Combustion Design

[racket]

Hi Tony

LOL...........Jetspecs is intended for complete novices so that they can get an engine to work , the pressure drop hasn't come into the calcs , most turbos used will have much more than a 5% miss match between comp and turb , Jetspecs is a crude device and was never intended as a scientific design tool .

I used this Paper to design my first flametube journals.sagepub.com/doi/abs/10.1243/PIME_CONF_1968_183_245_02

Cheers
John

[finiteparts]

Hi Tony,

First, I can tell you that neither one is "correct". These are approximations based on data and test over a range of conditions...

Secondly, they don't seem too far apart to me. Jetspecs is giving you geometric areas and the method that I am presenting here is providing you effective areas. So if I calculate the Cd needed to make them match, I get a Cd ~ 0.53. For the readers that don't know what "effective area" is, it is the geometric area (Ageo) of the hole multiplied by a discharge coefficient (Cd). Because the flow experiences losses as it enters and exits an orifice (the hole in the liner), the area that the flow actually "sees" as it goes through is smaller than the geometric area.

Aeff = Ageo*Cd

So I was going to get to sizing the holes after I covered sizing the primary zone, but I can switch the order without any loss of clarity. The Cd that you specify is all on you. You have to do the work to determine which Cd value applies to the holes that you produce in your liner.

I will go into this in more detail later, but there is data out there on all kinds of hole shapes that you can use to select the proper Cd for you design. Here is an example from some early gas turbine work done by NACA...notice they use Cp instead of Cd.

ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19930084890.pdf

So if we look at the first 0.275 inch primary hole called out by Jetspecs, we find a similar size sharp edged hole Cd curve...



So you can see the peak Cds are around 0.6, thus a 0.53 isn't that far off.  Once we work through how to calculate the flow parameter, you can pick the Cd that applies to the hole sizes that you will select based on the depth of jet penetration required.

I will work through the sizing for my combustor later tonight or tomorrow when I have some free time. I will also provide some more links to some other sets of flow data for plunged style holes.

- Chris

[finiteparts]

John,

What did you mean by this, "... most turbos used will have much more than a 5% miss match between comp and turb..."?

[racket]

Hi Chris

I was just trying to say , don't be too pedantic about getting down to the last percentage point with calcs when using a turbo for the engine , if the turb stage is oversized as can happen with some diesel turbos ( CI engine) then the airflow will be a lot higher than if the turb stage is a bit tight as can happen with SI engine turbo , this large variation in flows for a given comp inducer size will then change the pressure drop across the flametube holes .

Lets use your EFR6758 comp map for instance , at a 2.4 PR , it could flow anywhere from 15 to 50 lbs/min depending on what the flow areas are in the turb stage and downstream .

If flowing only 15 lbs min due to a restrictive scroll A/R then the "design" flametube hole area will be oversized and pressure drop percentage will be low , but if flowing 50 lb/min then the percentage drop will be much higher .

My apologies for being too brief , still getting over an "unscheduled" stay in hospital , so not contributing much .

Cheers
John

[CH3NO2]

Hi Chris,
Wow, with Cd factored into the Lefebvre "effective area" the area's come together very nicely.


Hi John,
I hope you feel better soon!

Tony

[CH3NO2]

Jun 2, 2017 17:50:02 GMT -5 racket said:

Hi Tony

LOL...........Jetspecs is intended for complete novices so that they can get an engine to work , the pressure drop hasn't come into the calcs , most turbos used will have much more than a 5% miss match between comp and turb , Jetspecs is a crude device and was never intended as a scientific design tool .

I used this Paper to design my first flametube journals.sagepub.com/doi/abs/10.1243/PIME_CONF_1968_183_245_02

Cheers
John

Hi John,

How was your experience designing your first flame tube and did it work as intended?

Has your idea of creating the optimum flame tube evolved differently from this method?

Thanks,
Tony

[CH3NO2]

Hi Chris,

I had some questions on page 3 of Primary Zone Sizing.

