Random Technical Questions

[CH3NO2]

Hello Everyone,

Is the flow capacity of a turbine wheel determined by its exducer choked flow area?
In the case of the GTX5533R turbine wheel it has a choked flow area of 7 square inches. Does this set the flow capacity?

With a radial inlet guide vane assembly feeding into the turbine inducer and the guide vane total choked flow area is 4.2 inches squared, does the NGV become the regulator of flow rate? If yes, it would seem a NGV choke area of 4.2"^2 would work very nicely with the 91mm compressor of the GTX5533R.

Plot shows mass flow rate Vs pressure ratio through a 4.2"^2 NGV using these gas conditions:


T(K)    T(F)    ENTHALPY   ENTROPY   CP/CV   GAS      RT/V
1159  1628       -16.55     1798.66    1.3054  31.826  0.094

SPECIFIC HEAT (MOLAR) OF GAS AND TOTAL= 8.448 8.417
AVE MOLECULAR WEIGHT OF 28.888

[CH3NO2]

So I've been digging into the fundamentals of radial inflow turbines and radial inflow nozzle guide vanes. Right now I'm just trying to figure out how to determine the flow capacity of a turbine, the resulting power and how to design a NGV around it.

I was previously working on a GT6041BL turbo but without a compressor map for it and the unknowns that go with it, I've decided to put that turbo aside for a later date. I dont really want to cut my teeth on a build without a compressor map. So I've decided to take a shot at doing a build with the GTX5533R turbo. At least they come with a compressor map. The other good thing about this turbo is that they come in a range of different compressor inducer diameters. If necessary, it will make changing out one compressor wheel to another easy to do. Just change out the CHRA with a couple of V-band clamps and its done.

As a starting point I picked up a 85mm GTX5533R with a 1.3 A/R housing. The initial turbine swallow calculations look like it should work well with a sorethumb combustor. However, I've been itching to look into the possibility of doing a reverse flow combustion chamber design with a radial inflow NGV. So I took the 5533R apart and took a bunch of measurements. I even brought it to a local scanning facility to get some of the measurements I couldn't get manually.

The scan is crude. It was done with a hand held scanner but its more than adequate to provide the necessary numbers.


ibb.co/NV5krxV

[CH3NO2]

If using the 85mm compressor with the 1.3 A/R scroll it should make a workable gas turbine, however during all the studying of the scroll I measured the radial annulus going to the turbine inducer. 2in^2… exactly the same as the turbine inducer. But I had to think about that for a second.

The scroll inlet is only 3.5” diameter and only about 2/3’s of the inlet is usable flow area due to the way the inlet is contoured. Then it hit me by surprise. The flow is choked at the scroll inlet!?!

Wow. That was interesting. I had no idea it worked that way but was happy to have learned something new. Never really gave it much thought but assumed the choke point would be right up against the inducer. 

So the measurements and analysis continued.   How is Mdot and power determined for a given turbine and how is a NGV designed around it?

The book I am using is Principles of Turbomachinery, and it can be found here: 160592857366.free.fr/joe/ebooks/Mechanical%20Engineering%20Books%20Collection/TURBOMACHINES/Principles%20of%20Turbomachinery%202.pdf

Power output of a radial turbine? 
I have seen the power calculated two ways:
1. The first way is from the specific work (w) of the rotor blades moving at a certain rpm. P=m_dot*w and w=U_2^2 and U_2=R_2*Ω Ω=rpm*2pi/60 where m_dot is our massflow, U_2 is the tip speed, R_2 is the turbine inducer radius and Ω is the angular velocity.
2. The second way is from the isentropic work of the gases, completely independent of rotor size or shape. Instead it relies on the combustion chamber pressure, temperature and efficiency of the turbine. P=m_dot*w and w=Cp(T_01-T_03) and T_03=T_01*[1-η_ts(1-(p_01/p_3)^-(γ-1)/γ)] where Cp is the specific heat of combustion gasses, T_01 is the combustion chamber temperature, T_03 is the exit stagnation temperature, η_ts is the total-to-static efficiency, p_01 is the chamber pressure and p_3 is the outlet pressure (atmospheric pressure in our case).
My question on these two different ways of calculating power stems from the fact that these have often calculate to different numbers. So, I would like to know if my following logic is correct.
In the case where the second method for power comes out to be greater than the first method, it means our turbine is producing more power than what the rpm condition would produce, therefore the turbine would want to speed up. Thus, if we want to achieve a steady state, we want to try and get both methods of calculating power to equal each other. This can be done by lowering the chamber temperature (by lowering the fuel flow). The same is true in the opposite case where the first method is greater than the second method, indicating the combustion is not enough to keep the turbine spinning at the design point rpm, and thus would slow down.


