dieselguy86
Veteran Member
Joined: September 2014
Posts: 186
|
Post by dieselguy86 on Dec 18, 2022 21:26:57 GMT -5
Looking great as ever John, I think they'll really make a difference. Looking forward to the new numbers.
Testing that monster is probably getting expensive by now lol
-Joe
|
|
|
Post by racket on Dec 19, 2022 0:10:38 GMT -5
Hi Joe
She's coming together :-)
I'll silicone it into the bellmouth tomorrow , just waiting on some paint to dry on the steel tube centre section where I welded things .
The industrial centrifugal air compressors use variable ones on their machines , so did a bit of copying .
LOL.......it is a tad thirsty , so shorter runs in order , hopefully the fuel burn rate will improve once the comp effic rises .
Cheers John
|
|
|
Post by racket on Dec 19, 2022 16:47:30 GMT -5
Hi Guys A bit more progress, just need to wait until the first bead of silicone to set before doing the other side , but looking sorta what I wanted Cheers John
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Dec 20, 2022 10:31:31 GMT -5
John,
Brilliant! much better than just a restriction. If they were variable you could tune the comp work to match the turbine. One of the WW2 aircraft used a variable inlet vane to fine tune the compressor behavior for altitude. I think it was a German aircraft, but can't recall.
Monty
|
|
|
Post by racket on Dec 20, 2022 15:37:32 GMT -5
Hi Monty
LOL........variable vanes might be just a tad too difficult for me to make :-(
But I did think about it after looking at the ones on some of the industrial air compressors where they're used to reduce flows when process air demands are low .
Currently my inducer tip angle of attack is up to ~14 degrees to the inflowing air , now even an aircraft wing doesn't like that high an angle , so I've made the IGV to hopefully bring it back to somewhere between 0 and 5 degrees, its a bit hard to know just how effective the vane turning will be as the liturature talks about the airflow not following the variable vanes all that accurately .
LOL...........but its gotta be better than what I've had :-)
I'm pretty certain the wheel is a "copy" of one of the newer GEN11 type wheels which are flowing at ~15 lbs.min/sq in of inducer rather than the "normal" ~12 lbs , so potentially ~5 lbs/sec , a lotta air ;-)
I'll remove the IGV from the jig today , the silicone should be solid enough by now , it'll be interesting to see how it looks against the wheel
Cheers John
|
|
|
Post by finiteparts on Dec 20, 2022 20:46:32 GMT -5
Hey John, Sorry to see you're having trouble with this new impeller. I know you will work through it. I was trying to think up a way to estimate the airflow through the engine and thus allow us to work out the vectors on the inducer, etc., but it is a bit of a challenge to get correct data at the inlet. I dd think that maybe a wall static pressure could get you a rough idea of the local inlet flow speed and then maybe we could assume the inlet plane is roughly similar to the local static/Mach number. But, in trying to figure out your rpm's from the video and I was able to get some good numbers, so I thought that I would share the info. As I have done in another post, I used a Spectrum Analyser app on my phone to back out the RPMs from the blade passing frequency (BPF) and it works really well. I was initially concerned that the splitters would add to the noise, but I get very distinct peaks that correspond to the main blade count (7 on your current impeller). The app is named "Spectroid" and with the cursor and tweaking the settings, I can get a sharp indicators of the BPF. So I checked it against one of your videos where you show your tachometer measurement and it looks like you get 36,900 something, and I get 37165 rpm, which is likely due to the limited number of "bins" that the FFT uses. But very close in my opinion, when you are doing it post-test, from a video which is shot on a microphone that could have it's own response factors. Now, I did have issues with the videos that are shot looking at the gauge cluster, which I think may be related to the amount of reflected noise..maybe. But if you use the videos that are shot by your "mobile" camera it provides a real clear line showing the BPF peak. I have a waterfall plotted and you can watch the BPF peak move smoothly with the speed changes. So, you can have a good speed measurement after the test without having to use you tach multiple times. With this, I checked