CFM - BHP and intercooler sizing.

[Evodelta]

Hi Guy,

Just a quicky (hopefully!)
Am I right in the assumption that a turbo engine needs to consume 150Cfm to give 100Bhp?
To explain a bit more clearly where I am coming from, this is to do with some intercooler sizing calculations I am doing, some cooler makers rate their coolers as '900cfm' so this should (in theory) flow enough air for a 600bhp car, correct?

900 / 150 x 100 = 600bhp.

Not of use to me, but still worthy of note in this post, does the same cfm - bhp rule apply to an N/A engine? I guess it should do.

Many thanks,
Martin.

[Acki]

Choice your intercooler after the pressure lose..

[Guy Croft]

Hi Martin

150cfm to give 100Bhp?

Unless I have misinterpreted what you¢ž¢re saying - no.

For a start it depends at what test depression the airflow is measured. I use 10" and 150cfm would typically be enough say, for a 16v n/a Vauxhall to give 240bhp or so.
If you pressure-charge then the power's going to increase sure, but by how much is rather indeterminate. Most turbo matching begins with 'how much power do you think you'll have' - a rather vague prediction without actual same-spec dyno records to fall back on. 150 cfm ¢‚¬Å“ well I can tell you that with the right cams and turbo it¢ž¢s plenty for an Integrale 16v to achieve well over 600bhp - given enough boost.


How much air does an engine need to develop 100bhp? For a start there isn¢ž¢t really a thing called ‹Å“head airflow in terms of what a head¢ž¢s producing. It depends if you are citing the ‹Å“bare port flow¢ž¢ with no valve fitted which is a fixed value, or the airflow under lift - which varies as you open the valve and where a mean (average) value is pretty useless. And the reality is you can cite bare port flow but it is the area under the valve-lift/cfm curve that really matters. You¢ž¢re looking for decent gains over standard at the various lift points and if you get it ¢‚¬Å“ you generally get an increase in power.

The response to improved flow characteristics varies from engine to engine. 110 cfm on a 2 liter 8v might be good for about 170+bhp (85bhp per liter) On a 1300 SOHC 8v you might get 100bhp per liter ie: 130bhp. I did a SOHC Peugeot 205Gti last year that gave a definite 174bhp on a 1600 motor with airflow of 116 cfm @ 10‚ at full valve lift. The logic being of course the more air the head flows the better it will work and the bigger the engine it more it needs.

If you double the depression on the flow bench from 10" to 20" bit like boosting at 1 bar (absolute pressure ratio of 2:1) would become about 1.4 x 150 - not double, ie: 210 cu ft/min. What does this tell us? If you¢ž¢re lucky the relationship to actual engine air demand might be a good one. It depends if you¢ž¢re looking at max power or max torque. At max torque if it¢ž¢s a good engine you¢ž¢re going to see 100% volumetric efficiency, at max power probably a lot less. Eg: 2 liter Vauxhall XE with max power 248 bhp at 8300 rpm, actually ingests per cylinder - about 316cu ft /min. And if the max torque was 185lbf ft at 6000 the total engine airflow would be 214 cu ft /min. (less air than at max power because the cylinders are not filling so well at high speed, but ¢‚¬Å“ provided the unit holds her torque well enough - you¢ž¢re getting more power anyway because the engine is cycling much faster).
If you bias max boost to the max torque rpm and pressure charge at 1 atmosphere (bar, 14.5 lbf/sq in) then that's going to become 428cu ft/min. 2 bar will be 3 times as much going thru the intercooler = 642 cu ft/min.

In terms of power one might reasonably expect that if the engine were 240bhp n/a with a head flowing 150 cfm at 10", boosting at 1 bar pressure (depending of course at what rpm the boost was produced) that would be a bit like doubling the flowbench test pressure and so multiply the n/a bhp by 1.4 and expect 336 bhp.
And so on.
Running two bar boost would be the equivalent of testing at 30" depression and the flow would be 150 x 1.7 or so = 255 cfm and maybe predicted bhp 1.7 x 240 = 408bhp.Very inexact, I¢ž¢m probably totally wrong about this, miles off the mark, and in reality the power depends on many things apart from head flow.

