Turbofan fluid mechanic and thermodynamic calculations

[sgtusmc]

I have recently developed a multi-computer networked simulation of a turbofan jet engine test cell control room, including a real-time simulated jet engine "model", PLC control logic, video playback of a engine at various states, and a throttle for maneuvering. I currently calculate all necessary engine parameters using 3-5 order polynomials based on a throttle lever angle from the physical throttle I built for the simulation (i.e. shaft speeds vs. TLA, TGT vs. TLA, P30/T30 vs TLA, etc), using actual engine data collected from our tests.

It looks great according to my techs that are currently using it for training (start-up, shut-down, accel/decel). For version two, I've been trying like mad to generate a generic engine model that has, as inputs, the environmental conditions (ambient pressure/temperature), mechanical parameters (shaft speeds - 3 spool) and physical dimensions of the actual engine (rotor tip/hub dimensions, compressor blade angles, etc). The model I'm attempting to create will then, as outputs, have fuel flow, P30/T30, TGT, mass flow, thrust, among others.

After collecting a a library of literature, I know I have everything in front of me I need, but not being a calc, trig, or even algebra guru, I've fallen short and must admit defeat. I have yet to complete a model that doesn't end in an unrealistic output.

A simple example of what I'm currently working on, the axial compressor. Taking one stage (rotor/stator), and given the hub/tip dimensions, blade angle, rotational velocity; can I calculate the pressure/temperature rise? I'm also stumped on what input velocity to use, given this engine is tested on the ground attached to a stationary test stand.

Finally my question; can anyone tell me (blunt is fine too, I can handle it), if I'm just marching down an endless path or if it's possible to calculate all engine parameters having only the environment, throttle angle, and physical engine dimensions as inputs?

Forgive my ramblings, but after 2 months of investigation I have generated so many questions I hardly know where to start.

Thanks for any help!

 

[AlephZero]

Your comments seem to imply you are working on testing real engines. The engine manufacturers already create this type of engine model for simulating aircraft performance amd supply them to customers (i.e. airlines)

You would seem to have a legitimate use for the models, so I suggest you talk to the engine manufacturers. You may have to do some work to make the models run in real time, instead of simulating a typical aicraft flight plan or whatever.

I don't think developing a realistic model on our own is a very feasible, especailly for "transient" behaviiour (as opposed to steady state conditions), and for startup and shutdown which can be way outside of "normal running" conditions. You would need a lot of engineering data that is not in the public domain.

 

[Vadar2012]

Zero is right mate, that wouldn't be a very feasible endevour, especially since you want a transient simlations. Not ganna happen, the information that you would require is unobtainable, except from the manufacturers for a price. Just talk to the them, they'll sort you out.

The equations to describe the stuff you're after are also complex. The trig relations for the blade angle and whatnot is the easy part.

 

[sgtusmc]

After re-reading my post at work today, I realized I made the mistake of asking for everything in the form of one question. My apologies, I will try and clarify.

Your comments seem to imply you are working on testing real engines.

Yes, I work for one of the big three manufacturers testing our development civil turbofans.

The engine manufacturers already create this type of engine model for simulating aircraft performance amd supply them to customers

Correct, and I have access to them, as well as direct data collected from our tests. The issue, however, is that the engineers using those models are using them on software we just don't have on hand in the test cells (mine is in the middle of nowhere). Several of my friends/colleagues could get me that information upon request. Currently our engineers are VERY busy developing our latest engine. Not that they wouldn't stop for a few minutes here or there to help me with a simulation that would help us test their creations, but asking them for every little thing here and there would be a bit frustrating for them. I'm a former metrologist and a long time programmer, but I'm in no way a math or physics expert. I do however wish to figure this out for myself so I can better understand what's happening "under the hood" so to speak. I'm big on teaching what I know, not what others can simply read. Didn't get the nickname and tattoo "Hard Way" on my arm in the Marines for nothing.

I don't think developing a realistic model on our own is a very feasible, especailly for "transient" behaviiour

I have already created a model that does mimic the engine to within 1% of the real thing, but I'm using unrealistic relationships between parameters with nth ordered polynomials with respect to throttle lever angle. I know I can use different relationships, like TLA to fuel flow, fuel flow to TGT, or any combination that matches a trendline on a spreadsheet for that matter.

And that all works perfectly, for starts, shutdowns, even accel/decels from GI to MTO... but, how then would I model a surge/stall, bleed valves opening (causing sudden increase in shaft speed)? One could argue that I should simply "play back" the data from what we collect at my work, and yes I could (tried that), but it over complicates things having to track playbacks in addition to polynomials and somehow melding them together into a harmonic "system", and still maintain flexibility and modularity.

I'm not looking to analyze an engine to death, just a more simplistic formulation of real-world physics put into objects (object-oriented programming BTW), that can be connected together inside a training simulation to show a rough estimate of what an operator might experience during failures, faults, and normal operation. 3D is not considered here, and rough estimates is more than enough to satisfy my needs.

I'll finish with a scenario I should have started with in my first post. Starting with the environment, I know the pressure, temperature, and axial velocity (0 m/s for ground test). Feeding the p/t/v into a fan at a rotational frequency (with of course blades that have dimensional properties of length, width, angles at arbitrary points over the radius), what would the velocity be after the fan has accelerated it? Taking this now accelerated velocity, delta temperature/pressure and feeding it into a axial compressor object, or even a bypass object (nozzle). Inside the axial compressor object, a similar process occurs where the p/t/v would be calculated at each stage.

Hopefully from this you can see what I'm after. I'm not looking for exact representation, but it is my belief that separating objects into their respective areas of concern, will make dynamic changes in the simulation reflect what would happen.

Truth be told, my curiosity, once sparked cannot be extinguished. It's no longer about just training my guys at work to test jets. I learn something everyday, and accomplishing this would be a great challenge and huge accomplishment for me personally.

Okay, I'm off my soap-box. Again, thanks for the reply and hopefully this clears things up a bit more.

 

[Vivek03]

Your concept of making a real working model by your self is really appreciable... but it seems the cost will be more. so i guess u should make a small model first to be sure of it. and if u consult to some engineers then that would be more helpful if u having your pocket full...