Air Flow thru rads

[ian.stewart]

Im having a discussion elsewhere about heat transfer and coolant speeds, and a few questions have raised their head that im not sure about, Its a long time since I have worked with fluid dynamics, that is if what I remember about fluid dynamics is relevant??
1, There has to be a speed where the flow of water thru the rad is to fast to allow good heat transfer to the air, or is it benificial to slow the water thru the rad by restriction to aid cooling??
2, Do fans, electrical or mechanical, pushers and pullers, help with cooling at speeds above say 10mph, which for arguments sake is the fans max speed of air flow??
3, What percentage of airflow that hits a typical ally rad actually passes thru the rad compared with airspeed of the car, and I assume is it possible to reach a point of max airflow that a rad will flow.
Ian

[JohnD]

Ian,
Recently, I came across a similar assertion to your Q1, that the coolant should not be pushed too fast through the engine, or it would not work efficiently. I can't understand that - surely it is the area available for the transfer of heat, the conductivity of the container and coolant that are critical. How can the velocity of the coolant be a factor?
I think of it like a conveyor belt. As long as you can load stuff and remove it fast enough to keep pace with the supply, it doesn't matter how fast the belt runs. Is this a valid analogy?
John

[ian.stewart]

Heat takes its time to transfer, its not an instant thing

[graham bahr]

above 10mph cooling fans do not help cooling, infact they actually reduce it as they restrict the natural air flow through the rad.

[JohnD]

Ian,
Agreed on time to transfer heat. However, when the flow velocity is high, then a unit amount of coolant may have less time to take up heat, but more coolant will pass in unit time.
Imagine a water bath, with a cooling coil, and that the bath is at constant heat. Run the cooling coil slowly; coolant comes out hot. Run it fast and the coolant will come out less hot, but the volume of fluid passing through the cooling coil will have been greater, so the amount of heat taken out will be the same, or greater.

I'm asking for advice too, but I can't see how flow velocity will cause 'inefficient' heat transport. Please show me how I'm wrong.

John

John

[dtype38]

Originally Posted by JohnD

Imagine a water bath, with a cooling coil, and that the bath is at constant heat. Run the cooling coil slowly; coolant comes out hot. Run it fast and the coolant will come out less hot, but the volume of fluid passing through the cooling coil will have been greater, so the amount of heat taken out will be the same, or greater.

Nearly right The amount of heat taken out will be greater! Ok, basic theory. With the higher flow rate, the temperature rise in the cooling coil will be lower, so the average temperature difference between the coolant and the bath water will be higher. Also, with liquid flow, generally speaking the faster flow will strip away more of the insulating boundry layer against the surfaces, giving better heat transfer. These two combined mean that faster flow nearly always gives better cooling.

I say "nearly always" because in practice there are complications. As you increase the flow around a closed system, the pumping power required increases and the friction in the flow increases. Both these put heat back into the coolant, so there comes a point when the energy put in for more flow outweighs the extra cooling.

More practically... long before that happens, an average engine coolant pump will reach a point where running it faster will be counter productive because the blade tips will stall and flow will dramatically reduce.

Even more practially... engine cooling systems are rarely a simple loop. The efficiency of the radiator is dependent on the shape of the header tanks and where the connections are. The flow round the block/head is extremely complex and can contain back-flow and eddy currents. Let alone any heater connections, bypasses, and recirculation circuits.

All of this means that the speed of flow chosen by the original manufacturer will tend to be the best compromise for maximum efficiency of the system. Any increase in flow could well lead to a drop in efficiency overall, even though the radiator itself may, in theory, be working slightly more efficiently. I believe, in many cases, this has lead to the idea that the coolant "needs time" in the radiator to shed heat. Sorry guys, been in thermodynamics for 20 years... it doesn't

[johnny yuma]

GREATEST danger if coolant moving too slowly,or stops,water will vaporise to steam around engine hot spots.I would think move the coolant at least as fast as manufacturer intended,or faster if it is sustainable.Dont forget to fully cowl the rad or air will bypass it.

