|
Post by racket on Jul 9, 2020 22:51:06 GMT -5
Hi Chris Yep , it gets kinda hard to produce enough jetpipe pressure from our DIY turbo based engines, though on 19/12/2000 I did see a total pressure of 15 psi with my TV84 engine but only with very high temperatures in the jetpipe and a 244deg C comp discharge temperature at a P2 of 38 psi , a combination of an oversized scroll housing and very high temps , thrust was good at >110 lbs through a ~82 mm dia jet nozzle which was some 6 mm greater than "normal" . My recent test of the 12/118 engine was getting some good jetpipe total pressures ibb.co/j81Y69r , 0.86 Bar - 12.6 psit at a 3.5 PR and "acceptable" jetpipe temps , if I hadn't started running out of fuel and been able to get the readings at a 3.8 PR I probably would have been running a choke jetnozzle. Hopefully with a bit more "refinement" I might just get there next time :-) Love the graphs Cheers John
|
|
|
Post by finiteparts on Aug 12, 2020 18:38:31 GMT -5
One of the challenges in designing a turbocharger based gas turbine has been the selection of a good turbocharger. What defines a good turbocharger for a GT application is how well the compressor and turbine will be matched at an operational point that is essentially a very off design condition for a turbocharger designed to be used on a IC engine. Even when we have access to the compressor and turbine maps from the manufacturer, there are a large amount of unknown variables that go into finding the operating line. In an attempt to simplifying this process, I am working on a method that creates a operating line on the compressor map that is "morphed" till the turbine map properties overlay the turbine map. In order to generate the turbine exit parameters, the compressor work is calculated, added to the mechanical work absorbed and then balanced in the turbine in a similar fashion to most hand calculations that you see in the textbooks. As you can see here, I have generated a steady state operating line on the compressor map of a GT6041. I have recently picked up a nice GT6041 and I am exploring how I might use it in the future. As you might have noticed, the operating line does not run across the peak efficiency islands and sits toward the surge side of the map. Obviously, this is not ideal. The good news for me is that I have a different A/R compressor housing with the same compressor wheel. My smaller A/R housing should move the peak efficiency island up and to the left slightly, putting my turbo set-up in a more favorable condition. The matched turbine line is shown here: It should be mentioned that this operational behavior is predicted for a fully opened turbine exit. As soon as a convergent nozzle is installed, the operating line will shift further to the left, potentially too close to safely operate. Even if the operating line was to have enough margin to the surge line for steady state operation, there could be problems when you try to accelerate with the engine entering surge. I am working to correct the calculations to solve for the static pressure after the turbine (Ps5) so that the operating line can be overlaid on the compressor map with the nozzle installed. Once I work through the process, I will share more. Also, if you cannot tell, the process is to generate the curve in an Excel worksheet with the axis scaled the same as the map and then overlay the compressor or turbine map. This way it makes it somewhat quick and easy for most people. - Chris
|
|
|
Post by finiteparts on Aug 12, 2020 18:55:02 GMT -5
Oh, I also wanted to mention that this run shows the operation over a turbine inlet temperature (Tt4) range of 1150 deg F (at PR just above 1.01) to 1650 deg F (at a compressor pressure ratio of 4.5). This is essentially a sweep over a "throttle parameter" (Tt4/Tt0) range of 3.1 to 4.021.