So you can see anyway that a 900cfm intercooler is pretty capacious. It¢ž¢s not going to give a catastrophic pressure drop but this can be worked out.
It's vital to remember that the intercooler is there to cool the charge and working out the size is quite an impressive undertaking. The best firms use computer programmes to do it, I know how using Excel but it's fairly laborious. The main thing is that the matrix is large enough to bring the charge temp down, eg dropping from 150 deg C to 40 deg C might need about 15kW heat rejection on a 16v 2 liter Integrale unit at 1 bar boost.

The flow potential of the intercooler is a function of its necessary size as heat transfer unit not the other way round.

Best to talk to a manufacturer really, see who's really clued up on this. Do a net search. There is a lot of guesswork in that business.

Hope this helps, happy to follow-up if need be.

GC

[Evodelta]

Thanks for taking the time for such a detailed reply Guy, I'm working on a reply too, but have been snowed under with a load of 'stuff' just lately.

Martin.

[Evodelta]

Hello Guy,

Well I read your reply and I don't get it, surely there has to be a direct relation between air consumed and power output?

I'd read this somewhere but couldn't for the life of me remember where, I've just retraced my steps and found it: http://www.rbracing-rsr.com/turbotech.html
It clearly states:

It's all about airflow whether it is expressed in cubic feet per minute (cfm) or pounds per minute (lb/min). In any case it takes about 150 cfm per 100hp....so for 300 horsepower you need a turbo that puts out about 450cfm.

I also just dusted off my copy of Corky Bells Maximum Boost And there on P.55 is a graph for estimating internal flow area of an intercooler, if you follow the scales across it clearly shows that 150cfm is making 100bhp.

Clearly something is wrong here??

As far as intercooler sizing is concerned I think I should open up a new thread for it sometime, it is an interesting subject which I have been studying for some time, but until my own design has been tested cannot speak too authoritatively on it.

Martin.

[Acki]

I don't understand why this is interested for making a intercooler.
There you have to find the way between cooling the air AND low pressure loses because of the small pipes which you are need to have a good cooling and when you have a good cooling the next problem is that the cooling needs energy...

[Evodelta]

I'm sorry, I don't understand what you mean.

[Rich Ellingham]

Martin, those statements relate to a starting point for turbo selection, if you could not guess a starting point for turbo selection based on HP and air flow, it would take a long time to select a turbo. They however are not talking about Head flow, which as Guy has explained is more complicated then the just the flow number or ability as there are issues concerning the quality of combustion, which related to numerous other factors, port velocity, fuel atomisation, injector positioning, chamber shape, compression ratio etc.

The turbo selection (estimation) equations rely on knowing the volume of air pumped by an engine at a given speed, estimating the Volumetric efficiency, as you will see in Corkys book (not as good as Graham Bell's book) adding in the Desired Pressure ratio then allows you to plot an engine speed dervied airflow on one map axis and then the desired Pr on the y-axis. These equations assume little about the whats happening in the combustion chamber, but more towards the air flow the engine theoretically could pump - Even then this is a compressor based equation, and no garauntee the turbine can spin the requird speed or go into over speed; hence the ultimate need for a turbo speed sensor, which perhaps is out of scope for club level projects.

rich

[Evodelta]

Hi Rich,

True, the quote from the RB racing site is for sizing turbos, but still the facts quoted should remain the same. The quote from the Bell book is squarely aimed at intercooler sizing.

I guess I'm trying to simplify things to make intercooler sizing easier to work out, as that is what I have been doing. In my studies on the topic over the last year I have found a HUGE variation in what one person says about size and another, using both manufacturers and theorists figures.

Some use the CFM rule to quote size, others will size an intercooler by quoting BHP.
Either I'm missing something or there is a lot of poorly constituted info being quoted in intercooler circles, for example:

Using the calculation in AGB or CBs book you will find that Lancia have sized the cooler on an integrale perfectly, it is a cooler matched for the output of the engine with only a little in reserve (it will go to about 220Bhp before significant flow restrictions are met)
An integrale specialist makes a replacement cooler which has less charge air face core area than OE, (but is longer) Are they offering value for money?

At the Autosport show on Sunday I saw two cars made by a well known company, they claimed power outputs of 500 bhp yet the coolers were not much bigger than the OE part on an integrale and they were laid flat on the back of the rear mounted engine where they must have received very little cooling air, for sure, with a big turbo you can force the air through something which is too small, does this mean that if these guys put a more sensibly sized (and positioned) cooler on their cars they would realise a big power increase? Or was the power rating quoted unrealistic?

