02 October 2015, ©. Leeham Co: After the article about the role of bypass ratio on a turbofan’s efficiency, we now look at other aspects of civil turbofan engines that are worth some light. It’s about how the engine OEMs create different versions of the same engine to cater for different aircraft variants.
The aircraft OEMs create different size variants from the same base model of aircraft by means of stretches. There is no better example of that than the Boeing 737. Over the years it has had more than 10 major versions. For the present in-service series, 737NG, there is three official variants, from the -700 to the -900ER. Originally it also had a smaller -600 variant.
These require engines from 20klbf to 27klbf. How this is achieved and what it means for engine characteristics and reliability is the focus of today’s Corner. We will also compare it to a typical long range engine, the Rolls-Royce Trent 1000/7000, which powers the Boeing 787 and Airbus A330neo.
Aircraft turbofans are real flexible engines. To cover the different size aircraft in the 737NG or 787 range, the engine OEMs make one engine that they then adapt to the different aircraft by simply spinning them more or less fast. This give them the different ratings the aircraft OEM needs. The different ratings are achieved by rating plugs, which are inserted into the engine’s control computer, the FADEC. The plug enables another version of the engine by changing the look-up tables which governs how the FADEC controls the engine.
The engine’s thrust rating is measured on the engine’s Static Take-Off Thrust at Sea Level and standard ISA temperature (15°C). For the 737NG series, it starts with operators specifying the 20klbf version (CFM56-7B20) for the -600 and finishes with the -27klbf version for -900ER (CFM56-7B27). For the 787, you would order a 64.5 klbf or 69.9klbf Trent 1000-A for a 787-8 and the strongest version of 78.9klbf Trent-K for a 787-10.
The way the engines OEMs create the different variants is by injecting more and more fuel into the combustor. This means the combustor temperatures increase and the fuel-air mixture has higher temperatures when hitting the turbines. This means more horsepower gets generated in the turbines, which drives the compressors and fan harder. The engine spins faster, pumps more air through and thereby generates more thrust. Stretching an engine like this is called “throttle push”.
There is, of course, a limit to how far this can be done and there are drawbacks for creating stronger variants with throttle push. The hotter temperatures in the core (both compressors and turbines get hotter at higher RPMs) will shorten the time-on-wing for the engine.
Whereas a 20, 22 or 24klbf CFM56-7Bs will stay on wing for the whole life of its Life Limited Parts (LLPs, the blades+ discs of compressors and turbine which need to be changed after 20,000-30,000 engine cycles), the higher rated versions, 26 and especially 27, will have to be taken off for performance restoration repairs to the hot sections at about half time or even less if the maximum take-off rating is used often, i.e. take-offs are made with little or no thrust de-rate.
The result is that the maintenance costs for the higher rated engines are higher in addition to a higher engine price from the engine OEM (engines are priced in proportion to thrust rating).
The throttle push principle of raising the engine’s thrust means that the engine parameters change for each thrust variant. The harder the core drives the fan the more dominant it gets in the engine. The engine’s pressure ratio (PR) increases as the compressors spins faster but its by-pass ratio (BPR) diminishes.
The 20klbf CFM56-7B20 has BPR of 5.5 versus 5.1 for the 27klbf CFM56-7B27 version. Pressure ratio is going from 23 to 29 for the stronger version. As the engine spins faster it generates higher compression ratios. For the Trent 1000, the values go from BPR 9.6 to 9 and PR 39 to 46 for the 64.5klbf engine to the 78.9klbf variant.
The fact that the values in a Turbofan vary with throttle position is very clear to see in this GasTurb simulation I did for the Trent 1000:
It shows the engines working lines (engine parameter values at different throttle positions) when a 787-8 cruises at FL350. A 787-8 then needs about 8 klbf-12klbf thrust per engine dependent on cruise weight and altitude. At a cruise thrust of 10klbf, the BPR is at 11 (red scale and curve) and the PR at 34, the burner exit temperature is at 1400 Kelvin and TSFC at 0.53 lb/lbf and hour. As can be seen, the values vary greatly with thrust and therefore engine RPM.
