Supercharger question

How much effect does the engine have on the discharge pressure of a centrifugal supercharger?

Will a supercharger give the same mass flow rate of air when attached to engines of different capacity or configuration?

For example.
Rolls-Royce Merlin RM-17SM
Bore: 5.4"
Stroke: 6.0"
Capacity: 1649ci/27.0l

2620hp @ 3150rpm, +36psi boost (ie, 50.7psi (abs)).

Using my very basic calcs I found the engine 2.45m³/s of air (at normal temp and pressure).

Could that supercharger have been fitted to the Rolls-Royce Griffon?

Bore: 6.0"
Stroke: 6.6"
Capacity: 2239ci/36.7l

Max rpm: 2750

For the same air delivery at 2750rpm the boost required would be +28.1psi.

Could you fit the RM.17SM supercharger to the Griffon and reasonably expect 2600hp?

The RM.17SM used a 12.7" diameter first stage impeller and 10.7" diameter second stage impeller, with MS gear ratio of 5.79:1 and FS gear ratio of 7.06:1.
How much power would be required to drive the supercharger to get that 2620hp (achieved in MS gear, presumably at sea level)?

The Griffon used larger supercharger impellers, but I don't have access to that information at the moment.

The answer is yes and no. Yes a centrifugal compressor can be applied to a different size engine within reasonable limits. No the compressor will not be the optimum size for both engines. Even the optimal compressor for a given engine will only be optimal at a specific operating point, so if the rpm or boost move way from that point, there would be another compressor that would be a better match. I am sure there were many different superchargers selected for Merlins alone - for exactly that reason.

OTOH the Griffon would probably make the 2600 hp. The Merlin supercharger would probably only drop a couple of percent efficiency to run the same flow at 8psi lower boost (a vertical drop on the compressor map - example linked below.) Apart from the compromised efficiency at max power (more heat load to intercooler and more blower drive power) other areas of engine operation (esp' higher rpm lower boost) may be even worse.

The inducer diameter is the best dimension to use as a guide for compressor flow capability.

I will have a go at estimating the blower power consumption later.

http://en.wikipedia....nt_comp_map.PNG

Edited by gruntguru, 19 April 2012 - 04:05.

The RM.17SM supercharger made a lot of heat, requiring ADI to control temperatures. Most UK Merlins did not require ADI.

The Griffon's supercharger was 13.4" and 11.3". MS gear was 5.84:1, FS gear 7.58:1.

has anyone tryed really over compressing / heating the charge air
then cooling with inter-coolers
and then allow the charge to expand = cool to near start temps

say they start with 100 deg air
heat it up near 1000 deg / 10 bar by compression
inter-cool back to 400-500 deg
and lose some bar 1 to 2 or a lot more ?

then simply allow the charge to expand/cool to 100 but at 3-4 bar ?

yes pumping losses would be a factor
but does that idea have other problems ?

Problems include:
- Lots more work (power drain) in the compression process.
- Additional materials costs @ 1000 deg
- Expensive compressor and drive
- More heat to reject

A similar concept has been used to cool intake air below ambient.
1. Compress air to higher pressure than required boost
2. Cool to near ambient in an intercooler.
3. Expand air through a turbine. (this recovers some work and also achieves a lower temperature than expansion in a difuser.

This would be a useful cycle for the hillclimb car using a GT to provide boost, mentioned in another thread here. Intercooling below ambient can produce more power within the thermal stress and detonation limits of the engine. This could be achieved quite simply by adding a turbocharger between the existing compressor and the engine and using the air expansion cycle instead of exhaust to drive the turbine.

Doesn't sound very good for an engine supercharger.

has anyone tryed really over compressing / heating the charge air
then cooling with inter-coolers
and then allow the charge to expand = cool to near start temps

say they start with 100 deg air
heat it up near 1000 deg / 10 bar by compression
inter-cool back to 400-500 deg
and lose some bar 1 to 2 or a lot more ?

then simply allow the charge to expand/cool to 100 but at 3-4 bar ?

yes pumping losses would be a factor
but does that idea have other problems ?

ray b,

The process you describe is known as "turbocooling", and it's been well studied.

The thing you need to remember when considering any process like supercharging or turbocharging is the net gain or loss in work it produces. Intake air charge cooling produces thermal losses and flow losses. There is also the trade-off between intake air pressure and density across the heat exchanger. Whenever possible, it is always more efficient to perform/extract work on the intake/exhaust gas within the cylinder itself, rather than externally. That's why combustion cycles like Miller & Atkinson look so good in theory, and also why GDI is so effective.

slider

Doesn't sound very good for an engine supercharger.

Wuzak,

The tangential discussion of supercharger centrifugal compressor performance over the past 50 years seems like a very interesting technical topic to me. So it's a bit saddening to see a lack of participation.

