Jet power management

[punkalouver]

To quote an article.

"...the first movement of the throttle in a jet does not provide much power because it isn't linear. The last movement of throttle, the last inch, almost doubles your power. So you can move the throttle three to four inches and not much happens"

Have you found this to be accurate for all jet engines that you have flown?

[Old Smokey]

punkalouver,

What you describe was very true for earlier Jet engines. With these engines, there was a pretty-much direct mechanical link between the Thrust Lever and the FCU. There's no reason why the mechanical response couldn't have been modified to provide a more linear response.

In more modern engines, where the Thrust Lever position is transmitted to an Engine computer, the response is much more linear. (Note that I said 'much more', it isn't quite linear, but much closer).

It's quite noticeable with trainees from earlier engines, that to achieve a little more break-away thrust, they advance the Thrust levers an inch or two, and give the terminal, other aircraft, containers etc. a good dusting off.

Regards,

Old Smokey

[XPMorten]

The Thrust/N1 curve is logaritmic.

Take the classic 737 as an example, CFM56-3 engine.

If you increase N1 from 80-90%,
you will add about 5000 lbf thrust.

If you increase N1 from 30-40%
you will add about 1000 lbf thrust.

(ISA, sealevel etc)

Cheers,

M

[Old Smokey]

What XPMorten says about the relationship between engine speed and thrust is absolutely true, in fact, that relationship has hardly changed over the decades. Engine thrust is APPROXIMATELY proportional to the engine speed raised to the power of 3.5, versus speed squared for a propeller aircraft.

What HAS changed in more modern engines is the relationship between TLA (Thrust Lever Angle) and Net Thrust. The Engine Speed Vs Thrust is pretty much unchanged.

Regards,

Old Smokey

[hawk37]

Smokey, is there any explanation as to why this 3.5 exponent hasn't really changed much over the years? I'd have thought the use of fans would drastically change the curve's shape.

So can I apply this to a spreadsheet, if that's what you've done, and derive RPM versus thrust for the range of turbos from turbojet to turbofan?

Hawk

[Old Smokey]

hawk37,

The motive force to drive the fan comes from the Jet engine core, and the 3.5 exponent (approximate) relates to the jet core. If the core output, both in terms of it's own direct contribution to thrust, and the contribution in terms of power applied to the fan, is raised to this exponent, then the overall thrust output follows in similar fashion.

Yes, you are correct to a degree, it's not as high as 3.5 for a high bypass engine, but not a great deal lower than the 'straight' jet engine. (If it did, the difference between N1 speed and N2/N3 speed at higher altitudes (where the N2/N3 is operating at much higher RPMs, would be much greater - It's not).

Regards,

Old Smokey

[tallsandwich]

Thanks to the contributors for such an interesting and well informed thread!

[barit1]

XPM & Smokey have it right, but WAIT! There's MORE! ;)

Early engines had (mostly) a linear throttle-to-rpm control characteristic - it's simpler mechanically.

But by the late 50's it was clear that this increased the pilot's mental workload a bit, and the systems guys decided the "speeder spring cam" (that links the throttle handle to the governor speeder spring) needn't be linear - it could just as well be shaped to give a large rpm change rate near idle, and a more gentle change near the top end. In so doing, the throttle-to-thrust curve can be made very nearly linear. Pilots like linear response! I think most engines built in the last 40 years have the "smart cam" or its electronic equivalent.

There's another aspect to this - the dynamic response of the engine. Early engines were notoriously slow to accelerate. If you were low and slow on approach and gave it some throttle, it seemed like forever before thrust started coming up. That made the early throttle-to-thrust characteristic SEEM even worse! Again, modern engines have a much friendlier response rate so it's no longer a big issue.

[John Farley]

An interesting topic, but as has been said no big deal these days.

Pilots in general have coped with some pretty non linear thrust versus throttle position arrangements over the years without (I suspect) many of them even realising how much they were compensating through training, experience and skill.

