The Relative Merits of Centrifugal and Axial Compressors (March 1951 Article)

[overscan (PaulMM)]

I bought a copy of the Journal of the Royal Aeronautical Society, March 1951 with an interesting article on "The Relative Merits of Centrifugal and Axial Compressors in Aircraft Gas Turbines" by de Havilland's E.S. Moult (presenting the centrifugal compressor's merits) and H. Pearson from Rolls-Royce (presenting the axial compressor's merits). The artic;e is quite interesting, but the General Discussion is fascinating, with a stellar array of participants.

 

[NUSNA_Moebius]

I find it fascinating how fast the Germans ditched centrifugal for mostly pure axial designs and I wonder how it played into thrust performance at higher speeds despite early British centrifugal engines having a clear sea level static thrust advantage. Aside from obvious frontal drag of centrifugal engines, the Germans knew where turbine engines needed to go to achieve higher speeds and save weight as engine sizes scaled up.

Are there any examples of engines with centrifugal compressors in any supersonic aircraft aside from the F124/F125 family?

 

[charleybarley]

Maybe this clears up, once and for all, why Germany went for the axial.
The RLM (German Air Ministry) asked industry to achieve 410 lb/sq ft to minimise nacelle drag. This information comes from an article by Heinz Hoheisal in on-line book "Aeronautical Research in Germany - from Lilienthal until Today" page 226.

 

[charleybarley]

I bought a copy of the Journal of the Royal Aeronautical Society, March 1951 with an interesting article on "The Relative Merits of Centrifugal and Axial Compressors in Aircraft Gas Turbines" by de Havilland's E.S. Moult (presenting the centrifugal compressor's merits) and H. Pearson from Rolls-Royce (presenting the axial compressor's merits). The artic;e is quite interesting, but the General Discussion is fascinating, with a stellar array of participants.

For free Flightglobal Archive has Flight 14 December 1950 with a couple of pages on the RAeS debate. The website is currently down for upgrade. I have the first 2 pages which I printed some years ago but there is more which I cannot check right now of course.

 

[overscan (PaulMM)]

Check your messages, charleybarley.

 

[Zoo Tycoon]

I find it fascinating how fast the Germans ditched centrifugal for mostly pure axial designs and I wonder how it played into thrust performance at higher speeds despite early British centrifugal engines having a clear sea level static thrust advantage. Aside from obvious frontal drag of centrifugal engines, the Germans knew where turbine engines needed to go to achieve higher speeds and save weight as engine sizes scaled up.

The Brits developed the axial flow gas turbine during WW2 ref the Metropolitan-Vickers F2 which first ran in 1941. It first flew in the Gloster F9/40 DG204 on 13 Nov 43 in underwing pods very similar to the Me262. Its thrust to weight was in excess of the both the Jumo and BMW engines. It went on to be the genesis of the Sapphire and larger Avon’s.

The reason why it didn’t make it into WW2 was Metropolitan-Vickers, in particularly there thermodynamic/Aerodynamic experts were heavily committed to an even more important project, development of the worlds first Uranium Isotope separation plant, which was delivered to the MS Factory Valley (P6), located near Rhydymwyn, in Wales during 1942. Industrial size Uranium isotope separation by gaseous diffusion is in essence a large capacity pump/compressor or series there of. After this, despite a brief burst of gas turbine work which saw the F2 developed into a multi shaft turbofan (known as the F3) . The team relocated to the US, I think Boston early 44, to support the development of the equipment which went into Oak Ridge as part of the Manhattan Project.

So German gas turbine strategy got axial flow into squadron service, whilst Brit gas turbine strategy got centrifugal flow into squadron service and enough HEU for an atomic bomb.

 

[steelpillow]

I wonder if Frank Halford was in the audience that evening in 1951. He was the architect of all the de Havilland engines; piston, turbine and rocket alike, as well as the last few in the Napier line. (His story is told by Douglas R Taylor in Boxkite to Jet: The Remarkable Career of Frank B. Halford, Rolls-Royce Heritage Trust, 1999.)

UK work on the gas turbine as an aero engine goes back at least to AA Griffiths at the RAE in 1926 but was not followed up at the time. Whittle knew that high performance demanded axial flow, but chose the centrifugal design as his starting point for its lower technical demands and ease of manufacture. The RAE responded to his arrival on the scene by beginning work again in 1937, but on the axial type and I seem to recall that this led on to the Metrovick work.