Hopefully you can see why I am confused....

Martin.

P.S. Well done on finding your missing power, onwards and upwards eh? ;-)

[Rich Ellingham]

Martin I believe you hit the nail on the head yes there is far too much poor info coming from firms who can tig weld Aluminium and set themselves up as so called 'specialists'.
What C Bell has in his book is good enough advice, as you know on the Integrale a large cooler will block the radiator. This is not too much of a problem if the intercooler is not too thinck in the core such it stalls the air (like the std its intercooler, not a problem with a good duct and not in front of the radiator) and if the cooling fins are not so dense. However the intercooler will give up heat to the air passing through the ambient face, meaning the radiator will have a harder time. You will know these things from CB's book.

But, as regard to size take heed of the early example in his book of the Nissan 300zx versus the Porsche 911 turbo. The turbo you are using and the rPand airflow you are asking for will dictate how much heat will be produced by that turbo. Thus is you ask high boost and airflow of a small turbo it will be inefficient heating the air far more then a larger turbo; this means you intercooler will need to be larger to recover this inefficient unwanted heating of the air.
If you have a large turbo that runs closer to its maximum efficiency at the rpm and boost you expect to drive the car at then perhaps you need less interms of intercooling. remeber a race car will be at or near max RPM all the time, do not ignore the thermal load on all items in the engine, and you wouldnt want the turbo compounding the heat. On a road engine things are slightly different, as average RPM will be far lower, and the driving style will also be different.

As regard intercooler rating, its nonesense really, a large turbo on a large engine could make big BHP with no intercooler if the compression flow etc were all compatible without detonation. So rating a cooler by BHP is nonsense. In regards the airflow, perhaps that is more relevant, but I would imagine many coolers would flow a lot of air, but an estimation of charge face surface area would be useful as you could approximate that to throttle body size (round tube flow) as an estimation of 'flow bhp potential' (GC could probably give flows for a tube of say diameters 40, 50 60, 70 80 mm diameter). However the internal design with or without turbulators affects the drag on the air, this drag is ok for heat transfer, but bad for pressure drop, which I belive to be the main 2 factors in an IC. If it has long tubes with low density turbulators or short with high density, then similar heat rejection properties should be possible, although charge faces will be differing in area.

In my view I'd go for the largest I/C I could fit in a space such that it did not kill flow to the radiator and the rear face had an exit route of lower pressure to ensure flow through the core. Construction I believe the extruded type just look right in the ends compared to the tube and fin - but tend to cost more. the end cap should not be of a tortuous design (ref RS500 style coolers) ans also encourage even flow along the charge face.

Have a look at the docking site for ideas. http://www.dockingengineering.com/ thats real motorsport equipment.

rich

[Guy Croft]

I've read that helpful web link you posted Martin.

That '150 cfm for 100 bhp' - is that supposed to suggest the extra airflow needed to generate an extra 100 bhp by turbocharging?

Whether that is the turbo free-air-delivery (FAD) needed for the extra 100bhp or (ie: venting to atmosphere) or engine demand, an accurate prediction is a bit more complicated than just that.
That is based the prior knowledge of achieving 100% volumetric efficiency on one cylinder with a normally aspirated engine of your type - and the bhp it develops at max torque, then multipying by the number of cylinders, and then saying we are going to boost at 1 atmosphere (1 bar gauge) which will essentially double the cfm. That is the (cfm) airflow at 1 bar pressure the turbo must be able to generate at engine full throttle - max torque - and is also the intercooler/throttle plate flow.

I am building a computational table for this, but it is not ready yet, when it's ready I'll be able to tell you what your airflow regime is and the boost too, but in the meantime if you want - consult an intercooler specialist (in my view some of the best are actually in Australia though I have lost the link to the best - who actually publishes his thermodynamic calcs).

I've nothing against informed discussion but don't try and deduce intercooler design here, this is an engine website, ie: specialises in engine techniques - we cannot do everything; anyway it's way too complicated for 99% of users. Yes I know how it's done but I don't build them so it's not really important to me.