How is Turbine Nozzle Guide Vane (or stator) angle α calculated?
In chapter 9 (pg 313) α_2 is defined as α_2=asin(U_2/V_2) with U_2=sqrt(Cp(T_01-T_03)), or U_2=R_2*Ω Ω=rpm*2pi/60.
I have only found V_2 to be calculated one way, that is in the choked condition: V_2=sqrt(γ*R*T_2) with T_2= T_01 *(2/(γ+1)).
Using all of the above equations we can find α_2, and then use A=m_dot/(C_d*sqrt(y*ρ*p_01*(2/(y+1))^(y+1/y-1)) to find the choked flow area. Then when we put all of this into Solidworks to make a vane design, we cannot possibly get this small of a gap, because the turbine inducer is just too large a diameter and has too high of a tip height. Thus, the flow could not be choked, and thus the velocity V_2 used to find α is also wrong, and therefore the α angle put into Solidworks is also wrong, telling me that perhaps this is still possible to choke the flow.

For example, if I were to put α_2=80° into Solidworks with 12 vanes wedge at 15° per vane and per gap between vanes, then the flow area equals the choked condition, but if I input the the value I found for the vane angle α_2=58.4° then my area is almost 3 times as large, and clearly not choked. Table 9.1 on page 329 shows that a typical value for α_2 is between 68 and 78°, so this might indicate the 58.4° is simply a wrong number, if so, are any of my equations wrong?
Now this might be explained by a concept in an earlier chapter, with reference to an axial turbine.

On page 100 Figure 3.18, it shows a parallel guide vane with α_1 as the angle of the vane itself, this is what we are trying to find. Then α_2 is defined as the angle of the flow stream and NOT the stator vane angle. The following examples show us that the difference of α_1 and α_2 varies from 2-10°. This difference from α_1 to α_2 seems easy enough to calculate, but they all depend on the pressure in the interblade section between the stators and the rotor blades, so how might I find that pressure, or is this slight 2-10° difference from α_1 to α_2 not worth worrying about?

[racket]

Hi Tony

It gets a bit more complicated due to the fact that, depending on comp efficiency, we don't need to always choke either the NGV or exducer to provide the required gas deflection for the horsepower required , and its probably more efficient not to choke one or the other but have less than "sonic" throughout the stage , the stage max effic of 74% is kinda low for the 5533 , a lotta pressure energy is "wasted" .

Your scanned exducer angle is rather "tight" at 67.5 degrees ( 22.5 degrees) for the 7 sq in exducer choked flow area given , I'd be inclined to do a manuel check of the wheel dimensions , the Free Flow area of my TV84 turb wheel 110/96 mm was only 4.8 sq inches

Cheers
John

[CH3NO2]

My final questions deal with the nozzle guide vane shape themselves. I have not found any literature that deals with the shape of vane design. So it was arbitrarily chosen 15° between each wedge and 15° on the wedge itself. The biggest question here is where should α_1 be located? Should it be on the left or right hand side of the wedge, or should it be halfway between the two?

You may be asking "Why are you using wedges on the NGV?"
1) It helps a bit when trying to manage the flow area going into the choke point.
2) I would like to flow air through them for cooling and for passing evaporator tubes through them.

Tony

[racket]

Hi Tony

The turb scroll choke throat should be at the end of the "tongue' close to the wheel where the gases enter it .

If you measure the area at that point and divde by the distance from the "Flow area centre" to the shaft centre , it should give your A/R .

Cheers
John

[CH3NO2]

Feb 13, 2019 20:01:33 GMT -5 racket said:

Hi Tony

It gets a bit more complicated due to the fact that, depending on comp efficiency, we don't need to always choke either the NGV or exducer to provide the required gas deflection for the horsepower required , and its probably more efficient not to choke one or the other but have less than "sonic" throughout the stage , the stage max effic of 74% is kinda low for the 5533 , a lotta pressure energy is "wasted" .

Your scanned exducer angle is rather "tight" at 67.5 degrees ( 22.5 degrees) for the 7 sq in exducer choked flow area given , I'd be inclined to do a manuel check of the wheel dimensions , the Free Flow area of my TV84 turb wheel 110/96 mm was only 4.8 sq inches

Cheers
John

Hi John,

That's great to hear it doesn't have to be choked.   The way things were stacking up, it was looking like it would be difficult to get the NGV to choke.   However, without a choke point, how is flow rate determined or set?   Where should it be set?  
And how is it possible to vent 50psig to atmosphere without a choke point popping up some where in the system?

