monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Nov 19, 2018 9:32:35 GMT -5
Assuming there is no flow separation or excessive flow gradients within the compressor wheel, a larger compressor exducer area (tip heigth) yields diffusion inside the wheel flow path producing a lower exit velocity. This would mean a smaller stationary diffuser diameter could be used. At least this is my understanding of it so far. Tony EXACTLY!! Very different from what the turbocharger folks are after. As far as the other stuff goes, I'm just looking at the cycle design with a constant TIT, no component detail design yet. All of this is hampered by the requirement to use an off the shelf turbine of course.
|
|
|
Post by finiteparts on Nov 22, 2018 19:55:20 GMT -5
Hi Chris, If I'm reading the Swain paper correctly, it seems the purpose of axial trimming is to reduce the pressure rise without reducing the flow range. And this is only if done within a limit. I'm only beginning to understand this science so I may be missing the point of axial trim (other than to suppress shock formation at the inducer leading edge). The CFD test results show axial trimming can quickly be counterproductive to efficiency, mass flow and pressure ratio. On the subject of axial trimming Swain states on page 71 "the purpose is to reduce the pressure rise without reducing the flow range." If the purpose of axial trim is to reduce pressure ratio, why would anybody in propulsion want to do it? Thanks, Tony
I think the big picture that you should look at is that Swain starts out with well designed impellers that then get trimmed...thus the reduction in performance. I do tend to agree with you though. I think in general that we would probably not want to axially trim a good impeller that we had a map for....but...the aftermarket guys don't seem to do any real work on estimating performance or testing in any way. Cloned impellers may or may not be correctly sized...they may be enlarged versions of a OEM impeller and as such might need to be modified. Additionally, of you wanted to design for a higher pressure ratio at the design point, you might need to axially trim the impeller to move the peak efficiency island up...but this should only be done with a good set of calculations estimating the change in the output vectors to make sure that you are actually correcting the impeller to the head coefficient design point.
- Chris
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Nov 22, 2018 21:27:28 GMT -5
Hi Chris, If I'm reading the Swain paper correctly, it seems the purpose of axial trimming is to reduce the pressure rise without reducing the flow range. And this is only if done within a limit. I'm only beginning to understand this science so I may be missing the point of axial trim (other than to suppress shock formation at the inducer leading edge). The CFD test results show axial trimming can quickly be counterproductive to efficiency, mass flow and pressure ratio. On the subject of axial trimming Swain states on page 71 "the purpose is to reduce the pressure rise without reducing the flow range." If the purpose of axial trim is to reduce pressure ratio, why would anybody in propulsion want to do it? Thanks, Tony
I think the big picture that you should look at is that Swain starts out with well designed impellers that then get trimmed...thus the reduction in performance. I do tend to agree with you though. I think in general that we would probably not want to axially trim a good impeller that we had a map for....but...the aftermarket guys don't seem to do any real work on estimating performance or testing in any way. Cloned impellers may or may not be correctly sized...they may be enlarged versions of a OEM impeller and as such might need to be modified. Additionally, of you wanted to design for a higher pressure ratio at the design point, you might need to axially trim the impeller to move the peak efficiency island up...but this should only be done with a good set of calculations estimating the change in the output vectors to make sure that you are actually correcting the impeller to the head coefficient design point.
- Chris
I'm not so concerned about axially trimming the wheel as I am with changing the trim of the inducer/exducer. Balancing the in diffusion in the wheel vs stationary diffuser for the application. I'm probably going down a rabbit hole....
Monty
|
|
|
Post by finiteparts on Nov 22, 2018 22:22:35 GMT -5
Haven't had time to digest the paper yet....but basically to match the compressor to the turbine/cycle. In my instance any additional PR over about 3.5 only results in robbing power from the fan, and eventually chokes the turbine exducer or the core nozzle. The core nozzle must match the pressure from the fan in the mixer, so that limits how high the core PR can be since I only have 1 turbine stage. Is this fuel efficient....NO!! but it does not require additional turbines, and concentric shafts...which are expensive and heavy.
