Leeham News and Analysis
There's more to real news than a news release.
Leeham News and Analysis
- Global instability poses questions for aerospace industry December 9, 2024
- How Trump tariffs affected, and could affect, Airbus, Boeing and Embraer December 6, 2024
- Start-up hybrid plane OEM MAEVE teams with MHIRJ for sales, more December 5, 2024
- Production increase delays hurt Airbus costs December 2, 2024
- Books to consider for Christmas November 29, 2024
Bjorn’s Corner: New engine development. Part 21. The High Turbine.
August 23, 2024, ©. Leeham News: We do an article series about engine development and why it has longer timelines than airframe development. It also carries larger risks of product maturity problems when it enters service than the airframe. We discuss why.
In our journey through an engine, we have reached the turbine part, where we will dig deeper into the high-pressure turbine. This is the most stressed part of the engine and has a major influence on engine performance and durability.
Figure 1. Our example engine, the LEAP-1A cross-section with the high-pressure turbine marked. Source: CFM.
The High Turbine
The high-pressure turbine, or the High Turbine/HPT, as engine insiders call it, is the engine’s most advanced technology area. Every degree that the working temperature can be increased without a fall in durability ( = time on wing before the hot parts need repair) creates a smaller, more efficient core and, by it, engine.
This means that the temperatures that the high turbines, and especially the first high turbine, can endure are critical parameters in engine design and characterization. Engine OEMs are, therefore, tight-lipped about the temperatures their engines can handle, as it tells so much about the engine.
We have given the entry temperature to the high-pressure turbine nozzle in Part 20 for the LEAP-1A 32 as 3,800R/2110K. The critical parameter is, however, not T4 but rather T41, the temperature entering the first rotating part of the high-pressure turbine.
T41 is often called TIT and sometimes TET (Turbine Inlet Temperature, Turbine Entry Temperature), but we will use T41 as TIT/TET could mean T4, which is the exit of the combustor and entry into the first turbine nozzle. The power extraction from the gas flow is, however, in the first rotating turbine, where Station 41 defines the input condition.
T41 is the most critical temperature because the parts (discs, blades) endure centrifugal stresses in addition to the high temperatures, pressures, and flow speeds. It’s also a part where cooling flows to the internals of the blades are more difficult to arrange (Figure 2).
Figure 2. The cooling of a turbine stage. Source: GasTurb.
The Nozzle/Stator vane can be entered from the outside and central cooling flows, whereas the rotating disk and blade can only have the cooling channeled from the engine’s central cooling flow, and the entry into the blade has to be through the blade root.
High-Pressure Turbine blades
The development of ever-better aircraft jet and turbofan engines has been tightly linked to the development in T41 temperatures. Figure 3 shows the development of long-range engine T41s. It’s a Rolls-Royce graph from before 2010. I have extended the T41 development until today. We are now at around 2,000K for long-range engines and about 150K below for shorter-range (i.e., higher flight cycle) engines.
Figure 3. The development of T41 over time. Source: Rolls-Royce and Leeham Co.
The graph shows the different cast alloys and techniques used for turbine blades and how these have improved the possible T41s, together with cooling and coating techniques.
Figure 4 shows how turbine blades have been cast over the years, progressing from conventional casting through Directional Crystal casting (DS) and finally Single Crystal casting (SC). Blade creep and, thus, prolongation happen in the grain boundaries. By first minimizing the grain borders in the radial direction and then making the blade a single crystal, the creep resistance of the blades was increased.
Figure 4. Turbine blade casting techniques. Source: Rolls-Royce, the Jet engine.
The manufacture of these blades and their accompanying nozzle guide vanes is complicated. This video from SAFRAN, GE’s CFM partner that makes the low-pressure parts of the LEAP, shows how the LPT nozzle vane sections are done. These are less demanding blade sections than the HPT sections. You get a good understanding of the manufacturing techniques used for the turbine section of a modern engine. It’s the most complex manufacturing of an aircraft part I’ve come across.
The graph in Figure 3 shows that the progress should be due to using Ceramic Matrix Composites (CMCs) from about 2000. It did not quite progress that way. We will dive deeper into how the T41 improved to 2,000K next week.
it is already enough complicated to integrate a CMC static part into a HPT module as any direct contact between CMC and alloy structural part must be avoided, so don’t expect to see a CMC HPT blade anytime soon, it is purely not sustainable for the HPT dove tail disk!!
interesting.
How is the issue handled? plain ceramic coatings / ceramic “shims” ?
Would going the blisk way make any sense?
i.e. reduce interfacing by way of integrating CMC parts into one piece?
