Intercooler CFD Testing

[proxlamus©]

LoL..

 

Jeff.. I love you

 

 

astonishing data and WOW on those welds! very pretty =) did you weld that bad boy up yourself?!

[MONZTER]

I knew I saw your design from some where else, the book "Maximum Boost" has a section on intercoolers, I copied this.

 

Me too, Thats where I got the idea from. Good book. most of my parts I have been building are based on the "ideal" concepts from that book

[Lunar240z]

Hey! I did this too last year. And yeah i got the idea from maximum boost as well.

Highly, highly reccomended book.

 

 

 

You can see the cuvature of the divider on there in marker.

[Evan Purple240zt]

^nice!^

[Bernardd]

Great post mr monzter. I'm about to weld two GN cores to make a similar cooler as what you've got there. I'll do the same with the baffle if it's ok with you of course.

[240hoke]

You have been doing some absolutely stunning work. I cant imagine what your car will look like when you are finished!

 

Keep it up, im looking forward to whatever your next project is!

 

I need to get off my arse and start learning cosmos better, the flow analysis between the intake and intercooler have been Saweeet to look at.

[MONZTER]

Some people have asked about the flow characteristics of some of the intercoolers that are being used a lot, mostly the Treadstone ones. So I made some quick models from pictures I saw (so, no the results are not totally accurate, just for the concept and comparison. Remember there are no turbulators in the tubes so the pressures will be different and the flow results may be different.) here are some quick test.

 

From what I can tell the cast endtanks that blend smoothly into the inlet and outlets are a big plus. I would stay away from the square edge cut and weld like I bought. The sharp corner on the cut and weld ones seem to be contributing to the recirculation flow, the smooth cast tanks do seem to have a problem. Again this is only from my rough testing, don’t take this as a rule.

 

So I wanted to look at core sizing, and the fact that everybody runs huge intecoolers rated for 1200-1400 CFM when there only pushing 700 – which is good for probably 450hp.

 

I am showing below the following: 793 CFM 550hp flow 20psi boost. The plot is velocity from 0-1000 inches per second and I am only showing the range above and below this to help identify where the flow is going. From the picture you can see how little of the core is actually being used. I am not saying air is not flowing, it’s just going way slower than the front half.

 

Click on the pics to enlarge"

 

My modified version 6x18.5x3.5

 

 

Treadstone 6x18.5x3.5

 

 

Treadstone 6x25x3.5

 

 

Now same test at ridiculous flow 1200 CFM for both 18.5 long cores and 1492 for the 25†core (what the max flow is rated at)

 

 

 

 

You can see better use of the core and more uniform velocity when the flow is closer to its suggested rating.

 

So what does this mean? My best guess is that we are using intercoolers that are bigger than necessary. What is bad about this -

 

 

I would like to know your opinions about intercoolers that are way bigger than the flow they need. What are the plusses and what are the minuses?

[hughdogz]

Another great post, but I don't get it, lol.

 

Is this result opposite to what was in Maximum Boost? Their plots showed maximum flowrate occured at the far end of the inlet end tank, where your lower velocities are. With your plots, high velocity should be high flowrate since mass flow rate = (density * velocity * area) assuming 20 psi across the entire end tank.

 

For a larger intercooler core pros, you have higher thermal capacity. This has the benefits of resistance to heat soak since it would take more heat to change the average temperature. Plus it has a larger frontal area for more cooling area to be available. Generally speaking of course...as you've just shown it also depends on the IC efficiency. I'm sure you already know a heat exchanger efficiency is a measure of the ability to change incoming air to ambient (and how little the pressure will drop).

 

The only thing I can come up with for cons are perhaps packaging and blocking cooling air to the radiator. I'm curious if there is generally a bigger pressure drop. Definitely more volume to pressurize, but I don't know if this will be noticeable in increased "turbo lag". Even with a bigger IC, mass flow rate is the same in or out.

 

If you can modify the end tank to straighten, distribute and direct the flow evenly, it seems you're increasing the flow and thus efficiency of the IC and getting the larger frontal area too! If you get a couple more psi drop you can crank up the boost to compensate, as long as you're not already maxing out the compressor efficiency.

[hughdogz]

*Bump*

 

Okay, come on guys...I know there are a lot members who know much more about IC design than me that could contribute to the discussion.

 

I would like to know if any of my comments were wrong. Anyone? *crickets chirping*

[ihatejoefitz]

There is some happy medium between flow and heat transfered. Slower air through the bars will result in cooler air. However, slower air would mean less flow. I think that an equation could be derived to show the equilibrium (perfect size), but in order to be effective it would have to consider the entire range of flow (inside and out).

[TimZ]

Check out the velocity. The air seems to stay more consistent in speed, I assume this is why there is a little more pressure drop.

 

 

Check out the numbers. An increase of 162CFM through the same core.

 

 

So what do you think? Does it look like it all makes sense?

As always, this is some very nice work, Jeff! This should really prove useful to many of us, and the mods you outline are pretty easily attainable by most of us.

