|
Post by finiteparts on Oct 21, 2018 15:48:56 GMT -5
Hi Monty I've designed for a flow of 3.3 lbs/sec for Anders at 5:1 PR with the X846, 3.6 lbs/sec will be getting too far to the choke region at only a 3.5 PR , efficiency would be poor along with SFC . 3.6 lbs/sec is what an Allison 250 C20 engine flows , thats a big "hole in the front" and potentially a very powerful engine . Some of the new GEN 2 Garrett turbos are producing huge flows per square inch of inlet , they're aiming for max flow rather than max efficiency ,............we can't have both . This comp flows 160 lbs/min at ~14.8 lbs/sq in/min www.garrettmotion.com/wp-content/uploads/2018/05/GTX553394mm_COMP.jpg , a reasonable efficiency at 3.5 PR , if the X846 was flowed at a similar rate it'd be ~3.4 lbssec , but you'd probably have problems fully expanding it and getting it out the turb exducer. Andy M's 10/110 engine with his X858 comp of 110 mm inducer at a lower PR than Anders has been designed for 3 lbs/sec , which just makes it through the turbine wheel, IF, theres several PSI of static pressure that will be expanded in the A/B nozzle. Your limitation is the turbine wheel exducer flow area , gases will be pretty well fully expanded down to ambient with maybe 800 ft/sec axial jetpipe velocity , you need to use that "restriction" as the basis for flow calcs . My 12/118 with the X831 of 118mm inducer has been designed for 3.6 lbs/sec at ~3.5 PR , but trying to get that much air/gas through the TV94 turb wheel is "challenging" , I really need a much larger turbine wheel. If you look at the Solar T62 engine its turbine wheel has an inducer of ~6.5" dia and an exducer of ~4.625" - 117 mm with minimal hub dia., it flows ~2.2 lbs/sec and produces ~150 HP ...........so look at your turb wheel as its going to govern the outcomes. Cheers John John, I didn't want to highjack Monty's post to talk about some of the new compressor designs, so I just quoted you and moved it here. I have been doing a little digging on the GTX and G-Series turbochargers and I have a few thoughts on the changes made to increase the flow capacities of the GTX compressors. The first thing that I noticed is that the inlet face Mach number is increased for the GTX turbos substantially. The general inlet face Mach numbers that we see on the Garrett GT-series, larger Holsets, etc is roughly between 0.44 and 0.5 at peak flows. Now when I speak of the inlet face Mach number, I am assuming only the area available to flow (i.e. inducer shroud area minus the inducer hub area). But the GTX compressors are getting inlet face Mach numbers between 0.64 and 0.71 at peak flows! So what are they doing to increase this flow range? Their stated approach is that they modified the blade shape to allow for the increased flow, which is just stating the obvious. So here are my thoughts. If we look at the inducer blade angles, it is apparent that they have decreased the angle (relative to the axial)...this is the logical first place to look, since they have all full chord blades (no splitters), the only way to increase the inducer throat area is to straighten the inducer blades somewhat to give a larger area normal to the incoming streamlines. To me this is apparent in this image... The extremely thin blades also offer lower flow disturbance to the incoming flow. They are amazingly well manufactured and the surface finish no doubt reduces the skin frictional losses. Another thought that I had is that the higher level of positional control the ball bearing cartridge offers, leads to tighter running clearances and thus lower tip losses. The other interesting feature is that they also are using the extended tips. So back to the discussions from a few months ago about the validity of these extended tips, jetandturbineowners.proboards.com/post/21393I still say it has a real and measurable effect and is not just "...marketing bullshit". When every high end turbocharger manufacturer is using it, there is more than just a marketing impact...especially when it is not cost neutral to incorporate it. My thoughts on the slightly lower efficiencies is that the higher through-flow suggests that the not only is the inlet face Mach number higher, but also to the inducer and passage velocities. These higher passage velocities generate higher boundary layer growth due to the larger velocity gradient and potentially, the discharge velocities are increased (at similar PRs, obviously they are up relative to other low PR comps when we are looking at PR's in the range of 5:1). You can see that the GTX like many of the newer turbochargers a going away from using splitter blades...this is often attributed to be due to the lower inducer noise, since the blade loading at the inducer is lower due to the presence of more blades, the pressure pulses from the difference between the pressure side and suction sides of the blades will be lower. Now I have my doubts that this is the reason that the GTX impellers are going to full blades, since turbo noise is usually a positive for the performance guys! My thought is that the full blades give better control of the inlet flow and due to how thin the blades are, the blade loading has to be below some threshold so as not to flex the blades too much and drive low cycle fatigue life issues. As usual, they claim everything is patented...but I can't seem to find the patents on any new compressor aerodynamics or other technology. It is probably just like Borg Warner saying they patented the extended tip technology that I could never find. So unfortunately no good info there. Have you seen anything else exciting about these compressors? A final neat bit of info is that some of the G-Series turbos (G25-600 maybe?) are using Mar-M for their turbines. They didn't state the exact alloy, but I suspect Mar-M-247 due to its good castability. They are saying that the turbines are rated at 1900F continuous inlet temperatures! I can't wait to see this in some of the bigger turbos. - Chris