[graham bahr]

Originally Posted by johnny yuma

GREATEST danger if coolant moving too slowly,or stops,water will vaporise to steam around engine hot spots..

hot spots maybe, but then thats down to poor internal design, forgetting airlocks, you wont get steam, thats why we have pressure caps, by keeping the system under pressure the boiling point of the coolant is raised, just in the same way lowering pressure reduces its boiling point, water boils at only 40c at the top of everest due to the very low pressure its under

[Al Weyman]

An old trick to slow coolant travel through the rad was to cut about a standard thermostate and just leave the outer ring intact, some companies actually sell a disc shaped like this but an old stat will do the trick.

[greenamex2]

For info some vehicles do rely on their fans above 10MPH. My motorbikes fans frequently cut in and out up to 25-35MPH depending on how warm it is. I know this because the fans are pretty noisy and I can hear them even at these speeds.

dtype38 - I am considering fitting an ally swirl pot/header tank to my kit kit car. Are they worth it? It's midengined and whilst the previous engine was fine (once I gave the air into the radiator somewhere to go!) it is about to get a doubling in BHP.

[johnny yuma]

Originally Posted by graham bahr

hot spots maybe, but then thats down to poor internal design, forgetting airlocks, you wont get steam, thats why we have pressure caps, by keeping the system under pressure the boiling point of the coolant is raised, just in the same way lowering pressure reduces its boiling point, water boils at only 40c at the top of everest due to the very low pressure its under

understood,but you will still get steam,just at a higher temperature.A 13LB cap will still release the steam[and water]otherwise the engine will,or already has,blown a head gasket.

[retro_msport]

water flow related effects on coolant temp... Ben Bowlby (sp) used to have a thumb screw restrictor on his coolant pipe in his supersport, he used to adjust this when racing to keep his engine running at the Dynos proven max power fig. Or put simply he squezzed the pipe or opened it out to change temp.

I think with rads and thermostats working correctly the ideal rad will be the smallest one that can cope (cool) the engine correctly, getting good airflow through the rads must be the most important factor, because no matter what you do with the flow if the air is stalled then you will have no cooling what so ever.

Ideally i would imagine that the area behind the rad should be a vacume, so that no restriction to positive pressure air infront of the rad is present. My thinking behind this is the temp differential between the two should be as high as possible... water hitting the rad at say 100'c and the rad being at say 20'c is better than a rad that has poor airflow causing it to stay at for arguements sake 60'c

[Tim Falce]

Originally Posted by dtype38

Nearly right The amount of heat taken out will be greater! Ok, basic theory.......and lots more boring stuff.....blah, blah, blah

So, does it matter how fast the water and air travels through the radiator and why does my engine overheat when I push it to high revs and won't cool down when I stop racing, I.E. after a race, on tickover, the temp will rise significantly before it cools.

[thebear]

Originally Posted by falcemob

So, does it matter how fast the water and air travels through the radiator and why does my engine overheat when I push it to high revs and won't cool down when I stop racing, I.E. after a race, on tickover, the temp will rise significantly before it cools.

Because of "Latent Heat". You haven't removed enough of the existing heat. A fast idle will do the most good. That is why after the checkered flag one should slow and continue around the course at a reduced speed on a `cool-down" lap.

A prior post mentioned "Boundary Layer". That is what forms at slow fluid speeds and the flow at that point is "Laminar" flow as one would find in aircraft wings. The ideal flow for heat exchange is "Viscous" thich is turbulent. The constant turbulence moves the center of the flow stream to the surface where it has more opportunity to release heat. The dwell time of the fluid is inconsequential at this point. Laminar flow acts like a thermal insulator.

Another post mentioned using a disc as a restrictor in place of a thermostat. That is only workable in a narrow range of ambient temperatures. Discs with different diameter holes are required to optimize the passage of coolant and subsequent heat transfer. Cavitation can also be caused in a fluid passing thru an orifice such as a hole in a disk.