your old compressor from your 6/9/2022 test and it appears that to get to 3 bar on your gauge, you hit 64,688 rpm (7547Hz), which has you at a tip speed of 1945 ft/s (assuming a 175 mm tip), which is pretty sporty but not to the 2100 ft/s that Garrett goes up to on it's forge/milled 7075 in the F1 applications...but we all know that it was not the same quality as a Garrett forged/milled wheel. Then I checked your 12/16/2022 test and I estimated that you hit the 3 bar on the gauge at 58,662 rpm (6844 Hz), at a tip speed of 1763 ft/s which is more in line with the general recommendations for 2618 T-6 Al...so with 7075 you should still have some headroom. You can see a screenshot of the BPF line in the waterfall plot on the bottom and the cursor lined up to it at 6844 Hz. I am working through a NASA paper on a flowmeter design that could get us within a few percent accuracy on the mass flow, which I think can be made from a fiberglass inlet section with static pressure taps, but I am still working through the details and also on you current setup, it could be challenging to install. A few thought on the results above...it makes sense that the newer impeller gets up to the 3 bar pressure at a lower rpm as opposed to the older one, since it has less backsweep. The increased backsweep in the older impeller means that it does more diffusion in the impeller and thus"unloads" the diffuser. Thus the newer impeller will work the diffuser a bit harder and thus be less forgiving of non-ideal diffuser angles, areas, etc.. Thus, my thought is that the diffuser might be causing your lower compressor efficiency. If you go with the logic in the Casey and Rusch paper on matching impellers to vaned diffusers, it would suggest that if you get to a higher pressure at a lower tip speed (or Mach number which is what they used in the paper), then you would INCREASE the diffuser throat area...but, by necessity you had to reduce the throat area...so I think you might be limiting the through-flow because of the diffuser. Also, since you impeller throat area increased, that would require an increase in diffuser throat area to just keep it at the same critical area ratio as before, without even trying to shift it up to match at a lower tip speed. So it may doubly impacted by the new impeller. As for the diffuser inlet angle, it is hard to say without knowing the mass flow. You might try to use oil or ink to get some traces of the airflow in the diffuser to "see" if you may be experiencing any large separation zones. Since the kinetic energy in the flow coming out of the impeller can account for more than 50% of the energy in the compressor system, if the diffuser doesn't recovery it well, it can really impact the overall compressor efficiency...and on top of that, vaned and wedge diffuses offer more potential for losses than vaneless diffusers, so they are notorious for being challenging at higher PRs. The other thing that I did want to mention, is that the IGVs will reduce the mass flow, but it will also reduce the pressure ratio of the stage, due to the reduction in absolute swirl change across the impeller. I was trying to figure out how to calculate this, but again, without the mass flow, it is tough to get a good answer. My guess is that you will have it tested quicker than trying to calculate things out and thus the real world physics will tell us the answer! ha! Good luck! Chris
|
|
|
Post by racket on Dec 20, 2022 21:25:04 GMT -5
Hi Chris LOL.............she's back on the engine and loaded onto the trailer ready for a test this afternoon hopefully :-) I'll keep you posted on results . With such a "high" inducer tip angle of 34 degrees ( 56 from axial) , to get any sort of reasonable alignment between blade and airflow the inlet airspeed is going to be rather high with a tip speed of 1300 ft/sec, tan 34 degrees is 0.6745, 1300 X 0.6745 = 876 OH!!!! ..............she needs IGVs :-) A few more pics Cheers John
|
|
|
Post by racket on Dec 20, 2022 23:50:43 GMT -5
Hi Chris Todays testrun youtu.be/1FH7AMVG4HAStrange T2 readings as well as the surge slot reading Cheers John
|
|
dieselguy86
Veteran Member
Joined: September 2014
Posts: 186
|
Post by dieselguy86 on Dec 21, 2022 0:18:52 GMT -5
Surge slot reading looked pretty rock solid to me. Your previous test I could watch the needle move, this one didnt seem to move much at all. P2 looked "bouncier"? T2, when comparing it to your Dec 16th test, looks like it dropped 80c?