|
|
|
Post by racket on Oct 21, 2018 18:54:09 GMT -5
Hi Chris Yep , the new Garrett wheels are works of art , getting some almost unbelievable flows per sq inch of inlet . With inlet airspeeds ~700 ft/sec and inducer tip speeds of >1,300 ft/sec the relative speed is up ~1,500 ft/sec , they must be getting some shock compression simply to be able to get the mass flow through the smaller inducer throat area . The inducer tip angle on the X846 wheel that Anders is using is set at a lowish ~24 degrees ( 66 from axial) and still has a relative Mach number of ~1.3 at design speed , the new Garrett wheels are really pushing the flows , but the old Garrett GT6041 comp with an "ancient" design is getting 80% effic whilst these new wheels are only mid 70's............ah, tradeoffs. I'm looking forward to the comp maps for the new largest Garretts "coming soon" www.turbosbytm.com/gtx5544r-gen2-106mm-inducer , Tony and I were taking about it a while back , should be "interesting" :-) Cheers John
|
|
|
Post by turboron on Oct 23, 2018 4:17:53 GMT -5
All, what about the turbines? The newer Garretts seem to have improved the scrolls several points in efficiency.
Thanks, Ron
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Oct 23, 2018 7:03:31 GMT -5
Ron,
Regarding no splitter vanes. I'm guessing the increased number of blades in the inlet is to keep the diffusion factor under control in the inducer. I run into a similar trade off in the design of a fan with solidity. Increasing the number of blades reduces the diffusion each passage must accomplish. Higher diffusion increases loss. Trying to get max diffusion and flow means trading off against efficiency. One of the other trade offs: I can increase mass flow into the fan at a given wheel speed. the angle of attack of the blades must increase and efficiency suffers, but I can fit more power in a smaller space. Modern fans and compressors are reaching diffusion factors around .6. I try to stay in the .4 realm. Above that and the airfoils start to matter. Since fuel burn and thermodynamic efficiency are of secondary importance to the IC performance turbo world, Garrett has elected to trade power density against waste heat out of the inter-cooler! And the power from the exhaust is essentially free.
Patents....yes they patent everything these days. Most of it isn't worth the paper it's written on. I found a lot of 2 cycle patents when I was working on a two stroke powered fan. Some of the patents were for concepts that I had read about in books published decades ago. No way the patent would hold up in court, but they can scare somebody with a letter from a lawyer and the threat of 100's of thousands in court costs to get it thrown out. Nothing in that wheel design is non-obvious to someone who designs these things. I think most of Garrett's patents are for the ball bearing arrangement in the center housing. The marketing dept. is of course going to play fast and loose with that.
I've enjoyed this thread. Some of the links for papers listed in the beginning are unfortunately not working anymore.
Monty
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Oct 23, 2018 7:18:49 GMT -5
Hi Chris Yep , the new Garrett wheels are works of art , getting some almost unbelievable flows per sq inch of inlet . With inlet airspeeds ~700 ft/sec and inducer tip speeds of >1,300 ft/sec the relative speed is up ~1,500 ft/sec , they must be getting some shock compression simply to be able to get the mass flow through the smaller inducer throat area . The inducer tip angle on the X846 wheel that Anders is using is set at a lowish ~24 degrees ( 66 from axial) and still has a relative Mach number of ~1.3 at design speed , the new Garrett wheels are really pushing the flows , but the old Garrett GT6041 comp with an "ancient" design is getting 80% effic whilst these new wheels are only mid 70's............ah, tradeoffs. I'm looking forward to the comp maps for the new largest Garretts "coming soon" www.turbosbytm.com/gtx5544r-gen2-106mm-inducer , Tony and I were taking about it a while back , should be "interesting" :-) Cheers John John,
I wouldn't be surprised if they are pushing choke on the throat in the inducer. That would be another reason to go with full blades and no splitters. Full blades allow a stable shock in each passage. A similar approach to matching the choke point of the exducer and diffuser throat could be used to get all three areas optimized at the same flow. My guess is Garrett has extended the Casey/Rusch method to all three and used CFD refinement from there.
