Kaveri & aero-engine discussion

[madhu]

madhu, Please run it by the good doctor and see if he agrees.

Raman, I have checked the calculation against a textbook problem and it well agrees with it. one thing that is different from reality (industrial practice) is Cp and gama value varies with temp. I have taken only one value for cold section and one for hot section. however I feel this will not change a lot. may be its affet is +/- 3~4% only.

[ramana]

Exactly. So it means the sfc for a thrust of 48KN is 1.92.
Long ago I used to do similar calcs in College
And with slide rule

[madhu]

here is the calculation with AB. I have assumed the Tab to be 2000. I believe the current technology is around 2000~2100K

[Zynda]

Madhu, which books which did you use as reference? If your calculations are matching established hand calc procedures, then at least it verifies. For rough calc, an error of around 5% is acceptable.

[ramana]

So Madhu, Now that you did the calls what is your prognosis?
Is this design on path the achieve goals?
What needs to be done to get there?
What are unintended consequences of we adopt those measures?

What else we should know?

[srai]

What were the original design goals? 52kN (dry) and 81kN (wet)?

[madhu]

Madhu, which books which did you use as reference? If your calculations are matching established hand calc procedures, then at least it verifies. For rough calc, an error of around 5% is acceptable.

I cross checker it with the problem from Gas Turbine Theory by Saravanamuttoo ,‎ G.F.C. Rogers,H. Cohen‎. prob no. 123, ex. prob 3.2. and also I have referred from Aircraft Propulsion by V. Babu prob no. 155, ex. prob 3.1.

[Zynda]

^^Thx. I am not a propulsion person...just asked out of curiosity.

[madhu]

So Madhu, Now that you did the calls what is your prognosis?
Is this design on path the achieve goals?
What needs to be done to get there?
What are unintended consequences of we adopt those measures?
What else we should know?

ramana, if I do the calculation to get to 52kN we need to change the moment. the momentum is a function of mass and velocity. since the nozzle is choked I cannot change the velocity. hence I need to change the mass flow from 78kg/s to 85kg/s or I need to change the TET to 2000. I doubt that even the competitors do not have so high TET.

[JayS]

I am certain to a large degree that Amirkhan has 2000K TET or even slightly more.

How are you calculating SFC Madhu..? And whats the unit..? lb/lb-hr ?? Kaveri has 0.78 lb/hr lb dry sfc and 2.0x lb/hr-lb wet sfc.

Also can you point out in the sheet (by color coding) which values are assumed and which are calculated..?

[JayS]

So Madhu, Now that you did the calls what is your prognosis?
Is this design on path the achieve goals?
What needs to be done to get there?
What are unintended consequences of we adopt those measures?
What else we should know?

ramana, if I do the calculation to get to 52kN we need to change the moment. the momentum is a function of mass and velocity. since the nozzle is choked I cannot change the velocity. hence I need to change the mass flow from 78kg/s to 85kg/s or I need to change the TET to 2000. I doubt that even the competitors do not have so high TET.

Nozzle flow is function of throat area (exit area in this case), mass flow rate, Total pressure and total temperature. You can achieve more thrust for same mass flow rate by increasing temperature and pressure post A/B and adjusting nozzle exit area. How is your T0 and Po after A/B and before nozzle fixed..?

[madhu]

Nozzle flow is function of throat area (exit area in this case), mass flow rate, Total pressure and total temperature. You can achieve more thrust for same mass flow rate by increasing temperature and pressure post A/B and adjusting nozzle exit area. How is your T0 and Po after A/B and before nozzle fixed..?

when we use A/B we need to increase the area. the way I have done is simply used a factor. in reality the factor will be sqrt(Tab/To6). this will give me a factor of just 1.25.
Let me know if you are pointing something to me.

[madhu]

How are you calculating SFC Madhu..? And whats the unit..? lb/lb-hr ?? Kaveri has 0.78 lb/hr lb dry sfc and 2.0x lb/hr-lb wet sfc.