A set of Whittle drawings landed pointedly on Halford's desk in 1940. He and Geoffrey de Havilland were taken to witness a flight of the Gloster "Whittle" prototype jet plane. The centrifugal impellers were familiar to him from his work on superchargers and the urgency was extreme, so he pressed ahead with the Goblin.

The German axial types were always difficult to manufacture, unreliable and short-lived in operational use, while the British centrifugal types (R-R and DH) proved the opposite, Halford declaring that it had taken him 22 years to achieve 1,000 hr between overhauls for his piston engines but just 8 years for the Goblin. Whether the Metrovick axial could have raised its game in time, had they stuck at it, must be a moot point. (To be fair to the comparison, I believe that the UK designers had access to better high-temperature alloys than the Germans - a worm in the back of my mind suggests these might have been developed in the USA? They would have greatly improved reliability and performance.)

Halford gave his own lecture on "Jet Propulsion" in May 1946, to the Royal Society of Arts, predicting 15,000 lb thrust axial engines for transatlantic travel. It too seems to have been quite influential. But he was not able to turn serious design attention to the axial type until 1950, resulting in the de Havilland Gyron series.

 

[steelpillow]

I find it fascinating how fast the Germans ditched centrifugal for mostly pure axial designs.

This seems to have been part of the command structure's lust for supremacy.
Their engineers knew perfectly well that a centrifugal type would be easier to develop but even before the war were being forced into the "we mush have more power!" axial mould. One of the first production projects to be undertaken was a large and powerful BMW type, but it never made it out of the development lab, while its smaller siblings overtook it and went into production. Occasional efforts were made by the odd designer to go centrifugal after some axial project or another failed, but they were always stopped in their tracks.

 

[edwest4]

I find it fascinating how fast the Germans ditched centrifugal for mostly pure axial designs and I wonder how it played into thrust performance at higher speeds despite early British centrifugal engines having a clear sea level static thrust advantage. Aside from obvious frontal drag of centrifugal engines, the Germans knew where turbine engines needed to go to achieve higher speeds and save weight as engine sizes scaled up.

The Brits developed the axial flow gas turbine during WW2 ref the Metropolitan-Vickers F2 which first ran in 1941. It first flew in the Gloster F9/40 DG204 on 13 Nov 43 in underwing pods very similar to the Me262. Its thrust to weight was in excess of the both the Jumo and BMW engines. It went on to be the genesis of the Sapphire and larger Avon’s.

The reason why it didn’t make it into WW2 was Metropolitan-Vickers, in particularly there thermodynamic/Aerodynamic experts were heavily committed to an even made important project, development of the worlds first Uranium Isotope separation plant, which was delivered to the MS Factory Valley (P6), located near Rhydymwyn, in Wales during 1942. Industrial size Uranium isotope separation by gaseous diffusion is in essence a large capacity pump/compressor or series of. After this, despite a brief burst of gas turbine work which saw the F2 developed into a multi shaft turbofan (known as the F3) . The team relocated to the US, I think Boston early 44, to support the development of the equipment which went into Oak Ridge as part of the Manhattan Project.

So German gas turbine strategy got axial flow into squadron service, whilst Brit gas turbine strategy got centrifugal flow into squadron service and enough HEU for an atomic bomb.

Do you have more information about British atomic bomb work? And this idea that some men were transferred to the US during the war? Great Britain didn't detonate its own atomic bomb until 1952 and history tells us the US was unwilling to share development information.

 

[msxyz]

High temperature Nichrome alloys were known also in Germany before the war. It's just that, by the time they were needed in actual aircraft, the German industry (and access to strategic raw materials) was collapsing fast. The Jumo 009 was redesigned to be made of simple sheet steel. Not that the projected useful life of the aircraft itself was much longer than that of the engines. The Luftwaffe was outnumbered, so with the exception of few battle hardened veterans, most of the pilots were rookies that had to learn how to survive fast on their own or die trying, no matter how good the plane was.

It's interesting to note that the German engineer who designed the Jumo engines went on after the war to design the highly successful Lycoming T53/T55 engines still used today. The Lycoming T5x employ a mix of axial and centrifugal compressor stages (If I recall correctly the Germans described the combination of the two technologies in a same compressor as a 'Diagonal' compressor). The T55 also became the core of one of the first high bypass turpofans (the L502), very quiet and very efficient even for today standards.

Engineers at Heinkel also played with the concept of combining an axial compressor in front of a centrifugal one. Basically the axial part acts as a low pressure stage and the centrifugal part acts as a high pressure stage.

