Comprehensive IC Water Pump/System Test Data

[Department Of Boost]

I just looked at that attached pdf. I think I've gone full tard. I'm not even sure I'm looking at it right. I'm reading that the best result they got was roughly 7.5gpm with the FloJet on a full system???

 

[Racer47]

The plot is pretty busy. The 2 upward curving black lines are flow resistance estimates with the left one being the high end of expected resistance and the right one is the low end.

The legend is not laid out exactly in order. The top 3 curves are actually the bottom 3 in the legend. So the highest curve is the WP29 at 14 volts. 2nd is WP29 measured. 3rd WP29 at 12 volts.

Then you start back over at the top of the legend. So the Orange line with round markers is the Jabsco 50860. Then from there on down in the legend is mostly in the same order on the chart.

The Poly. means its a polynomial curve fit to the data point shown.

 

[Department Of Boost]

So those are all estimates?

 

[Racer47]

Here is some of the same data but its easier to read (page 3 of the pdf). Maybe I should have posted this one first. This plot shows that the 2 Pierburg pumps are the best. The CWA50 is the GT500 pump. The CWA100 is what Mercedes uses.

Somewhere in the MBWorld link is another link to a Cadillac forum when a guy measured the actual heat rejection. The take away there is that you run into diminishing returns fairly quickly by just increasing flow.

So going from the 50 to the 100 shows about a 1 gpm increase (from appox 7 to 8 following the black curve). And looking at the heat rejection data, that 1 gpm will probably result in only a few percent increase in actual heat removed from the system.

So the added cost of going from the 50 to the 100 is probably not going to result in any meaningful IAT reduction for my KB and AFCO system. I bet the Mercedes systems are bigger and flow more water and see a real benefit from the CWA100's.

 

[Racer47]

So those are all estimates?

no its supposedly measured data using a paddle wheel style flow meter

 

[weather man]

Somewhere in the MBWorld link is another link to a Cadillac forum when a guy measured the actual heat rejection. The take away there is that you run into diminishing returns fairly quickly by just increasing flow.

Sorry, not believable.

 

[Racer47]

Sorry, not believable.

thanks for that scientific answer based on all of your research, read it here for yourself

http://caddyinfo.com/wordpress/how-...ressure-versus-pump-output-and-flow-motorama/

if you think its wrong, provide your data to prove it

 

[Racer47]

Using the chart in the link and also attached here, if I go from 7 gpm to 8 gpm (or 26.5 liter per min to 30.3) the result is from about 21 kW to maybe 21 3/4 (its hard to read exactly) that is a 3.6% improvement at best.

 

[weather man]

thanks for that scientific answer based on all of your research, read it here for yourself

http://caddyinfo.com/wordpress/how-...ressure-versus-pump-output-and-flow-motorama/

if you think its wrong, provide your data to prove it

My answer is from years of reading many sources. The article is interesting, but a mess.

 

[weather man]

Using the chart in the link and also attached here, if I go from 7 gpm to 8 gpm (or 26.5 liter per min to 30.3) the result is from about 21 kW to maybe 21 3/4 (its hard to read exactly) that is a 3.6% improvement at best.

It raises some interesting questions. What was the temperature drop with each Laminova core in the series. Is there a relationship between pressure and ability of heat exchanger to do work?

 

[Racer47]

I don't have all the answers but it seems very realistic to me. I've been an engr in the automotive industry for 25 yrs. I see this kind of data all the time. There is usually a point of diminishing returns. You can't just keep doubling one factor in a system and get keep getting double returns for it.

It seems very logical that installing bigger and bigger water pumps will at some point stop yielding real improvements. If all we needed was a bigger water pump to get the cooling needed, than we would not need to bother with bigger intercoolers and bigger heat exchangers. It all works as a system. This is the first data I've seen for actual (or estimated) heat rejection from an intercooler system (not some industrial cooling tower data). I think it makes perfect sense. Most everyone that messes with supercharger cooling always ends up with improvements that were less than expected. I think this data shows why that is.

Everyone is free to do whatever they want to. All I wanted was to see if I should spend more money on a bigger pump. And I think that answer is no. I'll get the CWA50/GT500 pump and call it good enough. Going from the stock KB/Bosch pump to the GT500 will very likely result in a significant improvement. However, spending twice the money on a bigger pump is very unlikely to result in much if any additional reduction in IAT.

 

[v6tungsten]

Best option for going cheap is 2013 gt500 pump and intercooler

 

[weather man]

I don't have all the answers but it seems very realistic to me. I've been an engr in the automotive industry for 25 yrs. I see this kind of data all the time. There is usually a point of diminishing returns. You can't just keep doubling one factor in a system and get keep getting double returns for it.