- Chris
|
|
|
Post by finiteparts on Aug 12, 2020 22:11:27 GMT -5
Quick post to show the impact of having a nozzle verses an open exhaust. I was able to get these two cases run, but it was not a very efficient process. I am going to work on streamlining the worksheet so that I can change the nozzle throat area and see the impact between an open exhaust curve and the one with the nozzle. Here is the first pass at this: It can be easily seen that the largest impact occurs at the mid-section of the curve. This is due to the lower left point being bound by the fact that at no speed, it is always a PR =1. At the upper right side, it is bound by the fixed pressure ratio and max Tt4 (=1650 F) that I am imposing. Additionally, I will mention that to get the nozzle case to converge on the turbine map curve, I had to increase the min Tt4 as compared to the previous fully open exhaust run. Since I had no idea of how far the curves might diverge due to the back-pressure presented by the nozzle, I think this is pretty interesting. I have to admit, that my notional expectations were that the difference between the lines would have been much farther. Once I get this more flushed out, maybe I can use this to create pseudo-transients such as acceleration and deceleration curves by assuming a steady state change in temperature that might approximate the energy push needed to create the unbalanced torque required for the acceleration. Always something new to add! My personal opinion is that the curves are probably not very accurate at the lower pressure ratios, because the lack of speed lines keeps me from fully matching the turbine and compressor operating points. BUT, I do think this is a very good approximation for our level of design, especially if the design point that you are shooting for is near the top of the map, which I am. I will post more on the process as I get it flushed out more fully. -Chris
|
|
CH3NO2
Senior Member
Joined: March 2017
Posts: 455
|
Post by CH3NO2 on Aug 23, 2020 20:45:23 GMT -5
Hi Chris,
The GTX5533R 98mm compressor wheel would make a nice matchup to the GT6041 turbine. It would be RPM limited to a 4.5 PR but otherwise would perform much better.
It's a shame Garrett doesn't make the 6041 turbine anymore.
Tony
|
|
|
Post by racket on Aug 23, 2020 22:54:40 GMT -5
Hi Tony
Theres still the industrial GT6041 wheels being made for the CAT 3512B engine , E and E turbo have them for ~$US230 , 130 mm inducer with 113 mm exducer ~22 mm tip height on the inducer , E and E # TW-0226
Cheers John
|
|
|
Post by finiteparts on Sept 4, 2020 14:12:46 GMT -5
John posted an ebay link (by the way, thanks John for posting those!) to a pair of ABB turbos for another reader. When I looked at them, I was very interested. I gave it some thought over the course of a week or so, and finally shot the seller an offer. I was half expecting them to not take it, so when they accepted, I was excited. I wasn't sure about the size of the turbos from the photos and I was hoping that the serial number was giving a hint into the size, which unfortunately it wasn't. The S/N was HT57..., so I was hoping that they were TPS-57 turbos, but when I got them, it turns out that they are TPS-44s. Still, awesome turbochargers for the money, especially, when the list price of one turbocharger from Caterpillar is over $27000 US and I got 2 for less than $850. I took the compressor housing off of the first one and snapped some pictures to share. Here is the TPS-44 compare to my GT6041. Interestingly, both can be found on the Cat G3500's engines. The ABB turbos are off of a G3156B engine. I will post more later. Enjoy, Chris
|
|
|
Post by racket on Sept 4, 2020 17:50:25 GMT -5
Hi Chris
Great buy , I'll be interested in more pics and some wheel sizes .
They've gone to a good home ;-)
Cheers John
|
|
|
Post by finiteparts on Oct 3, 2020 20:00:43 GMT -5
In my ever increasing collection of all things turbomachinery related, I got an amazing score! I picked up a ABB TPS57 turbine wheel for cheaper than it costs to get a standard sized turbocharger turbine. Now, this does not look like an ABB part, but more likely an aftermarket copy...but still an awesome piece to have in the collection. It is f'ing huge! Just think about the amount of inertia this thing would have spinning at 41,000 rpm! It is heavy and very hard to pick up with one hand. Enjoy! Chris
|
|
|
Post by finiteparts on Nov 29, 2020 15:09:31 GMT -5
