|
Post by finiteparts on Jun 20, 2016 15:41:23 GMT -5
Hi Patty,
After pondering on your question, I did some digging on what really is the correct pressure to specify at the free jet boundary and it turns out that the isentropic assumptions struck again. If the jet was issuing such that it was isentropic, then the total pressure would remain constant and the jet would drive to issue at sub ambient static pressures that slowly work themselves to match the total pressure.
But, the real world is not isentropic. The free shear layer that forms around the issuing jet is highly dissapative and thus the total pressure is reduced, not constant. Since fluid boundaries cannot support pressure differentials across them (back to the ballooning statement), the local static pressures have to be equal. Thus the static pressure in the jet has to be constant and equal to the ambient pressure.
An additional point of confusion is that in reality, the ambient pressure is such that total and static are equal due to the fact that we assume it to be a quiescent fluid and as such velocity is zero.
So yes, the static pressure at the exit of a real turbine or nozzle will equal the ambient pressure, with the attendant losses (entropy generation) being attributed to the reduction of the total pressure along the streamlines.
Sorry for the confusion. While I did mention previously that in the real world the total pressure would be reduced, I did not anticipate that the loss of total pressure would be so large. It is interesting how a simple assumption that the flow will be close to the isentropic case can really cause such large discrepancies. Additionally, it is quite bothersome that so many texts (I searched quite a few that I have on hand) just breeze over this point without providing a clear discussion or simply jump straight into the sonic or supersonic aspects of free jets.
~ Chris
|
|
|
Post by racket on Jun 20, 2016 16:53:18 GMT -5
Hi Patty
Could you please provide the calculations.................even using the 0.3 kgs/s at 3:1 at 900C the throats at still way too small , that combined with the "oversized" turb exducer at 54% larger than comp inducer is undoubtedly causing problems .
A larger NGV throat area will allow more flow , even if at a tad lower efficiency that optimal , but will "fill" the turbine flow passageways and provide better power transfer.
Cheers John
|
|
|
Post by madpatty on Jun 20, 2016 20:59:58 GMT -5
Racket,
At 3 PR and 900 degrees celsius(1173 K) TIT
Compressor flowing 0.3 Kg/s at 3 PR and 123000 rpm at an efficiency of ~ 70%.
Required Temperature drop across the turbine stage comes to be 138 degrees and hence the PR across the turbine comes to be 1.998 PR.
A) Velocity triangle (assuming radial entry at the turbine wheel inducer) and NGV discharge angle of 21 degrees to the tangent.
Turbine tip speed = Tangential component of gases = 489 m/s
Therefore Radial component = 188 m/s
Hence velocity needed out of the throat = 524.14 m/s
B) Temperature equivalent of velocity at throat = (V^2)/(2*Cp) = 103.279 K Static temp at throat = 1173 - 103.279 K = 1070 K Static PR at throat = 1.97 (2.85/((1173/1070)^4))
C) Hence Density at throat = 0.65 kg/m3
Total velocity = 524 m/s
Hence area required at throat = 0.3/(0.65*524) = 880 sq. mm
Boundary layer adjustment with this area puts me right in the ballpark with 960 sq. mm.
Am i doing it right?
Cheers. Patty
|
|
|
Post by racket on Jun 20, 2016 23:20:44 GMT -5
Hi Patty
In short .....NO .
You need to first workout the entire energy drop across the stage which includes the power for the comp as well as the "exhaust" power exiting the wheel .
Your 1.99 PR for driving the comp is correct , but because theres normally several hundred feet per second of velocity out of the wheel so we need to add on another 0.2 PR to get our total PR across the stage of perhaps ~2.2 in your case .
Now your current PR up to the throat is only ~1.5:1 to produce your ~525 m/s gas velocity .
But if we start with a 3:1 PR from the comp , and lose 5% across the flametube we only have a 2.85 PR going into the NGV , divide that by our ~1.5 PR leaves us with a 1.9 PR at the throat .
As we need a total PR of ~2.2 :1 to power the comp and provide the velocity out of the exducer , we also need a fair pressure drop across the turbine wheel , BUT , because it is fairly "oversized" for the amount of gases going through it the velocity will be low , there could be "recompression" in the larger flow passageways of the exducer.
I usually design for a choked NGV throat , that is a PR of ~1.9 in the NGV , because our engines with relatively low PRs need a total PR across the turb stage of >2.2:1 .
Now this is how I'd approach your design to get a ball park figure .