Granted, not many people still living have an intimate knowledge of the aero performance of specific RR Merlin/Griffon supercharger impellors. As for me personally, all I know is that these impellor designs had terrible aerodynamic performance, and thus tended to create lots of compression heating. They were mostly radial blades with ruled surfaces. A modern digital 3D CFD aero design would give far better performance, just from a single stage.

Of course, that doesn't mean to imply that 50 years ago guys like Stanley Hooker didn't know what they were doing.

slider

I suppose that CFD is about all you've got for optimizing compressor aero design; there are probably not many wind tunnel analogs for superchargers. Aircraft turbine blades exhibit today the same sorts of complex 3D compound shapes seen in F1 cars that have undergone thousands of hours of iterative development in windtunnels. Are those turbine designs purely CFD developed or do the engineers have other tools like LDV and PTV for getting empirical data?

as turbo's turn 100k rpms or more

are there speed limits on other types of superchargers ?

or is the OP Q just turn up the speed by different gears

They all have a speed limit. The smaller the wheel diameter, the higher the rpm limit. Ususally the compressor will be designed to produce peak efficiency well below the mechanical speed limit. Operating speed largely dictates the pressure ratio (boost) available whereas flow bears little relationship to speed (unlike positive displacement devices.)

Wuzak,

The tangential discussion of supercharger centrifugal compressor performance over the past 50 years seems like a very interesting technical topic to me. So it's a bit saddening to see a lack of participation.

Granted, not many people still living have an intimate knowledge of the aero performance of specific RR Merlin/Griffon supercharger impellors. As for me personally, all I know is that these impellor designs had terrible aerodynamic performance, and thus tended to create lots of compression heating. They were mostly radial blades with ruled surfaces. A modern digital 3D CFD aero design would give far better performance, just from a single stage.

Of course, that doesn't mean to imply that 50 years ago guys like Stanley Hooker didn't know what they were doing.

slider

Slider,

I was looking for a good picture of a RR supercharger impeller, but haven't succeeded except for this diagram.

http://www.flightglo...ger-cutaway.jpg

And this model of a Merlin XX

http://www.enginehis.....lin XX/2i.jpg


It looks as though the lips on the intake area are bent, but otherwise they are straight. Later superchargers also had rotating intake guide vanes.

Sir Stanley Hooker's first contribution to the field was to discover that the governing equations for superchargers being used at the time were wrong.

Bristol Centaurus supercharger impeller

A modern digital 3D CFD aero design would give far better performance, just from a single stage.

At its rated altitude (Full Supercharger gear) of 27,000ft the Merlin 113's supercharger provided +18psi boost, and a pressure ratio of over 6. Could a single stage centrifugal compressor do that now?

At its rated altitude (Full Supercharger gear) of 27,000ft the Merlin 113's supercharger provided +18psi boost, and a pressure ratio of over 6. Could a single stage centrifugal compressor do that now?

I haven`t heard about higher than 3.5-4bar from Car turbos.

At its rated altitude (Full Supercharger gear) of 27,000ft the Merlin 113's supercharger provided +18psi boost, and a pressure ratio of over 6. Could a single stage centrifugal compressor do that now?

Wuzak,

A centrifugal compressor designed using modern 3D CFD codes would logically give better performance than a centrifugal compressor designed over 6 decades ago using some hand calculations, lots of intuition, and lots of rig testing. As for the performance of the centrifugal compressor itself, it doesn't know whether it's being driven by the crankshaft (ie. as a supercharger) or by an exhaust turbine (ie. as a turbocharger).

The design of a centrifugal compressor for a turbocharger would probably be a bit different than one for a mechanically driven supercharger. The turbocharger compressor design would make some compromises for turbine matching, and would likely have a smaller diameter and run at higher speeds. The supercharger compressor would have a larger diameter and run at lower speed.

The design of the Merlin compressor presented numerous issues that the typical automotive turbocharger design does not have to deal with. The Merlin compressor had to operate at a wide range of air density and temperature, and it also had to deal with the density and latent heat effects of the fuel being added to the intake air ahead of it.

Modern turboshaft engines use centrifugal compressors for the HP stage specifically because they can achieve high single-stage P/R's.

Regards,
slider

Modern turboshaft engines use centrifugal compressors for the HP stage specifically because they can achieve high single-stage P/R's.

Yes, but what P/Rs do they give?

You did state that a single stage modern compressor would outperform an old R-R two stage compressor. While it would be true for the lower altitudes, it probably wouldn't work for higher altitudes.

But then, of course, you could change to a multi-stage modern compressor.

Bristol Centaurus supercharger impeller

Beautiful casting. Bet it wasn't as shiny when it was new.

We have a sectioned Merlin 85 outside my office at work. The impellers aren't fully exposed but I will take a photo and post it.