 

[red admiral]

Do you have more information about British atomic bomb work? And this idea that some men were transferred to the US during the war? Great Britain didn't detonate its own atomic bomb until 1952 and history tells us the US was unwilling to share development information.

Lots of the work was done in the UK and then ideas, materials and people transferred to the US. It was very much an international effort. The US unilaterally clamped down postwar making it noforn - yay Allies. Best thing i've seen on the UK's side of the story is in The Birth of the Bomb by Clark.

 

[edwest4]

Do you have more information about British atomic bomb work? And this idea that some men were transferred to the US during the war? Great Britain didn't detonate its own atomic bomb until 1952 and history tells us the US was unwilling to share development information.

Lots of the work was done in the UK and then ideas, materials and people transferred to the US. It was very much an international effort. The US unilaterally clamped down postwar making it noforn - yay Allies. Best thing i've seen on the UK's side of the story is in The Birth of the Bomb by Clark.

Thank you. This is something I will look into further.

 

[Zoo Tycoon]

I recommend you read Pulitzer Prize winner “The Making of the Atomic Bomb” by Richard Rhodes which is widely regarded as the definitive history.

Details of the early Uranium isotope separation work can be found here;-

M. S. Factory, Valley - Wikipedia

en.m.wikipedia.org

The work and personal transferred to the US after the Aug43 Quebec agreement;-

Quebec Agreement - Wikipedia

en.m.wikipedia.org

 

[charleybarley]

Are there any examples of engines with centrifugal compressors in any supersonic aircraft aside from the F124/F125 family?

I guess the F124 is a bit of a cheat in a 1950's era debate as the cf is the 8th stage and therefore, as shown in the F124 wiki article photo, doesn't seem to increase the engine diameter. There is still a debate today (which the cf obviously won in the F124) but I think it must be a bit different from the early days as it's about the highest pressure stages.

 

[Richard N]

A bit about the German engineer who led the Jumo jet program.

Anselm Franz - Wikipedia

en.wikipedia.org

"Anselm Franz (January 21, 1900 – November 18, 1994) was a pioneering Austrian jet engine engineer known for the development of the Jumo 004, the world's first mass-produced turbojet engine by Nazi Germany during World War II,[1] and his work on turboshaft designs in the U.S. after the war as part of Operation Paperclip, including the T53, the world's first helicopter turboshaft engine,[1] the T55, the AGT-1500, and the PLFIA-2, the world's first high-bypass turbofan engine."

 

[NUSNA_Moebius]

Are there any examples of engines with centrifugal compressors in any supersonic aircraft aside from the F124/F125 family?

I guess the F124 is a bit of a cheat in a 1950's era debate as the cf is the 8th stage and therefore, as shown in the F124 wiki article photo, doesn't seem to increase the engine diameter. There is still a debate today (which the cf obviously won in the F124) but I think it must be a bit different from the early days as it's about the highest pressure stages.

Having a final centrifugal stage works well in a "wasp waste" axial configuration since it allows for a return to full diameter creating better mechanical advantage that comes with a wider turbine. Increases the air's pathway through the engine though while not allowing as much volume per second IIRC. Then you have the scaling limitations of weight with a centrifugal impeller, hence the F124 seems at the practical limit for centrifugals, at least for low bypass turbofans. I've seen no other engines of it's type and beyond it's size with a centrifugal stage. Similar performance engines like the AI-25 are fully axial. The F124 is surprisingly supercruiseable too.

Kind of surprised the J85 is full axial, though I bet flow rate issues negated the use of a centrifugal stage since transonic capability was the goal from the get go.

These are all gueses BTW, feel free to school my dumb ass lol

 

[Zoo Tycoon]

Supersonic on a centrifugal;-

DH108 (of sorts)​

Dassault Mystere 4 (first 50 should have been)​

North American YF93 (if performed as advertised)​

 

[charleybarley]

Not supersonic, but nevertheless getting up there in core size, the Safran Silvercrest ( 11,000 lb) was a surprise apparently in having a cf stage. It didn't make the final hurdle, operability at altitude, accel times I think. This failing was the end of the Falcon 5X business jet.
It has twice the pressure ratio, 38:1, of the F124.

 

[charleybarley]

The choice is not limited to axial or radial as there is also in-between known as diagonal or mixed-flow as used in 1942 in the He S 011.

https://www.steamforum.com/pictures/historical%20evolution%20of%20turbomachinery.pdf

Although this idea may be thought to have no merit today it is used on the PW600, see https://patents.google.com/patent/US6488469B1/en for the usual very enlightening explanation for the current state of the art found in so many patents.