2.85 PR into the NGV , divide by 1.9 gives us a PR of 1.48 at the throat at a static temp of say ~1,023 K ( 1173 K for TIT) , density of ~32 cu ft/lb , multiply by our 0.66 lbs/sec gives us a flow of ~ 21 cu ft/sec , our ~150 C temp drop should produce a gas velocity ~1900 ft/sec , therefore we'd need a flow area of ~0.011 sq ft , so if we add on a tad for boundary lets use 0.012 sq ft -1.73 sq ins, which gives us a total throat area of 1116 sq mms or 74.4 sq mms for each of the 15 throats , which if 12mm height to match the turb inducer tip height would mean a throat width of 6.2 mm .
BUT , ........don't you just love buts , we're assuming your turbine wheel is a good match for your 0.3 kg/sec- 0.66 lbs/sec flow , but as its >50% bigger in the exducer area than your comp inducer it could mean a more appropriate flow would be somewhat larger .
The power required by the comp wheel requires a certain amount of gas deflection across the turb stage , an "oversized" exducer will restrict gas velocity if mass flow is a bit below what it'd prefer and in the process reducing gas deflection , the usual result is that we need to run higher than required temps to compensate , this then exacerbates the rest of the turb stage design .
I'd be measuring the actual flow passageway area at the exducer to get a rough idea of its throughput so as to determine if its going to compromise gas deflection .
Generally speaking , as long as you run a choked NGV there should be enough energy imparted to the turb wheel to drive the comp regardless of the exducer configuration , my 12/118 engine with nearly all its exducer removed is able to power the comp
My gut feeling would point me towards a throat of >6.5 mm possibly 7 mm , and then redo the design numbers for a mass flow ~1/8th more , say 0.34 kgs/sec - 0.75 lbs/sec to see how the exducer flow area would process it .
Hope this helps
Cheers John
|
|
|
Post by racket on Jun 20, 2016 23:40:15 GMT -5
Hi Patty A quick look at the Garrett GT4088 www.turbobygarrett.com/turbobygarrett/turbocharger#GT4088R which has a turb wheel with just a 1mm difference in sizes produces a flow of ~0.85 lb/sec - 0.38 kgs/sec , roughly 25%more than your 0.3 Kgs/sec. Yep , bigger NGV throats required. Cheers John
|
|
|
Post by madpatty on Jul 14, 2016 20:47:24 GMT -5
Hi Racket,
I had been experimenting with the reversed flow combustor lately.
I kept everything else constant/same, including the NGVs and designed a new combustor (in reversed flow configuration) around it.
And the result was quite good.
Max P2- 30 psi Max TOT- 606 degrees celsius.
What I have found out is, though the idea of the combustor cross-section area = 3x inducer area is right, but as the inner flametube diameter increases(due to OEM bearing housing used by me) the annular width (or the portion where combustion has to take place) decreases.
This may be the cause of high TOT problem.
The annular width was just 22mm in my case.
So it appears there may be a limit to the minimum annular width you can use for good combustion.
Cheers.
|
|
|
Post by madpatty on Jul 14, 2016 20:56:16 GMT -5
Racket, The post that will be of your interest. During my early tests with this new combustor design, I was also having hotspot issues (same as in your case). Before- Then I checked the fuel(gas) injectors and found that some of them were throwing part of the fuel in the space between flametube and outer casing. I elongated the fuel injectors so that they are now injecting completely inside the flametube. The problem is more or less solved now. After- Cheers.
|
|
|
Post by madpatty on Jul 14, 2016 21:13:53 GMT -5
There is a minor hotspot issue at higher P2s, which i am sure now is not because of the fuel being injected at wrong places. The flame tube shows the flame crossing the tertiary zone at some places. Good combustion - Bad Combustion- Ther gas injectors in my case are pointing towards the inner flametube. Can this cause a problem or should i open up the primary holes in that bad combustion area. The flametube has 70 mm annular width. Primary Zone- 18 x 6mm holes in outer can 6 x 6mm holes inner can Secondary zone- 6 x 9.5 mm holes in outer can 3 x 9.5mm holes in inner can Tertiary Zone- 6 x 15mm holes in outer can. Cheers.
|
|
|
Post by finiteparts on Jul 14, 2016 22:21:29 GMT -5
Hi Patty,
One way you can check to see if the hot spot was due to the fuel injection problem is to polish off the oxide layer and re-run it to see how it re-heat stains. Rolls Royce used to due this before thermal paints were available. They talk about using this polishing technique as a means to design the industrial version of the Olympus engine's combustor. They had the different oxide colors mapped to actual temperature ranges though.