Obviously quite a big step forward.
GE has already tested ( 2015) CMC blades in the turbine, yes it was the LPT for a military engine
RR, Safran and Pratts are working in this area- rotating CMC- too
Theres multiple layers of proprietary coatings, these probably solve the ‘contact’ with alloys issue.
https://www.geaerospace.com/news/press-releases/defense-engines/ge-successfully-tests-worlds-first-rotating-ceramic-matrix-composite
“Because the rotating turbine blades made from CMCs are one-third the weight of conventional nickel alloys used in the high-stress turbine, they allow GE to reduce the size and weight of the metal disks to which the CMCs system is connected. “Going from nickel alloys to rotating ceramics inside the engine is the really big jump. But this is pure mechanics,” said Jonathan Blank, general manager of CMC and advanced polymer matrix composite research at GE Aviation. “The lighter blades generate smaller centrifugal force, which means that you can slim down the disk, bearings and other parts.”
switchover uncooled to cooled turbine ( stator, rotor ) ~1960
The Junkers JUMO 004 had this feature from the get go.
( forced by scarcity of high temp materials. )
Same applies to the BMW 003 engine.
Yes, and the reason was that Germany didn’t have access to nickel-based alloys such as Inconel. A lot of Inconel’s development was made to support the Whittle jet engine developments in the 1940s.
“Prior Art”
( like in a range of other research domains fortunately no issues with patents )
The scarcity is a bit of myth for the Jumo 004. The real reason seamed to be because high temp steel alloys were easier to manufacture and used both folded and deep drawn *hollow* steel alloy blades ( half the weight of solid blades) used instead of stamping them out. They tried Tinidur first then used Chromadur alloy
The mass production of around 6000 engines overcame the short engine life – no possibility of overhaul anyway
The axial flow engines by Jumo and BMW were done separately from Ohain and Heinkels team who were still doing centrifugal
The real issue was the random failures that you could not get the 50 hours out of.
It was an amazing job under dire conditions (self inflicted).
It also was the real way forward as far as Architecture went. Centrifugal was easier then but also fatter and area rule had not been applied.
In WW2 , even 15 hours – especially if towed out to runway by tractor- was close to the physical life of the pilot/plane.
10 hours was worst case and some have said engines could last 100 hours with ceramic blades in the last months of the war (needs checking)
https://hushkit.net/2022/10/13/10-myths-about-the-first-jet-fighter-the-messerschmitt-262/
Remember the fuel capacity was limited , so 10x 1 hr missions would be most many pilots could achieve with the limited training for that last year of war.
Even the Merlin with all its development needed complete overhaul at 100 hrs…or sent to the tank factory to be reconditioned for the lower load use.
There were forged turbine blades on early jets, Nimonic and similar. Then overtaken by cast blades. Lots of high speed milling with ceramic bits today. In theory they can be sintered in blade shapes.
Latest Boeing embarrassment:
https://apnews.com/article/boeing-spacex-nasa-astronauts-starliner-e4e81e5a6c23dee2f8f72260ddea011c
Hotter and hotter, stronger and stronger; these blades sure do work hard.
I’ve always wondered if a couple of tricks have been missed. A lot of the load on the blades is radial forces. The blade is difficult to make, because the end of the blade necessarily needs to get very close to the turbine housing inner sealing surface, but mustn’t touch it.
So trick #1. If the first stage turbine blades are so hard to get right, give them an easier time by not requiring them to reach all the way to the inner surface of the housing. Deliberately let a fair proportion of the gas pass by the blade end, and handle it in the 2nd stage turbine instead. The energy that is taken out of the gas flow by the stubbier first stage means the 2nd stage isn’t hit by the full force of the gas flow either. I dare say the gas dynamics could be a bit messy, but that might simply require careful design of the blade to minimise it.
Trick #2. If the goal is to get the blade endpoint close to the inner sealing surface, why not support its radial load with an air cushion? If the blade tip had a circumferential end plate (like a lot of them do?), and had some gas sealing structure on the outer surface of that (like a lot of them do?), blow highly pressurised air in from the inner sealing surface on to the end of the blade. The closer the blade end gets, the more support from the air cushion it’s going to get. Just use very high pressure air so that the blade’s radial load / expansion can never quite close the gap. That could also be useful cooling.
This would be a reversal of ideas I think they already use, where they control the radius of the inner sealing surface (inflating / deflating a sealing structure?). Well, instead of taking the sealing surface to the blade tip, allow the blade tip to come to the sealing surface but make sure that the air pressure and surface area is there to keep it from actually touching. I don’t know where the high pressure air could come from. It’d need to be quite high to overcome the radial loads on the blade. But, the seal against the hot gas flow would be really good. A very neat arrangement would be if compressor outlet air pressure was adequate.
