Bjorn’s Corner: Geared turbofans

By Bjorn Fehrm

January 13, 2017, ©. Leeham Co: The time has come to go through the reasons why some turbofan engines are designed with a gearbox between the fan and the low pressure shaft.

The principle design is shown in Figure 1. It’s a graphical representation of a geared turbofan from the engine analysis software GasTurb.

Figure 1. GasTurb principal representation of a geared turbofan. Source: GasTurb.

The base idea is to have the low pressure spool of the engine to run at a considerably higher RPM than the fan.

Speed is good

The reason that one wants the low pressure spool to run faster is that the low spool’s compressor and turbines get more efficient when the blades have a high tangential speed. The physics behind this are the same as for an aircraft wing: lift increases with forward speed.

When the RPM and the diameter of the blades bring the tips of the blades to supersonic speed, the aerodynamics get more complicated. The supersonic aerodynamics are connected with shockwaves, which can degrade efficiency.

Turbofan engine designers therefore try to keep only the tip of the blades in the supersonic region. As the fan has a very large diameter, the RPM should stay below 5,000 RPM for single aisle engines (typical fan diameters of 50- to 80-inch) and below 3,000 RPM for wide-body engines ( fan diameters from 75- to 135-inches).

At the same time, the booster and low pressure compressor have diameters of one third of the fan diameter. Consequently, you would like to see low spool RPMs of 9,000-15,000 RPM to get the blade speeds in the right window. The way to get there is to have a gear ratio of 1:3 or more between the fan and the low spool.

Geared Turbofans

A gear between the fan and the turbo-machinery is nothing new in the aircraft engine domain. Turboprops have such a large diameter for the propulsive blades that a gearbox is necessary between the core and the propellers. It was when the high RPM cores made for turboprops or APUs are used for turbofans that a geared fan became necessary.

A gearbox is always a complication and a component that can give trouble. Therefore it was avoided for a long time on jet and turbofan engines which were designed from scratch. The designers added more stages to the compressors and turbines to compensate for the lower efficiency and they increased the diameters of the stages as much as possible. With the very high bypass ratio engines of today the mismatch has become so large that geared designs will be more common.

The first geared turbofan for civil use, Garett AiResearch’s (today Honewell’s) TFE731, was developed from the Douglas DC-10 APU, Figure 2.

Figure 2. TFE731 derived from an APU by Garret AiResearch 1970. Source: Garret.

It had quite a few gearbox problems in the first variants. The fan is a large gyro and the gearbox needs to be decoupled from the forces this generates. Garret learned this the hard way. The TFE731 is primarily used on business jets since its certification in 1972. It’s still popular today for smaller business jets.

Another early geared turbofan was designed by Lycoming from the T55 turboshaft helicopter engine. Once again, the high RPM of the core made a gearbox necessary for the ALF502, Figure 3.

Figure 3. ALF502 turbofan designed by Lycoming. Source: Lycoming.

This engine started life on some military projects but found wide use from its 1980 certification on the BAe 146 regional jet and the Challenger 600 Business jet.

I wrote “The first geared turbofan for civil use” above. The reason was that the first geared turbofan used on an aircraft was for a military aircraft. I learned to fly jets on the aircraft, the SAAB 105 jet trainer. It was equipped with a geared turbofan made by SAFRAN’s Turbomeca division. The engine (Figure 4) was based on the Bastan turboprop engine.

Figure 4. The first geared turbofan, Turbomeca Aubisque. Source: Swedish air museum.

The engine ran for the first time in 1961 and was used on the SAAB 105 between 1966 and 1990.

Pratt & Whitney GTF

Pratt & Whitney (PW) realized the advantages of making a clean-sheet turbofan with a geared fan (a GTF for Geared Turbo Fan). PW also saw the problems Garrett had when applying a gearbox with a fan. PW’s application went far beyond the 7000hp that had been the highest gearbox rating historically. PW needed 30,000hp for its new single aisle engine family.

The company therefore took its time and researched the use of a planetary gearbox through different projects for 20 years. Certification of the first PW Geared turbofan, the PW1500G engine for the Bombardier CSeries aircraft, occurred in February 2013. First entry into service was on the Airbus A320 with the PW1100G, Figure 5.

Figure 5. PW1100G geared turbofan for the A320 family. Source: Pratt & Whitney.

The gearboxes used for turbofans are of the planetary type using Epicyclic gears, Figure 6. Such gearboxes are compact and can transfer high torques with high efficiency. The input from the high RPM low spool comes from the middle shaft (the yellow sun gear) and output to the fan can be taken through a rotating ring gear (pink) or planet gear carrier (green).

Figure 6. Epicyclic gears used in planetary gearboxes. Source: Wikipedia.

