush washer seal. - Fuel starvation under cornering on CR21, solved by the HydraMat. - HydraMat packaging. Small sizes are not readily available, which makes it hard to fit under the seat.
SolidWorks for cell geometry and line routing; all fuel line is modelled in SolidWorks and sourced from BMRS. MoTeC M1 Tune for injector characterisation and density correction. Hub dynamometer for the E85 against 93 octane comparison. 24 hour gasoline immersion testing of DP420 bonded to aluminium (no surface defects or dissolution; bond failure took up to 4 months).