Reject the heat the engine puts into the coolant, so water temperature stays below its limit through an endurance event.
A single-pass louvered-fin water-to-air heat exchanger with tanks top and bottom. Coolant enters the top tank through a -16 AN fitting, passes down through the core where air pulled through the fins carries the heat away, and leaves through the bottom tank. Inlet and outlet temperature bungs let the actual heat rejection be back-calculated from logged data, and a bleed valve on the top tank lets trapped air out.
Sized by an NTU-epsilon sweep in MATLAB starting from the heat rejection measured on the previous year's endurance. Core length, width, thickness, fin density, water mass flow and air mass flow are all swept until the candidate rejects significantly more than the maximum seen. The core is then leaned forward 45 degrees, which trades mass flow for much less drag and better packaging with the undertray, duct and sensors.
Sidepod packaging, undertray clearance, and the requirement that the radiator be serviceable without removing the undertray. Air mass flow available is set by the duct and shroud, not by the core. From the design reports and one-pagers: From the CR22 Cooling System Design Report: dissipate the heat generated by the engine; fit the majority of the cooling system into the sidepods; the rising cost of aluminium forced reuse of the previous year's radiators; the system must run pure water; and electrical components must not pull excessive amperes. Rules: T.5.4.1 water cooled engines must use only plain water with no additives of any kind. T.5.5.1 any cooling or lubrication system must be sealed to prevent leakage. T.5.5.2 the vehicle must be capable of being tilted to 45 degrees without leaking fluid of any type.
NTU-epsilon results at 0.343 kg/s air and 1.11386 kg/s water: maximum heat dissipated 35.2554 hp, NTU 4.15306, water inlet 105 C, outlet 99.407 C, delta 5.593 C. Assumptions: steady state, single-phase water, constant air mass flow from CFD at the 32 mph average endurance speed, 105 C water inlet, 27 C air inlet, constant fluid properties at inlet temperature. Known discrepancies: the model is a single-state output compared against a transient system, the overall heat transfer coefficient is inaccurate from fouling and flow rate inconsistency, air speed is assumed constant, and fan-airflow interaction is excluded. Also simulated as a porous region in ANSYS Fluent with Darcy-Forchheimer coefficients from flow bench testing (viscous 1,919,749.637, inertial 52.945). From the design reports and one-pagers: Radiator sizing is done with transient heat transfer methods, either NTU or LMTD. The team uses NTU, referenced to Bergman, Fundamentals of Heat and Mass Transfer, 8th ed., chapter 11 section 4. A MATLAB script iterates to steady state and sweeps core length, width and thickness. CR22 assumptions, stated as approximate at the time: overall heat transfer coefficient U taken as roughly 20 from table 11.2, acknowledged as not exactly correct and needing more testing; all values calculated at 120 C; air temperature assumed 83 F. CR22 two parallel radiators, 9.5 x 15 x 2.5 in each: water flow 55 L/min, air flow 616 CFM per radiator, target heat dissipation 45 hp, expected total 45.9937 hp, engine inlet 77.45 C, engine outlet 68.1967 C, total temperature delta 18.5065 C, offset from target 0.9937 hp. The report's own closing assessment is worth carrying forward: "This was my first designed cooling system, the initial analysis needs lots of work and a lot was assumed by young me. I assumed bigger was better this design cycle and have learned otherwise." CR25 and CR26I acted on that by downsizing to a single radiator.
Aluminium core and tanks. -16 AN (1 inch) inlet and outlet fittings. From the cost report drawings and CAD: From the cost report drawings: Water-to-Air Heat Exchanger (Water-to-Air Heat Exchanger).
The radiator is not using its full surface area to cool. Recognised in the Powertrain knowledge transfer notes with no solution recorded. Peak water temperature reached 113.5 C on a mock half endurance against a 115 C maximum and a 126 C boiling point at system pressure.
CR26I added 3D printed weld jigs for the radiator mounting tabs, which improved the reliability of radiator location and therefore its integration with the ducts, shrouds, fans, bodywork and floor.
Overall 16 x 9 x 2.375 in. Core 9 x 13 x 1.5 in. 14 fins per inch, louvered, 0.32 in long, 0.0051 in thick. 2 in tanks top and bottom, single pass. Angled 45 degrees. Rejects an average of 24 hp. Wet weight 7.21 lb, 3.68 lb lighter than CR22's dual radiators, and the change reduced total water weight by 10.12 lb. Mounted on the left side of the car. From the design reports and one-pagers: CR22 radiator, made by Dubz Custom Radiators: aluminium construction with AN-16 connections, single pass water-to-air, 10 x 9 x 2 in core with 2 in tanks top and bottom, tube width 0.078 in, fin width 0.0058 in, roughly 7 lb each when wet. Cost about $500 three years before the report, roughly $1000 at the time of writing, which is why they were reused rather than replaced. CR26I single radiator (see the main entry above) replaced this pair. From the cost report drawings and CAD: Dimensions called out on the cost report drawings. These are the figures printed on the drawing, not a labelled specification; open the drawing itself for what each one refers to. Water-to-Air Heat Exchanger (mm): 106.1, 45.5, 330.0 Source: Cost report drawings: CR26E/Powertrain - EV/Design/Cost Report Drawings/Cooling/Water-to-Air Heat Exchanger.pdf Quantity per car, as costed: 1
Nothing in CrimsonStorage covers these yet.