By the time the P1127 (Harrier forerunner) started in 1960 it was clear that a linear relationship between the position of the left hand and thrust variation was really needed if you were to be able to control hover height accurately and without real skill. We settled for 0.1g per inch for the hovering range of the throttle handle and crammed the rest of the RPM range up into the length of slot left. There was some debate about whether the extremely sensitive wing-borne range that resulted was a major disadvantage for close formation but in the end people just accepted it.

[hawk37]

Smokey, I tried to apply your 3.5 exponent to some approximate data for the Garrett 731 engine I had, generally called a medium bypass fan. Using N1 of 97.7% and 3700 lbs, I then produced the resulting Thrust vs N1 data. Unfortunately it was not all that close. An exponent of 2.4 seems the closest, however the middle values have the most discrepancies.

But then it occurred to me that the N2 section, a completely free spool from the N1, would for the same thrust have completely different speeds from the N1. Not having N2 data, I can’t take my calculations any further.

My interest is peaked in seeing how close a formula can come to predicting the thrust that an jet engine produces, based on rpm. Sea level and 15 deg C for starters.

Perhaps I should point out that my "data" I have for the 731 comes from the engineering pages off a level 2 simulator at Flight Safety

Do you have any representative data you could provide, for any engine at all?

Enicalyth?

Many thanks, Hawk


A= 0.062092802
exp= 2.4

N1 % EFN lbs Calculated Thrust
97.7 3705 3705
96.8 3658 3624
96 3614 3552
95.2 3575 3482
93.9 3475 3369
92.3 3347 3232
91.3 3255 3149
89 3059 2962
85.8 2748 2713
81.7 2350 2412
76.3 1937 2047
70.5 1554 1693
64.9 1293 1388
60.2 1084 1159
55.1 891 937
49.8 695 735
43.2 514 523
31.1 240 238

[john_tullamarine]

Your problem lies in trying to oversimplify the sums ... RPM is a major contributor but not necessarily in a simply encoded manner.

The airframe manufacturers generally use a detailed computer program based on engine OEM code to generate the thrust data .... trying to put it into a simple fn(RPM) relationship just isn't going to cut it, I'm afraid .. other than in respect of a rough first approximation.

[barit1]

Tip o' the hat to JF! Thanks for that really interesting writup for a unique application. Some good skullwork went into that system! :ok:

[hawk37]

John, I realized that one cannot get any exact correlation. As Smokey made a point to say, approximate.

However it was this “approximate” relationship I was hoping to observe a little closer, and perhaps play around with some formulas to come up with some realistically close Thrust VS rpm calculations. To this end, I’ve recalculated values for Thrust = A * RPM^3.5 + B, where A and B are constants, however they’re not as close as I’d hoped. Changing rpm from % N1 to % N2 didn’t help at all.

Of course, my thrust values may not be all that accurate to start with. And that is why I’m asking if there are any generic, but reasonably accurate, RPM versus Thrust figures out there for any engine.

Do you have any sample computer programs/OEM engine data? I realize they may be proprietary, but surely the older 737’s performance can hardly be a Boeing secret anymore!

Anyone with ideas?

Hawk

[barit1]

I've resurrected a few nearly-defunct brain cells, and to the best of my memory, the thrust-N1 curve on a well-known engine goes something like this:

1) At low power - typical approach settings - the curve is approximately linear with a low dF/dN slope.

2) Near TO, the curve is again nearly linear with a high dF/dN slope.

3) There's smooth curve connecting the two linear segments.

Hope this make some sense - and that my memory isn't playing tricks on me.

[enicalyth]

hawk37, G'Day

I am still on extended world tour though I have found an employer. It means I am a bit divorced from my library and follow KO Sally from one Wonder of the World to the next going OOooh and AAAaah a lot.. If darling daughter 3 would pick up the phone it would help. I could tell her to put the phone down, pull out the green Cat 5 and stick in the Orange one. That way I'd be connected to my server instead of recorded messages from various beaux. [Who the heck is Dennis and why was it such a wonderful night? Wonderful evening I can handle, but NIGHT!!!???].

