Intercooler modelling programs

[oldcorollas]

even for a given construction, there are too many complexities to generalise it..

you could simplify it to be that the actual material of both has same conduction and transfer to air, then work out the internal and external surface area of each design, air volume and velocity (assume same pressure), residence time, then calculate based on the effective "thickness" of air that needs to be cooled by each unit of surface area.
but then you have more turbulent flow with the internal turb things (it doesn't matter if they don't have contact, thats not what they are there for), which will increase the....i forget.. reynolds number? the turbulence term in heat transfer equations....

anyway, as you said, less residence time but more surface contact (and probably better turbulence) but how much more area?

What's more important..area of contact for heat exchange or time in contact with surface?

both
BUT, as time increases, the effectiveness decreases, since the rate of heat transfer decreases as the temperature difference decreases... so a 1m long tune may not be a heap better than a 30cm one.

heat transfer rate decreases with temp difference decrease as a sort of inverse square(???) (like this, but this graph is heat transfer thru a piece of copper between cups of warm and cold water http://www.vernier.com/caliper/fall0...raph_large.gif )..

heat flux is a function of the transfer coefficient, the area and time, on top of the above temp graph, ie the temp difference.
so you need to do some calcs to plot the temp versus time, for a given surface area, assuming same heat transfer coefficient, and then adjust that for the resience time due to the different lengths...

still way out and fudged, but without knowign specific data, which is all based ont he geometry adn turbulnce, is abotu the best you can do.

should be a lot of literature/knowledge on industrial heat exchangers fromt he past century?


edit: oh and mass is very important, since you are working on time based events, so the heat capacity will determine how fast the alloy heats up and how hot it will get (and thus how hot the air remains, balanced by the heat flux in and out...
so a heavier IC will maintain temp difference to internal air for longer = more heat lost, but not have a big temp difference to the outside, and lose heat slowly from alloy.
conversely, a thin/light IC will heat up faster initially (at inlet end) but will more effectively lose heat to surroundings..
so you can model that difference as well

[JustenGT8]

Oh well, saves me the brain effort of tackling the calcs then as i described above, the tube and fin will flow more and i think this potentially holds more hp gains than a temp drop eg 1hp per 1.5 CFM gain versus 1% gain per 10 deg drop (common rule of thumbs).

Aiming for 450rwhp i doubt i'd get much more than a 10 deg temp difference but with 170% flow increase (potential) you'd think CFM could go up quite a bit?

[JustenGT8]

but then you have more turbulent flow with the internal turb things (it doesn't matter if they don't have contact, thats not what they are there for), which will increase the....i forget.. reynolds number? the turbulence term in heat transfer equations....



edit: oh and mass is very important, since you are working on time based events, so the heat capacity will determine how fast the alloy heats up and how hot it will get (and thus how hot the air remains, balanced by the heat flux in and out...
so a heavier IC will maintain temp difference to internal air for longer = more heat lost, but not have a big temp difference to the outside, and lose heat slowly from alloy.
conversely, a thin/light IC will heat up faster initially (at inlet end) but will more effectively lose heat to surroundings..
so you can model that difference as well

Not sure i agree with that 1st statement as the turbs' need to transfer their stored heat to the tube and then to ambient air...the ARE site mentions this as being very important and i'd have to agree.

If just an on boost heat sink then not so much perhaps? as they can then just dump the heat back into the off boost air, but in steady state at the track, which is as close as you'll ever get, where you need the actual heat transfer to ambient, it is crucial.

and your last point is one of the tradeoff i'm aware i'm making. I'll loose the big heavy heat sink bar and plate, which does a very good job at absorbing the initial doses of heat. And instead i'll have a poorer heat sink that hopefully performs better under more constant heat loads at the track?

[oldcorollas]

what is your current IC inlet temp?
is the water injection pre or post IC?

[trdee]

Gonna agree with you there justen. At the track you will always have lots of air flowing through it, so as a rule of thumb it is better to have something with a higher rate of heat transfer than just a large heatsink. Initial heatsoak can be combatted with your water injection setup, and really, it's not a drag race, so who cares if it heatsoaks in the pits? More important to have it cooler when on the track

[JustenGT8]

That's a prob Stew as i don't have much hard data.

I have tested final inlet temp on road. It was approx 30 deg ambient...cruising temps got as high as 50 deg. As soon as the foot went down, inlet temps dropped like a stone with increased airflow and even at sustained peak boost didn't get above 35 deg. WI is post IC and pre TB.

Same for multiple runs. You could see the IC dump heat back into the intake charge off boost and temps would climb again but as soon as there was any decent airflow temps would drop again.

The setup is very good for inlet temp control and one of the factors why i could run 16psi on a 10:1 engine i guess?

Actual flow and pressure drop is unknown other than obviously pretty good as has supported 400rwkw.

On the stock inlet the 12psi wastegate springs delivered 12psi manifold...not a true test of IC pressure drop but a good indication that the pressure drop is small. Interesting with the higher flowing short runner inlet manifolds, MAP dropped to more like 10psi? Maybe the compressor wheel couldn't supply enough airflow at that PR?

