Can a water pump have too high a gpm rating?
 

Had someone ask me a teaser question yesterday:

"Can the water pump move the coolant too fast through the system to allow the radiator to effectively lower the temperature?"

His argument was: "At high flow rates the coolant is not in the radiator long enough to allow the temperature to drop as much as if it moved slower"

Part of the argument was to never remove the restriction of the thermostat even in a race car!!

I am pretty sure I know the correct answer. But interested in seeing what the forum techies think of the argument he presented.

I dont think so. I just replaced mine with an electric water pump which moves about 3 times faster than the stock pump and the car stays cooler now than it did before.

old wives' tale. The only issue you can ever run into is that the cooling system is so restrictive you are essentially dead-heading the pump and it works itself to death or cavitates like crazy. But otherwise, as volume increases for the same cooling system, the velocity of the water will have to increase and when it pushes thru the radiator and across cylinder walls it has a greater scrubbing effect for cooling.

It is a "Reynolds number thing" is you are willing to dig deep into how convection works between the fins and passing water - the higher the velocity, the more turbulence at the boundary layer, which increases the convection coefficient (h). Car "radiators" are really not cooling by radiating, they are convecting.

That said, there is a limit to how high "h" can be - theoretically math says it can be infinite, but all liquids will reach a maximum, beyond that is just wasted energy, so yeah, you could get more pump than you would need (and you won't find one on the market for a car BTW so no worries here) but you would have other issues, like blowing hoses off, severe cavitation in the pump itself (this you do have to be careful of if you have a really restrictive cooling system).

BTW, most electric pumps have a constant volume quoted like 35 gpm or 55 gpm, whereas a belt driven pump will start low and increase, say from 25 gpm at idle to over 100 gpm on race pumps. Most guys don't realize this and put E pumps on their cars and then they overheat on the road track - it is because at 7000rpm the belt pump would have been flowing 100gpm, not 35 :thumbs: Similarly, in traffic the 55gpm E pump will tend to cool the car better.

I have a 55 gpm Meziere on my race car and it will get hot on a hot day at the track but so do I so I just come in and we both cool off :lol: - nice feature to the E pump is I have a toggle switch to turn it off and on so I can leave it on with the fans in the pits and aid the car in cooling down :thumbs: If I were competitively racing my car and running sessions in the hour range or longer, I would go with a belt driven pump setup.

To answer your question, yes, it is possible to move coolant through a radiator too quickly for the coolant to dissipate heat sufficiently. That being said, I too, run the Meziere on my ProChargerd, modified C6. The 55 ghp is great as it is at that rate constantly and when sitting in traffic, my cars actually cools, and quickly. The reason mechanically driven WP fail to cool sufficiently at idle is that they are pushing coolant at a rate of 4 gph. I think I heard that the C6 stock pump moves coolant at 29 gph at 6000 rpm. My set up uses the Meziere, a Ron Davis radiator, and an LPE oil cooler. I never have cooling issues.

You have to worry more about cavitation than the liquid flowing through the radiator too quickly.
On vtec Honda motors the impeller design is less efficient than the non vtec siblings because the motor revs higher. With the same non vtec impeller design, cavitation would occur at a high rpm and cAuse problems.
That is with an rpm dependent pump such as a mechanical pump. With a constant rpm electrical pump you don't have to worry about that.

I am assuming you had this issue and aren't perpetuating a tale, so can you give us the specifics on the setup you had that was flowing the water too quickly? The capillaries in radiators tend to be pretty common sizes, and none are large enough to create a perfectly free stream thru them down the center away from the boundary layer, so given constant cross section and constant length in a given radiator, I am interested to know the pump volume that was determined to be too much.

I'm impressed....learn something every day on the Forum! :thumbs:

So the simplistic concept that the coolant does not stay in the radiator long enough is BS.

I would sure like to see some specific numbers on potential improvement in cooling vs increased flow rate.

I track my TVS2300 Callaway and am always looking for a way to knock down the temps.

If putting in a high flow pump would help take 10 degrees out of the ECT then I would consider doing so.

CyberGS : You have any practical math to back up the Reynolds number concepts!

Best I can figure from a quick search is that the Reynolds number (Re) is directly proportional to the volume of the fluid flowing in a pipe, and the heat transfer coefficient (h) for the cooling fluid to the pipe wall is also directly proportional to Re. Hence the cooling (heat transfer) from the coolant to the air flowing through the radiator is also directly proportional to the flow rate (gpm) of the coolant pump.

