HX coolant too fast to cool?

[frankiee]

Is it possible to have coolant go thru an engine too fast and have the coolant not pick up enough heat so the engine overheats?
I think it would be like a simple heat exchanger.
Is it possible to have the coolant go so fast that the heat exchange is lessened?

 

[rmw]

Possible; yes. Probable; doubtfully.

How fast are we talking about?

rmw

 

[IRstuff]

Not even possible, I think. A fast coolant would simply maximize the temperature delta across the exchanger, therefore, there some be even more cooling than with a slower, hence, hotter coolant. A trivial check is to see what your coolant temperature is when the engine is overheating.

It's more likely that something else is wrong; your exchanger may be grossly fouled.

TTFN

FAQ731-376

 

[rmw]

My first post assumed you were talking about the water going through the engine.

Heat transfer takes place at a rate which has time as a factor. If the time of the water in contact with the hot surface is inadequate, less heat transfer takes place. The rate is also affected by factors like velocity and turbulence. The higher the velocity, the more turbulence you have which improves the heat transfer rate. (The reason you sling your finger through the air when you burn it - speeding up the cooling rate.)

But they are not linear. Doubling the water flow through the engine (velocity) does not necessarily mean that you double the heat transfer rate or the duty.

So at some point, theoretically you can blast the water through so fast (if physically possible with your pumping system) that it picks up virtually no heat. Easy to theorize, hard to do.

So, unless someone has sneaked out there and added several pumps to your water circuit, I would look for other more realistic factors, starting with fouling.

rmw

 

[frankiee]

This is in theory only.
Started off as a debate about an engine overheating but now has engineer A (me) saying it is theoretically possible to have velocity so fast in a closed system that the cooling capacity can be reduced.
Engineer B says that as mass increases then the heat transfer increases and therefore there can not be a reduction in heat transfer no matter how fast the coolant is going thru the system.
The heat transfer would be radiation plus convection plus conduction.
I origianlly thought this would be a very easy question to answer but now I am learning my heat transfer knowledge is lacking.
Anybody got a formula or model or difinative answer to say either way?
Just in theory
Thanks

 

[speco]

franklee,

I have to go with Engineer B on this one. Here's why:

Let's assume that the amount of heat that you need to remove from the engine is a constant, say 50,000 BTU/hr.

If the cooling through the engine is say 5 gpm, that gives you a temperature rise through the engine. Using water properties(to simiplify the math), that works out to a 20 degree rise.

Now, if we double the water flow rate to 10 gpm, the temperature rise is only 10 degrees. This affects two things. If we also assume the inlet water temperature is a constant, this increases the temperature difference between the coolant and the engine. It also greatly increases the heat transfer because the film coefficient of the water is increased (or the film resistance is reduced).

The same thing happens in the radiator.

In the raal world, the system will achieve its own equilibrium, but is affected by whatever controls are used such as thermostats.

A secondary consideration is the coolant itself. Pure water is a great coolant, but causes problems with corrosion and freezing. Pure antifreeze doesn't freeze, but is a not a very good heat transfer fluid, because the viscosity is high, the specific heat is low, and the thermal conductivity is low, all compared with water. That's why most systems use a mixture of approx 50% of each. Its freezing point is very low, it boiling point is high, and it contains some corrosion inhibitors.

Regards,

Speco

 

[MintJulep]

The heat transfer from the block wall to the coolant is a function involving a bunch of dimensionless numbers named after dead guys - like Prandlt, Nessault and Renyolds.

Most, if not all, of those numbers are a function of velocity.

In a real engine block, none of these numbers are a constant as the coolant passes through the block.

So the definitive answer is: It depends.