Yes, but what P/Rs do they give?

You did state that a single stage modern compressor would outperform an old R-R two stage compressor. While it would be true for the lower altitudes, it probably wouldn't work for higher altitudes.

But then, of course, you could change to a multi-stage modern compressor.

Wuzak,

Sorry, I didn't mean to imply that a modern single stage CF compressor would outperform the Merlin two stage arrangement. I only meant to compare the performance of a single stage from each.

A good modern single stage CF compressor design will give a P/R of at least 4. That number could easily be higher, but there many compromises made to balance the requirements of efficiency, flexibility, complexity, compatibility with other stages, etc. The use of variable nozzle geometries also would help greatly.

Here's a good tech paper on a modern turboshaft engine using a centrifugal HP compressor:

www.grc.nasa.gov/WWW/RTT/docs/Welch_AHS65_final.pdf

Regards,
slider

Sorry, I didn't mean to imply that a modern single stage CF compressor would outperform the Merlin two stage arrangement. I only meant to compare the performance of a single stage from each.

Fair enough. The modern one should handily outperform the one from the 1930s/40s.

A big leap in performance was found when the Merlin XX went into production using a Hooker desigend supercharger. Over in the US compressor performance was also improved when, one by one, the engine manufacturers started designing their own impellers, rather than leave it to General Electric as tehy had previously.

A good modern single stage CF compressor design will give a P/R of at least 4. That number could easily be higher, but there many compromises made to balance the requirements of efficiency, flexibility, complexity, compatibility with other stages, etc. The use of variable nozzle geometries also would help greatly.

I think R-R determined that with the existing designs a P/R of greater than 2.5-3 would be more efficient with a two stage design.

That is not to say that a single stage supercharger couldn't get a higher P/R, and in fact I think many of them went up to 3.5-4.

Sir Stanley Hooker's first contribution to the field was to discover that the governing equations for superchargers being used at the time were wrong.

Whittle makes similar claims about jet engine compressors in his book "Jet".

Whittle makes similar claims about jet engine compressors in his book "Jet".

I don't know about that claim. A.A.Griffiths wrote a paper in the mid 1920s about turbine design, and described a turborpop engine (with multi-stage axial flow compressor), using aerodynamic principals. After Whittle wrote his paper he sent the paper to the Air Ministry, who then forwarded it to Griffith for comments. Griffith corrected an error in Whittle's calculations and criticised the jet concept.

I don't know about that claim. A.A.Griffiths wrote a paper in the mid 1920s about turbine design, and described a turborpop engine (with multi-stage axial flow compressor), using aerodynamic principals. After Whittle wrote his paper he sent the paper to the Air Ministry, who then forwarded it to Griffith for comments. Griffith corrected an error in Whittle's calculations and criticised the jet concept.

Woozy - I have to confess to having made a slight mistake - Whittle's similar claims were about jet turbine calculations - not compressor calculations.

A big leap in performance was found when the Merlin XX went into production using a Hooker desigend supercharger. Over in the US compressor performance was also improved when, one by one, the engine manufacturers started designing their own impellers, rather than leave it to General Electric as tehy had previously.

Stanley Hooker was definitely a bright guy. I remember reading his autobiography, "Not Much of an Engineer" where he talks about getting hired by RR. He was given free reign to work on anything he wanted to (can you imagine that happening at a large company today?). He chose to study supercharger aero design/analysis and fortunately he turned out to be quite good at it. One of those rare happy twists of fate- the right guy at the right place at the right time. While Hooker is the most well known of the early compressor designers, there were others like Sanford Moss at GE.

What gave GE an edge in turbochargers and early turbine engines was their extensive knowledge of high temp metallurgy, rather than aerodynamics.

Regards,
slider

What if you went the other way - stuck the Merlin's supercharger on a smaller engine.

For instance, if the two stage supercharger was adapted to the Peregrine (5.0"/127mm bore x 5.5"/139.7mm stroke, 1296ci/21.2l capacity)?

What if you went the other way - stuck the Merlin's supercharger on a smaller engine.

For instance, if the two stage supercharger was adapted to the Peregrine (5.0"/127mm bore x 5.5"/139.7mm stroke, 1296ci/21.2l capacity)?

Mating a Merlin 2 stage, 2 speed CF supercharger to a Peregrine recip engine could certainly be done if one were so inclined. But due to the high level of systems integration (fuel system, charge cooling, manifolds, drive system, etc.) on these engines it would be a huge effort. The engine would certainly run and the supercharger would likely compress the intake air, but how well the mash-up would perform is a big question.

Of course, there are a couple things about the Merlin supercharger system that could make the adaptation easier. For example, the ratios used in the gear drives could be altered as well as the engage/release points of the shift clutches.



slider