|
|
|
Post by racket on Dec 21, 2022 0:19:57 GMT -5
and another youtu.be/7LPQO5O7BAMInteresting numbers coming out , I think I need to tighten up the jetnozzle . At 3.5 Bar P2 it appears as though there little air exiting the surge slot as the pressure reading is dropping back to non running position , whilst at lower power settings the surge slot is "working", I need a much better/finer gauge on it. TOTs are "cooler" , but those T2 readings on the red thermo are very strange , they were up at 240 C last test at only 3.2 Bar whereas they're 60 degrees cooler this time at a higher P2 . Fuel pressure wasn't as high this time either even though I was at 3.5 Bar whereas last test the pump was flat out at 3.2 Bar . LOL............there were differences , just need to fathom out whats going on :-( Cheers John
|
|
|
Post by racket on Dec 21, 2022 0:24:13 GMT -5
Hi Joe
I swapped the P2 pickup from behind the diffuser to the P2/T2 pickup in a diffuser channel outlet subjected to higher velocity flows , an earlier test with the P2 there produced a very bouncy reading, its better this time .
Yep , those T2 readings have me baffled .
Cheers John
|
|
dieselguy86
Veteran Member
Joined: September 2014
Posts: 186
|
Post by dieselguy86 on Dec 21, 2022 8:47:30 GMT -5
Hi John,
Sounds as tho you've "unloaded" the rotor? Less lbs/sec of flow so the turbine isn't needing to make the extra power + extra fuel? If t2 is accurate then it was running at 76% efficiency at 3 bar. That's up ~10% from before right? It sounded to me like it was idling at higher rpms also.
|
|
|
Post by finiteparts on Dec 21, 2022 13:13:40 GMT -5
Congrats John on the good results...even though they need a bit of unpacking to understand.
What I back-calculated for the rpm is as follows:
P2= 2 bar, 5997 Hz, 51422 rpm = 3 bar, 6891 Hz, 59065 rpm = 3.5 bar, 7184 Hz, 61577 rpm
Which is interesting because these are lower rotational speeds than the previous test without the IGVs. First question is what were the ambient conditions for these two tests? A cooler, lower humidity day may allow it to operate at a similar corrected flow...so we need to get the corrected conditions to accurately compare them.
I had a question on the data when I was trying to calculate the stage efficiency. The T2 is reading total? and the P2 is reading static? I was trying to calculate the total to total, but I think the data may be giving you a total to static efficiency-ish...sort of mixed data and may be tough to get a accurate number....but still we be good to get a comparison to other tests and a good indicator to how the performance is changing.
Glad to see that it worked so well.
- Chris
|
|
praendy2203
Veteran Member
Joined: August 2020
Posts: 160
|
Post by praendy2203 on Dec 21, 2022 16:06:19 GMT -5
Hi John,
looks and sounds very good. All values in limits now😊
Incredible, what a too good wheel can do.
Andy
|
|
|
Post by racket on Dec 21, 2022 16:20:23 GMT -5
Hi Chris
P2 and T2 ( top reading on the red thermo) are both reading from the same position and should be "total", mounted in a "pitot" at the outlet of a diffuser channel ., only problem is the T2 reading , at 188.5 C at 3.5 Bar P2 on a 22 deg (295 K) ambient day ( same as last testrun) gives a comp effic of 96% which is impossible , something has gone wrong with the red thermometer :-( it was reading much higher during the last testrun.
The T2 minus T1 thermo is also a bit sus I feel as there could be abberant air temps at the sensor with hot air off the flametube close by as well as the T1 sensor having sunshine on it , not easy getting good data "in the field" .
I'm going to setup a water manometer on the teststand to measure the surge slot exhaust pressures more accurately , from the readings on the gauge as well as my "flapping string" it appears as though at 3.5 Bar theres little air exiting the surge slot , if this is the case then my IGVs are "restricting??" flow at the high power setting , on turbos used on IC engines if the turbo is a bit undersized then air gets sucked in through the surge slot at full throttle .
The test has thrown up some curly questions , but also some "results" , I'll have to have a think on the thrust numbers today, they appear to be down on last test , but with a choked turb exducer and possibly uncontrolled expansion downstream because of an oversized jet nozzle area , theres room for improvements :-)
This "race horse" of a comp wheel is testing me .
Cheers John
|
|