SFC = ma/(B+1)*(cpair*T04-Cpair*T03)/(eff_fuel*H-Cpcomb*T04).
and TSFC = Tcore/SFC

all units are in SI

[ramana]

So Madhu, Now that you did the calls what is your prognosis?
Is this design on path the achieve goals?
What needs to be done to get there?
What are unintended consequences of we adopt those measures?
What else we should know?

ramana, if I do the calculation to get to 52kN we need to change the momentum. the momentum is a function of mass and velocity. since the nozzle is choked I cannot change the velocity. hence I need to change the mass flow from 78kg/s to 85kg/s or I need to change the TET to 2000. I doubt that even the competitors do not have so high TET.

Excellent. Hope you don't think I am giving you extra work.

I want to make you an analyst and not a number cruncher.

I think we can forget increasing the TET to 2000 as that means the hot section turbine blades need exotic materials which have their own problems.

And it doesn't matter what US does as they have the infrastructure to support what they do.

So how to increase mass flow for Kaveri with that nozzle? Can that be changed?
It appears to me that this design is constrained by geometry and the early design choices.
We know the F404 has the thrust levels needed.
We know the M-88 has the same thrust levels needed.
Can I bother you to analyze the M88 and if possible F404 from Maitya's spreadsheet and tell us what's different and what's same?

[madhu]

How are you calculating SFC Madhu..? And whats the unit..? lb/lb-hr ?? Kaveri has 0.78 lb/hr lb dry sfc and 2.0x lb/hr-lb wet sfc.

SFC = ma/(B+1)*(cpair*T04-Cpair*T03)/(eff_fuel*H-Cpcomb*T04).
and TSFC = Tcore/SFC

all units are in SI

here is the correction
SFC = 1.72 kg/s [ma/(B+1)*(cpcomb*T04-Cpair*T03)/(eff_fuel*H-Cpcomb*T04)]
TSFC = 0.1301 kg/hr-N [Tdry/sfc*(3600/1000)]

so as per you we need =0.78*0.102 = 0.0796 kg/hr-N [lbm/lbf/h -> kg/hr-N: (1*0.4536)/( 1*4.45) = 0.102]

my error, sorry

[madhu]

Can I bother you to analyze the M88 and if possible F404 from Maitya's spreadsheet and tell us what's different and what's same?

Maitya's spreadsheet as I know has only M88 and not F404.

[ramana]

Could you repost the corrected numbers in the graphic?

So sfc is 1.72 kg/s with the AB on

This is used momentarily during take off and during fighter combat maneuvers.
What's the dry thrust sfc which gives us idea of endurance?

OK. Lets see M88 for now.

Folks can you gather F404 data for madhu?

[JayS]

Nozzle flow is function of throat area (exit area in this case), mass flow rate, Total pressure and total temperature. You can achieve more thrust for same mass flow rate by increasing temperature and pressure post A/B and adjusting nozzle exit area. How is your T0 and Po after A/B and before nozzle fixed..?

when we use A/B we need to increase the area. the way I have done is simply used a factor. in reality the factor will be sqrt(Tab/To6). this will give me a factor of just 1.25.
Let me know if you are pointing something to me.

Never mind. I was thinking something else while typing.
I am thinking why your SFC value is coming so different. If all SI units, your unit for TSFC should be kg/kN*sec, right..? Then 0.044kg/kN*sec is almost 2x of the Kaveri's published TSFC of 79.xx kg/kN*h. Or am I doing something wrong here..? I see you have posted some corrections in numbers.


Have you given a thought on cooling flow off-take from HPC last stage, which bypasses the combustor and gets dumped back to LPT after cooling HPT..? The off-take can be very significant portion of HPC mass flow output. What impact will this have on whole calculations..?

Its been many long years I have done these kind of calculations so apologies if I sound stupid.

[ramana]

JayS those are changes after getting the baseline correct.
First lets all agree on the baseline data.

[madhu]

So sfc is 1.72 kg/s with the AB on

this is without AB. i have used T04 values in my formula.

JayS those are changes after getting the baseline correct.First lets all agree on the baseline data.

thats good point. one question here to gurus. what kind of GT is kaveri. wikky just tells it as "leaky' turbojet" .

the bypass ratio that can be supported, even with a modest fan pressure ratio, is only about 0.16:1, which means the engine is a "'leaky' turbojet" like the F404

my calculations are based on pure turbo-fan. if Kaveri is mixed flow turbofan then we need to use one more constraint of pressure equivalence between cold flow and hot flow. not sure how to do it as most books do not talk about it.

