Theoretical limit of intercooler size
09-14-2021, 07:57 PM
#1
Newb
Thread Starter
Join Date: Jan 2021
Location: San Mateo, CA
Posts: 35
Total Cats: -35
Theoretical limit of intercooler size
Hi all, I'm doing some research into intercooler design and want to do some custom fab work.
I've read plenty about the evils of the FFS kit and how much shade is thrown around here because of the relatively low power to cost ratio. I'm designing an intercooler for the cold side mount so it can make better power for those who already have one. Swapping some parts out, modifying others, adding new ones, etc. I'm not entirely familiar with the intricacies of air to liquid intercooler design, but I've been in contact with some companies trying to work out specs. The main thing I'm getting stumped on is this:
What is the relationship between intercooler thickness, fluid temperature, and fluid rate in regards to charge cooling?
For example, with a given air to liquid intercooler size, flow rate, and design, does doubling the fluid temperature to charge temperature delta also double the pre to post-intercooler charge temperature delta? Would doubling the flow rate have a similar effect? And what are the theoretical limits of these kinds of gains? Increasing flow rate infinitely can't infinitely cool the charge, but could decreasing coolant temperature do so?
What I've settled on so far is a small, full copper, liquid intercooler sandwiched between the supercharger and the manifold, being cooled by a separate coolant loop and rad with a very high flow rate pump.
I've read plenty about the evils of the FFS kit and how much shade is thrown around here because of the relatively low power to cost ratio. I'm designing an intercooler for the cold side mount so it can make better power for those who already have one. Swapping some parts out, modifying others, adding new ones, etc. I'm not entirely familiar with the intricacies of air to liquid intercooler design, but I've been in contact with some companies trying to work out specs. The main thing I'm getting stumped on is this:
What is the relationship between intercooler thickness, fluid temperature, and fluid rate in regards to charge cooling?
For example, with a given air to liquid intercooler size, flow rate, and design, does doubling the fluid temperature to charge temperature delta also double the pre to post-intercooler charge temperature delta? Would doubling the flow rate have a similar effect? And what are the theoretical limits of these kinds of gains? Increasing flow rate infinitely can't infinitely cool the charge, but could decreasing coolant temperature do so?
What I've settled on so far is a small, full copper, liquid intercooler sandwiched between the supercharger and the manifold, being cooled by a separate coolant loop and rad with a very high flow rate pump.
Reply
0
0
09-17-2021, 11:52 AM
#3
Junior Member
Join Date: Dec 2017
Posts: 57
Total Cats: 8
The general stuff:
The principle you're looking for is called the Heat Transfer Coefficient.
https://en.wikipedia.org/wiki/Heat_transfer_coefficient
The biggest think to keep in mind is ΔT is what matters. Maximize ΔT.
As far as practical principles are concerned:
The only significant portion of the core that does any actual cooling is the first 1/2" of core. It is imperative to maximize the surface area of the core. The remaining mass of the core acts as a heat sink, absorbing thermal energy until it can be conducted to the front of the core and be transferred to the incoming air. You can test this by adding thermocouples to your intercooler core in the 8 corners, doing a full throttle full boost pull, and logging temperatures. The bigger the core, the bigger the heat sink and longer periods you can load the core without giving it time to normalize.
Using liquid as an intermediary works in the exact same way. Water can absorb approximately 4.25 times more heat per unit mass than air can, and thus can act as extra heat sink capacity. Because no heat exchanger is 100% efficient, you lose a little efficiency converting the charge air to water. Ideally you would have a 20% larger surface area for a liquid/air exchanger than an air/air exchanger per thermal mass. So if you require a core with a surface area of 20x10 inches (200 sq inches), you would require a liquid air core that's (200/4.5)/.8 or 55 sq inches to have the same cooling capacity.
Regarding flow, increasing flow through the system will land you on a higher flowing and more efficient spot on the pump efficiency map. Depending on your pump curves, a pump with a higher head and lower flow ratings can get you more net flow vs a high lowing low head pump. This is why "ford cobra" pumps work better than rule "sump" pumps.