It is common that the topic of how we can increase performance comes up on this forum and I thought that it might be good to "illustrate" the design space that we have when using turbocharger based components. Once again, I will use my evaluation copy of GasTurb to generate some very illustrative plots. I cannot say enough good things about GasTurb and I am glad that I snagged an evaluation copy before they started to add in the time-bomb that is on the new eval versions. For the first part of this, I will look at the non-afterburner turbojet and how compressor pressure ratio and turbine inlet temperature impact the generated thrust. In this first image, the yellow shaded region illustrates the notional upper-bound of off-the-shelf compressor PRs. The red shaded region show the zone where the turbine inlet temperature exceeds our typical design limit of 1650 F (actually, I have it set to 1675 F, but I didn't want to redo the plot...so...). As you can see, if we are just going for the highest thrust that we can get, then there is a peak thrust production at a PR ~ 5.75 to 6.0. This suggests that the desire to add a second stage compressor section to push to a higher pressure ratio is of marginal value. The complexity required to add the second stage is rather large and if the benefit is not that large, it might not buy it's way into the design. Secondly, we can see that there is still more core efficiency that can be achieved as we go up in compressor pressure ratio, but in achieving further efficiency, the thrust suffers. Also note, that these plots were made for a core flow of around 2.9 lbm/s, compressor efficiency ~ 76% and a turbine of ~ 82%. Changing these values as you can guess, changes the resultant peaks. Now for the second part of this, let's look at how this plays out for a free turbine arrangement. Again, we see that for our typical turbine inlet temperature of around 1650 F, there is a peak specific power achieved at a compressor PR ~ 5.75 to 6.0, and once we surpass this compressor PR, the specific power falls. In this plot, the color contours are showing the SFC and it can be seen that at the peak spec power, going higher in PR does reduce the SFC, but only marginally...definitely not worth the added complexity of a two stage compressor. Now, I do realize that not everyone will have access to a Garrett GTX turbo...so for some of you it might make it more appealing to build a two stage compressor, but you must also remember that cast compressor wheels are unable to cope with the higher temperatures that are experienced at higher PRs. I hope this was informative and helps to set expectations on what is worth trying on our turbocharger based engines.
|
|
|
Post by racket on Nov 29, 2020 15:25:45 GMT -5
Hi Chris
Nice representation :-)
What mass flow are you using in the first graph, to correspond with the projected thrust numbers ??
Cheers John
|
|
|
Post by finiteparts on Nov 29, 2020 22:11:11 GMT -5
Hi John,
I must have been updating the post while you were typing. If you look at the end of the third paragraph, you will see the relevant information.
- Chris
|
|
|
Post by racket on Nov 30, 2020 0:30:40 GMT -5
Hi Chris
Thanks for that .
Thrust numbers workout just right with the 12/118 engine for its ~3.3 lbs/sec at 3.75 PR .
I like your 170 lb "design point" for the GTX ;-)
A very nice representation of our "limitations", or more precisely, the hard facts of turbine life when we turn a turbo capable of supplying a 1,000HP IC engine into a GT and the thrust level is barely enough to move a kart at a reasonable speed.
Cheers John
|
|
slittlewing
Senior Member
Joined: November 2017
Posts: 458
|
Post by slittlewing on Nov 30, 2020 6:01:16 GMT -5
Great information and graphs, thanks for the post Chris.
|
|
|
Post by finiteparts on Dec 5, 2020 17:53:30 GMT -5
John, Actually, I wasn't using that design point to represent a Garrett GTX turbo. I had been playing with the ABB turbos in GasTurb and it was really a leftover from those runs. I was trying to say that the GTX turbos seem to have the highest compressor PRs out there, so even if you used one of them, there would be a peak..but a peak "close" enough to the theoretical peak that it wouldn't make sense to do a second compressor stage. But I do realize that the wording might have made it seem like that...and as you and I know, the challenge for many of the larger performance turbos is the hot side flow limitations. For fun, I threw in the compressor specs in for a GTX5544 with the 102 mm and checked if the turbine side could swallow the design point mass flow heated to 1600 F. Of course it can't, even with the largest A/R housing listed of 1.4. A quick check shows that the A/R = 1.40 housing can only flow 125 lbm/min at an inlet temperature of 1600 F and a inlet total pressure of 52 .4 psia. To achieve the 200 lbm/min physical flow the that the compressor is supplying at the notional design point, it would require a turbine corrected flow of 111.8 lbm/min and an A/R ~ 2.24 housing. So even for such a large off-the-shelf turbocharger, this is an unlikely case (upper limit of performance). To see the maximum thrust that we "might" be able to achieve, I had it with an afterburner...Here is the results table and the station numbering diagram from the GasTurb run. Enjoy, Chris
|
|