Due to the thin air at FL350 weighing much less than at sea level (about a quarter of the seal level weight) the engine needs to spin at around 90% of max RPM to generate the 10klbf cruise thrust. During decent the engine is throttled back to around 60% of max RPM and the BPR raises to 15-20, PR is around 15. The engine can not be throttled back more than around 60% as the compressors need to supply bleed air for the aircraft air condition and de-ice, the engine’s compressors also have problems working reliably at still lower RPM.
As can be seen from the above one shall always ask at what conditions an engine OEM states his engine parameters.
I have always wondered: What is the SFC penalty of making an engine family with multiple possible thrust ratings, versus a single-member family optimized for a single rating?
With the CFM56, for example, what would SFC be at both max BPR and max OPR? Obviously its architecture would have to be changed so that the high spool spins faster and/or the low spool slower…
actually the penalty for making a family is not that large. The reason is that the cruise thrust variation for different variants for e.g. 737NG is less than the take-off thrust variation. Also when you design an engine the lower part of the SFC curve, called the cruise bucket, is normally relatively flat, i.e. different cruise thrust needs don’t carry a large penalty.
As we have a simulated diagram for Trent 1000 I checked the variation of cruise thrust for the 787-8 to -10 in our aircraft performance model, it is around 1,200lbf between the 787’s so with a correctly designed cruise SFC bucket you would see almost no SFC penalty between variants.
For several engines, CFM56 included, the max continuous and max climb ratings are the same for the two to three strongest variants, only the 5 minute max take-off rating and therefore higher turbine temps differ. The rest of the flight profile is run with about the same values as for a smaller/lighter variant of the aircraft.
in summary I think the fuel consumption penalty compared to an optimized engine for each thrust need would be fractions of a percent.
Thanks, interesting. I guess if cruise thrust doesn’t vary much that makes sense.
Maybe you could address in another post the issues of frame/engine optimization and how that affects SFC. I believe RR and Boeing have emphasized the importance of this issue lately.
Thanks Bjorn, very interesting!
A bit off-topic but engine related and therefore, I may find the right persons to answer; as all musing about A380neo and Airbus engine options (Trent 7000 or ADVANCE or GP7200-PIP etc.), I wonder if a “Trent 9000” – hence a T-1000 core with the T-900 fan, would make sense, what it would achieve and if it is affordable? Thanks!
Actually you don’t need to go to the trouble of changing the fan of the Trent 1000, the cruise thrust need of the A380 is around the same as for a 787-10 and the take-off thrust is less (72klbf versus 78klbf) so one could take the Trent 1000TEN as it is. You would only gain around 5% TSFC however which is to low. A Trent 900 fan would be installing old technology with quite some effort, then better to take the ALPS CFRP fan and fan case.
But ultimately the effort of installing and flight testing is so large that only a new engine makes sense, the Trent 1000 is 10 years old. I bet on the Advance being chosen when it finally gets done (RR is developing Advance as we speak).
The dry engine weight for the Trent-900 and Trent-1000 is, respectively, 5, 408 kg and 6,246 kg (NB: Not including fluids and nacelle). Hence, the dry engine weight differential is 3352 kg for 4 engines. Assuming that the combined weight of the propulsion system, nacelles and pylons are at last 50 percent greater than the combined engine dry weight (i.e. 1.5 x 3352 kg = 5028 kg), the reduction in the manufacturer’s empty weight for the A380 – outfitted with 4 x Trent-7000 engines – would be slightly more than 2 percent. Thus you’d be looking at a further 2-3 percent reduction in fuel consumption on top of the 5 percent reduction in TSFC. If sharklets could be added to the wings as well – IMJ the total reduction in fuel consumtion could approach 10 percent.
Bjorn leehamnews reported that the advance engine apprears to be heavy the advance for the a380 and airbus was discussing with EA for a GP7200 PIP. https://leehamnews.com/2015/03/09/notes-1-from-istat-2015/
With these data i have the following questions?