I'll see if there is a way to verify the choked flow area of the turbine exducer.
What we did was to scan the whole turbine.  Most of the scan was very noisy but we were able to find one section between two blades where we could model in a perpendicular surface and measure that area.  0.7"^2 per passage and there are 10 blades.

drive.google.com/open?id=1YvNQY3CY5DxXMSK1QkJmSMl4qP3n8mbM

[CH3NO2]

Feb 13, 2019 20:06:07 GMT -5 racket said:

Hi Tony

The turb scroll choke throat should be at the end of the "tongue' close to the wheel where the gases enter it .

If you measure the area at that point and divde by the distance from the "Flow area centre" to the shaft centre , it should give your A/R .

Cheers
John

Yep, that's right where the choke point is at.  The end of the tongue.

It's published as having a 1.3 A/R but for good exercise I'll measure it to manually find the A/R.

[finiteparts]

Tony,
One of the primary problems with your power calculations is that you are assuming that there is no residual swirl...which means that you are forcing the calculations to have a potentially false boundary condition. The power is due to the change in the swirl through the turbine stage and the exit vector is only axial at a single point on each speed line.

Good luck,

Chris

[racket]

Hi Tony

I'm getting confused here :-(

Are you wanting full expansion through the turbine stage , or only partial expansion with the remaining portion expanded through a jet nozzle or freepower wheel ??

On a side note , your GT6041BL has the same comp wheel ( #700248-0001) as the "Performance" GT6041 that I used in my kart build , though the turb wheels have different part numbers .

Cheers
John

[racket]

Hi Tony

I did some rough numbers for the 85 mm unit and your NGV angle is going to need to be down lower so as to provide ~19-20 gas flow angle to the inducer .

NGV throat area ~3.7 sq ins which if divided by your 0.871" tip height( this seems a bit large , was expecting maybe 18 mm- 0.75" ) makes for ~4.25 linear inches of throat, so 12 throats 0.35" X 0.87 " ( 9.4 mm X 22.1 mm) , theres been an allowance ( 10% ) for "boundary" .

With a 4:1 PR at 74% we've got an ~190 C temp rise needing ~ 162 C drop through the turb , but because of the "lowish" T I T and the poor turb effic of 74% we'll be using ~2.4 PR of the 3.8 PR ( 5% combustor loss) entering the turb stage.

Because theres exit velocity from the wheel I've used an overall PR of ~2.7 across the stage , now sq root 2.7 is only 1.65 , but I've gone for ~1.75 for the NGV because the turb wheel inlet gets some carry over energy from the "radial" velocity going in ( ~550 ft/sec??) .

3.8/1.75 gives us 2.17 PR at the throats and a density of ~21 cu ft/lb so ~42 CFS going at ~1,800 ft/sec .

I haven't done the exducer calcs because I'm a tad concerned about that flow area .

Cheers
John

[CH3NO2]

Feb 14, 2019 16:23:50 GMT -5 racket said:

Hi Tony

I'm getting confused here :-(

Are you wanting full expansion through the turbine stage , or only partial expansion with the remaining portion expanded through a jet nozzle or freepower wheel ??

On a side note , your GT6041BL has the same comp wheel ( #700248-0001) as the "Performance" GT6041 that I used in my kart build , though the turb wheels have different part numbers .

Cheers
John

Hi John,

Yes, full expansion through the turbine.  No nozzles or freepower turbines for now.

As for the GT6041 and the GT6041BL:

The 6041 compressor has an exducer tip height of 12mm.

The 6041BL compressor has an exducer tip height of 9mm.

I've seen that you have posted several times that this change in Inducer/Exducer area ratio suggests a potential increase in compressor pressure ratio.   This may be true but the degree of change is unknown.  Also, a few weeks ago, Finiteparts Chris, posted a scholarly document about the change in compressor performance when reducing tip height.  In all the examples shown in the document, the pressure ratio and mass flow went down in every example tested.

Chris mentioned that the reduction in performance was likely due to the reduction in tip height being down from the optimum design point.   IE - there may be cases where a reduction in tip height can improve performance.

This can all be true but its still an unknown factor in the GT6041BL.   So it's hard to predict how the compressor map has changed from the reduction in tip height.

But here is one other thing I found out about the 6041BL.   I was talking with a used turbo supplier here in the US.   We were talking about the 6041BL.   He said there are many of these turbos available for sale.   He said they are available because they were all pulled off of brand new, unused, Caterpillar engines.   The unused Caterpillar engines were repurposed to a new application that required more power so all the 6041BL turbos were removed and better turbos were reinstalled.  He never said what the new turbos were and I didn't ask, but just hearing about a large number of 6041BL's being "scrapped" and replaced with a better turbo left me questioning the 6041BL's tip height and what it may mean to its compressor performance.   

