Contemplating an HX82 Turbojet

[monty]

John,

I need the inputs for his engine in order to give you something. Which turbine (clipped??), TIT, comp, efficiencies plus mass flow.

Monty

[racket]

Hi Monty

OOPS , no need for his numbers , you mentioned you might be doing something wrong , so I thought I'd compare where you currently are with what I had for Anders as I designed for 5:1 for that engine which only used the smaller F Trim wheel , same turb inducer width though , so potentially the same NGV , 72,000 rpm for a comp tip speed of ~1900 ft/sec , 24-25 degree inducer tip .

Cheers
John

[monty]

John,

I can't compare without the inputs.

My #s are assumptions based on a choked exducer of the turbine I have and some guessed efficiencies. I don't know whether it's correct and won't until I build and test or have some good data to enter and compare.

For a mass flow of around 3.3 lbm/s and 68krpm PR 5 I'm getting 200 lbf thrust. NGV angle around 30 deg and exit mach .78. Inducer tip angle around 35deg. I'll adjust when I get the wheel and measure the throat area and tip angle.

Monty

[racket]

Hi Monty

Mmmm , OK , 5:1 needs ~1900 ft/sec , so 68 K means a comp ~162mm dia

Turb at 129 mm at 68 K has tip ~1500 ft/sec .

M.78 at NGV exit would mean gas velocity ~1560 ft/sec at 30 degrees

That makes for a pretty bad looking velocity triangle at the turb wheel , the turb tip are going to have to accelerate the gases as they hit the front of the blades making for less gas deflection and probably not enough power production for the comp .

5:1 at 78% effic will produce a ~215 deg rise in the comp , and require ~132 HP/lb to compress,

132 HP and some rough calcs ,..............using a blade velocity of 1300 ft/sec (exducer tip speed) , can't take the inducer tip speed into account because of the strange gas angle , 132 = 1300 X gas deflection divided by (32.2 X 550) = ~1800 ft/sec gas deflection , but we only have ~1600 ft/sec, with a choked exducer at 35 degree tip :-(

You'll run into temperature problems trying to get the horsepower production up ...............bummer , been there www.youtube.com/watch?v=J29zgn7xv6g :-(

Cheers
John

[monty]

John,

Are you accounting for the work from radius change of the gas? Your diameters and tip speeds are larger than my sources are recommending. For PR 5 a tip speed ratio of just over 1.6 is sufficient with a proper inducer. 1.62*sqrt(R*T01)~1530ft/s. The inducer would need to be small to maximize the speed ratio between inducer/exducer tip. So a large inlet angle~45deg would be needed.

The same is true on the turbine in reverse. The inlet angle needs to be negative for the flow into the inducer to behave. I'm actually pushing the limits of what is needed to keep efficiency up. About -20deg. I'd prefer -30ish.

Monty

[racket]

Hi Monty

Maybe take a look at this comp map www.garrettmotion.com/racing-and-performance/performance-catalog/turbo/gtx5533r-gen-ii/ ...... 85/133 mm so a low 40 Trim , but will need ~1850 ft/sec tip speed for 5:1 PR.

With the turbine stage, to have that "lowish" Mach number being supplied its like fitting a large A/R housing , but they run into problems once our PRs start to rise , we then need to start tightening up the A/R to be able to get to higher rpm .

The T351 Paper Link I sent you www.semanticscholar.org/paper/Performance-of-a-High-Efficiency-Radial-Axial-Rodgers-Geiser/3cbfe4b716432d7e8bdf5d425009c1f29fcc2cbe probably explains things best , its turb wheel is larger in diameter to the comp , the turb tip speed is very high , great for power production , the high tip speed allows a high spouting gas velocity from a choked NGV , great for power production , whilst maintaining the 0.7 ratio for efficiency , the large exducer diameter/area with low exit gas velocity makes for high efficiency.................our turbo based wheels are all "wrong" , but thats all we've got to work with so we need to simply wear the inefficiencies as best we can :-(

Cheers
John

[monty]

John,

I agree the turbines are all wrong for what we are doing. But they are still radial turbines and the math is the same. I'll check my spreadsheet, but without a known case I can run through it, I got nuthin'. The cycle numbers are good. That's easy enough to test against a known solution. The vector math could have an error somewhere.

However, I disagree on the inlet triangle. It looks perfectly fine to me (though I'm pushing it a bit on the shallow side). I assumed choked flow exducer and sonic in the exit frame zero swirl which is conservative. The flow will go supersonic with higher pressure ratios. The turbine won't reach its limit until the axial flow is at M1. Still my work balance is fine. The NGV angle is based on satisfying the work condition. That's why I asked if you are accounting for the radial velocity change. The flow is not going to hit the turbine face, it will flow in correctly, because the turbine is in the rotating frame. Every single paper or book I have on radial turbines discusses this. The inlet is not designed as a reaction (I mean impulse) stage. That's why it's so important to match the specific speeds of the two wheels. Gas deflection isn't how the inlet should work. If it was, the inlet blades would be scooped towards the NGV. Not possible because of structural limitations...

Anyway, no sense arguing. I'm going to check my numbers again. I need to include exit swirl to get the most from the turbine. Right now I have assumed M1. While I'm incorporating that, I'll do some error checking. If I don't find anything, there's only one way to find out.....