In my experience, the best pump for a street car is the Davies EBP40. It has an exceptional head/flow curve, draws less than 10a, and is compact so it can be easily packaged. One EBP40 flows the same as 3 cobra pumps in series (which increases head) while drawing less power. I would at some point like to play with a Shurflow gear-headed high pressure washdown or ballast pump. They are powerful, but a little bulkey.
Now to your example
Are you asking if you go from a 100° HX ΔT to a 200° HX ΔT, does charge temperature go from 50° ΔTt to a 25° ΔT respectively? The answer is not quite. The relation between HX and fluid medium temperatures is non-liner because heat transfer is not linear. However, within the practical limits of our applications, you can assume it to be so. Keep in mind, however, your fluid temperature will be a constant, assuming your have an adequate radiator and pump. The only variable will be your charge temperature.
LIke I explained above, increasing flow increases ΔT, but only up to the point that the incoming liquid doesn't increase in temperature as it exits the core. The theoretical limit of this in practical applications (LOL) is going to be pump efficiency vs flow instead of the heat exchanger. It will take so much pump power that the pump itself will heat the water temperature to above that of a lower powered pump. Decreasing coolant temperature works, but the question is how? Refrigerant systems consume more engine power to operate than they can theoretically add to the coolant on a closed system. A common practice in drag racing is to ice the reservoir down between passes to lower the waters temp. This only lasts a very short amount of time, literally 1 pass on the drag strip, before the ice melts and needs to be replaced. These guys also don't run a radiator on the system either. You could potentially add an evaporative cooler before the heat exchanger, but that's not going to really be practical. It would be hilarious to see going down the road though! You could add a liquid nitrogen or dry ice loop, but again, you're eventually going to run out of heat sink capacity there too.
Another thing to consider is you don't really want to use a copper heat exchanger in this application. Yes, they are more efficient, but now a days the only copper exchangers you're going to find either are junkyard pullouts from old ford trucks or are going to be a tube and fin construction instead of a multiple tube or bar/plate design with internal turbulators inside the coolant passages. Plus you cannot easily weld the core to flanges and run a gasket or o-ring for proper charge air sealing AND you run in to issues with galvanic corrosion. It's not really worth the added headaches vs an aluminum pre-manufactured heat exchanger.
The principle you're looking for is called the Heat Transfer Coefficient.
https://en.wikipedia.org/wiki/Heat_transfer_coefficient
The biggest think to keep in mind is ΔT is what matters. Maximize ΔT.
As far as practical principles are concerned:
The only significant portion of the core that does any actual cooling is the first 1/2" of core. It is imperative to maximize the surface area of the core. The remaining mass of the core acts as a heat sink, absorbing thermal energy until it can be conducted to the front of the core and be transferred to the incoming air. You can test this by adding thermocouples to your intercooler core in the 8 corners, doing a full throttle full boost pull, and logging temperatures. The bigger the core, the bigger the heat sink and longer periods you can load the core without giving it time to normalize.
Using liquid as an intermediary works in the exact same way. Water can absorb approximately 4.25 times more heat per unit mass than air can, and thus can act as extra heat sink capacity. Because no heat exchanger is 100% efficient, you lose a little efficiency converting the charge air to water. Ideally you would have a 20% larger surface area for a liquid/air exchanger than an air/air exchanger per thermal mass. So if you require a core with a surface area of 20x10 inches (200 sq inches), you would require a liquid air core that's (200/4.5)/.8 or 55 sq inches to have the same cooling capacity.
Regarding flow, increasing flow through the system will land you on a higher flowing and more efficient spot on the pump efficiency map. Depending on your pump curves, a pump with a higher head and lower flow ratings can get you more net flow vs a high lowing low head pump. This is why "ford cobra" pumps work better than rule "sump" pumps.
In my experience, the best pump for a street car is the Davies EBP40. It has an exceptional head/flow curve, draws less than 10a, and is compact so it can be easily packaged. One EBP40 flows the same as 3 cobra pumps in series (which increases head) while drawing less power. I would at some point like to play with a Shurflow gear-headed high pressure washdown or ballast pump. They are powerful, but a little bulkey.
Now to your example
For example, with a given air to liquid intercooler size, flow rate, and design, does doubling the fluid temperature to charge temperature delta also double the pre to post-intercooler charge temperature delta?