-With a too heavy advance engine and the trent 1000 together with the XWB making no financial sence has the a380neo program stalled and
how much can EA improve the GP7200
The PIP for the EA7200 would not bring the required 10% better aircraft fuel consumption therefore it is no longer in the picture of an A380neo.
I will agree with steve. The advance and the xwb is too heavy for the a380 while the T1000-TEN and GENX are outdated. I believe a combination of the 2.5% improvement EA says it can offer and 5% from aerodynamic improvements is the best solution for the aircraft.
Thanks for the article Bjorn. Very interesting. I’ve got 3 questions.
1) Is the ratings LUT plug purely s/w, or is there actually a physical ‘plug’ inserted?
2) What is the historical basis for the OEMs selling engines based on thrust?
3) Is the price per unit thrust a roughly linear plot or some other plot?
Here the answers as best I can:
1. It is a physical plug but it contains numerical information nowadays.
2. They price per market value as all pricing shall be made. The airframe OEMs do the same, a lower MTOW A330 regional cost less than the real thing at 242t.
3. It seems to be linear within an engine series.
Thank you Bjorn.
Re 2) pricing by thrust sounds artificial to me, rather than supply & demand market led pricing. I’m guessing not something that would have occured back in the piston engine era with lots of competing mfrs & products. Could it have started with Pratts early jet age JT3 dominance?
A small correction; blades are normally not Life Limited Parts. Engine spools, disks and rotating seals are. The GE90 Composite fan blade was initially Life limited to approx 30 000 cycles. This is no problem for a long range Aircraft but if the LEAP-series and PW1100G get the same Life limit it will effect operators eventually.
Thanks for the clarification, when I say blades I really mean compressor or turbine drums/spools with discs and blades mounted to a balanced unit, in your parlance spools.
Interesting article. Going on a slight tangent, can you comment on the reason that the same engine family can have very different looking nacelles depending on the aircraft it is fitted to – a couple of examples that come to mind is the CFM56 on the A320 vs. the higher thrust CFM56 on the A340-200/300 or the RB-211 on the 747-400 vs. the RB-211 on earlier version of the 747.
Does the design of the nacelle (long vs. short) have any meaningful impact on overall engine efficiency?
Yes it has, long nacelles or so called mixed units increases the engine efficiency a couple of percent. If it gets done or not depends to a part if the engine OEM also has the nacelle responsibility but also on the mission type. A short haul CFM56 has a short nacelle for weight and simplicity, the long haul A340 CFM56 has a mixed nacelle for efficiency.
Thanks Bjorn for the excellent article.
As you said, the parameters such as BPR, OPR, FPR (fan pressure ratio that is seldom mentioned in discussions but is an important parameter determining the fan stream speed and hence the propulsive efficiency), TSFC, specific fuel burn … do not mean much, unless accompanied by the conditions at which they are valid. Descent is usually not a critical design point for an engine and so the engine parameters then are not as important as at OEI TO and TC.
To run GasTurb12, you need component efficiencies, which is information OEMs closely guard. How do you manage to input those values? Educated guesses or “inside” information. Just curious!
Hi Kant, indeed the OEMs don’t like to tell you the shaft hp to thermodynamic transfer efficiency but there are surprisingly good reports on the matter on the internet with current data. As with all our data, be it for our proprietary aircraft model and/or our GasTurb work, it is a mixture of collected open info and inside information.
You did not discuss derated engines that are used to power shrink versions of the aircraft. Derating is an advantage since the engine life is prolonged because of lower TETs, but the disadvantage is that the aircraft is carrying an unnecessarily heavier engines.
Also the quoted thrust is usually sea level static (SLS), nothing to do with TO, as you yourself have made clear on previous occasions, since the ram drag at TO could reduce the net engine thrust by as much as 20-30% depending on the takeoff speed.
It is a bit touch and go what a variants is and what is a de-rated engine. Engine OEMs have told me they design for the highest power engine variant and then make sure the other lower powered variants have the correct characteristics via FADEC tuning. De-rates beyond initial planning like the RB211 for 757 get very long time-on-wing.
Does the 787 gain an advantage in descent phase or taxiing phase because it can be throttled to less than 60% since bypass air is not used for generators? If so , how much can it be throttled back and is savings significant?