In either case, the other gear member has to be fixed to hold the gearbox. If the output is taken from the ring gear (pink), gear ratios are around 3:1 for practical geometries and from the planet gear carrier (green) around 4:1.

The PW GTFs use a ring gear which is connected to the fan; therefore the gear ratio is 3:1, Figure 7.

Figure 7. PW’s gearbox. Source: PW

31 Comments on “Bjorn’s Corner: Geared turbofans

  1. I read somewhere that the A320 with the geared P&W engine is lighter by more than a tonne than the CFM-LEAP equipped equivalent.

    The article doesn’t go into the potential weight savings of adding a gearbox, the fan rotates slower so the fan blades can be lighter, the fan case doesn’t need to be as strong to contain the blades and not as many LP turbine stages are needed since they are more efficient at the higher RPM.

    A big problem is cooling the gearbox. A 30,000HP transmission operating at 99% efficiency will generate 300HP of heat.

    I don’t understand why turbofan gearboxes don’t have similar issues to helicopter gearboxes, which transmit far less power.

    • The total difference, installed engine with nacelle and pylon, between the PW1100G and LEAP-1A on the A320 series is below 0.2 tonnes. The PW1100G has a larger fan which increases nacelle weight. Thus, while there are weight savings on the LPT and compressors, the global weight saving is not large. The big gain is in cost for compressor and turbine blades, partly offset by the cost for the gearbox with its oil system.

      A gearbox poses several additional challenges; mechanical integration, reliability and heat are among these. The engine also gets longer. The heat is a major consideration, efficiency therefore has to be better than 99%.

      Part of the long time in bringing the PW gearbox to service was perfecting the separate oil system that should transport the heat from the gearbox to the oil cooler in the bypass stream.

      • while the GTF nacelle is slightly larger (only 3 inches inside diameter I believe) P&W should have been able to build it lighter than GE since with slower turning, lighter fan blades than the LEAP, the total energy that needs to be absorbed in a blade out situation should be significantly lower (50% lower fan RPM would reduce angular energy by 75%, not accounting for blade weight).

        this should result in a corresponding 75% weight reduction in the containment structure given the same materials.

        Did P&W choose to use cheap and heavy materials for their fan case compared to GE? or did their nacelle team overengineer the fan case?

      • Gents,

        Regarding the heat coming from the gearbox; it is not a problem.

        The fan drive gear system (FDGS) has a dual-stage fan-driven oil pump
        that works with the journal oil shuttle valve (JOSV) to form an integrated
        oil system, supplying oil to the FDGS journal bearings during normal,
        windmill, and negative-G conditions. A variable oil reduction valve
        (VORV) supplies pressurized oil from the engine oil system during
        normal operation. The VORV controls the amount of oil supplied based
        on engine power settings.

        The dual stage pump continuously draws oil from a dedicated auxiliary
        reservoir and compartment sump located in the front bearing
        compartment. The pump is installed in the no. 1 and no. 1.5 bearing
        support, and provides the journal bearings with an alternate oil supply
        during engine windmilling and negative-G conditions. It is turned by the
        fan drive shaft and rotates at fan speed.

        The front bearing compartment contains two sumps that return oil back
        to the main oil tank from the components in the front bearing
        compartment. The sump, at the bottom of the front bearing
        compartment, also acts to supply oil to the journal bearings during
        windmill conditions. In normal operation the sump oil is scavenged and
        returned to the oil tank by the fan-driven oil pump.

        A secondary sump collects oil from the FDGS via an oil collection gutter.
        And as I said already, this secondary sump in the front compartment supplies oil during negative-G events. In normal operation, the fan-driven oil pump scavenges the oil and returns it to the oil tank.

        The thermal management system, consisting of three heat exchangers,
        ensures that engine oil and fuel temperatures are maintained within
        limits.

        Sounds complex, but it’s not… and it is very effective !

        …and the Turbomeca Aspin 1, in 1952, was the first geared turbofan… never put in production !

    • The gearbox has to be able to transmit the full power of the engine without melting, but that’s only take-off. Most of the time it’d be run at much reduced power settings, whilst the aircraft is cruising.

      So there’s maybe room in the cooling equation to allow heat to build up in the gearbox for that short period of time, and cool it down later on. Still, 300HP is a lot of heat being generated!

      But that makes me wonder what temperature they design it for. Peak efficiency would be achieved when everything has warmed up just so. Cooler or hotter, there’d be more loss and more wear. So does one design the box so that sweet point is at take-off power, or during cruise? Or does one simply put a lot of effort into the oil system so as to be able to thoroughly regulate the gearbox temperature no matter what?