Are you asking if you go from a 100° HX ΔT to a 200° HX ΔT, does charge temperature go from 50° ΔTt to a 25° ΔT respectively? The answer is not quite. The relation between HX and fluid medium temperatures is non-liner because heat transfer is not linear. However, within the practical limits of our applications, you can assume it to be so. Keep in mind, however, your fluid temperature will be a constant, assuming your have an adequate radiator and pump. The only variable will be your charge temperature.
Would doubling the flow rate have a similar effect? And what are the theoretical limits of these kinds of gains? Increasing flow rate infinitely can't infinitely cool the charge, but could decreasing coolant temperature do so?
Another thing to consider is you don't really want to use a copper heat exchanger in this application. Yes, they are more efficient, but now a days the only copper exchangers you're going to find either are junkyard pullouts from old ford trucks or are going to be a tube and fin construction instead of a multiple tube or bar/plate design with internal turbulators inside the coolant passages. Plus you cannot easily weld the core to flanges and run a gasket or o-ring for proper charge air sealing AND you run in to issues with galvanic corrosion. It's not really worth the added headaches vs an aluminum pre-manufactured heat exchanger.
Reply
4
4
09-17-2021, 01:33 PM
#4
Junior Member
Join Date: Apr 2018
Location: Central Texas
Posts: 356
Total Cats: 93
Hi all, I'm doing some research into intercooler design and want to do some custom fab work.
I've read plenty about the evils of the FFS kit and how much shade is thrown around here because of the relatively low power to cost ratio. I'm designing an intercooler for the cold side mount so it can make better power for those who already have one. Swapping some parts out, modifying others, adding new ones, etc. I'm not entirely familiar with the intricacies of air to liquid intercooler design, but I've been in contact with some companies trying to work out specs. The main thing I'm getting stumped on is this:
What is the relationship between intercooler thickness, fluid temperature, and fluid rate in regards to charge cooling?
For example, with a given air to liquid intercooler size, flow rate, and design, does doubling the fluid temperature to charge temperature delta also double the pre to post-intercooler charge temperature delta? Would doubling the flow rate have a similar effect? And what are the theoretical limits of these kinds of gains? Increasing flow rate infinitely can't infinitely cool the charge, but could decreasing coolant temperature do so?
What I've settled on so far is a small, full copper, liquid intercooler sandwiched between the supercharger and the manifold, being cooled by a separate coolant loop and rad with a very high flow rate pump.
I've read plenty about the evils of the FFS kit and how much shade is thrown around here because of the relatively low power to cost ratio. I'm designing an intercooler for the cold side mount so it can make better power for those who already have one. Swapping some parts out, modifying others, adding new ones, etc. I'm not entirely familiar with the intricacies of air to liquid intercooler design, but I've been in contact with some companies trying to work out specs. The main thing I'm getting stumped on is this:
What is the relationship between intercooler thickness, fluid temperature, and fluid rate in regards to charge cooling?
For example, with a given air to liquid intercooler size, flow rate, and design, does doubling the fluid temperature to charge temperature delta also double the pre to post-intercooler charge temperature delta? Would doubling the flow rate have a similar effect? And what are the theoretical limits of these kinds of gains? Increasing flow rate infinitely can't infinitely cool the charge, but could decreasing coolant temperature do so?
What I've settled on so far is a small, full copper, liquid intercooler sandwiched between the supercharger and the manifold, being cooled by a separate coolant loop and rad with a very high flow rate pump.
Reply
0
0
09-19-2021, 11:58 PM
#5
Newb
Thread Starter
Join Date: Jan 2021
Location: San Mateo, CA
Posts: 35
Total Cats: -35
This is some awesome information, thanks for the lead gooflophaze, and thank you so much jiinxy for the thorough explanation. I'll definitely post back here with any additional progress I make, hopefully this pans out to something really cool and unique.
Yeah, I don't buy the whole E-Cool thing. It kind of makes sense and I'm sure it'll help slightly since I'm running E85, but I want significant headroom. I do like how it plays nice with using the stock ECU and an additional injector, but I really don't want to rely on it to cool the charge.
Reply
0
0
Thread
Thread Starter
Forum
Replies
Last Post