Its been many long years I have done these kind of calculations so apologies if I sound stupid.

never mind. even i am learning. just got bored of doing FEM always. so thought of doing some calculation.

[madhu]

here is the update sheet.
I think the current calculation is not applicable to kaveri type of turbofan. these are mixed flow and we need to do some iteration to get real mass flow in fan and compressor. will try if I can do it.

[JayS]

SFC = ma/(B+1)*(cpair*T04-Cpair*T03)/(eff_fuel*H-Cpcomb*T04).
and TSFC = Tcore/SFC

all units are in SI

here is the correction
SFC = 1.72 kg/s [ma/(B+1)*(cpcomb*T04-Cpair*T03)/(eff_fuel*H-Cpcomb*T04)]
TSFC = 0.1301 kg/hr-N [Tdry/sfc*(3600/1000)]

so as per you we need =0.78*0.102 = 0.0796 kg/hr-N [lbm/lbf/h -> kg/hr-N: (1*0.4536)/( 1*4.45) = 0.102]

my error, sorry

I calculated the conversion factor between lb/hr*lbf to kg/hr*kN:
1 lb/hr*lbf = 101.978 kg/hr*kN

As per specs Kaveri should have 0.78 lb/hr*lbf = 79.54 kg/hr*kN = 0.0796 kg/hr*N

OK Now I see we are on same page.

WRT your updated sheet. Interestingly you get higher SFC for M88 than for Kaveri. Given higher OPR, BPR and TET you should have seen significantly better SFC for M88 than Kaveri with this simplified model.

Re, mixed flow and all. Just follow the streams. Keep track of hot and cold streams separately and they mix calculate avg properties considering 100% mixing efficiency to start with. The cold stream starts mixing with the hot stream right before the A/B and continues to do so until just ahead of the nozzle. This is so that it can be used as cooling blanket for the exhaust tube and nozzle parts. Else they would not stand the reheat temperatures. It is difficult to figure out what would be the temperatures just ahead of the nozzle when all mixing is complete, but I suppose we can use known wet TSFC values to calibrate the post A/B temperatures and pressures. We can use all known info as constraints and use the unknowns as varying parameters (efficiencies compressors and Turbines for example). This kind of exercise may help expose weak points in the model and tell us where fine tuning is needed. Excel's iterating calculation feature can be used I suppose for this. I once made a full 1D design sheet for stage by stage of Turbofan as an assignment in college days. Unfortunately I have lost the file. But given system level specs and some assumed parameters like efficiencies, it would give full stage by stage fluid state.

[SaiK]

4 problems described

http://www.delhidefencereview.com/2017/ ... ssion=true

The key problems encountered by the Kaveri design, according to sources who have formerly been associated with the program, are:

Unacceptable levels of fan-blade flutter risk – It seems that the Kaveri intake may need some redesign to reduce the chances of stall inducing self-excited vibrations (flutter) being experienced by the engine’s duct fan blades.


Reheat oscillations – Kaveri prototypes currently experience significant combustion oscillation in their augmentors/afterburners. This also has an impact on specific fuel consumption during reheat.


First stage low-pressure compressor blade vibration – The Kaveri’s first stage low-pressure compressor is also experiencing worrisome levels of rotor blade vibrations at the moment.


The issues delineated above have been deemed rectifiable by those in the know. But it seems outside consultancy support will be needed for the same. That, is a story for another day.

revisiting these problems, I understand the repeated words - "vibrations", and very precision engineering issue. can someone explain what is reheat oscillation and why is it a problem? (is this shock wave diamonds?)

[JayS]

Madhu,

You can use this image to cross check your calculations. This gives T and P at each module exit.

https://goo.gl/images/bFNcHd

Also you can do away with inlet efficiency. The published specs are almost always static uninstalled thrust values. As such you have assumed too low value for Kaveri vis-à-vis M88.

PS: Also, as per DRDO site, the Fan PR is 3.4 and HPC PR is 6.4. I have Kaveri's brochure at home I'll cross check numbers in that. But multiple sources show these figures so it should be correct. You have take Fan PR as 4 and with 21 OPR, HPC PR would be lower in your calculations.