 

[frankiee]

Oh great MintJulep!
Just as I was going to post to engineer "B" and eat crow telling him that he is right and I am wrong, you offer me a glimmer of hope that it is in theory possible.
That is what I was thinking that in a real world engine block that there would be so many calculations to make that it would be easier to test then calculate a certain condition.
Shear, pressure drop, laminar flow, turbulent flow, surface film, vapour bubbles, viscosity, cavitation,.....and likely more variables that I have not heard about.
It would be unpredictable chaos inside and engine block.
Now I don't know what to do.
I guess it is not that important right now.
But, how important is the pursuit of knowledge to me?
How important is it to me to admit I might be wrong?
I think I will post engineer "B" that I can not find any proof of my theory but that I still think it can happen.
I figure he will claim victory.
I will continue my research when I have more time.

Thanks for any and all input for and against.
Any future facts would still be welcome.

 

[willard3]

As Engineers, we always assume things are steady state and they're not......it's a simplifying assumption. Combustion in an otto cycle is not steady state and neither is the cooling.

Things are not homogeneous and isotropic, either, it's a simplifying assumption.

 

[TERIO]

Here is a thought:

When you move fluid through a duct there is head loss due to friction. This energy is dissipated as heat. The head loss increases with increasing velocity, so maybe at high enough flow rates the coolant would heat up enough due to this viscous dissipation that it would not be able to take out as much heat.

 

[Compositepro]

Heat transfer will never go down with an increase in coolant velocity. Frictional heating is a separate issue and not really a practical consideration except for energy efficiency (i.e., don't use a 100 hp pump where one hp will do).

 

[whammett]

The only way friction would come into play as a temperture rise and the collant velocity rises is if you were pumping a slurry or something with a very high viscocity.

"I came, I saw, I made it better."
-Ode to Industrial Engineers
Will ChevronTexaco Corp.

 

[electricpete]

If I understand correctly, there are two separate heat exchange processes:
Process 1 - Coolant picks up heat in engine engine
Process 2 - Coolant loses heat in radiator

So far, I think all the focus I think is on process 1, where the effect of increased coolant flow is clearly beneficial (neglecting friction) if we look at system 1 in isolation (assume constant supply supply temperature at the inlet of the engine.)

If we look at process 2 in isolation, increased flow means increased radiator outlet temperature (assuming constant radiator inlet temperature).

Putting them together, one would certainly intuitively think that increased flow helps. Because we are not looking at maximinizing heat transfer per unit mass of coolant, we are looking at maximizing heat transfer per time. Faster flow should help provide a better thermal link between the two systems (engine and air).

Nevertheless, there are non-linear effects to be considered as alluded above, and sometimes those defy intuition. It is not 100% transparent to me that it is always a benefit to increase flow (even with friction neglected).

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[zekeman]

From the well known “dead men’s” correlation
Nu= Re^.8*Pr^.4
hD/k=(rho*V*D/u)^0.8*(Cp*u/k)^0.4
V- velocity
rho =density
u=viscosity
D duct diameter
k thermal conductivity
At a minimum of logic, if you make V large enough, h will become almost proportional to V^.8 and (Tw-T) will approach (Tw-T0), so you have
Q =P*V^0.8*(Tw-To)
P= constant
Which mathematically grows without bound but practically, is limited by the limits of the correlation and pumping power, but in any event could be much larger than any of experience..

Tw= engine block temperature
T0 = initial cooant temperature
x = position along duct
T = temperature of coolant at any position along x

Now , the more interesting proposition is:

Does the heat transfer to the coolant always increase with increasing V? i.e.