      • Cruise, that’s where you run the longest and where efficiency has a leveraged affect (PW was going to use variable vanes at one point but got what they wanted without it, lots of tools to play with)

        Rumors are the RR 3 Spool does better in long cruise vs the two spool.

        That seems to be variable as the GP7000 was beating the RR on the A380 handily before TC decided to make his move.

  2. The issue of heat in the gearbox will be an issue for engines for widebodies.

    So I think RR will continue with the 3 shaft solution for some time on large engines and hope to get back into the small engine market with its UltraFan

    Will PW try a geared fan for large engines? Why the question? The compressor on the GTF is fabulous, all blisk and very high compression. Therefore, do they need to take the risk or will they go three shaft. I do agree they are back!

    Will GE change from two shaft to geared fan and/or three shaft? I think they will have to, for there is growing evidence that they will be outperformed by RR and PW

    To add to the article, a lot of work is being done on shock waves. So expect fan speeds to increase as this work matures, thereby reducing the need to increase the fan diameter, in doing so mitigating drag and weight issues with large fans

    • .. thereby reducing the need to increase the fan diameter, ..

      you increase fan diameter for increased area covered to accelerate more mass to less over speed increasing propulsive efficiency.

      With fan rpm locked 1:1 l to LP turbine rpm blade tip speed on the fan limits rpm for the LP turbine. you have to significantly increase LP turbine diameter to bring LP (blade) speed into a range that allows acceptable work extraction per stage.

      • The issue with fan efficiency is the shock wave when the fan goes supersonic. Shock waves cause a dramatic increase in drag. So the fan is made large but run at a slower speed to reduce the drag effect of the shock wave. The issue with a large fan is three fold: 1) base drag, 2) a less efficent LPT (the PW GTF has a 3 stage LPT because it runs at 3 times the speed of the fan whereas a LPT running at the same speed of the fan typically has 5 stages), and 3) an increase in weight. Consequently a lot of research is being done on shock waves to reduce their effect.

        • Running the fan a little bit faster and more supersonic, but still directly attached to the LP turbine, how does this compete with a geared fan?

          The geared engine would still have an LP turbine operating at a much higher RPM and therefore better efficiency, it would be quieter, lighter and probably have a higher bypass ratio than a conventional engine with a more supersonic fan.

          Both types of engine have big engineering hurdles to be overcome, but the gains for a geared engine are far greater than for a more supersonic conventional engine are they not?

          And, at the end of the day, any advance that you can make on a conventional engine can be engineered into a geared engine. That goes for supersonic fans

    • It looks very much like RR are aiming their GTF efforts at large engines, not smaller engines.

      Ultrafan, if they can make everything work, could be amazing – variable pitch, carbon-titanium, geared turbo fan. That’s throwing literally every dreamed-of trick in the book of ideas at a single design.

      RR have done 3 shaft to death, although their Advance concept involves finishing it off with carbon-titanium fan blades. As well as knocking off some weight, this should allow them to achieve the ultimate in aerodynamic efficiency. You can easily mould CF to any shape you want, whereas hollow Ti takes more coercion to shape it. It would be a bit pointless indeed on the part of GE or PW to go 3 shaft without any other innovation when RR themselves are looking to move beyond Advance to UltraFan.

      But yes, GE are beginning to look somewhat behind the times. They did buy PW’s gearbox partner, which maybe betrays something of their intentions. GE are seemingly not as public about their R&D direction as RR (for example, see http://www.rolls-royce.com/products-and-services/civil-aerospace/products/future-products/ultrafan.aspx), but it doesn’t mean they’re not working on something new.

      GE do have CF fan blades, but this technology could probably do with a refresh; the root of one of GE’s blades looks quite fat in comparison to one on RR’s Advance, which is going to have all sorts of implications for airflow as it enters the compressor. I know that RR have worked quite hard to find a way of making thin CF blades, which brings improved aerodynamics in the fan and probably ends up weighing less too.

      With GE having bought PW’s gearbox partner, I think a betting man would put his stake on GE doing a GTF, not three shaft. However, even if they do that, they’re going to have to contend with RR’s UltraFan.

      Anyway, it’ll be interesting to see what they all do next. As Bjorn has pointed out previously, strap 4 of RR’s Advance engines onto an A380 and there’s a neo, right there, and it’d be quite hard to beat it. Boeing might have to be sure to keep RR on board and willing, lest they do steal a big march in efficiency. Boeing won’t want Airbus having an exclusive for that.

      • “But yes, GE are beginning to look somewhat behind the times. ”

        At what level of behind or ahead will the GE engine for the 777X materialize?