You have to sum the power off-takes of bleed air and shaft hp which gets converted to something like 500kW of electrical energy for the 787. In sum there is no large difference if you take your airframe power as bleed + electricity or only electricity. The big debate is whether all electrical is more efficient than a mixed power off-take, something we don’t decide in this article.
There is in fact a larger difference between the two engines, the power off-take on the classical two spool GEnx-1B is from the low spool whereas it is from the intermediate spool from the Trent 1000. Due to stability concerns the two spool have to be kept at a higher RPM at descent idle and the Trent 1000 gain its short haul fuel consumption gains from a lower idle power setting (and not from the generally rumored better climb efficiency).
This conversation make me wonder about solutions such as TaxiBot of IAI or WheelTug. Do you see these becoming mainstream at any point? Any other type of solution for taxiing?
I do, it makes no sense to have big turbofans to taxi for e.g. 30 minutes on the large airports. The solutions must be better integrated and weigh less before it becomes common however.
Bjorn, thanks for excellent article. I hugely expanded my aircraft knowledge thanks to your technical articles. In many of them take-off is most challenging part.
-It requires bigger engines, since more thrust is needed than on cruise. Heavier engine-more fuel used.
-It causes more wear on the engine. More maintenance cost.
-Some aircraft are take off limited, e.g. 737-9 or 787-10 in Dubai. (I was astonished how 7000nm aircraft can not make a 4000nm trip in bad conditions)
This is surely crazy idea, but what do you think. If R&D and introduction costs is not considered, could assisted take-off (electromagnetic, steam or even horses 🙂 ) be viable option for the industry? Can savings outweigh cost? Or should engines have the same thrust anyway, because of go-around and performance needed for climbing (especially in case of engine failure).
It would have to be a reaction devise as the critical point for a One Engine Inoperative (OEI) take-off is not when the aircraft is on the runway but a failure directly after lift-off. The aircraft in take-off config must be able to climb on one engine directly after liftoff at the so called safety speed V2. Early jet aircraft had straight jet engines which were marginal for take-off and climb, therefore take-off rockets (JATO) were strapped on, they fell off after the assist. Their complication did not justify them once the high bypass engine was available, these have high thrust at low speed and fixed the OEI take-off problem.
Thanks for that description. The flatness of the fuel consumption curve is eye opening. Proportionally going from 8 to 12 is like going from a 100′ 100K aircraft to a 150′ 150K aircraft at cruise. Powered by the same engine, with only a 2% fuel efficiency penalty at the extremes, that covers a lot of territory with one engine design.
I guess you can infer the advantages of the new ceramic engine materials which can operate at higher temperatures from above information. Presumably, you could push a smaller engine harder, to gain efficiency?
The higher temperature in the turbines allows a smaller core to serve a larger fan, i.e. increase BPR and decrease specific thrust and therefore propulsive efficiency.
Thanks Bjorn for an interesting article.
It’s an interesting thing to balance fuel consumption, combustion temperature, NOx emissions, thrust, bypass ratio, service life. That’s a lot of things to optimise for!
RR’s three spool engines allow them to achieve a better balance between all those things.
Adding in a CF fan (make blades lighter, and any shape they like), variable pitch blades (which is sort of like changing the blade shape in flight), and a fan gearbox (which is kinda like adding a fourth shaft) is going to bring many more options for achieving optimal performance.
If they can make, oh, say two of those things real then they’re in a good position.
I’ve not heard what GE are doing, but they’re in danger of being out developed by RR. I don’t see how they can sensibly evolve a two shaft architecture to compete. Perhaps they’re not planning on doing so.
Can you see a future where just one of the manufacturers dominates? And how well would that go down with the various regulators around the world?
Perhaps I could make a few comments on the subject of mixed or unmixed exhausts. The A320 family are powered by either the CFM56 ( unmixed ) or the V2500 ( mixed ) and the A330 by the Trent 700 ( mixed ) , CF6-80E ( unmixed ) or PW 4000 ( unmixed).