Did I get it right?

PS> Not clear that the heat transfer from the engine internal coolant surfaces is also improved in a direct relationship to the rate of flow. But even if it is less than direct it is not likely to be non-monotonic so a higher flow through the engine is likely better than not.

PSS> So how do I get a feeling for the OEM coolant flow rate? Have never seen any spec on this.

Your wrong. Peter Bryant,the designer of the Shadow Can-Am and Formula cars states in one of his books that they kept putting biggger and bigger radiators on the first Shadow Can-Am car and all that was wrong was that the water pump was moving the water to fast and it did not stay in the radiators long enough to cool.

Perhaps his Re was way to high. I think CyberGS is correct.

The fallacy in the concept that the water is not in the radiator long enough is that it also is not in the engine long enough. Point being that having the coolant in the engine for a shorter period of time limits the temperature rise of the coolant and hence the the radiator has proportionally less temperature to drop the coolant when it returns to the engine.

I firmly believe that if it were not for the improvement in heat transfer coefficient the cooling system would preform the same regardless of the rate of flow within reasonable bounds.

i think and know CyberGS is wrong, i have personally had many cars that i took the tstat out of and the over heated, the stat stuck car was heating so i removed it, then it ran hotter so i took the stat center out and it ran cool, it slowed the flow down, so even if the flow is high with the stat it controls flow without the stat no control of the flow proving that if the coolant moves into the rad and out before it can cool, and this has happened many times on different cars i have had as i prefer them to run cool in the 160 range, what happens is the engine keeps heating it and the rad doesn't cool it so the temp goes up the engine never gets cooler until you shut it off

Makes perfect sense...if water passes through the radiator at such a high rate it can't possibly dissipate heat properly.

Dennis:

Take a read of these two simple references on the topic and see if you still draw the same conclusion.

http://en.wikipedia.org/wiki/Heat_transfer_coefficient

http://en.wikipedia.org/wiki/Reynolds_number

I will stick with Dittus-Boelter correlations.

So does this make sense as well?

..if water passes through the engine at such a high rate it can't possibly collect heat properly.

Think about "Wind Chill Factors", another Reynolds Number related phenomenon. The faster the wind the colder it feels !! I dont think you can argue that fact.

Same is true in convection cooling, the faster the flow the better it transfers the heat !!

Now you're in my area. You can possibly move air to fast across an evaporator coil with less cooling effect....slow the cfm and temps will decrease. However, fan speed is not a factor here. One question for you. Does it seem reasonable for hot fluid to pass through a convector/radiator at a slower rate giving more time for heat to dissipate compared to speeding through so fast there is little time for dissipation and re-entering a already heated engine?

Dan i don't have to read it i am basing my opinion on my own experience's as i have been racing since 1971 and have seen it first hand many times, changing pumps bypassing hoses, heater cores top tanks and side tank 1 core to 3 cores and the problem was no stat it over heated proving flow is the problem, my 8th grade teacher use to say and drill it in my head all the time, don't believe anything you read and only half of what you see:thumbs:

common sense should prevail, take an electric stove ceramic cook top, turn it on hi the element stays lite, the water boils, put it on low or medium it pulses off and on, but the water still boils it just takes longer

the reason the car overheats with no thermostat is because the pump cavitates. the housing of the thermostat blocks off a bypass port on most motors.

and the stock pump is only efficent at a small range of rpms, somthing like 60% at 800 rpms 90% at 1800 rpms and only like 20% at 6000rpms, at too high of an rpm, the pump just keeps spining the same water around and around.

-carl

The math in general is simple, but it is a collection of equations. Generally what you summed up is true but it really does go deeper than that to be honest. You start out determining the "hydraulic diameter" of what you are flowing thru and there are various equations for this based on the shape of the geometry, then you calc a Nusselt number based on the Reynolds (Re) number (which you have to calc and where a velocity term exists) and get hold of tables for the coolant fluid for various temps to get a Prandtl number.

This is way simplified and probably a bit too loose lipped, but it is close enough for this discussion...