Mar-M wheels.....YES PLEASE!!! A TV94 sized Mar-M wheel would be nice.
The only bad thing is as they optimize for automotive performance it moves away from the optimum design point for us- if we are trying to get fuel consumption down. If we want more thrust at the same size and don't care about fuel burn, I guess it is a positive.
Monty
|
|
|
Post by finiteparts on Oct 27, 2018 20:48:39 GMT -5
Ron, Regarding no splitter vanes. I'm guessing the increased number of blades in the inlet is to keep the diffusion factor under control in the inducer. I run into a similar trade off in the design of a fan with solidity. Increasing the number of blades reduces the diffusion each passage must accomplish. Higher diffusion increases loss. Trying to get max diffusion and flow means trading off against efficiency. One of the other trade offs: I can increase mass flow into the fan at a given wheel speed. the angle of attack of the blades must increase and efficiency suffers, but I can fit more power in a smaller space. Modern fans and compressors are reaching diffusion factors around .6. I try to stay in the .4 realm. Above that and the airfoils start to matter. Since fuel burn and thermodynamic efficiency are of secondary importance to the IC performance turbo world, Garrett has elected to trade power density against waste heat out of the inter-cooler! And the power from the exhaust is essentially free. Patents....yes they patent everything these days. Most of it isn't worth the paper it's written on. I found a lot of 2 cycle patents when I was working on a two stroke powered fan. Some of the patents were for concepts that I had read about in books published decades ago. No way the patent would hold up in court, but they can scare somebody with a letter from a lawyer and the threat of 100's of thousands in court costs to get it thrown out. Nothing in that wheel design is non-obvious to someone who designs these things. I think most of Garrett's patents are for the ball bearing arrangement in the center housing. The marketing dept. is of course going to play fast and loose with that. I've enjoyed this thread. Some of the links for papers listed in the beginning are unfortunately not working anymore. Monty Monty, I assume that you meant to reply to me on this, since Ron actually didn't comment on this part of the topic.... I am glad you enjoyed the content. I just went through and checked the links...I had to delete a few of the pdf versions of the books and I highlighted them as such...the other links appear to be ok. I would tend to agree with you on the diffusion in the inducer, but just for clarity, the term "diffusion factor" is typically referring to the stage, not the sub-regions of the impeller. Yes, if you look at any impeller, the role of the inducer is to "onboard" the flow to the impeller passages and this is challenging to do with minimal losses, especially when the inlet relative flow is high subsonic or sonic. Just as a refresher, the term "transonic inducers" in centrifugal compressors refers to the fact that a portion of the inducer span is operating with supersonic relative flow speed while there is an inner portion of the blade span that is seeing subsonic relative flows. The leading edge is very thin to minimize the streamline curvature as the flow splits around the blade and the blades begin to controlled the flowfield. The local effect of the leading edge is referred to as the over-velocity ratio, meaning that the local velocity over the leading edge relative to the bulk flow relative velocity...and of course you want to minimize this. This is often a sign of a cheap, copied impeller...as the original manufacturer like Holset or Garrett will take great pains to produce an elliptical leading edge profile to minimize this overvelocity ratio, while the cheap copies will have round leading edges. If you look at the old Boeing high pressure ratio compressor design work (like the one in my avatar) they discuss this and as I recall, the elliptic leading edge gives a local flow acceleration to roughly 1.25 x freestream velocity, while the round leading edge gives roughly a 1.5 X the freestream velocity...and these were even referencing very thin leading edges. That is one reason that I have a