Yes, I agree that too shallow of an annulus area would be problematic. When you size the liner holes for a certain pressure drop, you are really just trying to make sure you get a good dilution jet penetration. With a shallow liner volume, those jets probably impinge on the other side and cause axial flow blockage as opposed to cross-passage mixing.
I don't think I understand what you mean when you say that the injectors are pointing towards the inner liner. Are they pointing at a slight angle or are they pointing radially inward? And when you say gas are you meaning propane or vaporized fuel (gas flow)? What does your fuel injector look like?
Good luck!
Chris
|
|
|
Post by madpatty on Jul 14, 2016 22:30:14 GMT -5
Hi Chris,
The injectors are single hole radial injectors.
6 of them covering the annulus with 1.5mm hole each.
Each hole is pointing radially towards the inner liner in tge primary zone.
Cheers.
|
|
|
Post by racket on Jul 15, 2016 1:15:22 GMT -5
Hi Patty
Something isn't making sense here ..............LOL, you've suddenly jumped sideways on me :-(
With a very narrow 22 mm annulus width you'd need multiple fuel injection points , probably each no further apart than 25 mm from its neighbouring injector so as to get some sort of overlapping, an annular flametube can be imagined as multiple tubular flametubes the annulus width , its one of the reasons why I run 18 evaporators .
Cheers John
|
|
|
Post by madpatty on Aug 12, 2016 0:09:32 GMT -5
Hi Experts,
I was wondering if there is any whirl component in the flow exiting the compressor scroll of the turbocharger.
If there is, then in which direction the air wants to rotate.
Cheers.
|
|
|
Post by racket on Aug 12, 2016 3:54:02 GMT -5
Hi Patty
There can be considerable "helical" flow in the scroll , different scroll designs can have different flow characteristics, I'd need to see the scroll design to comment of what might happen
Cheers John
|
|
|
Post by madpatty on Feb 28, 2019 11:45:13 GMT -5
Hello All, Hadn't posted in this thread for quite a while now. But much has been going on in the background. I have been experimenting with different turbojet configurations in last few years. In my last creation I was able to get the engine upto 4 PR with decent TOTs of ~650 degrees. Few pics of the latest engine- I had to do all sorts of arrangements for equal fuel delivery to all fuel injectors and I had to use external fuel supply so I came up with this Common rail sort of an arrangement. Space wasn't a constraint at least for testing purposes and once engine is running fine the fuel manifold can always be re-made with a sleeker cleaner design. Now the problems- While testing and close to full power, that is, at 4 PR the TOTs were about 650-660 degrees C. But looking at the videos the turbine color seems like turbine is running at close to 800+ degrees C which thermocouples are not registering. I am using 2 thermocouples at 12'o clock and 6'0 clock position. Thermocouples are best quality K-type from Omega rated up till 1200+ degrees C and thermometers are from Fluke so it's a pretty reliable setup. Also both the thermocouples were within 18 degrees of each other. The image below maybe too dark to see but one thermocouple is reading 640 degrees and other is 658 degrees C. I also tested the thermocouples by heating them together and registering the temperatures on the fluke thermometers. The tip colors and temperatures seem to agree exactly. But I am concerned about the turbine color at full power. It maybe because it was pitch black outside. Can you guys give any expert opinion on this or share what your turbine color looks like at full power and temperatures(if anybody has ever tried to look)? Thanks.
|
|
CH3NO2
Senior Member
Joined: March 2017
Posts: 455
|
Post by CH3NO2 on Mar 1, 2019 13:36:02 GMT -5
Hi Patty,
That's a good question. It seems like you have quality thermocouples giving you good numbers if they are in agreement when tested over a burner. From that, I would tend to trust the thermocouples. But yes, if the turbine is glowing orange like shown in the picture... it looks hotter than 660C.
I can't speculate on how it would work, but, could it be that both observations are correct at the same time? Maybe some other factor is at work between the turbine and thermocouples?
A big change in pressure at the turbine's exducer?... Thermocouples being heat sinked at their coupler?... Dilution?... Air getting sucked into a leak path at the thermocouple's sleeve?... Without a nozzle to provide back pressure it may be possible. IDK. Just throwing out some speculation.
BTW, that's an awesome build.
Tony
|
|