|
Post by finiteparts on Sept 14, 2014 13:13:34 GMT -5
Mark,
I tried to go around an find that thread and couldn't locate it. Could you post a link in here? I would like to see the combustor layout on that also. Somewhere, I read a paper on that engine design or one very similar done by Garrett...if I can find it again, I will post a link.
Thanks!
Chris
|
|
|
Post by finiteparts on Sept 14, 2014 13:50:31 GMT -5
I also took some additional images of the NGV section shown here...
Notice that the leading edge shape of the nozzle guide vanes is quite large and rounded. This helps the nozzle accept flows at all kinds of incidence angles more effectively. Using flat plates as NGVs could have bad effects if the flow enters at a large enough incidence angle so that the nozzles experience large regions of flow separation, which essentially become flow blockage in the nozzle.
Also notice the complex shape that they machined in to accelerate the flow inside the nozzle passage. In these images, you can see the location of the shock waves/expansion waves forming at the trailing edge. These shock waves are what "chokes" the flow. Since you are expanding in the turbine stage, you usually strive to "choke" in the NGV section, but it is possible to choke in the turbine and the NGVs due to the different speed of sound in either location (since the gas temperature is reduced in the turbine, the sonic velocity also drops). Below, you can see the effect of the bent vane as the shock and the upstream "horse-shoe" vortex seem to be modified as compared to the other passages.
Here you can see the cooling holes just upstream of the NGV section...diameter = 0.125 inch X 66 holes Also, the compressor impeller... Note how "early" the splitter blades are introduced. Usually, the splitter blades are farther down in the impeller flow passage. Also, this impeller is from a later generation T-62 and is made out of 410 SS due to service life requirements. Since the compressor and turbine are connected, the compressor absorbs heat and since low cycle fatigue life is reduced as the temperature rises, a more capable material is required over the Al castings. Finally, I also took some images of a later generation T62 combustor that utilizes pressure atomized injectors through venturi styled inlets. Enjoy, Chris
|
|
|
Post by racket on Sept 14, 2014 16:51:29 GMT -5
Hi Chris
Very interesting witness marks :-)
When I had my T62 apart I felt the NGV to turb tip constituted a divergent duct to produce supersonic velocities , it normally ran with a >1730 ft/sec tip speed , so would have needed some pretty high gas speeds just to catch up with the blades , but T I Ts weren't that high , so a limited sonic velocity ............it had me a bit perplexed when I measured it up , NGV had 11mm height vanes whilst turb had ~13mm tip height.
Cheers John
|
|
hurdman
Member
Joined: July 2014
Posts: 11
|
Post by hurdman on Sept 14, 2014 19:32:59 GMT -5
Chris, Thanks for the quick-turn on the t62 numbers. I'm doing the maths now...
Thanks again Danny
|
|
|
Post by finiteparts on Sept 14, 2014 21:40:08 GMT -5
Hi John, The design tip speed for the T62-11 is 1538 ft/s, but they do comment on test to show that even after 250 hrs at speeds of 1900 ft/s there was no measurable material growth (creep).
The uncovered region of the nozzle exit can sometimes form an expanding section relative to the flow. As you know, if there is a throat where the flow reaches sonic velocity, an expansion behind that throat would lead to supersonic flow velocities.
As for the passage differences, in this engine the NGV exit and the blade inlet match up quite nicely. Was yours a -11 variant?
~ Chris
|
|
|
Post by racket on Sept 15, 2014 1:05:40 GMT -5
Hi Chris
It was a T-62T-32 , -13 variant maybe , rated rpm 61,091 , turb wheel 6.516" dia , 1737 ft/sec tip ............I no longer have the engine with me , only a manual , a friend has the engine, so can't verify "numbers" , only some scribblings in a book I have...........it did have the venturi injection bits .
The exhaust diffuser was a nice item , the cone had a 15 degree included angle , area ratio was ~2:1 between inlet and exit .
Cheers John
|
|
|
Post by pitciblackscotland on Sept 15, 2014 1:40:00 GMT -5
|
|
turbotom
Junior Member
Joined: June 2011
Posts: 58
|
Post by turbotom on Sept 18, 2014 17:32:05 GMT -5
Chris -
sorry but I disagree with your statement that the traces on the NGV originate from shockwaves forming from transonic flow. This is rather just the effect of dust blasting the surfce of the NGV. This type of "grit blasting" can frequently be found in engines with radial turbine wheels and vaned NGVs. Actually, this is a big problem that finally caused several attempts to produce ceramic radial turbines to fail.
Any debris that gets drawn with the flow of hot air into the NGV, is accelerated towards the turbine wheel. But due to the higher inertia of the debris, it doesn't reach the same speed as the gas (which approaches the wheel with a tangential velocity component approximately equalling the circumferential speed of the wheel tips). Since the debris is slower, the leading blade faces of the wheel impinge the debris and "kick" it at high speed back towards the nozzle vanes. This may happen several times until the debris is ground to dust so it finally can exit the turbine wheel with the gas flow, or until the debris is absorbed by the ductile surface of the blades or vanes. This results in the typical "blast patterns" on the downstream suction faces of the nozzle vanes and the adjacent walls. Usually, there's a sharp transition from "unblasted" to "blasted" surface where the vane surface exits the "shadow" of the preceding vane.