Trick #2a. As above, but mount the turbine blade so that there is some radial slackness in the fir tree somehow. That is, deliberately allow the blade to move radially a little, under load, and rely on the end tip air cushion to push back on it.
Tricks #2 and #2a are interesting to me because you no longer need a material with strong dimensional stability across temperature. It can expand all it wants, you can adapt (especially #2a). If that opened up use of other materials that have good resistance to high temperature, but are going to expand too much, they could become usable.
Another aspect of Tricks #2 and #2a is that if the blade end can be kept very close to the inner sealing surface, you do actually get quite a lot of heat transfer. The heat transfer is far higher than you’d expect, given that they’re not actually in contact. This trick is used in some electronics cooling systems, where a rotating heat sink is very very nearly touching a part to be cooled but not quite, but is still able to take a lot of heat out of it. So heat could be soaked out of the blades through their ends as well as through their roots.
Another aspect is that it could be that it’s the turbine housing / seal that partially gives way under the load from the air cushion, not the blades being pushed back. In fact, there’d be an equilibrium between the blade radial load being born by the air cushion and the tension around the circumference of the turbine housing as the radial loads are transferred into it and stretch it. Get that right and you could have a very happy seal indeed!
Patents pending!
#1 Tip losses are exactly that: losses.
(you’ve performed work on compression, now you dump it back to the front of your work job. )
#2 “flying heads” are working for hard disk drives.
but imagine the list forces you’d have to create to counter the centrifugal forces working on individual blades.
#1 Er, it’s the velocity of the gas flow that turns the turbine. There’s not actually much pressure in the turbines, or much pressure drop across them; the tail end of the combustor is essentially fully open to the ambient air pressure, with a few turbine blades in the way with some very large gaps between them.
As for the idea, all that matters is that by the last stage of the turbine one has tapped off as much power as wanted. Tapping it with a few deliberately leaky stages is not as weight efficient (because one would likely need more stages), but if the pay off is that the gas temperature can be higher, it might work out.
#2 yes the lift forces would be high if the air cushion were to support the whole radial load. But it isn’t doing that, it’s simply distorting the inner seal surface outwards to accommodate whereever the blade tips have expanded / stretched / moved to. The lift forces required are not as great as what you’re thinking, though the greater they are the better the blade end tip seal against the gas stream. The really cool thing is that, for a given air cushion feed pressure, more force can be pushed back against the blade by increasing the area of the blade tip circumferential plate. The blade gets heavier, but the added radial load isn’t on the blade root or fir tree.
” Er, it’s the velocity of the gas flow that turns the turbine. There’s not actually much pressure in the turbines, or much pressure drop across them; the tail end of the combustor is essentially fully open to the ambient air pressure, with a few turbine blades in the way with some very large gaps between them. ”
You may want to reconsider that assumption:
https://leehamnews.com/wp-content/uploads/2024/06/Gas-Turbine-cycle.png
(borrowed from Bjorn’s previous article ( 20 )
the pressure differential is significant, on the up side and on the down side 🙂
… and overall pressure ratios are increasing all the time.
We can split hairs if necessary, depending on whether or not one considers the 1st stage inlet nozzles to be part of the combustor or not.
That chart shows the 1 stage inlet nozzle accelerating the flow and reducing the pressure of the air hitting the turning turbine blades themselves. The pressure is already down 40PSI before it hits the turbine blades, and the gas speed has doubled.
That’s born out (in less detail) by the other charts on that article.
But yes, I was wrong about the bit about there being no pressure drop across the blades (there clearly is!)
But given that it is gas velocity that matters, #1 still works.
There are two types of turbine: impulse turbines converting fluid kinetic energy to work, and reaction turbines converting both pressure and kinetic energy to work.
impulse turbines work after a stator has turned gas pressure to speed.
Whichever way you look at it work done is achieved by decreasing pressure .. step/stagewise
It is pretty simple to divide the power extracted by stage 1 and 2 of the HPT to maximize efficiency. The volume increase as you lower pressure and hence T2 blades needs to be made bigger or have the axial flow airspeed to increase. The stage 1 cooling airflow is added for the T2 blade to swallow. You normally use HPT Case Cooling air to control the HPT blades tip gaps and it is common to use it during cruise. In theory you can have shrouded HPT blades to reduce tip leakage but they need to be cooled and it is a costly design (RR).