[dtype38]

Guess I shouldn't have admitted to knowing anything about this.... but here goes:

Denis, a swirl pot is only needed if you have ongoing issues of getting air stuck in the block, rad, or some high part of the system. This tends to happen for a couple of reasons. If you have a "full flow" header tank (ie in series with the main circuit) then some of the air in it can be drawn into the flow and find its way into the block or the rad. A swirl pot should be able to separate out the air and put it back in the tank. The other reason is if you have a remote expansion tank and it is not at the highest point, when the pump stops, any air entrained in the system gets stuck in a high loop of pipe or the like. This then has to circulate the system when you restart before finding its way back to the tank. The swirl pot makes sure it is positively separated out and can't get into the pump. You can, of course, use a swirl pot with a rad cap on as the header tank if you can get it high enough.

On my car I've taken out the full flow header tank and put in a remote one instead. I've then run a 1/4" line from the highest part of the system (inlet manifold rail) to the top of the tank and run a 5/8" line from the bottom of the tank to the water pump inlet. This makes sure that there is always a small flow from the top of the system through the expansion tank to draw out any trapped air. If there are any other high pockets in the system, say the rad header tank, then just run another 1/4" line from the top of that to the top of the tank as well, and any trapped air will be able to escape.

Retro, interesting about the coolant restrictor. I think that works because engines tend to work best near their maximum working temperature, and overcooling them actually loses a bit of power, but I haven't done any measurements myself to prove that. That ties in, though, with only having just big enough a radiator to keep the engine within limits (besides the weight of course).

As for radiator airflow, I've certainly found that it is by far the most important issue in keeping the engine cool. On my old XJS, having the packing around the radiator come adrift and allow air to bypass the rad = instant overheating. I have found also that getting plenty air into the radiator isn't worth a hoot if it then has no place to go. The flow will just stall and give very little cooling at all. Making a definite exit path at least as big as the inlet path, then directing it into a low pressure area (say behind the wheel arches or out the middle of the bonnet) is just as important. On my car I had a full duct from my "whalemouth" sealed snug to my radiator, with three small bonnet vents for the air to get out. The engine used to run very hot and even with the fan running could overheat standing still in a hot paddock. I then threw away the duct, all the inner wings, cut away the bulkhead and put in a wider tunnel, and cut four big holes in the top of the bonnet. The idea was that a howling gale would run through my bonnet no matter how slow I was going.... last time I ran it at Lydden hill in 35 deg C ambient, it just topped 90 deg C on the water temp for the first time this year. Errr.... think I might have overdone it a bit there Does prove the point though.

Falcy, yes it does matter how fast the water and air go. Both should be as quick as you can arrange without cavitating the water pump or blowing your bonnet off. As for your other issues.... My best guess for those symptoms is that you're running lean at full load. That would make it run hot while you're racing, but the high coolant and air flow would just keep the thing under control. Soon as you slow down, the system can't keep up with the amount of heat that's built up in the block/head and the temp rises before it falls. Another possibility would be silt build up in the block causing a "dead spot" in the flow. You desilted that thing recently? I'm sure it's going to be something fairly basic like that, cos your rad is pretty much the same size as mine and you aren't making masses more power, so it should cope easy. Might be worth us comparing block, header and and mainfold temperatures straight after a race.

[graham bahr]

re swirl posts,

my radiator man, who also manufactures them for aircraft as well as cars, states that unless you have problems bleeding air out of the system, or have problems with localised overheating your better off without a swirl pot, as they can actually airate the water in the system.

how right he is i dunno, but every race car which i've had him involved in has had a normal non swirl pot cooling system, sometimes with seperate header tank and sometimes built into the top of the rad, but always an uprated rad by him and no cooling fans have performed perfectly with no overheating.

[dtype38]

I like your radiator man Graham