I shall see what I can dig up.

Oooh look at this pyramid, just like the one over there.

Listen if all these ancient civilisations were that much cop then why....

Okay I'll get me coat

The "E" and KO Sally

[XPMorten]

Barit,

I have CFM56-3 data.
The thrust/N1 curve is logarithmic all the way.
Did some testing and amazingly
the furmula below is pretty accurate :O
(a tad high in the mid N1's)

Thrust (lbf pr engine)) = N1^2,2

- no constants needed.. .

I suppose older low bypass engines have less thrust in the low N1's and therefore need a higher EXP and
some constants.

Another note, by the time you reach rotate speeds,
you will have lost about 15% of your static thrust..

Cheers,

M

[hawk37]

XP, any chance you could post your data, ie N1 and Thrust figures. Unless you can see a way to post an excel graph or a scan of the data you have. Or PM me.

I see you must have Thrust vs speed data too, given your 15% factor.

Enicalyth, any thing you can dig up would certainly be helpful. Enjoy the world tour!

Hawk

[XPMorten]

Hawk,

Need to protect my source.
Drop me an e-mail morten(at)xplanefreeware.net

With regard to the thrust speed data, that you can
find on NASA's Engine Sim (http://www.grc.nasa.gov/WWW/K-12/airplane/ngnsim.html).
Use the Turbofan CF6 model.
There is also a theoretic way of calculating thrust at alt/speed if you have SFC and FF data (not that I remember how:p )

Cheers,

M

[enicalyth]

XP, G'day!



There is also a theoretic way of calculating thrust at alt/speed if you have SFC and FF data (not that I remember how )

You mean this? 200 lines of QBasic to derive CDc and one page of Excel to derive the rest?

http://i28.photobucket.com/albums/c220/enicalyth/CDc_jdb.jpg

http://i28.photobucket.com/albums/c220/enicalyth/tsfc_jdb.jpg

By recommending Tom Benson's applet which neither asks for nor deserves approval for real-life use you run the risk of crossing the line between cyberspace and meatspace.

I mean for the best but I feel I still have to interject. The samples I illustrate are correct and open for verification. Tom Benson's applet is neither.

Best Rgds

the "E"

[hawk37]

Smokey.
I’m trying to work with excel to produce a ** reasonable ** formula for thrust versus engine speed for some ** approximate ** data I have for 2 engines.

I have 2 dilemmas that I was hoping you could perhaps give me some quick guidance about:

1. Would the equation be more accurate using the form Thrust = A * Speed^exp, or perhaps Thust = A * speed^exp + B ? Or am I just complicating things with the B constant?

2. Should it matter whether speed is N1 or N2? Or is one more likely to fit a logarithmic curve than the other?

Thanks for your time, Hawk

[pstaney]

Hawk37, Hesitant to reply as is my first post, but I'll try. I see no reason to have a Y intercept for a thrust versus rpm graph. Zero rpm = zero thrust.

Enicalyth, the TSFC graph seems understandable. What does the compressibility coefficient graph have to do with engine TSFC?

Stan

[hawk37]

Well, here is my Thrust Versus N1 graph for a variant of the Honeywell 731 engine, data taken from an FAA level C certified simulator. For this data, an exponent of 2.4 seems to be closer than an exponent of 3.5

http://i31.photobucket.com/albums/c382/severn/N1jcalc.jpg

Stan, I agree, no B value, or intercept should be required. And nor did it help my curve fitting.

Enicalyth's TSFC graph seems odd to me at one spot. Namely, minimum for FL390 is about .582, then for FL 380 it rises to about .583, and then at FL 370 it drops again to the minimum value for the whole graph, just below .582. I know rounding errrors can make a difference, but the precision of the data, reflected in the orderly presentation of the TSFC lines, leads me to believe there must be some other explanation.

Enicalyth, can you provide your insights? Or should I PM you?

Hawk