Just below design points W3 and PR is value 0.72. What is this number?

In the paragraph below design points it states "cycle calculations show that the fuel flow is...". Is this the Brayton Cycle?

Sorry for going O.T. and thanks for a great thread.
Tony

[racket]

Hi Tony

The FT worked as intended right from the start , I can't remember a lot about the actual designing of it as it happened >20 years ago, I have one roughly to scale pencil drawing .

Theres a note on the drawing that I added 23 X 5 mm dia holes below the main Primary Zone holes on 15/4/03 to counter some burning of the louvered overlap , there appears to be another note that I needed extra small cooling holes in the primary zone wall , the 0.5 thick SS sheeting wasn't up to the job without extra wall cooling air blanket .

Nope I don't think I'd change much .

Cheers
John

[madpatty]

Jun 4, 2017 17:08:41 GMT -5 CH3NO2 said:

Hi Chris,

I had some questions on page 3 of Primary Zone Sizing.

Just below design points W3 and PR is value 0.72. What is this number?

In the paragraph below design points it states "cycle calculations show that the fuel flow is...". Is this the Brayton Cycle?

Sorry for going O.T. and thanks for a great thread.
Tony

I think 0.72 is compressor efficiency and yes its Brayton cycle.

Compressor efficiency and pressure ratio gives you the compressor discharge temperature or temperature of air entering the combustor. You have a certain Turbine Inlet Temperature (1600 F here). We have the mass flow of air and then by simple calculations of heat required to be added by the fuel to get the Air upto the Turbine Inlet Temperature, we can calculate the amount of fuel required(obviously taking into account heating value of the fuel).

Let the experts comment and correct the errors, if any😉

Cheers.

[finiteparts]

Hi Tony and Patty,

Yes, that is correct, the compressor efficiency was set at 72% and yes, it is based on the ideal Brayton cycle.

John, I hear what you are saying, but I am not sure that anything here is too "detailed"....I feel like Lefebvre's calcs are pretty low level and using the best data, such as the NASA reports on dilution hole Cds is just a good practice.

As for the mass flow, yep I totally agree and I was going to reply back that I feel it is the designers responsibility to "know" where the turbo will operate and use his best judgement for selecting the design point. But...I realized that because I didn't know the turbine A/R, I had never calculated the operating line and as such, I was not doing exactly what I was assuming that a good designer would do.

So, after some poking around, I got some data from a friend that helped me identify that I have a 0.85 AR housing (orange curve). I used some Borg Warner data to back-calculate the housing throat effective area, which was around 1.15 in^2. Before I had this info, I did a few points with a 0.68 AR housing that I had seen on some other similar turbos (red curve). I was hoping the switch to the 0.85 housing would make things better, but unfortunately with this throat size and setting the turbine inlet temperature to 1600 F, I can't get this turbo to operate over a PR = 2.5.



I set up a program that balances the compressor and turbine work, pressure ratios and the throat Mach number at the throat of the turbine scroll, assuming that the Aeff = 1.15 in^2 from above. The pressure ratio balances required the nozzle pressure ratio to be 1.03, so that there was sufficient exhaust pressure to maintain exhaust flow, but low enough to minimize the exhaust kinetic energy losses.

As shown in the operating line curves, the temperature creeps up faster as the mass flow increases. The nozzle pressure ratio was maintained at 1.03, which required that the turbine inlet temperature increased as the compressor pressure ratio increased. This makes physical sense, since you need to have more available enthalpy to match the increased power demand of the compressor pumping. But, as the gas temperature increases, the throat Mach number would get closer to unity...the throat Mach number could be reduced by increasing the engine pressure ratio, but as stated above, the turbine inlet temperature needed to climb to keep the nozzle pressure ratio up.

This ultimately lead to the upper limit of the operating curve.

So, the initially selected design point is not attainable...and the design point will have to be re-scoped. I will recheck my cycle calculations before I reset the design point, but it will likely be:

mass flow = 0.45 lbm/s
PR = 2.5
T4 = 1675 F

This shouldn't change the calculation process much, but it should allow for a much smaller combustor.