This also shows that dirt or foreign objects are a much more problematic issue for radial turbines while they would be a "single-pass-through" thing in an axial turbine stage.
Cheers, Thomas
|
|
|
Post by racket on Sept 18, 2014 19:35:33 GMT -5
Hi Thomas
LOL.......I've seen that "brown dirt" in my FM-1 engines turb stage , burnt/ground up bits of turb tip , the NGV exits didn't look very nice afterwards :-(
Cheers John
|
|
|
Post by finiteparts on Sept 18, 2014 21:33:01 GMT -5
Thanks for the info Thomas and don't be "sorry" for disagreeing with me. That is the best part of this global forum, sharing ideas and information so that we are all getting more knowledgeable about gas turbines. I hadn't ever considered the blasting off of the NGV surfaces due to particle recoil from the turbine wheel, but that makes sense. I wasn't actually thinking that the blasted zone was the effect of the shock...I was unsure of the mechanism that would drive such an effect and your theory definitely fits the bill nicely. I was thinking the dark to light transition marks more forward of the "blasted" zone were due to the shock wave, since the throat would be at a minimal cross sectional area, not in an expanding area downstream of the vane trailing edge. This may have been hard to see in the images, so I marked up a few to show you what I was thinking. The pattern in this passage with the bent vane appears much more pronounced, which I thought might have been due to tripped flow condition causing slightly more blockage....just a thought.... I know that I am showing the shock region as if it was in a single plane, even though a lot of research has shown that due to boundary layers, and secondary flows in the passages, the shock formation can be a highly curved 3-d surface that doesn't necessarily "cover" the entire passage. Any thoughts? ~ Chris
|
|
turbotom
Junior Member
Joined: June 2011
Posts: 58
|
Post by turbotom on Sept 19, 2014 3:08:53 GMT -5
Chris - I have to admit I didn't even notice the small clue in your original photo unless you now marked it. I checked back the pictures and there actually may be a hint for a shock wave at the throat of the NGV which is probably much better visible "on the real thing". But anyway, if there is, it should be a very weak one. A classic calculation shows that a gas speed of M 0.82 should be reached at the throat. Depending on the reaction of the turbine, this could also be a little higher so mild shocks may actually be possible. Usually, you will find shock waves in compressor diffusers, here's a nice photo of a transonic one with multiple shocks, starting even in the vanless space directly after the impeller (running at 600m/s = 1960ft/s) Cheers, Thomas
|
|
|
Post by finiteparts on Sept 22, 2014 21:29:41 GMT -5
Thomas, That is an awesome picture, thanks for sharing that! I really like the marks and they are just where you would expect them to be. For those out there that haven't seen some of the flow conditions at the entrance to the diffuser, here are some schlieren images that Boeing did back in the 1960's showing the shock formation in the vaneless space and entry into the diffuser of several diffuser designs. ( these can be found here: www.dtic.mil/docs/citations/AD0384923 ) The detatched bow shock is pretty apparent just upstream of the diffuser vane and represents a normal shock (transition from M>1 to M<1). The second wave seen inline with the vane leading edge is likely due to expansion waves starting from the leading edge and creating another density jump. The final ones in the passage occur at the diffuser passage throat and are due to the local curvature reaccelerating the flow to sonic conditions and then shocking the flow back down to subsonic speeds. I really wish there was some video of this...I built a cheapo schlieren imaging set-up a few years ago just for fun, it might be fun to make up the compressor housing like the Boeing guys did and capture some video...you know, because I don't have enough projects! ha!!! Thanks again for the cool picture! ~ Chris
|
|
|
Post by racket on Sept 23, 2014 0:37:56 GMT -5
Hi Chris
That comp wheel is running very close to the diffuser vane , any wonder they were having "problems" .
Cheers John
|
|
|
Post by finiteparts on Sept 23, 2014 22:50:49 GMT -5
Hi John,
What "problems" are you referring to?
The compressor and DI-1 diffuser arrangement in the lower figure was running at a pressure ratio of over 9:1 at a corrected airflow of 2.3 lb/sec...quite an accomplishment in my book! At the highly supersonic discharge speeds of these impellers, the streamlines are very shallow and thus the diffuser leading edge radius ratio is in the 1.05-1.06 range due to the need to balance the frictional losses in the vaneless space (between the rotor and the diffuser vanes) and the entrance losses into the diffuser passages.
The program target was a pressure ratio of 10:1 at an adiabatic efficiency of 80% for a mass flow of 2 lb/sec. They succeeded in getting a pressure ratio of 10.6 to 1 at an adiabatic efficiency of 72%. Not bad when the small turbos we are using are only getting around 80% efficiency at MUCH lower operating targets and they include over 40 years of "newer" design technology. Their work is really something to read about.
There is a great ASME paperback book out there called, "Advanced Centrifugal Compressors" that covers Boeing's work, Pratt's work on 12:1 compressors with pipe diffusers and some other odd stuff. Definitely worth the money if you can find it cheaper.
~ Chris
|
|
|
Post by racket on Sept 24, 2014 0:17:05 GMT -5
Hi Chris
I meant having shock wave "problems" , with the air leaving the wheel at "high" supersonic velocity theres very little velocity reduction with a 1.05-1.06 radius ratio , the air would have most likely still been supersonic at the vane leading edge , hence a shock wave .
Cheers John
|
|