It seems very logical that installing bigger and bigger water pumps will at some point stop yielding real improvements. If all we needed was a bigger water pump to get the cooling needed, than we would not need to bother with bigger intercoolers and bigger heat exchangers. It all works as a system. This is the first data I've seen for actual (or estimated) heat rejection from an intercooler system (not some industrial cooling tower data). I think it makes perfect sense. Most everyone that messes with supercharger cooling always ends up with improvements that were less than expected. I think this data shows why that is.

Everyone is free to do whatever they want to. All I wanted was to see if I should spend more money on a bigger pump. And I think that answer is no. I'll get the CWA50/GT500 pump and call it good enough. Going from the stock KB/Bosch pump to the GT500 will very likely result in a significant improvement. However, spending twice the money on a bigger pump is very unlikely to result in much if any additional reduction in IAT.

I agree that real world reliable data is scarce. I have read a lot of college papers that kind of nibble around the edges.

I think my skepticism is based on the fact that moving to much coolant through our car's real world cooling loop, and specifically the intercooler, to the point of no gain would be a pretty tough thing to do.

After all, we pretty much read nothing but guys fighting iat2 temps. You don't read about to many saying they have no room to improve.

 

[Department Of Boost]

Using the chart in the link and also attached here, if I go from 7 gpm to 8 gpm (or 26.5 liter per min to 30.3) the result is from about 21 kW to maybe 21 3/4 (its hard to read exactly) that is a 3.6% improvement at best.

This will be entirely system dependent. I'm going to use extremes to make my point here.

System A

-Massive IC. 400cu in
-Low blower discharge temps. Let's say 8psi (190ish deg F)
-Massive HE. 2000cu in
-Low ambient temps. Let's say 55deg F

System A doesn't have a very high heat load at the IC and a whole bunch of HE with "cold" air blowing through it. You will have low water temps and not very much temp differential at the IC and HE. System A is about twice the size that is actually possible.

System B

-Standard size IC. 200cu in
-Reasonably high blower discharge temps. Let's say 18psi (310ded F)
-Average "big" HE. 500cu in
-Normal ambient temps. Let's say 85deg F

System B has a high heat load at the IC. It transfers a lot of heat into the water. Almost enough to boil it (this is common). The water going to the HE is therefore pretty darn hot. But the HE is not very big and it doesn't strip much heat out of the water. So the next time the water goes though it's hotter.

As the water gets hotter you want to move it faster. It will keep a higher temp differential at the IC, which is the first order of business. This will also make the water running through the HE hotter. Which despite common thinking is a good thing. The hotter the water is in the HE the faster heat transfer takes place.

If you add water speed to system A you won't see much of a IAT decrease because you already have a system that is doing it's job well. That is a fantasy system though. On system B you will see a bigger IAT decrease because there is so much to gain in efficiency at the IC and the HE with faster moving water.

The key to sussing this stuff out is to understand temp differential and that the wider the differential the faster (as a percentage) heat transfers. Once you look at everything through the lens of heat transfer it's a lot easier to think your way through what will show improvements and what won't.

Here is a good visual representation of heat transfer:

 

[Racer47]

I understand heat transfer. Here are the equations for reference

http://formulas.tutorvista.com/physics/heat-transfer-formula.html

And I'm not saying you are wrong. I agree that it all works as a system. What I am saying is that with a given real system and under the same real world conditions, continuing to increase water flow will yield smaller and smaller results. If increasing water flow alone was so effective, car radiators would be half the size and the water pumps would run twice as fast.

Here are 2 more another examples that illustrates exactly what I mean.

This one is worth reading at least the first half of the article.
http://www.achrnews.com/articles/128199-german-water-how-flow-rate-affects-heat-output

This is about cooling molds but its the same concept with the same results.
http://www.ptonline.com/columns/leverage-your-cooling-power

 

[Department Of Boost]

I understand heat transfer. Here are the equations for reference

I was directing that at everyone, not necessarily you specifically.

And I'm not saying you are wrong. I agree that it all works as a system. What I am saying is that with a given real system and under the same real world conditions, continuing to increase water flow will yield smaller and smaller results.

I absolutely agree. But real world testing on real world A2W IC systems with a dizzying number of HE, IC, pump and line dia combinations has show me that even at 20+gpm the returns, while diminishing, are nowhere close to "diminished".

If increasing water flow alone was so effective, car radiators would be half the size and the water pumps would run twice as fast.

Counter argument for fun:

Not necessarily. It's a whole lot less expensive to run a bigger rad than put a high dollar pump in. And the manufacturers can also avoid having to deal with things like cavitation issues. Oh yeah, high pressure/high flow water pumps take more power to drive. That hurts mileage. The OEM's are killing themselves to improve mileage.

