Bjorn’s Corner: New engine development. Part 13. The compressor.

By Bjorn Fehrm

June 28, 2024, ©. Leeham News: We do an article series about engine development. The aim is to understand why engine development now has longer timelines than airframe development and carries larger risks of product maturity problems.

To understand why engine development has become a challenging task, we need to understand engine fundamentals and the technologies used for these fundamentals.

We have covered the need for the axial speed of the core air to decrease and increase depending on the needs of the core engine’s sections. Now, we will start to look at the different parts of the core in more detail. We start with the compressor.


Figure 1. The gas turbine cycle and its parts. Source: Rolls-Royce: The Jet Engine.

The radial compressor

There are two types of compressors: centrifugal and axial flow compressors. We start with the centrifugal compressor, which was the first type that was universally used.

It has two parts. The Impeller and the static diffuser+collector, Figure 2.

Figure 2. The parts of the radial compressor. Source: Rolls-Royce, the Jet engine, and Solar turbines.

The impeller rotates and throws air that enters the center outwards through centrifugal action. As the air enters the diffusor, it slows down, which, as we saw in Part 12, means the pressure is increasing.

The centrifugal compressor is simple and robust, with few moving parts. It also has a high stage pressure gain, up to 5:1. Its drawback is that it takes radial space and it’s cumbersome to lead the air from a stage’s diffusor to the center of the next stage’s impeller (Figure 3).

Figure 3. The dual radial compressor of the PW100 turboprop engine. Source: Pratt & Whitney.

Radial compressors were universally used in the superchargers for aircraft piston engines like the Rolls-Royce Merlin and are used for turbochargers for all types of engines.

They are also used in smaller gas turbines, either as the only compressor stage type (Figure 3) or as the last compressor stage in engines with axial compressors when the dimensions get too small for late axial stages to be efficient (Figure 4).

Figure 4. The Pratt & Whitney PW500 business jet engine. Source: Wikipedia.

The axial compressor

The axial compressor works on the same principle: rotating stages with blades that increase the air’s speed and fixed stator vane stages that slow the air and increase the pressure (Figure 5). The stator vanes also correct the direction of the air as the rotating blades deflect the air at an angle. The stator vane then slows down the air and changes direction to hit the next rotating stage at an optimal angle. The last stator vane directs the air straight back into the combustor.

Figure 5. The axial compressor. Source: Rolls-Royce, the Jet engine.

The axial compressor is compact and can easily be daisy-chained. Its drawback is a low stage gain, typically 1.1 for early designs and up to 1.5 for the latest designs. The low stage gain comes from the need to keep the flow around the blades and stator vanes attached to avoid local stalls. Because the trailing edge of a blade or vane is at a higher pressure than the leading edge, a tired boundary layer will easily grow to a trailing edge stall and then grow to a complete stage stall.

You talk about compressor aerodynamics working uphill and, therefore, being demanding, whereas turbine aerodynamics works downhill and, therefore, is more robust aerowise and resistant to disturbances.

Another problem is the gradually increasing temperature of the air. This forces the last stages of high-pressure ratio compressors to be made of special temperature-resistant nickel alloys. Compressor air is used for cooling many parts of the engine, but it can’t cool itself. For this, you need slightly higher-pressure air, and this higher-pressure air has an even higher temperature than the stage you want to cool.

The temperature of late stages in a compressor is a major problem for engines with high intake pressure gains, like supersonic engines. It limits the compressor’s pressure ratio. Supersonic engines, therefore, have relatively modest pressure ratios, as they want to use the free intake pressure gain to the maximum. Compressor material limits the final stages to about 700°C.

Because of their compact form, all larger engines use axial compressors. By adding stages, the pressure gain of the previous stages is multiplied by the pressure gain of the next stage. The initial axial compressors had many stages and low pressure ratios. An example is the GE J79 jet engine for the F-4 Phantorm, F-104, B-58, and Convair-880 airliner. It had a 17-stage compressor with a 13.5:1 pressure ratio. Today’s compressors achieve this with five stages.

Several problems crop up when compressors get into many stages. We will dig into the problem areas of multi-stage, high-pressure-gain compressors in the next Corner.

18 Comments on “Bjorn’s Corner: New engine development. Part 13. The compressor.

  1. Demo radial compressors have shown over 10:1 in pressure ratio, however they need to be able to handle transients in lots of different condition. One large radial compressor famous case is the Safran Silvercrest compressor. An axial compressor benefits from variable vanes and bleed valves in transient conditions. You need to test at Arnold Army Center or flight tests early.

  2. Because of their compact form, all larger engines use axial compressors.

    I thought axial compressors were favored over radial compressors because of their higher efficiency?

    Also, wasn’t the final-stage radial compressor popular because it made the engine shorter?

    • I am just an amateur if even that in regards to jet engines, but the radial compressor added diameter to the engine.

      Ergo, planes can be made longer, area rule involved, so a fat engine in a fighter is not a desirableness thing.

      At least back in the 50s fat engines were not for airliners either though now they are definitely huge diameters, but that gets into the use of the Fan for most of propulsion vs exhaust gasses.

      • The famous Mig 15 used a radial compressor in its engine which was derived from the engines RR sold to the Soviets, hence its fat shape.