        • The issue with the GE9X is the numbers don’t add up. 10% better than the GE90 and 5% better than anything else. The TRENT XWB is 10% better than the GE90

      • RR hasn’t done three shafts to death. New materials will mean three shafts will get even lighter, even though they are already the lightest large engines around. New materials and new aerodynamics will mean compressor and fan speeds will increase

        It took PW 15 years to make a 30,000Ib thrust geared fan. A 100,000lb thrust geared fan will happen, but don’t hold your breath!

        • You miss the real aspect and that is opportunity.

          Boeing did not let PW bid on the 787 so they could not get in on that thrust level.

          Its only as an opportunity comes up you can compete there. Boeing is stuck on GE.

          The only possibility is an NEO in the higher thrust class, that leaves the 777x out, A350 is set for some time.

          PW didn’t go GTF to be in a single nitch.

          They have plans that scale that engine up.

          They have plans that use a different gear type.

          There is no impediment other than implantation of the design. RR obviously has theirs where they believe it will work.

          GE can’t access PW gear system as it will be firewalled off as a proprietary project.

  3. Perhaps the next step for the GTF is a 45K engine for a clean sheet slightly larger aircraft.

  4. Bjorn,
    In your explanations you said that that the highest gearbox rating historically
    had been 7.000 hp, but I believe that there were a few exceptions for turboprop engines. I recall that the Kuznetsov NK-12 for the TU-95 (and even the commercial TU-114 ) achieved in the initial versions 12.000 ehp, and later 14.795 ehp for the military aircraft. And more recently the engine of the A400M has a very high rating as well. I know that the rating at the gearbox is lower than the total ehp output, but still probably higher than 7.000 hp.
    Rgds,
    Mario

    • Hi Mario,

      The text was about turbofans, I’ve made it more clear. And yes the turboprop market has seen higher power gearboxes, you gave the examples.

    • hi
      you are right and as an exemple the TP400 turboprop provides 10700 SHP

      • That is 7…8MW not 20..25MW 😉

        One could argue that the NK 12 has two power paths ( counter rotating props.)

        • Dont you mean Contra rotating ? Counter rotating means propellers on each wing rotate in the opposite direction to the other wing. It was the method used on the Wright Flyer but its rare as usually they all rotate clockwise ( as seen from behind)

          • The A400M has both contra-rotating props on each engine and counter rotating pairs on each wing. A first ?

          • NK14 / TU-95 has counter rotating prop pairs.
            TP400 is single prop, counter rotating pair per wing.

  5. The gearing and high speed LPT is not that simple. At off design operation the LPT does not take up that much heat putting strain on the LPT structure, the gear and its bearings must hold the fan in place if the thrust bearing seizes and shear off the shaft (RR fan catcher), the gear takes axial space making for a core mounting of the aft mount and its inherent problems CF6-6/-50. For higher thrust geared fans you have the gear oil cooler and its redundancy to solve. But you gain bypass ratio with better propulse efficiency and TSFC, a smaller and cheaper/lighter LPT and LPC/Booster. There are rumors of PW1100G engine changes in India.

  6. GE engines are more fuel efficient than RR in all class they compete in even on A330 CEO. You can verify it from Mr fehrm. You do not think GE will have engine in development beyond GE9X. They have advance turbine for all kind of application than the entire budget of UTC and RR.

  7. Napiers in the 60s were developing a disc form of infinate vairable gearbox
    has any one esle picked this idea up

    • All the IV type transmission designs are either too heavy, too complex, inefficient or fall short on other design metrics.
      ( or a mix of all detractors.)
      Hydrodynamic transmissions (Voith) are popular in diesel engines though there diesel electric ( actually hybrid with supplemental batteries ) is on the fore.
      The fully mechanical ones rely on friction. That requires internal forces exceed useful moment transmission by a magnitude or two.

  8. Hi Bjorn,

    Thanks a lot for this very intresting article! I am working on journal bearings in wind turbines and the oil filtration size is a big question for us every time.

    I was wondering: the application of journal bearings in the GTF, Ultrafan and co means that the filtration size is different on these engines? Or in the aerospace industry you already have a sufficiently small filter for the gears and bearings? Since the journal bearing consumes a lot more oil than a bearing and needs (usually but I’m not sure it’s the case here) a smaller filtration size that could be quite problematic for the oil pumps.

    And just for my curiosity : the Lycoming ALF 502 had journal bearings or roller bearings?

    Thanks for your time,

    • Roller bearings I worked R&D in Connecticut on the ALF-502
      Pulled many planetary gear sets out of those engines. Sun gear
      and planetary wear mostly

  9. “7000hp that had been the highest gearbox rating historically. ”
    The Soviet NK14 turboprop engine transfers better than twice the power ( ~15000shp, late 50ties ).

Leave a Reply

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