Velocity of the fluid is the key - so, if you take a pipe and a known flow rate, then increase the flow rate thru that pipe, the velocity had to increase. You can hold volumetric flow constant and reduce the diam of the pipe, but then you start playing a game between how quick h increases v. how much surface area you lose to convect the heat.

Anyway, Re is calc'd based on the length the fluid travels which would be constant in the radiator capillaries, there are two fluid "constants" (they vary with temp of the fluid, but let's say constant for sake of discussion) and then velocity. The only thing you can vary then is the velocity. As velocity of the fluid increases, the Re increases, which as you noted is proportional to convection coeff (h), meaning higher V = higher Re = higher h. At minimum you want the liquid turbulent - until the flow is turbulent the h will not increase well.

In math, of course, you can get values that approach infinity, but in reality water will exhibit an h that maxes out regardless of how much velocity you shoot for. Takes a fair amount of velocity to get highly turbulent tho'.

As for in the engine, regardless of the location or geometry, more turbulent, higher velocity coolant can be better, unless you have such a large hole that most of it is passing thru in laminar flow thru the center. There are enough turns and rough surfaces inside of an engine, the water should have a hard time staying laminar at all, but nonetheless, more turbulence is better.

Guys will often refer to the phenomenon as "scrubbing" the heat away. We will nearly design a maze for coolant to flow thru in lasers to "trip" the coolant to aid in getting the water turbulent. When I first started at my latest job they had a laser that would run away and overheat... I was stuck with a slow flow as they had a set gpm flow rate from the pump. I decreased the diameter of the passage in the cooling blocks (which means I am losing surface area for convection) but zig zagged it back and forth to increase the length and trip the coolant, with the smaller cross sectional area increasing the velocity so I could get a turbulent Re value. When I was done they had to actually turn the chiller temp up because the diodes started running too cool, below optimum range :thumbs:

If Bryant put a smaller radiator back in the car he essentially did what I just described above - reduced the cross sectional area for a given flow and increased the velocity of the water thru the reduced size and/or number of tubes, increasing the Re and therefore the convection coefficient. If I understood the comment correctly, the pump flow became a constant and he reduced the radiator size back down. Makes sense to me, bigger isn't always better. If he reduced the pump volume then my money is that he cavitated the higher flowing pump and mistook the heating issue as a flow rate thing.

For the flow rates we could get from any available water pump on market isn't going to flow too fast; if you have a system that causes cavitation with the new pump design that would be an issue.

Regarding a Tstat - again, I agree regarding a potential cause of cavitation. Some backpressure in the system is good, if you get too big of a hole to flow a given volume of water thru, the pressure goes to zero and it gets to be a bit like running downhill when you start to feel your legs getting away from you. My Meziere doesn't cavitate without one in the system, so my race car doesn't care if it has a Tstat or not. It isn't that the Tstat reduces flow (since it can't!) as much as it creates a restriction IMHO. If the pump flows 55 gpm, it is going to flow that whether you have a Tstat or not, your Tstat is just an area that the velocity of the water has to increase as it passes thru the restriction, the VOLUME REMAINS CONSTANT and the backpressure is key.

So to be clear, if your pump flows 5 gpm, it will do that regardless of Tstat or not, radiator size, whatever, assuming no cavitation. If it flows 105 gpm, it is going to flow 105 gpm as long as you don't cavitate.

But the OP asked if the gpm flow of the coolant can be too high, and given the pump capacities that are available, I say no. If you feel the need to run a Tstat, that is fine, but that has nothing to do with the pump flow rate - put a higher volume pump in your car and it should be more efficient, to a point of course, but it should be better. Put a monster radiator in it, not necessarily. There is always a point where good enough is just that.

You aren't going to find a thermal dynamicist that tells you low laminar velocity is better than turbulent, higher velocity.

Man, I am a wordy ba$tard :lol:

Thanks CyberGS.

I guess we will always have those that think the world is flat despite the math.

Being an engineer I am going stick with the math and try a larger pump.
Only question remaining for me is what is the flow rate of my stock pump? Would be nice to get some prediction of the cooling improvement that is based on change in flow rate.

PS: I suppose I could start a thread about 160 degree thermostats and how useless they are when the steady state temperature of the system is 230 under race conditions. But I won't because I am sure I will get a bunch responses from members that known they are better but have never tracked their car hard enough to draw a comparative conclusion. If I don't understand the science behind something I am not going to make the change.