I want to optimize the diffusion in the compressor and shrink the stationary portion of the diffuser as much as possible. Size works against me. Everything is a compromise.
Monty,
I am a bit confused about the comments on your core PR limit, which seem counter to the general design process for actual turbofans.
Since it is the static pressures that must match at the mixing plane between the core and the fan discharge, they are set by the velocities of each stream, i.e. the nozzle areas. Are you suggesting that having the core nozzle and/or the turbine choked is some form of limit? Because operating above the initial choke point is commonplace...in fact it is usually ideal because the engine operation then becomes a linear function of the engine temperature ratio. Jack Mattingly's book does a great job describing the engine op-line with the assumption of both the turbine nozzle and core nozzle choked. Just curious...
Also a higher pressure ratio in the core allows a larger mass flow for a given flow area. Because of the divergence of the constant pressure lines (isobars) on the T-s diagram, as you go to a higher system pressure you will have more available energy entering the turbine to drive the fan. The enthalpy also increases non-linearly meaning that you have a higher gas energy density to drive the turbine for a given flow area.
Just curious...
-Chris
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Nov 23, 2018 11:41:37 GMT -5
Chris,
No doubt you are correct, but the total work has to be produced by the turbine, and it can only handle a fixed temperature ratio. If I was designing a turbine to fit the cycle per the usual method, higher core PR is always a good thing. When designing the cycle to match the existing turbine.....I don't have that luxury. Max TIT is fixed, and total turbine stage deltaT is fixed. Therefore the work available must be divided between fan and compressor. At some point higher core PR takes away from the work available to the fan. It may help specific fuel consumption slightly, but not much. This is why no sane person makes a single shaft fan engine. It really is just a crappy single shaft turbo-prop.
Monty
|
|
|
Post by finiteparts on Dec 16, 2018 17:16:53 GMT -5
Hi Monty, I guess my day to day work on full scale, commercial engines had me quite disconnected from these low temperature, low efficiency turbomachines...so I did some calculations to help me better understand how these machines are impacted by component efficiencies and temperature limits. The 3.5 Compressor PR limit that you are seeing seemed a bit low based on what I was seeing in the literature for single shaft, simple cycle turbines...so I want to better understand the how and why of "optimal compressor pressure ratios". I built a quick model that evaluates the specific power of the compressor and the turbine, then combines them with mechanical losses and an assumed engine flow rate to determine the net power output. The assumed adiabatic efficiencies initially specified for each component are held constant over a range of compressor pressure ratios and the end result is a plot showing how the spec work and net power output vary as a function of compressor PR. I verified my model based on the work of several papers...a good example is: "Approximately Calculating Optimum Pressure Ratios for Various Gas Turbine Cycles", Ushiyama, I.J. and Matsumoto, N., ASME 76-GT-21, ASME Gas Turbine and Fluids Engineering Conference, Dec. 1975 So with the model coded up, I started to play around with the variables to see what the predicted optimal pressure ratio would be relative to what you are seeing. For the first pass, I chose what I would define as the upper end, but still achievable parameter values... - Compressor Efficiency = 79% - Turbine Efficiency = 82% - Mechanical Efficiency = 96% - Turbine inlet temperature (Tt4) = 1800 F*
* --- (for the turbine modeled (5 inch inlet, 60 krpm) this gives a relative total temperature of 1600 F, matching the metal temp limits of a standard INCO713LC cast rotor)
Also, all the plots assume 5% dP/p combustor loss and a 1.05 exhaust nozzle pressure ratio, which could probably be made smaller, but for a homemade engine this seemed like a safe assumption point.
The plot shows the compressor specific work as a blue line, the turbine specific work as a red line and the green line is the net power out based on an assumed engine mass flow rate of 2.0 lbm/s (engine mass flow required to get approximately 150 hp net). The engine temperature ratio is 4.36 with T4 set to 1800F (for an ISO standard day) and as can be seen, the optimal compressor pressure ratio was around 5.77.