GC

[sumplug]

One thing to note, Mounting an intercooler flat ontop of the engine is a no no. The heat from the engine plays havoc with it and they become inefficient.
I remember a magazine article testing all the leading intercoolers on an Escort Turbo, and only one worked properly. They measured the temp at each end of the cooler and some of the famous name's coolers were no better then the OE one, and that was poor. If i remember, it was Pace who's IC worked and dropped the temp right down.

Andy.

[Evodelta]

Rich,

yes there is far too much poor info coming from firms who can tig weld Aluminium and set themselves up as so called 'specialists'

Indeed there is, so why not do the calculations yourself, then hand over your design to these who cannot, but can make it?

I agree with much of what you have paraphrased, but:

rating a cooler by BHP is nonsense

I disagree, having read both Corky Bells (whose company make coolers) and AG Bells I would say they do too, as this is how they calculate the size, hence my original question because I think that CFM and BHP must be related. Interestingly although it is in CBs book, I haven't found any mention of the CFM - Bhp relation in AGBs book (as yet)

In my view I'd go for the largest I/C I could fit in a space

I would like to take a more methodical approach and would take the temps I wanted to cool, the amount of air that I needed to flow for my power output and the use of the vehicle into consideration, then design the cooler to this spec. If It didn't fit I would then think of what I was going to have to compromise to make it fit, ie: Shape the cooler to fit the bodywork or cut the bodywork to fit a no compromise cooler, also, consider water - air systems (chargecooling) if they are suited to your needs.

I had a look at Docking Engineerings site, but whilst their work looks very nice they certainly aren't giving any help or secrets away, there is also the most terrible spelling mistake right in the middle of their home page!



Guy,

That '150 cfm for 100 bhp' - is that supposed to suggest the extra airflow needed to generate an extra 100 bhp by turbocharging?

Interesting, could be.... ;-) I'll give it some thought and have a root around as I think it's fair to say we have reached the end of my limited knowledge in this particular avenue.
Your computional table sounds very interesting, I hope you keep us informed, finding a good intercooler specialist is very hard especially one that will part with the specs of their own products and you are right, many are abroad, I have one coming from an Australian company and one from the US.
I agree with your comment that there is a whole lot more to it than this though, but amateurs like myself have to start somewhere, maybe you can 'dumb down' the theories so people like me can understand them, It may not be important to you, but it is to me.


Andy,

One thing to note, Mounting an intercooler flat ontop of the engine is a no no. The heat from the engine plays havoc with it and they become inefficient.

In theory yes, but Subaru have been mounting theirs here for 10 years or more, of course they get around the problem by mounting a huge vent on the bonnet, mounting the cooler here means less pipework = less lag, this is also coupled with a very short tube length within the core which again, cuts down on lag, see the diagrams below, they show how you can get much more flow out of the same capacity intercooler simply by changing the way you send the air through it.
Heat soak on a top mount is only a problem when stood, on the move it is less so.

Martin.

[Testament]

Another reason for all this CFM talk is to try and simplyfy things, for a true engineering perspective you really want the comperssor maps shown in pressure ratio vs. mass flow (most garrett maps are) and then you can sue an empircal number to convert this to a power figure, since (ignoring detonation etc. etc.) if the engine can pass a certain mass of air you can burn a certain mass of fuel with that, the calorific value of the fuel being known, then by using an empirical figure for the engines thermal efficiency (maybe 20% for a SI car engine) to get a mechanical power output.

If you do all that according to the rules of thumb garrett use you get something around 9 or 10hp per pound/minute of air (SI engine using normal petrol fuel), which is generally a pretty good ball park figure by most accounts.

For sizing your intercooler you need to know the outlet temperature of the compressor which is a function of boost pressure and inlet air temperature. Then depending on the desired intercooler outlet temperature you can take your mass flow and temperature drop, determine the heat rejection rate and chose the heat exchanger that can satisfy these with an acceptable pressure drop.

But thats just the charge air side of things, it's a whole lot nastier when you take the external air flow and its pressure drop into account as well. Hence you see all the "boys" driving around with intercoolers taking up the entire front bumper. Very few people go through all the calculations, generally when modifying (street cars) people just fit something bigger than what is required and it works reasonably well, they just end up carrying extra weight and probably restricting the radiator airflow.

[Acki]

@evodelta: Very nice images. This option I never thought yet. Thanks!
Do you have more such images?

Here maybe a image what you can help ;)