Contemplating an HX82 Turbojet

[monty]

I'm out of ideas for my 5axis software problem other than writing checks I can't cash...so:

I have some pieces laying around. One is a Holset 3521034 Turbine, and the other is a X856 comp wheel.

My goals with this engine will be simplicity and efficiency. So many smaller engines are trying to cram everything into the smallest space possible. Instead, I'm going to design the diffuser components by the book, and let the sizes fall where they may. Within reason of course. I'm also going to allow the combustion chamber to be large enough to prevent any flame impingement. The goal is an efficient reliable engine without too much drama. We'll see how true to these goals I can stay!

Monty

[racket]

Hi Monty

Yep , I think you could do a bit more clipping of the turb wheel to get the angle out to 40 degrees to turn it into a jet engine wheel :-)

This will give you a bit more flow area to cope with that 30 degree inducer tip .

I think the comp will cope with a few degrees of incidence angle , bit like an aircraft wing .............heh heh , but too much :-(

Cheers
John

[monty]

John,

Build and test first, clip as necessary? Or just go for it? I'm inclined to test first and see what I find before getting too crazy with the grinder. First do no harm??

Monty

[racket]

Hi Monty

Yep , its your build, enjoy the experience :-)

See how things go when you do the numbers , the comp will be more than capable of supplying the hot end , so maybe work on it first to see what you can squeeze through , then go back to the front to see what you need to do .

Cheers
John

[racket]

Hi Monty

Looking at the comp maps of the latest Garret 106mm inducers , at 4:1 PR they flow ~3.3 lbs/sec at best effic , which is ~20% better than the old GT6041 comp , so they must have considerably higher inlet air flow speeds and probably increased inducer tip angles

Cheers
John

[monty]

John,

But....can I squeeze it out the back!!??. that seems to be the issue! Seems I vaguely recall working on this problem in the past for the fan. Whether the nozzle or the exducer is choked..

Not used to working backwards. Usually I do the cycle and then spec. the hardware. This way I have to do the opposite and figure out how they fit together. Going to require some number crunching.

Monty

[racket]

Hi Monty

LOL.....and to complicate things even further we don't know what pressure ratio we need to use across the stage as we can't be certain of the comps requirement or the efficiencies of each stage , happy number crunchin' :-)

I'll be interested in the outcome .

Cheers
John

[monty]

Oh to have a compressor and turbine map.....

Instead, you have to measure the components and reverse engineer everything using design theory. Design theory is imperfect. Otherwise maps wouldn't be needed. Blind man feeling around an elephant.

[monty]

Thanks to John's corruption of my brain, I have a plan to deal with the fact this compressor throat is 15% bigger than the turbine, and the inducer is designed for some ridiculous flow. These two are not going to get along obviously. So I will borrow from the turbojet engines of old, and simply place a wastegate on the gas generator housing. Having lots of bleed air allows for all kinds of fun. You can take a portion, cool it down and dump it at the base of the comp and turbine to keep them and the center housing cool. You can dump it to the outer jacket of an afterburner and use it to cool the inner liner. Or you can do that with a jet pipe without an afterburner. The bleed air mixes with the turbine outlet air and heats up, then goes out the main nozzle to make thrust. No "waste". The compressor is after all what makes the thrust. Not the turbine. Might as well use the biggest comp you can.....So a bypass engine it is!

[monty]

I have a spreadsheet for matching the diffuser to the compressor. It's based on a paper John posted a while ago. I fiddled with the mass flow to get the inducer angles happy at 66krpm. It's about 3.7 lbm/s corrected. Scott has apparently seen about 4.25 pressure ratio with his cart. I'm going to shoot for 4.5. My reasoning is a more efficient diffuser (hopefully) and the geometry of this wheel is slightly different than the X858. It should favor slightly higher pressures.

So what am I planning to do different? I have a couple books on diffuser/compressor design. I'm trying to use the "best practices" from both. One source has a plot relating aspect ratio of the diffuser channel to efficiency. The highest efficiency is an aspect ratio of 1. A square entrance. Not surprising really. Least hydraulic diameter of a rectilinear shape. This same source lists the optimum angle as 9 degrees. And the length to width optimum is 8. Another source lists optimum L/W as 5-6. Who's right??? IDK. So In the interest of keeping things semi reasonable on diameter, I went with around 6. This gives me an engine OD of 12in. Discharge angle is around 20deg. Luckily all this works out to 23 channels with a very slight adjustment to the channel width. 23 is a nice prime number safe to run with this 8+8 comp wheel. There is a slight pinch into the diffuser entrance. There is also a 10% pinch in the bend to the outer case. The idea here is to minimize diffusion in the bend to keep the flow behaved. Best practices here are a larger radius, and at least 1 annulus width to the beginning of the de-swirl stage. I haven't put much effort into the de-swirl vanes yet, so that will probably change. Here is what things look like. No regard for matching to the turbine because I'm going to use a bypass valve.

And now playtime is over. Maybe tomorrow I can work on some general configuration layout.

Monty

[racket]

Hi Monty

This is going to be interesting :-)

Cheers
John

[monty]

I did some cycle analysis on the bleed air idea......not gonna work. Turbine PR goes too high. So:

I got serious about matching the turbine...

I made a cad model of the blade surfaces so I could play around with different trims and see what the areas and deflection angles would be. The stock Holset seems to have about 8 in2 of throat area. 

Then I started looking at choked flow limits. From this I started trying to solve the equations for work, and get the inlet conditions to behave. I have decided you don't want a choked exducer. You want a choked nozzle...In the process of doing this I found what I believe is an error in a textbook. I went to visit a colleague who once worked for Pratt and Whitney R&D. He agreed with my assessment. So now I must re-write my spreadsheet. He gave me a parting gift that would solve my problem...

I'm starting to understand why the aerospace turbines look this way.

Monty

[racket]

Hi Monty

Bummer , that turb wheel spins the wrong direction ;-)

8 sq ins is a good area, better than the similar dimensioned Garrett G Trim

Cheers
John

[finiteparts]

That is a nice turbine. It looks like a Garrett GTC36 turbine. Do you have any info on it?

Also on the choked ngvs verses the rotors exducer, it should be highlighted that the rotor choke is speed dependent; since the local gas properties are in the relative frame of reference, the local acoustic velocity is dependent on the relative temperature, not the absolute static temp.

This means that predicting the choked flow is more determinant when you design for a choked ngv.

- Chris

[monty]

John,

Yeah....I'm not sure I can find a comp wheel to match that turbine...

Chris,

The relative frame issue was actually the error I found in the book I have.

As far as the turbine goes, nobody knows. It's been passed around when various people retired. It has "scrap" and an itar number etched on the back. The guy who originally had it used to work for Allison circa 1980's. It had a long tension member that retained it in whatever it came out of. Someone unfortunately cut that off. Bearing surface is spalled a bit, could be why it was scrapped, or it could be failed X-ray....who knows. Wouldn't try to run it in anything, but it makes a nice piece on the shelf.

Monty