I set the mechanical efficiency low ( ~ 4% loss) due to the fact that you are suggesting using a gear reduction to the fan and thus the mechanical loss should be high relative to a standard turbojet configuration...to get an understanding of the magnitude of this loss, at a PR = 5, the mechanical power loss is around 20 hp. This may or may not be relatistic, since most of the turbine efficiency numbers that we see actually have the mechanical efficiency already baked into them...so we may be slightly double dipping here, but I thought it at least "looked" reasonable. I also am assuming that your Overall Compression ratio (OCR) is what is being shown as the compressor pressure ratio in the plots.
Now, when I change the parameters to lower values closer to standard turbocharger values, we move nearer to your 3.5 prediction. For the second run, I chose these parameters...
- Compressor Efficiency = 74% - Turbine Efficiency = 80% - Mechanical Efficiency = 96% - Turbine inlet temperature (Tt4) = 1600 F* (yields Ttrel4 = 1452 F)
These reduced parameters move the optimal pressure ratio down to 4.08, but the big thing to notice is that assuming the mass flow doesn't change from the previously stated 2.0 lbm/s, this configuration doesn't have enough net power output to meet your fans 150 hp need...
and you would be required to increase the engine mass flow to around 3.1 lbm/s as shown below...
So after working through the calculations, it can be seen that increasing the efficiency and/or the engine temperature ratio (ETR = Tt4/Tt1) helps to push the optimal compressor pressure ratio to higher levels. The compressor specific work consumption begins to increase faster than the turbines specific work production, so that when you subtract the compressor spec work from the turbine spec work, it peaks at some PR and then begins to consume a larger proportion of the turbines output capability.
What kind of assumptions were you assuming?
Thanks,
Chris
|
|
CH3NO2
Senior Member
Joined: March 2017
Posts: 455
|
Post by CH3NO2 on Dec 17, 2018 0:06:49 GMT -5
Great analysis Chris!
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Dec 17, 2018 21:39:27 GMT -5
Chris,
Your numbers are very close to mine. Around 4 total PR, roughly 80% efficiencies, and 1650F TIT with ~3 lbs/s. All guestimates of course.
I would like to have 1800 TIT!!.....then I could increase the core PR or power output. I'm not sure I really trust my aftermarket wheel at that sort of TIT though....and if I increase the core PR much more, I'll need a titanium compressor wheel....so there's that i$$ue too.
My design lends itself to adding a recuperator at some point, to get fuel burn down, but the thrust would suffer. Currently the core produces about 1/4 of the thrust.....but you gotta start somewhere.
I will put thermocouples and pressure tranducers all over the thing. I've already got a test stand I can mount the engine on that will give me thrust and HP.
Just getting it to work at all is challenge enough!
Monty
|
|
|
Post by racket on Dec 17, 2018 22:39:43 GMT -5
|
|
|
Post by finiteparts on Dec 17, 2018 23:17:20 GMT -5
Monty, The typical quoted continuous gas temperature limit for cast Inconel 713LC is 950 C (1742 F)..if you want a reference on that have a look at Baines "Fundamentals of Turbocharging" or pull the paper "Exhaust Gas Temperature 1050 C" from Borg Warner's Knowledge library here: www.turbos.bwauto.com/en/press/knowledgeLibrary.aspxYou will see that once you get up near the 1800 F region, the creep rupture and tensile strength of 713LC plummet and you need to move to Mar-M-247 or the like. But, you also have to recall that the rotor "sees" the gas temperature in a rotating frame of reference, so the stationary reference frame gas temperature is higher than what the rotor sees. You have to calculate the total relative gas temperature and then do the heat transfer calculations based on the relative properties. Maybe this is an easier way to think about this...the total gas temperature is the static gas temperature PLUS the temperature that would result from bringing the gas to rest isentropically. If we were on the rotor, thus seeing that gas flow from the relative frame of reference, the gas velocity coming at the rotor is lower due to the rotational velocity of the rotor...so by default, the temperature that would result from bringing the gas to "rest" relative to the rotor, in the rotational frame of reference, would be lower. Is that clear? We usually just throw out the 1600 F number to keep a buffer on turbine entry temperatures for people designing engines that will have little control fidelity or instrumentation to know how close they are to the T4 limit...this gives a 150 F margin to the suggested steady state operating temperature for 713LC, so if their is a hot streak or just poor control of the fuel, hopefully they don't instantly experience high nickel emissions. The equation for the relative total temp (Ttrel) is: --> Ttrel = T + W^2/2Cp where W is the relative velocity and T is the local static temperature If you know the total temperature in the stationary reference frame, then Ttrel is: --> Ttrel = Tt + (W^2 - V^2)/2Cp where V is the local absolute velocity. Most books cover this, but I will recommend Erian Baskharone's book "Principles of Turbomachinery in Air-Breathing Engines", Cambridge Univ. Press, 2006 as an excellent resource. He worked at Allied Signal/Honeywell and brings a wealth of knowledge to the book that others seem to lack. He also does work through quite a few examples that help to clarify the material. I hope that helps, Chris
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Dec 18, 2018 0:22:36 GMT -5
Chris,
I understand the T0 situation, and I'm experienced enough to want that 150F margin....