They may not be using the best solution because of non engineering factors.

 

[Nutter281]

I'll try and simplify this topic as much as possible with some work I had previously done that is currently not be used - might as well hand a portion of it to the forum

As the water gets hotter you want to move it faster. It will keep a higher temp differential at the IC, which is the first order of business. This will also make the water running through the HE hotter. Which despite common thinking is a good thing. The hotter the water is in the HE the faster heat transfer takes place.

With all due respect, while squid has a very good breadth of practical experience in this field, most of the heat transfer explanations leaving his finger tips are perpetually dancing around anything that could resemble theoretical knowledge of the science. I have no idea where your graph came from but clearly it is two fluids in a parallel flow heat exchanger - not a single copper bar as the title describes and parallel flow doesn't exist between two fluids in a automobile coolant system (could also be two equal mass copper bars insulated and touching each other but that would be weird I can't think of why that would be relevant). That said, not everything you wrote or preach is wrong on this subject, let me explain.

The best advice on this subject that I can give is this: stop compartmentalizing the "temperature rise" of a fluid through one portion of a closed system, around here we support the belief in conservation of energy and mass. At the end of the day, the only thing that matters at steady state operation is the total thermal resistance from the IC to the outside air, and that thermal resistance is always reduced with increased coolant flow rate, though there is absolutely a point of diminished return.

It is best to consider the heat transfer path as a series of electrical resistors, I've illustrated below. To make the anology quite obvious (as I think most people can think in terms of electrical resistances), if you have 2 resistors in series: 100 Ohm and 100 Ohm, their combined resistance is 200 Ohms right? What if we take 1 of those resistors and cut it in half? 100 + 50 = 150 Ohms - that's a pretty good reduction in resistance. What if we keep halving that? 125, 112.5, 106.25, 103.13...... each time I am decreasing the resistance of the second resistor by a good amount, but the "Total Resistance" has hit a point of diminished return and their is an assymptote at 100 Ohms - no matter how "awesome" i make one of my resistors, eventually, the other resistance will be the only thing that matters.

In the illustration below, I have identified all of the thermal resistances between the IC and outside air. They stack in series and for sake of simple explanation, only those resistances with an "h" in the equation will improve with fluid flow rate, and of those, R1 and R6 will only improve with air-side flow rate - not coolant flow rate. So to drive the analogy home to the electrical resistance above, R3 and R4 are the 2nd resistor in the analogy that gets improved (reduced) with flow rate and R1, R2, R5, and R6 combine to be the first resistor that is unchanging.



Does increased flow rate reduce the overall thermal resistance? Absolutely, if R3 and R4 go down, the sum of R1 + R2 + R3 + R4 + R5 + R6 will also go down. Is there a point where reducing R3 and R4 no longer appreciably decreases the thermal resistance? Absolutely.

Below is the graph of thermal resistance vs flow rate for a common supercharger setup:




Below is a graph that shows the impact of coolant flow rate on IAC and total heat rejection. As you can tell, there is a region of strong 'gain' and then a region of 'diminished return'.





In conclusion, as squid correctly identifies, every system is different on the 'point of diminished return'. The related heat transfer physics provides an easy rule: the larger the heat exchanger or IC (and thus surface area) the greater the coolant flow rate required to hit flow saturation (or diminished returns). That is NOT to say that a larger heat exchanger will, by any means, perform worse than a small heat exchanger at identical flow rate, it is to say that the larger heat exchanger will reap more benefit from an 'increase' in flow rate.

Hope this helps!

BTW, I made the graphs above so please feel free to send me any questions related to their meaning.

Thanks,
Austin

 

[Nutter281]

Racer47, you are on the right track with your thinking. Just keep in mind the graphs you posted would be for relatively small liquid to air heat exchangers which is why they seem to assymptote at lower flow rates than what Squid is proclaiming. The necessity and benefit of flowrate is uniquely driven by the size of the heat exchangers in the loop.

Austin

 

[Department Of Boost]

I have no idea where your graph came from

google

but clearly it is two fluids in a parallel flow heat exchanger - not a single copper bar as the title describes and parallel flow doesn't exist between two fluids in a automobile coolant system (could also be two equal mass copper bars insulated and touching each other but that would be weird I can't think of why that would be relevant). That said, not everything you wrote or preach is wrong on this subject, let me explain.

I was simply trying to illustrate that when it comes to temperature differential and heat transfer there is a curve to it. I didn't grab that graph for anything but how the lines look. I didn't even look at it really.

 

[Nutter281]

LOL, I mean, I have to ask bro, what do 'you' think it means and how it applies to supercharger intercooler cooling dynamics?