PPS: Someone has uploaded Kaveri's brochure here: http://uvpcemecharocks.blogspot.se/2012 ... -full.html

[madhu]

Madhu,

You can use this image to cross check your calculations. This gives T and P at each module exit.

https://goo.gl/images/bFNcHd

i just cant believe that combustion outlet (T04) can go upto 2200C [~2400K]. do you think CM-247-LC(DS) HPT NGV can handle it for 4300 cycles?

Also you can do away with inlet efficiency. The published specs are almost always static uninstalled thrust values. As such you have assumed too low value for Kaveri vis-à-vis M88.

i have included in the formula and as ca = 0 it automatically do not consider it.

PS: Also, as per DRDO site, the Fan PR is 3.4 and HPC PR is 6.4. I have Kaveri's brochure at home I'll cross check numbers in that. But multiple sources show these figures so it should be correct. You have take Fan PR as 4 and with 21 OPR, HPC PR would be lower in your calculations.

so you are saying FPR=3.4, TET=2473K, OPR=21, ma=78 kg/s is giving me a dry thrust of 58.8 kN

[JayS]

Madhu,

You can use this image to cross check your calculations. This gives T and P at each module exit.

https://goo.gl/images/bFNcHd

i just cant believe that combustion outlet (T04) can go upto 2200C [~2400K]. do you think CM-247-LC(DS) HPT NGV can handle it for 4300 cycles?

Also you can do away with inlet efficiency. The published specs are almost always static uninstalled thrust values. As such you have assumed too low value for Kaveri vis-à-vis M88.

i have included in the formula and as ca = 0 it automatically do not consider it.

PS: Also, as per DRDO site, the Fan PR is 3.4 and HPC PR is 6.4. I have Kaveri's brochure at home I'll cross check numbers in that. But multiple sources show these figures so it should be correct. You have take Fan PR as 4 and with 21 OPR, HPC PR would be lower in your calculations.

so you are saying FPR=3.4, TET=2473K, OPR=21, ma=78 kg/s is giving me a dry thrust of 58.8 kN

Thats not combustor outlet temperature. It perhaps is max temperature inside the combustor (flame temperature). There would be a pocket of very high temperature region, surrounded by cold air which keeps the combustor from melting. Max TET for Kaveri is always mentioned as 1700K.

Change the PRs and see what happens. No need to change TET. FPR 3.4 and HPCPR = 6.4.

But you mentioned something which I have been trying to find. Lifing numbers. Do tell us what all you know, if you think its appropriate to share. I would be very interested to know module wise life in hours/cycles.

[ramana]

JayS, Can you dig up the definitive as tested values for Kaveri so we can compare madhu's calculation?

dry thrust, wet thrust, sfc at both values.
And design goals for same three parameters.

Thanks for taking the trouble.

[madhu]

Thats not combustor outlet temperature. It perhaps is max temperature inside the combustor (flame temperature). There would be a pocket of very high temperature region, surrounded by cold air which keeps the combustor from melting. Max TET for Kaveri is always mentioned as 1700K.

i did not understand what is the use of giving temperature in primary zone flame tip? other than fancy showoff to mango people to tell them how difficult it is.... it is of no use and can never be measured accurately ( we can calculated based on chemical kinetics theory).

Change the PRs and see what happens. No need to change TET. FPR 3.4 and HPCPR = 6.4.

as of now the excel is designed for FPR and OPR. i do not know how to use HPCR to calculate OPR. let me look into it if i get time.

But you mentioned something which I have been trying to find. Lifing numbers. Do tell us what all you know, if you think its appropriate to share. I would be very interested to know module wise life in hours/cycles.

there is nothing to hide as i do not work for any R&D labs. the general life numbers are
minimum design life/inspection interval for
a) cold parts : 8600 cycles
b) hot parts : 4300 cycles
c) A/B parts: 2100 cycles
d) creep life is normally around 800 hrs
d) accessories need 4000 hrs of life.
these are generic numbers and not specific to any engine. However they give us ball park numbers of what it is designed for.

[Rakesh]

Video: https://twitter.com/livefist/status/972 ... 20/video/1‬ ---> ---> As President Emmanuel Macron begins India visit, crucial $1-billion SAFRAN-DRDO deal on Kaveri jet engine ready to be signed.

https://twitter.com/livefist/status/972422572594876421 ---> Just so we’re clear, work on this has already begun with Safran Engines. The agreement is likely to finalised during this visit — the actual signing is a formality. This, as they say, is in the bag.