        • I think direct copy is the right term but the point is valid, makes for fat engines which is not good for fighters though hard to argue that the Mig-15 was not pretty stunning and the F-86 had a fat engine as well (later versions would outperform the Mig-15 but despite its issues, well flown it was a tough adversary)

          • RR Nene only used a single compressor stage and that was double sided centrifugal. Not complicated at all by later standards

            License was also sold to Pratt who built it as J42 , for the Grumman Panther and as J48 with larger diameter

            The Soviets even sold a licence for china to build their unlicensed Nene derivative Klimov VK-1 (higher airflow from a slightly larger diameter)

          • I know this gets off the direct topic but I think its of interest.

            P&W is designing a replacement PT-6. Nothing to speak of out there but my assumptions current tech and materials that is not back fitable into the current PT-6.

            PT-6 is been amazing both design wise and longevity.

          • I can add that it was interesting to see a historic cost for the ME-262 engines was cheaper per engine by more than half than a piston engine for a fighter.

            Of course they were not reliable and severely limited run hours but fascinating.

    • That is correct, easier to get high efficienty of a larger axial than radial compressor. However below a certian size the last stages of an axial compressor gets “too small” and you go for a radial compressor, often a combination of a few axial stages and then a radial compressor on smaller engines. The radial compressor diffuser is hard to design like what Sir Stanley Hooker discovered on the RR Merlin early compressors and he fine tuned the diffuser to get more pressure and less temperature rise.

      • Correct, but there was something else about a radial compressor (besides the higher pressure gain and ability to handle smaller air volumes) that made it better than an axial compressor as the last compression stage, right before the combustion chamber. The preference dates back to the 1950s on the Chinook helicopter’s T55 engine (or its immediate predecessor), but I never understood the explanation. I thought the advantage was something to do with being able to position the combustor closer to the opposite end of the radial compressor (somewhat analogous to sticking a cup mute into the bell of a trumpet or trombone) to make the engine even shorter. However, Figure 2 makes it look like there isn’t room for that setup.

        • For medium power gas turbines 800shp to 7000shp you usually stop the axial compressor stages as the blades becomes under a certain size and then use a radial compressor stage to boost the pressure for combustion, hence the combustor can start at a fairly large diameter close to the compressor diffuser outlet. Note the volume flow increases alot after combustion and you design the turbine blades pretty small at a fairly large diameter.

      • There is also an issue with “backflow” / leakage for high single stage pressure gains.

  3. You can quite easily get high pressure air to a lower temperature by cooling it, however getting it into the late stages of the HP compressor blades, where it would need to be, is much more difficult.

    A power cell running on liquid hydrogen or ammonia would be good at cooling HP air. Otherwise the free airstream could be used.

  4. Bjorn lays out succinctly the advantages and disadvantages of the two types of compressors. To elaborate a bit (with his permission):

    1. It is hard to scale up a radial compressor. Adding more stages is not only cumbersome but involves more stagnation pressure losses in ducts carrying air from one stage to the next. As such, there is an upper limit to the number of stages and the OPR.

    2. By far the principal deficiency of radial compressors is simply their frontal area. For the same air flow rate, a radial compressor, which depends on centrifugal forces, can have almost twice the diameter of an equivalent axial compressor.

    Imagine trying to build a radial compressor equivalent to the axial compressor of GE9x! Just impractical.

    That said, axial compressors are finicky, especially high OPR ones. They work fine under design conditions, but off-design conditions require careful design. But engine manufacturers, with modern CFD tools, have mastered the design of axial compressors to the extent that it is very rare to have an engine surge during its lifetime.

  5. The axial compressor is compact and can easily be daisy-chained. Its drawback is a low stage gain, typically 1.1 for early designs and up to 1.5 for the latest designs.

    By adding stages, the pressure gain of the previous stages is multiplied by the pressure gain of the next stage. The initial axial compressors had many stages and low pressure ratios. An example is the GE J79 jet engine for the F-4 Phantorm, F-104, B-58, and Convair-880 airliner. It had a 17-stage compressor with a 13.5:1 pressure ratio. Today’s compressors achieve this with five stages.

    1.5 to the fifth power is only about 7.6. You’d need two more stages of 1.5:1 gain to get the pressure ratio to 13.5:1 or higher. What am I missing?

    The glaring question to me is that if a radial compressor can have a 5:1 pressure gain, why wouldn’t you just string 3 of them together to get a 125:1 pressure ratio (5 * 5 * 5), especially since it takes 14 axial stages (3 low-pressure compressors and 11 high-pressure ones) for a powerful engine like the GE9X just to get to 60:1. Since radial compressors are more reliable than axial compressors anyway, it seems like there’s a missed opportunity here. I guess this will be one of the things that Bjorn will be covering in the next article.

    • OPR is measured differently and includes other effects as well as through the high compression stages. The inlet cowl pressure rise to the front fan and the fan itself must be responsible for a lot of the ‘missing’ OPR

    • Yes that is a more reasonable pressure rise for a compressor stage (blade+vane), you also have to handle engine transients, ice&rain, flutter, attachment wear, blade l.e. wear, acoustics in the channels between blades/vanes trigging vibrations, efficiency and if blisk design the loss of the damping the blade attachment gives. You strive to keep the casings with wear strips of plasma perfectly round even with housing flanges, borescope ports and bleed offtakes. Cooling the compressor helps a bit in efficiency but you want the blade/vane design reach close to 100% efficiency in incoming shaft power extraction work vs. pressure rise. Don’t know if they are over 90% nowadays.

Leave a Reply

Your email address will not be published. Required fields are marked *