hard time believing the claims of companies that copy impellers and CNC them out and sell them as "upgrades" to 7.3L, 6.5L, etc...without real mapping and testing. As you can imagine, if you have a local inlet relative flow at the inducer tip of M=1.15, then the difference in the local shock strength is really between a local suction surface velocity of M=1.4 and M=1.7'ish...which is a huge difference when the local shock is tripping the boundary layer flow and causing your impeller passage blockage to shoot up. Because the impeller sees an adverse pressure gradient everywhere after the inducer, if you trip the boundary layer, there is no real mechanism to stabilize that...so anything that happens in the inducer, only makes things worse through the rest of the impeller and then feeds the diffuser with poorly developed flow making it's job harder too. Additionally, after the leading edge effects, the surface curvature across the suction side of the blade causes a corresponding flow acceleration, for transonic or supersonic relative inlet flows, this acceleration needs to be minimized so as to minimize the local flow accelerating and strengthening any shock formations in the impeller. This is the reason that the camber of the inducer is very small or they are even straight for a portion of the passage after the inducer. The Boeing impellers had very long inducer sections to try to minimize the local rate of flow turning and thus the local flow accelerations. The GTX blades look to me to be trying to coax that flow as gentle as possible from the inlet angle to axial and I think the lack of splitters is due to leading edge vortex interaction from the main blades to the splitters being more of an issue on these higher pressure ratio impellers. I don't think there is really an increase in blades. The GT55 that I have runs 14 blades (7+7), while the GTX5533R has 13 blades...so I am not sure that your theory holds up. Usually, splitter blades are used to open up the inducer throat area, but allow control of the radial portion of the flow to reduce the blade loading after the throat. On the transonic impellers, the control of the shock wave is very important and as you state you do have a standing shock in the outer spans...and this shock can help develop some pressure rise, but the chance to do real harm to the boundary layer and hurt the stage more than it helps usually drives the design to try to minimize the shock strength. Thanks, Chris
|
|
|
Post by finiteparts on Oct 27, 2018 21:03:17 GMT -5
Hi Chris Yep , the new Garrett wheels are works of art , getting some almost unbelievable flows per sq inch of inlet . With inlet airspeeds ~700 ft/sec and inducer tip speeds of >1,300 ft/sec the relative speed is up ~1,500 ft/sec , they must be getting some shock compression simply to be able to get the mass flow through the smaller inducer throat area . The inducer tip angle on the X846 wheel that Anders is using is set at a lowish ~24 degrees ( 66 from axial) and still has a relative Mach number of ~1.3 at design speed , the new Garrett wheels are really pushing the flows , but the old Garrett GT6041 comp with an "ancient" design is getting 80% effic whilst these new wheels are only mid 70's............ah, tradeoffs. I'm looking forward to the comp maps for the new largest Garretts "coming soon" www.turbosbytm.com/gtx5544r-gen2-106mm-inducer , Tony and I were taking about it a while back , should be "interesting" :-) Cheers John John,
I wouldn't be surprised if they are pushing choke on the throat in the inducer. That would be another reason to go with full blades and no splitters. Full blades allow a stable shock in each passage. A similar approach to matching the choke point of the exducer and diffuser throat could be used to get all three areas optimized at the same flow. My guess is Garrett has extended the Casey/Rusch method to all three and used CFD refinement from there.
Mar-M wheels.....YES PLEASE!!! A TV94 sized Mar-M wheel would be nice.
The only bad thing is as they optimize for automotive performance it moves away from the optimum design point for us- if we are trying to get fuel consumption down. If we want more thrust at the same size and don't care about fuel burn, I guess it is a positive.