Assuming that only u and k are sensitive to Temperature, I get
h=A*V^.8*k^(.6)*u^(-.4)
A =proportionality constant
It turns out that the product terms on the RHS are increasing functions of T, so we can write the inequality owing to the fact that the coolant temperature T is increasing along x


h>A1*V^.8
Now writing the energy inequality
rho*area*cp*V*dT/dx=hL*(Tw-T)> A1*V^.8*L*(Tw-T)

L= circumference of duct
rearranging using another constant, B and since Tw is a constant we may write
-Vd(T-Tw)/dx>B*V^.8*(Tw-T)
Now since - rho*cp*V*A*d(Tw-T)=q is the flux*dx from the engine block to the coolant we can change this to
-dq/dx>c*V^-.2*q
-dq/q>c*V^-.2*dx
where c is another constant
Integrating from x=0 to x and q=q0 to q yields
q>qo*exp^-c(V^-.2)*x giving the flux at any position
To get Q, the overall heat transfer we must integrate once more to get
Q>q0*integral[exp^-c(V^-.2)*x] dx=qo*V^.2*{1-exp^-c*( V^-.2)*x}/c

Since q0=h0*Tw-To)> A1*V^.8*(Tw-To) and combining exponents
Q> A1*V^.8*(Tw-To)*{1-exp^-c*( V^-.2)*x}=

Q> A1*V*(Tw-To)*{1-exp^-c*( V^-.2)*x}
Finally we must prove that dQ/dV>0. i.e. that a positive change in V yields a positive change in Q.For this we go inside the
Integral [exp^-c(V^-.2)*x] dx
and differentiate with respect to V
yielding

integral[0.2*c*V^-1.2*exp^-c(V^-.2)*x] dx
It is seen that the integrand is always positive and therefore the result is positive proving
dQ/dV>0

DISCLAIMER:
I won’t take this to the bank as is, since I am prone to making errors ( hopefully not in judgment). I mostly used water as the coolant, although the temperature behavior of most of the liquids I looked at fit the assumptions.

 

[IRstuff]

That may be true on the hot side, but the reverse should be true on the cold side. The faster the coolant flow, the harder it will be for the coolant to lose heat to the external exchanger.

Moreover, the overall system is limited by the thermal resistance to the ambient air, so there may be a net change of zero.

TTFN

FAQ731-376

 

[zekeman]

"That may be true on the hot side, but the reverse should be true on the cold side. The faster the coolant flow, the harder it will be for the coolant to lose heat to the external exchanger.

Moreover, the overall system is limited by the thermal resistance to the ambient air, so there may be a net change of zero. "

We'll just blow the radiator fans faster, or buy an infinite radiator.

 

[ione]

We are talking about forced convection. The Nusselt number increases as fluid velocity increases and so does heat transfer coefficient: no way to reduce heat transfer with increasing flow velocity. Just my (and probably Wilhem Nusselt) opinion.

 

[Transient1]

I have to agree that when you consider the closed system of the engine and the radiator, that if the flow around the loop speeds up you'll reach a maximum heat transfer point were increasing the velocity of the fluid will not increase the heat transfer from the engine. Since at a certain point the heat picked up by the fluid cannot be removed fast enough at the radiator.

Even if you consider a theoretical system were fluid flows in at one constant temperature and flows over the engine and out then there will be a bounding point where the rate of heat transfer from the engine to the fluid reaches a maximum. As heat transfer coefficient goes to infinity the delta T between fluid and surface of engine goes to 0.

 

[frankiee]

I was thinking that it would have more to do with the fluid moving thru the engine block so fast that low pressures would build up around the jacket areas and form a boundary of air bubbles in certain areas where there in next to no heat transfer thru the bubble film.

I wonder if anybody has access to any HX, engine, boiler tests where the flow of process or coolant was analyzed.

Now I guess the question would be if the high flow rate would cause a more turbulant flow and therefore less chance of air bubbles against the cylinder wall, or more?

 

[IRstuff]

You seem to be talking about cavitation, which would not even be possible in a normal engine, since the fluid flows would need to be substantially higher than their normal rates, and the general geometries are not conducive to cavitation. Cavitation cannot occur solely due to fluid flow rate.

The only place that cavitation could routinely occur is in the water pump itself, on the impeller blades. Any place where the engine is supposed to be transferring heat to the coolant would not be designed to promote cavitation.

TTFN

FAQ731-376