Wow, this thread needs some equations!
Let’s start with a little background on heat transfer basics.
Radiation
Conduction
Convection
Heat is transferred to the block from the combustion chamber via radiation and convection. Heat then moves via conduction from the combustion side of the block to the coolant side of the block. Next, heat is transferred via convection from the block to the coolant. Once in the radiator, heat is transferred from the coolant to the radiator metal via convection. Then the heat is conducted to the cool side of the radiator tubes. Lastly, air passing over the radiator convects the heat from the tubes.
The transfer we're concerned with is the convection from the block and to the radiator.
The equation for total heat flux for a given area via convection is:
q = h-bar*As*(Ts - T∞ )
Where,
q = total heat transfer rate, W
h-bar = average convection coefficient, W/m^2*K
As = area of the surface, m^2
Ts = surface temperature, K
T∞ = bulk fluid temperature, K
As is constant for our example (an engine block). So, heat transfer is related to h-bar and the temperature difference between the block and the coolant. The delta-T is straight forward, but h-bar is not. Let's delve into that now.
For a flat plate with fluid running parallel to it...(This is not a completely accurate representation of an engine block. The actual h-bar will be different, but the
relationships will remain the same.)
h-bar = (k/L)*NuL
Where,
k = thermal conductivity, W/m-K
L = length of plate, m
NuL =
Nusselt number, dimensionless
Thermal conductivity of the fluid and the length of the plate are constant. So, h-bar is directly related to the
Nusselt nmber. Basically, the Nusselt number is related to how well the fluid is convecting heat. Bigger is better. This number is mostly
independent of temperature. As I will show later, it is mainly related to mass flow rate (or velocity) of the fluid. As mass flow rate goes up, the Nusselt number goes up. As the Nusselt number goes up, the average convection coefficient goes up.
For a flat plate with turbulent fluid flowing parallel to it...
NuL = .037*Pr^.33*Re^.8 (This relationship is derived experimentally)
Where,
Pr = Prandtl number, dimensionless
Re = Reynolds number, dimensionless
Where,
Pr =(mu*cp)/k
Where,
mu = dynamic viscosity, cP (centipoise)
cp = specific heat, kJ/kg-K
k = thermal conductivity, W/m-K
Specific heat and thermal conductivity are constants. Viscosity is related to temperature, the colder the fluid, the higher the viscosity, and the higher the Prandtl number.
Re = (rho*u*L)/mu
Where,
rho = density, kg/m^3
u = free stream velocity, m/s
L = length of plate, m
mu = dynamic viscosity, cP
Density and length are constants, so the Reynolds number is a function of flow velocity and dynamic viscosity.
If you use some algebra, the Nusselt number simplifies out to the following,
NuL =
(.037*cp^.33*L^.8*rho^.8*k^-.33)*
(u^.8*mu^-.46)
All of the numbers in red are constants for any given case that you solve for. The numbers in green are variables.
What you are left with is an equation mostly related to flow velocity (mass flow rate). If velocity goes up, NuL goes up.
The increase in flow velocity outweighs the increase in viscosity (which by itself will lower the Nusselt number) that might come about by a decrease in system temperature caused by increased flow rate or more efficient convection. Run the numbers if you don't believe me. The end-game is, increase in flow rate = increase in convection efficiency, regardless of temperature.
The other factor in the coolant heat transfer is the delta-T. The delta-T for a cooling system without a thermostat will be at some equilibrium point. Assuming the vehicle is at idle or some other constant load with a constant flow of air over the radiator...an increase in flow rate will decrease the coolant temperature. Whyso, you say?!
Let's go back to the original equation for convection.
q =
h-bar As (Ts - T∞ )
The constant load assumption means q is constant. As is also constant.
So, if you are at an equilibrium delta-T and you increase flow rate, you increase h (as I proved above). This requires that delta-T decrease to maintain the same q (heat transfer).
One thing you'll notice that doesn't appear in my derivations is time. Time is irrelevant in this analysis. Boo Boo Foo covered that part well enough.
My post isn't meant to state that running without a thermostat won't cause any vehicles to overheat. Things such as pump cavitation from too much flow can occur. This will cause the vehicle to overheat, but the owner will interpret the lack of a thermostat for the overheating. My post was simply meant to say that "dwell time" isn't true. If your car overheats without a thermostat/restriction, something else is at fault (that could be caused by high flow rate).
PS Boo Boo Foo, do you post on CC.com?