Off topic now, but just to illustrate in this regard, here is a patent which is a very good primer on the problems facing turbomachinery at high mach numbers. https://patents.google.com/patent/US3344606A/en

 

[NUSNA_Moebius]

The HeS 011's configuration is interesting. Maybe the designers figured that having the semi-centrifugal stage + ram air effect would give them a bigger pressure boost than going axial first?

As I mentioned earlier, there definitely seems to be a real practical limit to centrifugal engines when trying to scale thrust upwards. The Silvercrest's accel times are likely influenced by such a large heavy centrifugal rotor and I imagine it effects the actual exhaust speed regardless of pressure ratio. Trying to purpose build an engine specifically for near-transonic cruise (my assumption based on Falcon 5X and Cessna Hemisphere specs), and with a centrifugal stage to boot is pretty crazy.

 

[msxyz]

An axial compressor is harder to design correctly and, at certain regimes or in certain conditions is is prone to stall or surge. Other problems include the vibrations of the long shaft and vibration of the compressor blades themselves that, even if undetectable, it manifests itself as fatigue cracks in the blades roots. Early axial compressors were also quite inefficient; it took 20 years for the axial compressor to mature: even 'advanced designs' introduced in the '50s like the famed GE J79 had something like 17-18 stages for a modest 12:1 compression ratio!

About Heinkel designs, I've read that the team designing their turbojets developed the diagonal design as a mean to make their engines surge/stall proof and also to increase the compression ratio more than it was practical at the time using an all axial or all centrifugal design. Proof of this is the lower SFC of Heinkel turbojets. They were also lighter.

 

[Trident]

The diagonal or mixed-flow concept can also be applied on the turbine side, BTW - recent automotive turbocharger turbine rotors are starting to adopt such a configuration.

Like the Silvercrest it isn't supersonic, but the prime candidate for the largest flown engine with a centrifugal compressor has to be the Progress D-27. If the TP400-D6 gas generator as built is slightly smaller than the M88-based core considered initially, that would indicate the D-27 gas generator is roughly EJ200ish.

 

[NUSNA_Moebius]

There are a number of light turbofans using a double sided centrifugal rotor for compression one one side, and as a turbine on the other by using a reversed flow combustor.

 

[robunos]

Interesting, how is heat transfer from the 'turbine' to the 'compressor' side of the rotor managed in these engines ?
Also, AIUI, the secret of Whittle's success was that he realised, at the then current state of the art, that a centrifugal compressor was 'better' than an axial one, and that an axial turbine was 'better' than an inward-radial type. How has the inward-radial turbine design been improved in these modern types ?

cheers,
Robin.

 

[NUSNA_Moebius]

Interesting, how is heat transfer from the 'turbine' to the 'compressor' side of the rotor managed in these engines ?
Also, AIUI, the secret of Whittle's success was that he realised, at the then current state of the art, that a centrifugal compressor was 'better' than an axial one, and that an axial turbine was 'better' than an inward-radial type. How has the inward-radial turbine design been improved in these modern types ?

cheers,
Robin.

To be honest, I tried finding out what engine I've seen with a "centrifugal turbine" but failed to find it. However conceptually it would basically be similar to the compressor side using nozzles or vanes to direct fluid against the rotor blades from the outside going in. Not too different from an impulse turbine really, but would make for a short (but fat) engine using a single rotor for the turbomachinery that would help keep costs down while being rugged and probably fairly good with heat management.

Whittle's use of centrifugals was prob due to higher achievable pressure ratios and the already well understood centrifugal compressor designs used in super chargers.

 

[riggerrob]

The (Swedish) Lyungstrom brothers' first engine (1908) had a centrifugal turbine. Hot air exited the center gas passage, then flowed outwards between a pair of blade-studded discs. Blade long axis were parallel with the center drive shaft.
However, Lyungstrom only built a few centrifugal/radial turbines before concentrating on axial turbines.
More recently, a few helicopter inventors have proposed radial/centrifugal turbines, but none have entered production.

 

[riggerrob]

Dear Steelpillow,
Whittle's biography also mentions that he knew that annular combustion chambers were - in the long run - but started with can combustion chambers because because they were the biggest that he could test on existing (1930s) test stands.
First generation American jet engines burned their fuel in can combustion chambers. The second generation had can-annular combustion chambers and finally the annular combustion chambers of today. Engine designers need a decade or two to perfect cooling air flow before they could develop fully annular combustion chambers.