I just wanted to explain the new numbers that will be posted for the example calculations from here on.

Chris

[racket]

Hi Chris

Theres also problems with your turbine wheel exducer size as well , the ~51.3 mm dia is ~11% smaller in area than your 54 mm comp inducer , even if you could get a larger scroll A/R you might find that the turb exducer throat will be the limiting factor..............ideally the exducer area needs to be ~20% greater unless you want to clip back the exducer to "open" it up , a wheel of ~65/60 mm (ind/exd) would be about what you need for a turb wheel .

My 12/118 engine has a turb exducer area ~15% less than the comp inducer and I'm having all sorts of difficulties getting the gases through it despite the severe clipping .

Yep , without a decent turb map it gets difficult to do the design work for the combustor , we can very quickly get a long way off "design" with regards comp flow rates , which then affects the diffuser design , which affects efficiencies which then affects "actual" pressure drops across flametube , then NGV/turb wheel .........then everything further downstream .

I'd strongly recommend the Garrett turbos for anyone contemplating doing a complete design job for an engine because of the availability of the turb maps which will get us somewhere in the ball park .............flying blind ain't much fun , .....LOL, we hopefully live to regret some of our assumptions and guesstimates.

The Garrett GT3071R www.turbobygarrett.com/turbobygarrett/turbochargers/gt3071r is a similar size to your comp at 53/71mm vs your 54/67 , but it has a larger turb at 60/55 mm vs your 58/51 mm , it might be worth doing some of the Corrected vs Actual gas flows through its largest scroll A/R .

Cheers
John

[CH3NO2]

Hi John,

The turbo I am working with doesn't have a turbine map available for it. Is there a way I could map the turbine and/or compressor via test massflow, pressure and thermal measurements?

Thanks,
Tony

[CH3NO2]

Hi Chris and John,

Here is what I put together so far.  I entered the formulas into a spreadsheet for fast working of the numbers. Using the example EFR6758, a GT6041 clone on Jetspec specified diameters and a third example of the GT6041 using reference plane diameters extrapolated up to match the dynamic pressures to your EFR6758 example.

Given the differences in dynamic pressure at the reference plane, what is generally considered  minimum, maximum and optimum in this zone?  
What are the trades?
Is dynamic pressure the right metric or should it be absolute velocity?

The GT6041 clone has a compressor Inducer of 106mm.
Turbine Exducer is 119mm with a  1.39 A/R housing.
Exducer/Inducer area ratio is 126%.

[racket]

Hi Tony

You won't be flowing 200 lbs/min , that CT-106 map is "sus" , as I probably mentioned in our PMs.

The Garrett GT6041 with its 106 mm inducer only flows ~165 lbs/m at a 3.5 PR , I can't see it flowing 20% more whilst still maintaining a reasonable efficiency, besides , you won't get 200 lbs/min thru the turb stage, especially with a 1.39 A/R housing .

A 106 mm inducer only needs a 7.25" ID flametube ( 3 times inducer area ) and the outer can, which we usually try and find as a ready made "pressure vessel" , so not always "ideal" , needs to be ~9" ID to give a reasonable gap .

By all means do the calcs , but , as I mentioned in an earlier email , when building a turbo based engine there are a large number of compromises and unknowns , even when we have a map like the CT-106 one , which I'm very suspect of , and without a good turb map there are going to be lotsa guesstimate needing to be made .

Then we throw in your particular requirements of large bleed air flows , which will change the mass flow actually passing through the gap between FT and outer can altering any of your "full flow" calcs.............it'll be a "dogs breakfast" before you know it :-(

Cheers
John

[CH3NO2]

Hi John,

Yeah, I'll save the bypass experiments for later. Right now I'll just build it one step at a time and keep it simple & conventional. I have enough "dogs breakfast" on my plate for now.

Is there a feasible way to map out components of the performance(s) via test measurements?

Thanks,
Tony