Monty

[monty]

John,

I looked at the compressor map. A couple points. The inducer is too large and inlet angle too shallow. These turbo wheels are designed to hit pressure sooner than we want. There is a big engine in the way that wants air. And the designers want the compressor to build pressure against it. The turbine is throttled by a wastegate and powered by the engine drive pressure, not the compressor. The designers don't care if it starts blowing out the surge slot. They are mismatched for a jet, but correct for a turbocharger. The fast boost response, surge slot thing means making the inducer bigger than what is ideal for a jet. If you tried a jet engine comp design on a turbo installation, the inducer would stall and it would perform poorly.

For a jet, the inducer should be smaller ~75mm (to maximize tip speed ratio) and the inlet angle steeper ~45deg. You want the smallest diameter you can get away with while still keeping the throat large enough. That's why this type of wheel has a larger hub. To increase the throat area and match the inlet speed. The exducer tip ht is set to maximize pressure rise in the wheel without separation, and the diffuser should convert the tip speed to the final P2. If this is done that map will be compressed horizontally and expanded vertically. There would be no problem hitting PR 5 with a lower tip speed.

Have a look at this wheel if you don't believe me....I think the designer knew a thing or three. You might recognize the name. Look at the pressure ratio and the tip speed....

Since I'm building a jet, I'm going to use jet engine practice. At least as close as I can get with what I can buy. That will prove/disprove my approach.

Monty

P.S the picture got cropped somehow. That impeller was designed by Noel Penny Turbines. Got cut off.

[racket]

Hi Monty

LOL.................you've lost me :-)

Nope I haven't taken much into account , just kept it simple , I've never been able to reconcile the turb maths with reality , preferring to look at the available turbine maps to get an idea of whats happening .

Thats why I'm very interested in your project, you've got the capability to do the maths , I've had to go down the trial and error route which has been expensive at times :-(.

I'll drag out an Allison C28 comp I have and have a look at its "numbers", its been sitting around for a long time doing nothing :-)

Cheers
John

[monty]

John,

It's not exactly the most intuitive of things.....The axial stuff is a lot easier to understand. I have avoided crawling into the math because I just want to build something. But I finally got tired of being confounded by the darn things. Not saying I'm right. I still don't trust my numbers, but they are turning up things that look more and more like the parts I see in the big guys stuff...so I'm willing to burn up some Benjamins to test the ideas. Either way we'll learn something...or simply get more befuddled...

;-)

Monty

[racket]

Hi Monty

I got the C28 comp out



and did some measurements , 32 Trim wheel , ~5.5"ind 9.75"exd , 55 deg backsweep and a "high" 40 degree inducer tip .

Specs are suppose to be 7:1 PR at 51,000 rpm and 4.33 lbs/sec through the ~13.1 sq ins of inducer throat , ~20 sq ins of inducer flow area and ~10 sq ins of exducer flow area .

Now 9.75"dia at 51 K gives 2170 ft/sec tip speed for the 7:1 PR , now if pressure goes up at the square of tip speed 2170 X 2170 = 4708900 , 7:1 PR div by 5:1 PR = 1.4 , 4708900/1.4 = 3363500 , square root it for a tip speed of 1833 ft/sec at 5:1

The Noel Penny comp has some dodgy numbers if compared to the Allison , the NP has an 11.88 dia comp at 38950 rpm for a tip speed of only 2019 ft/sec yet has a much higher Pressure Ratio .

Its hard to find the truth sometimes , LOL, lets keep looking

Cheers
John

[monty]

John,

You can only use scaling laws to compare wheels of the same family. In other words what you have done there is correct for a scaled version of the Allison wheel. It will not work to compare a wheel with totally different geometry. The Noel Penny wheel was a research item. Tip ht too small for the leakage area. Efficiency was low. There are all sorts of reasons not to design a wheel like the Noel Penny wheel. Especially for turbochargers. This doesn't make the results invalid. The numbers aren't dodgy, the wheel was just designed for a different purpose. There are many more examples of wheels designed for propulsion applications. They don't look like turbo wheels for very good reasons. Just do a google image search. You will notice the trends.

Here is the map for the Noel Penny Wheel. This is what is possible when you design for the specific application:

Nobody wants the wheel we need, and trying to find it among the turbo offerings is proving to be futile. I'm going to measure the compressor that is as close to what I want as possible. Design an engine based on that, build and test. I can use that data to design a wheel for this specific purpose. Then we'll see how much it costs to make.

Monty

[monty]

Here is a more recent example:

NASA CC3

PR 4.5 Tip Speed ~1600

[racket]

Hi Monty

Yep , totally agree , the big locomotive turbos often have low trims ..............one thing that does bother me though is why ABB don't use them , one would think a turbo thats built for 40,000 hours and generally at a "design" output, would be wanting every last bit of efficiency considering the potential fuel savings .

Very low Trim , lotsa blades to minimise slip and plenty of backsweep for efficiency.

The closest thing I've see from a turbo is the TF15 comp wheel that came with my CHRA , low trim , lotsa blades ( 10+10) but unfortunately not a lot of backsweep , but I'd overlook that :-)

Cheers
John

[racket]

Just checked out the CC3 ntrs.nasa.gov/api/citations/20180001471/downloads/20180001471.pdf lotsa data :-)