Monty
Monty, If you look at almost any map, you will see that they are pushed to choke...in vaneless diffusers, this is 99% of the time at the inducer...thus the near vertical speed lines. Vaneless diffusers don't choke at the tongue of the scroll because that would imply that the discharge from the impeller would have to be super fast so that the flow remains sonic through the vaneless diffuser...which is very unlikey. Maybe you meant something different? You get shocks standing in both full bladed and splittered impellers...again, not sure what you meant here. You don't want to get choking in the inducer and the exducer of an impeller as this would mean a very poor ratio of relative velocities through the impeller and likely not much pressure rise. Backwards vane sweep is purposely done to reduce the relative discharge velocity and it is this reduction in relative exit velocity that has driven up the impeller efficiencies over time. If you reduce the diffusion completed in the impeller, then you are forcing the diffuser stage to do more of the heavy lifting, and diffusion in the rotating impeller is more efficient because it has the stabilizing effect of the centrifugal forces acting in its favor. The diffuser stage is more challenged when you feed it with higher speed flows from a poorly diffused impeller and that is exactly what trying to set the exducer area such that the exducer flow chokes is doing. The Casey and Rusch method is better thought of as a means of matching the limiting flows in the rotating and the stationary reference planes. I hope that helps. Chris
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Oct 29, 2018 8:43:24 GMT -5
Hi Chris,
That actually helps quite a bit. Thanks! Do you have any good resources for matching the impeller/diffuser interaction while changing the exducer height? It changes the supply value and output angles significantly. I would imagine there is trade/off there with how much diffusion happens in the impeller vs the diffuser. Probably something to do with specific speed I'm guessing. There is some hand waving about it in Japikse but no real data. The Japikse book also has a picture of a wheel that looks much like one of the new Garretts, next to one with splitters, and states the one with the splitters outperformed the one with full length vanes in both range and efficiency. I'm not any clearer on it really. But... I don't have to be-I just need a map!
I was very relieved to find elliptical leading edges on the wheel I just purchased. I just wish I had a map so I could quit guessing...and I wish I could quit having to switch between English/SI! All the graphs I have to work from have specific speed in English units.
Monty
|
|
|
Post by finiteparts on Oct 29, 2018 21:18:53 GMT -5
Hi Monty,
I am sure I have some good references somewhere, but I am taking a rotordynamics class and have homework to get done, while also dealing with a teething 15 month old...so I will probably have to wait tomorrow or later to go digging through the stack of compressor papers.
Japikse's Centrifugal compressor book is a great reference...I would suggest working through the two zone modeling as it really helps to better understand a more realistic flow that is being fed to the diffuser stage. Japikse's book uses Dean's work on wake-jet flows...he was a pioneer in the wake-jet style modeling of the centrifugal impellers and his work was then supported by the laser velocimetry work done by Krain and others at the time. To modify the b-width (the impeller exducer tip height) in any logical way, you are going to have to have a good predictive model to understand how you are going to impact the compressor stage performance. Japiske's book is good for this, but there are also other book and papers that can be helpful. Whitfield and Baines' "Design of Radial Turbomachines" is also a great reference. I will try to find more when I have time.
What exactly are you trying to achieve by trimming the b-width? As a general rule, changing the b-width moves the vertical location of the peak efficiency island, because you are changing the meridional impeller discharge velocity when you change it's height. If you look at the exit to inlet area ratio, EI, generally, a larger EI pushes the peak efficiency island to a lower pressure ratio. So if you don't change your inducer diameter, then reducing the b-width will likely push the peak efficiency island to a higher pressure ratio.
But, you don't have any idea where your peak efficiency island is, since you don't have a map...so...I would suggest you build a good mathematical model for a centrifugal compressor and calibrate it to a similar compressor that does have a map. Also, if I remember correctly, you are not trying for a "high" pressure ratio...maybe around 3:1...? If that is the case, you might want to look at using a low solidity style diffuser...at lower PRs, they are more forgiving and give a wider map width with almost the same efficiency as a channel diffuser.
I will try to get you more info as soon as I get some free time.
- Chris
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Oct 29, 2018 22:31:32 GMT -5
Chris,
No worries, take care of life. This is just a hobby!
I'm trying to understand the trade-offs so I can more intelligently select off the shelf solutions. I'm a design engineer. I really don't want to be a compressor design guy, but I need to pick one that's "good enough" And understand what variables to play with. Your post confirms my intuition. I'm looking for 3.5-3.7ish PR. Anything higher is a waste for my insane contraption. Much lower and the cycle efficiency takes a hit. I've got too many rats to kill to go off down obscure rabbit holes and get side tracked. I'll leave that to those that are passionate about those things...and I'm glad they are! My insanity lies elsewhere...
KTS should hire me to generate maps....it's not that hard...just another rabbit hole.