Anyways, what i have seems good but it's heavy....i want the weight reduced and moved further back. What i don't want is a new IC that is light but has cost me power

[JustenGT8]

Hmm latest brain fart. The x-sectional area of the pipe leading to the IC is 4560mm2...versus the 9750mm2 for the bar and plate.....maybe i'm worrying over nothing RE flow?

Although, if i can do the vertical core and get the dual entry my pipe flow then becomes 6350mm2.

Bugger me there are alot of variables to consider

[Supra967]

Haha, seems a rip off that most intercooler setups come down to 'buy a massive chinese cooler' hey? Plenty involved.

Go with the water coolers to really confuse things

[oldcorollas]

Not sure i agree with that 1st statement as the turbs' need to transfer their stored heat to the tube and then to ambient air...the ARE site mentions this as being very important and i'd have to agree.

If just an on boost heat sink then not so much perhaps? as they can then just dump the heat back into the off boost air, but in steady state at the track, which is as close as you'll ever get, where you need the actual heat transfer to ambient, it is crucial.

and your last point is one of the tradeoff i'm aware i'm making. I'll loose the big heavy heat sink bar and plate, which does a very good job at absorbing the initial doses of heat. And instead i'll have a poorer heat sink that hopefully performs better under more constant heat loads at the track?

thermal mass of the turbs? relative to the tubes, they should be lower, plus they will heat and cool with the air temp. ie initially they will help cool, then do nothing (during a power run) for heat loss directly, but the added gas/wall contact from turbulence is what they are there for.
it is better if they have good contact, but the amount of contact required depends on the ratio of heat transfer from air to alloy, and within alloy.
alloy heat sinks (like PC ones?) often have long thin alloy fins, with small cross-setional area, as the alloy conducts quickly.

if you are comparing an IC with and without turbs, then the transfer to ambient during long boost is not relevant, since the non-turb doesn't have them (or the turbulence) and the turb'd one they will be effectively thermally transparent durign boost anyway...

re: last point.. all things equal, a thinner may have a higher metal temp (at least at start of IC) but a more consistent temp over repeated runs around a circuit... but if the heavy one has enough cooling between runs to keep it's temp below the lighter one, then the heavy one might eb better


edit: extra last point... the lighter IC will rely more on airflow/car speed for it's cooling capacity... but how much difference in weight is there?

[oldcorollas]

Hmm latest brain fart. The x-sectional area of the pipe leading to the IC is 4560mm2...versus the 9750mm2 for the bar and plate.....maybe i'm worrying over nothing RE flow?

Although, if i can do the vertical core and get the dual entry my pipe flow then becomes 6350mm2.

Bugger me there are alot of variables to consider

for same x-sect area, a lot of small pipes will always have higher resistance to flow than fewer larger pipes...
the "effective" area of the IC may be much smaller...

but then again, think of cats.. they have 1-2mm squares and flow a lot, even for small total x-sect area... but their entry/exit fromt he "tubes" is a lot better too..


water coolers are easier as you really only think of one side of the heat transfer equation, since the water temp should always be near ambient, and huge thermal mass, and assume you have a big enough water to air elsewhere

[oldcorollas]

Haha, seems a rip off that most intercooler setups come down to 'buy a massive chinese cooler' hey? Plenty involved.

yes and no..
if you will only ever do a squirt here and there, or a single drag run/dyno queen pull, then a big farkoff heavy IC will most likely be better. heck, give ti a water jacket to increase thermal mass = even better....
cool that water.. then equals WTA

but for anything requiring more than one throttle application, the heat flux balance becomes much more complicated, and becomes much more science than knuckledragging
you won't see any highly developed turbo race cars with "china spec" cores.. BUT depending on their intended car speed, and time at speed, they may have calculated the optimum thermal mass to maintain most consistent temps, or best temp increase relative to car speed during accel...

and all will come down to empirical testing before you can tweak it

[Supra967]

but for anything requiring more than one throttle application, the heat flux balance becomes much more complicated, and becomes much more science than knuckledragging

Thats the problem though isn't OC? Your average back yardie/ once a month racer probably won't get into the nitty gritty of heat transfer when they can quite successfully whack on a big core and go for it. Even working out the heat flux is only half the story, and it becomes much more difficult to work out the actual heat transfer through the tubes/fins etc to compare designs.

You'd be best off logging temps for different coolers, and go with the lightest cooler for the desired outlet temperature under specific conditions.

[oldcorollas]

sure, if successful does nto involve doing 10-20 laps at full noise around a track.. unless they go stupidly large, and then they are wasting weight and response and front of car real estate.

yes, empirical testing of a bunch of coolers is what most people end up doing. or at least enough empirical testign to be able to predict behaviour under varying conditions.

[trdee]

ive designed 40,000L shell and tube heat exchangers for a peroxide plant before and even i would say fuck trying to calculate the figures on a front mount there are just too many unknowns.

[mattysshop]

PWR seem to have a calculator of sorts... when calling them to put twin GT35r's on a 419 chrysler smallblock, they recommended a close mesh tube and fin core, measureing at least (can't think off the top of my head) dimentions..... obviously folded and made custom end tanks to suit the application.... thought about just ordering the core from PWR and making your own end tanks, to suit where you want to run your piping?