Arouse the DAMPFHAMMER!
Thanks Dave for your recommendation concerning the temp. hub at peak power operation point . But I really wonder how an 700 hp modified K20 with an CWA200 can survive 24 h endurance races when a TC'd stock longblock K20 with around 500 hp also just survives. That 700 whp application made about 20-40 °C of temperature drop after the radiator. According the 10 °C-hub goal this would be a 1540 l/min flow flux, making 21 m/s on an ID 38 mm pipe or 1074 l/min according the 1.35 l/min/hp thumb rule. This flow fluxes would rise the coolant pressure drop over the engine up to a few bars instead of some 100 mbar's. Real world race engine coolant design might be different compared to an other specified (more solid) OEM design aspect. Just to give you an example:If you are saying that you have a temp rise of 25C from engine inlet to outlet at full load, I think you see my point that you are low on total coolant flow through the engine. The pump should be sized to a peak delta-t across the engine of less than half of that.
I worked 9 years as an develop engineer at an engine OEM, I wasn't much deep in cooling, I just know the typical Gensets- and Speed and Load-variable engine applications temperature hubs, which were around 12 °C. For the pictured genset below I've invented the combustion process idea for it, was responsible for the alternation of load team and for the development of the ignition system (pre-chamber combustion system). These engines have a 1-point operation: WOT (MAP = 4.5 bar), 1500 rpm and endure 63,000 h, which equals to running around 7.2 years at WOT (mean peak cylinder pressure 165 bar) running. Yes these engines need a best as possible equalized coolant temperature over all 8 to 20 cylinder units to achieve their lifetime goal for every cylinder unit. But I might think a very short I4 engine, which endures at maximum about 50 h of WOT without knock issue, would be ok with 25 °C temperature hub. Otherwise I would need 477 l/min (10 °C goal) or 419 l/min (1.35 l/min/hp goal) of coolant flow, which even the OEM pump is not able to deliver. Those pumps are capable of around 500 l/min at zero pressure hub and far less then 300 l/min at my system pressure and especially at rated engine speed (efficiency map of the pump is there at a worse point). Anyway, what is right or wrong, I think less is possible, otherwise a 600 whp Lotus Exige race application wouldn't be able to run only a CD150 without heating issues, but more is helpfully to maintain a better engine healtyness.
Picture of the 20 cylinder engine of the BR4000L64 series (3536 hp at flywheel at 1500 rpm ) incl. cooling circuit and generator of MTU
The picture shows the flow rate over the system pressure for several pumps and some system pressure curves of several engines with an early guessing of a system pressure curve of an K20-Elise-coolant circuit. Dave knows it, but for the others, where the system pressure curve crosses the operational working line of the pump, that's the point of maximum flow of the specific pump in that system. Be aware, don't take the system pressure curves for granted, these are idealize examples and is highly depended on your specific cooling system. But everyone can see Craig David or DeMeziere pumps are a joke in the wind. The real system pressure of an K20 engine is higher what they deliver...