Monty
|
|
Chuks
Senior Member
Joined: August 2015
Posts: 498
|
Post by Chuks on Nov 9, 2018 17:14:11 GMT -5
|
|
|
Post by finiteparts on Nov 18, 2018 15:44:11 GMT -5
Hey Monty,
I went through a pile of papers and found a few that focused on compressor trimming, but there were only a few that looked at axial trimming. I made the assumption that you were not attempting to trim anything other than the exducer b-width, thus "axial trimming". Rodgers and others had a few papers that covered flow trimming (which means that you are trimming the inducer and exducer to target a reduced flow range). But the following paper is free and does a great job covering the axial trimming.
The fact that you do not have a compressor map for your impeller means that you will need to do the full set of calculations on the impeller to understand if you want to trim the impellers discharge height or not...so, you unfortunately are forced to become a part-time compressor engineer.
If you go through Daniel Swain's thesis (or any of the other papers out there by Rodgers, Came or Whitfield) or you work through the vector calulations, you see that reducing the exducers axial height (b-width), you increase the relative discharge velocity due to the reduced exit area. This means that you have reduced the diffusion of the relative flow through the impeller passage, which is captured in the diffusion ratio, W1s/W2 (W1s = inlet relative velocity at the shroud, W2 = exit relative velocity assuming an even discharge velocity profile. Just as a stationary diffuser has a limit on how much the area can expand before the flow in the diffuser separates, so too does the flow in the rotating diffuser that is the impeller passage. That is why there are often suggestions on the upper limit for the diffusion ratio. Dean (1972) suggests a DR < 1.8 for a subsonic inducer and Rodgers (1977) suggests DR = 1.9 -2.0. So you might want to pick a DR and calculate the exit relative velocity from that. These DR are ok for use with a one-dimensional velocity calculation, but it should be noted that a more accurate approach would be to calculate an estimate of the "jet" velocity (from the two-zone model, i.e. wake-jet model). Finally, the decreased DR reduces the pressure ratio capability of the impeller stage, since you have reduced the recovery of the total pressure in the impeller, while the recovery in the diffuser hasn't changed.
The increased exit relative discharge velocity will also change the efficiency of the impeller. If the b-width is too large and you are getting to much diffusion through the impeller, the large flow separation could reduce the efficiency. So if that were the case, then reducing the b-width moves you to a lower DR, which could reduce the internal flow separation...thus giving you an increased efficiency. BUT...if you are already close to the ideal b-width, reducing it further increases the relative flow through it and thus increases the frictional losses. Also, reduced b-width means that a fixed axial casing-impeller clearance becomes a larger proportion of the blade height and thus the clearance losses increase.
Finally, the increased relative discharge velocity will impact the absolute discharge velocity. The discharge angle gets more radial as the DR is reduced and the entrance angle to the diffuser needs to be adjusted to meet this change. The increase in discharge velocity also has to be handled in the vaneless space between the impeller discharge and the vaned diffuser entrance, which may or may nor be hurt from the increased velocity. In one respect the increased radial velocity shortens the path length from the impeller to the diffuser, potentially reducing the frictional losses, but the increase in velocity may also also increase the same frictional losses. Related to that is also the fact that the incidence losses for the diffuser vanes, increase as the inlet Mach number is increased.
So there is no real easy answer...as with all engineering problems, there are always trade-offs.
I hope this helps,
Chris
|
|
CH3NO2
Senior Member
Joined: March 2017
Posts: 455
|
Post by CH3NO2 on Nov 18, 2018 20:11:03 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
|
|
monty
Senior Member
Currently being spanked by mother nature.......
Joined: September 2018
Posts: 400
|
Post by monty on Nov 18, 2018 21:31:21 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 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.
|
|
CH3NO2
Senior Member
Joined: March 2017
Posts: 455
|
Post by CH3NO2 on Nov 19, 2018 8:27:49 GMT -5
Hi Monty, A reduced compressor pressure ratio will reduce it's power requirement if it's efficiency remains constant. This part is easy to understand. However, for a given turbine and nozzle flow area, the choke point(s) are determined by mass flow rate, temperature, average molecular weight and the ratio of specific heats (Cp/Cv). Only so much can be flowed through the respective flow areas before choke. If avoiding choke is a requirement this variable is regulated by compressor inducer area (ie mass flow rate). So clearly there is a matching point to be made at the compressor inducer. Then there is the need to reduce stationary diffuser diameter. At a given mass flow rate, the stationary diffusers diameter is determined by its inlet velocity. Reducing the compressors exducers height increases it's outlet velocity... which requires a larger diameter stationary diffuser. 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
|
|