 

I do have a couple of questions, though. I'm a little confused as to how both the pressure drop and the flow changed between the two examples above. I would have expected to have seen one of the two held constant from run to run. Maybe I'm missing something - like maybe the two pressure references that you show are somewhere other than the inlet and outlet?

 

Assuming that I'm just missing something, I think if you dial back the flow to match between the two runs, you'll find that the pressure drop is actually less for the tanks with the dividers, since the flow increased by a larger amount than the pressure drop did.

[MONZTER]

Hi Tim,

Ya I guess you are right about the pressure drop thing and flow. Let me explain how I did the test. I let the boundary condition for the inlet simply be ambient pressure (around 14.7) and told the outlet boundary condition to be 25â€. I was trying to simulate flow as if on a flowbench. The new baffled design had better flow and the pressure drop was higher because of this. So I went back and re-ran both test. This time for the inlet boundary condition I gave it a volume flow of 392CFM (amount of flow pulled on the first test standard design) and the outlet a boundary condition of ambient ( so this is like if you blew 392 CFM of air in and just let the outlet be open. Look at the pressure drop now. The pressure drop for the standard original design stayed the same, as the flow was at the same, but look at how much the pressure drop decreased on the new baffled design.

 

What do you think about this test? Does it make more sense for checking pressure drop for a given design?

 

[TimZ]

Hi Jeff...

Yes that makes sense to me now - thanks for the clarification. For me, it's much clearer if at least one variable stays the same for the before and after. For instance, either keep the flow the same and measure the new pressure drop, or keep the pressure drop the same and measure the new flow. When both change, it's harder to tell what happened.

 

So it looks like theres no real downside here, huh? I'm going out to the garage to get my dremel...

 

:biggrin:

 

...did I mention that's some really helpful work? Thanks again for sharing.

[MONZTER]

No Problem Tim,

Hey, I have never seen pictures of your intercooler set up. I know I have looked at your engine set-up for inspiration, but never recall the intercooler. Do you have anything to show? or specs, like inlet outlet size, core size, flow direction. I want to know what helps 600+ HP obtainable with no detonation:lol:

[TimZ]

I had to go digging to find a few pics.

My intercooler is a pair of Cartech 3" cores, with 2.5" at the inlet and 3" at the outlet. Similar config to yours, but not as pretty...

 

The "fencing" around the cores was an attempt to increase the airflow through the core - the one that is on the front if the IC is intended to funnel more air through the core, the one at the back is there to give me something to fasten to so that I could seal it to the rest of the cooling package and keep any air that goes around the IC from interfering with the flow through the IC (can't find a pic right now - I'll try to come up with one shortly)

[MONZTER]

That looks great Tim, The end tank shape looks like the ones I modeled from looking at the Treadstone. They worked waaaay better than the square Spearco ones in the standard CFD test. I also like your idea for the fencing. is the intercooler that close to the radiator or is there some plates or rubber seals that hooks to the fencing? Also do you have the front fencing extend to the inlet above or below the bumper as in one big ducted system. I experimented with completely enclosing the gap between the intercooler and the radiator, so basically the only air the radiator could get was through the intercooler. The car overheated immediately, Oh well

[dr_hunt]

I was thinking about your boundary conditions. Are you putting the outlet at 25" (assuming you mean vacuum) and the inlet at 15psi? If so then I think that in reality the outlet pressure should be an iteration of multiple runs to see what the pressure drop for a given mass flow rate would be.

 

ie. run 1 with 15psi and 25" vacuum, pd would be x

run 2, 15psi and outlet pressure would be 15 - x, pd would be x2

run 3, 15psi and outlet pressure would be 15 - x2, pd would be x3

etc, etc.

 

After n runs, say about 15 iterations your calculations would be more real world dynamic conditions IMO.

 

Edit; The reason I say this is that what we want to look at is a system under dynamic conditions that mimic actual runing. And actual runnng the manifold pressure will be positive, not a vacuum, therefore the core outlet pressure shouldn't be 25". The mass flow capability of any core will be less when there is more pd. Also heat rejection and thus efficiency of the intercooler depends on residence time in the core. Further, too high a velocity of air will actually heat up the air to some degree. Of course all of this goes out the window if I'm reading your 25" wrong.

[TimZ]

Found a couple of pics of the backside:

[Midnitz]

the modifications totally make sense, I don't understand why Spearco wouldn't perform their own CFD analysis.

 

The first velocity analysis cross section is right on.....the consistent size as you progress away from the inlet creates a pressure drop....thus a drop in velocity....by adding your radiused end, you have diminished volume enough to stabilize velocity across the whole intercooler.

 

The plenum divider also killed any low pressure spots by splitting the volume into two....and making the forced airflow more directional, this also mitigates the possibility of unwanted turbulence.

 

Excellent work! Now if every supplier could excersize this level of dilligence!!!

[Midnitz]

Now get the intercooler and spray on some thermal dispersant coating to optimize efficiency.