What the change in tip height does on the compressor map is unclear.  So far I'm not getting a good feel about it.   Because I want to learn the subject of turbo machinery, I think I'd rather start with a turbo that has a clearly defined compressor map.   At least I can make calculations based on a real map with less guessing and unknown confusion involved. 

Tony

[CH3NO2]

Feb 14, 2019 19:00:09 GMT -5 racket said:

Hi Tony

I did some rough numbers for the 85 mm unit and your NGV angle is going to need to be down lower so as to provide ~19-20 gas flow angle to the inducer .

NGV throat area ~3.7 sq ins which if divided by your 0.871" tip height( this seems a bit large , was expecting maybe 18 mm- 0.75" ) makes for ~4.25 linear inches of throat, so 12 throats 0.35" X 0.87 " ( 9.4 mm X 22.1 mm) , theres been an allowance ( 10% ) for "boundary" .

With a 4:1 PR at 74% we've got an ~190 C temp rise needing ~ 162 C drop through the turb , but because of the "lowish" T I T and the poor turb effic of 74% we'll be using ~2.4 PR of the 3.8 PR ( 5% combustor loss) entering the turb stage.

Because theres exit velocity from the wheel I've used an overall PR of ~2.7 across the stage , now sq root 2.7 is only 1.65 , but I've gone for ~1.75 for the NGV because the turb wheel inlet gets some carry over energy from the "radial" velocity going in ( ~550 ft/sec??) .

3.8/1.75 gives us 2.17 PR at the throats and a density of ~21 cu ft/lb so ~42 CFS going at ~1,800 ft/sec .

I haven't done the exducer calcs because I'm a tad concerned about that flow area .

Cheers
John

Yep.  3.7"^2 is about what I came up with for the 85mm inducer too.

Now that the turbine wheel has been measured out how do I calculate it's swallow potential?

I'll recheck the turbine exducer choke area once again but so far it looks to be about 7"^2.   So lets assume for the moment that 7"^2 can be used for calculations.   
How is the flow potential through a turbine calculated?

I could just calculate flow through a 7"^2 orifice but I know that alone wont work.  It needs to go through a NGV first.  

So I did an example calculation using the 98mm compressor wheel. 
At a pressure ratio of 4:1 @ 180 lbs/min  I came up with a NGV throat area of 4.885”^2.

This choke area is sized to 3lbs/sec air and 0.0694 lbs/sec fuel included. 

With a NGV choke area of 4.885”^2 expanding to 7”^2 at the turbine exducer, I cant see a reason why we cant flow it all through the turbine.

But I dont really have a full understanding of how to calculate it all just yet.   Getting a little closer!

Thanks,
Tony

[racket]

Hi Tony

I took my GT60 info from here www.turbomaster.info/eng/catalogs/model.php?base=garrett&pagina=GT60 , the "better" turbo you mentioned could mean anything from a higher PR , cheaper , more readily available , better probably meant more appropriate for changed demands .

Compressor flows from a wheel will change with a change of the A/R of the comp scroll , also if a vaned diffuser is used , ............speeking with an experienced GTBA Member about this subject he told me that he'd given up using comp maps for the wheels when used in a micro engine .

I can appreciate your desire to use a rotative with good maps for both ends .

OK , back to the 5533 , full expansion , with or without power takeoff , without power takeoff then the TOT temps will be rather low 400-550 C , which will change required NGV throats compared to a "full throttle" engine scenario .

Cheers
John

[CH3NO2]

Feb 15, 2019 4:55:31 GMT -5 racket said:

Hi Tony

.... full expansion , with or without power takeoff , without power takeoff then the TOT temps will be rather low 400-550 C , which will change required NGV throats compared to a "full throttle" engine scenario .

Cheers
John

OK, that's good to hear there's lots of margin on the TOT.

I definitely plan on using it a couple of different ways.   
1)  With a nozzle for thrust.
2)  Without a nozzle for air bleed.

The calculations so far strongly favor using a 98mm compressor for bleed.   However, I'm not entirely sure if the turbine will allow it.  The isentropic equations say the turbine & compressor will work power wise, but they dont say if the turbine will actually swallow this much mass flow.   And doing it with a reasonable spouting velocity.

With 4.885”^2 expanding into 7"^2, it "seems" like it will work but I can't say with certainty.  Not yet sure on how to break down the interstage dynamics with the back pressure of the turbine.

Still researching the books.

Thank you!
Tony