The 787, A380 and A 350 are all powered by unmixed engines of high BPR. In certain circumstances, at lower values of BPR, there may be a slight gain in SFC with mixing, but at higher values of BPR unmixed exhausts are standard. This is fundamentally due to the fact that as BPR is increased FPR decreases, and the cold stream exhaust pressure is likely to be lower than the hot stream pressure; one (unpublished) rule of thumb is that the cold stream pressure should be about 7% higher than the hot stream exhaust pressure for good mixing. If you trace the history of the RB 211/ Trent family we find -22 unmixed, 535C unmixed, 535E4 mixed, 524 D unmixed, 524G mixed, 524L (Trent 700) mixed and all later Trents unmixed, showing the effect of increasing BPR leading to separate exhausts.
This can also be seen on the GE 90, GP7200 and GEnX. It should be noted that virtually all bizjets use mixed exhausts, the CF34 in the Challenger being an outlier; it was, of course, a civil development of the TF 34 designed for the Warthog ground attack aircraft. Bizjets cruise at higher Mach numbers than civil transports and require a higher rate of climb, resulting in lower values of BPR.
honored to have a Gasturbine professor on the blog. I wondered why mixed nacelles were no more present on modern long range engines, now I know. Thanks.
The proposed RR engine with VP fan is actually a two spool arrangement with a geared fan. At extreme BPR I think the LPT in a three spool would become very heavy, and a gearbox becomes necessary and this allows the booster stages following the fan to run at higher speed. Configuration appears to be similar to PW GTF with the addition of VP fan which will permit deletion of the thrust reverser.
Indeed RR Ultrafan is a classical two spool GTF. RR likes to mix the cards with calling the low spool an intermediate spool, this is deception marketing. Guess they do that as its RPM due to the gearbox is in the domain on the intermediate spool of a three shaft engine.
Herb, Bjorn, very interesting. I had assumed that RR would keep their 3 shafts with the gtf in order to get lots of options for optimisation. You’d think that if they could do the gearbox at all, it could be stuck on a 3 shafts engine quite easily.
It’ll be interesting to see if the weight and maintenance of the gearbox is worth it vs the weight and maintenance of third shaft.
I did read that their CF layup is most promising. It seems they can do thinner roots than GE, getting better aerodynamic use of the whole blade length.
Time will tell I guess!
There is no point in doing a three shaft engine if you have the luxury of a gearbox between the fan (which shall rotate slowly) and the low compressor and turbine spool (which shall rotate fast), the gearbox ratio makes sure all three components can run at their prefered RPM. So why then put a third shaft in? It has its price in mechanical complication and friction losses and you already paid another similar price with the gearbox. Why pay twice for no real gain?
The variable pitch fan of RR. I’m taking a see first approach on this. Where’s the air coming from & where is it going. Compressor stall anyone..
.. during thrust reverse..
Yes, that is something that I was thinking as I was reading this variable pitch on a ducted fan engine. If the fan pitch is changed and thus the airflow is reverse, does it mean that the fan will be sucking up the hot gas coming from the engine exhaust? Does it not create problem for the compressor as stated by keesje and potentially to the engine as a whole?
Bjorn, Thanks for kind remarks! I agree that a three spool does not need a gearbox. The RR UltraFan is talking about unprecedented levels of BPR and OPR, and the very low fan speed would require a very large LPT. With a gearbox and high speed LPT the booster stages would run at comparable speed to the IPC in a three spool. It may be worth recalling that IAE offered an engine called Superfan for the A340, incorporating a VP fan. This was never built , but nearly killed off the V2500.
As an aviation enthusiast, I greatly enjoy reading your insights and articles.
I had a question really on Engine ratings and why some aircraft who operate the same family of engine do not always offer the same ratings.
I am thinking more specifically here on PW1000G engine.
As an example, we know the C Series competes directly with the E2.
Yet the C Series offers a thrust rating up to PW1525G (25k of thrust), yet the E2 does not, it only lists on their website the 1923G (up to 23,3k)
Is there a reason for this, or does the actual aircraft design stop some aircraft from offering the higher thrust that is capable from the engine?