Monty
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Dec 18, 2018 9:48:36 GMT -5
John,
That is a fantastic paper. It does illustrate the farther you push the limits, the more challenging everything becomes. That turbine efficiency graph made me very happy!
My goal with this engine is to get something to work first! (a big enough problem for one guy in a garage!!) Then it can be refined and improved upon. Higher pressure ratio would be nice....but the turbine has to be capable of powering everything and dealing with higher TIT/PR. The gearbox has to be able to take the additional rpm. The compressor wheel needs to be either cooled or made of better more expensive stuff. There's only so many technical bears I'm willing to poke at any given time...
I gotta walk before I can run....right now I'm just flailing away on my back, whining and crying with a soiled diaper
Monty
|
|
|
Post by racket on Dec 18, 2018 15:17:16 GMT -5
Hi Monty
LOL..........we'll soon have you walking ;-)
One other thing you could try to help with obtaining higher T I Ts and which you'll need to implement is some bleed air to your turbine wheel hub area , I use 12 X 2.5mm holes feeding air into the space behind the turb wheel to prevent hot gases getting anywhere near the shaft , that air also must flow out over the hub providing some "blanket" cooling effects .
Cheers John
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Dec 18, 2018 19:39:57 GMT -5
Hi Monty LOL..........we'll soon have you walking ;-) One other thing you could try to help with obtaining higher T I Ts and which you'll need to implement is some bleed air to your turbine wheel hub area , I use 12 X 2.5mm holes feeding air into the space behind the turb wheel to prevent hot gases getting anywhere near the shaft , that air also must flow out over the hub providing some "blanket" cooling effects . Cheers John John,
I'm currently planning to have a labyrinth seal on the back side of the compressor, and bleed whatever air gets past that to the turbine base area. That way I get free bleed air, and cut down on the thrust load the front bearing has to deal with.
Monty
|
|
|
Post by racket on Dec 19, 2018 0:39:26 GMT -5
Hi Monty
Laby seals are OK , but you won't be able to use it and have the bleed go to the turb area.
Generally the static pressure behind the comp is roughly half your total pressure rise in the comp , lets assume 20 psi static and 20 psi dynamic velocity exiting the comp , now at the turbine end you'll have a similar division of pressure drops , so any pressure inboard of your laby seal will need to be at a higher pressure than the "intermediate" static pressure at the turb end for the air to flow rearwards .............it probably won't happen .
You might end up with a situation like I had with my FM-1 engine where I had hot gases going back into the shaft tunnel despite the shaft tunnel being fed with air from within the outer can .
It wasn't until I had "snokels" fitted to the shaft tunnel air supply ports that took air from the diffuser outlet at max static pressure that it cured the problem , I was attempting to use the "standard" RC micro engine type lubrication/cooling of the ball races .
I had a feeling that the gases with a radial inflow turbine wheel "recompress" behind the turb wheel , unlike an axial wheel where the gases are slipping past and tend to suck gases out from behind the wheel.
If you do go for a laby seal to unload the comp, you'll need to dump the bypass to atmosphere
Cheers John
|
|