[thammu]

Sorry folks, no good news. NDTV reports:

....Today's joint statement says ''the leaders noted ongoing discussions between DRDO and SAFRAN on combat aircraft engine and encouraged necessary measures and forward looking approaches to facilitate [an] early conclusion.''....

[Rakesh]

Read the tweets above your posts please....

[thammu]

Read the tweets above your posts please....

It was a reply to the tweet. The joint statement was issued after the tweet.

[JayS]

Thats not combustor outlet temperature. It perhaps is max temperature inside the combustor (flame temperature). There would be a pocket of very high temperature region, surrounded by cold air which keeps the combustor from melting. Max TET for Kaveri is always mentioned as 1700K.

i did not understand what is the use of giving temperature in primary zone flame tip? other than fancy showoff to mango people to tell them how difficult it is.... it is of no use and can never be measured accurately ( we can calculated based on chemical kinetics theory).

Change the PRs and see what happens. No need to change TET. FPR 3.4 and HPCPR = 6.4.

as of now the excel is designed for FPR and OPR. i do not know how to use HPCR to calculate OPR. let me look into it if i get time.

But you mentioned something which I have been trying to find. Lifing numbers. Do tell us what all you know, if you think its appropriate to share. I would be very interested to know module wise life in hours/cycles.

there is nothing to hide as i do not work for any R&D labs. the general life numbers are
minimum design life/inspection interval for
a) cold parts : 8600 cycles
b) hot parts : 4300 cycles
c) A/B parts: 2100 cycles
d) creep life is normally around 800 hrs
d) accessories need 4000 hrs of life.
these are generic numbers and not specific to any engine. However they give us ball park numbers of what it is designed for.

What is the definition of cycle here..? Some mission or one use of full throttle..? And what would be conversion of these cycles in equivalent hours of flight..? (We almost always see no of hr as life numbers for Mil jets). I am more familiar with Civil side. Not much fo the Mil jets.

OPR = FPR x HPCPR

[Vips]

SpiceJet inks $12.5 billion deal with Safran for aircraft engines .

No-frills airline SpiceJetBSE 2.25 % today inked USD 12.5 billion deal with French major Safran Group for CFM aircraft engines.

At current exchange rate, the deal is worth more than Rs 81,000 crore and is one of the biggest in the aviation sector.

SpiceJet and Safran Group, have now finalised the purchase of LEAP-1B engines to power a total of 155 Boeing 737 MAX planes, along with spare engines to support the fleet, according to a release.

CFM engines are manufactured by CFM International, a joint venture between Safran and General Electric.

The deal was inked at the Indo- French Economic Partnership signing ceremony here on the occasion of French President Emmanuel Macron visit to India.

Not something related to Kaveri, but something that can be leveraged on?

[suryag]

Could be interlinked in the sense that spicejet is existing today thanks to this government

[madhu]

What is the definition of cycle here..? Some mission or one use of full throttle..?
I am more familiar with Civil side. Not much fo the Mil jets.

cycle is just 0-max-0. in other words from 0 to full throttle back to zero. i think we are in different page. i was referring to fatigue life. creep is in hours and fatigue is damage accumulation. off-course i am not accounting for LCF-HCF interaction or creep-fatigue ingratiation. coz this will bring down the life further.

And what would be conversion of these cycles in equivalent hours of flight..?

there is no easy way. u need to study the mission cycle and do rain-flow counting to get to the life number. to see if LCF life is consumed or not.
very crude way of calculating is as fallows
A=0 to Max dry
B=idle to max dry
C = cruse to max dry
then calculate damage accumulation as Damage = A+B/4+C/40 to this add factors and check if it is more than desired cycles for a given component.

I am more familiar with Civil side. Not much fo the Mil jets.

what do you mean? i have worked in powerGen side of GT and we do cycle counting and creep in hrs. am i missing something?

[ramana]

Miner Coffin rule for cumulative damage lefe estimation.

[JayS]

Madhu,

To elaborate on my previous post.

For engines used on civil aurlines, life is counted in terms of missions which is same as cycles. One mission or cycle could be any number of hour ranging from say 1 hr to 8hr or even more depending on what kind of aircraft we are talking about. But in simple words one set of TO-Landing makes up one cycle, cruise can be of any length. And engines would have very high life say 50000 cycles.

But for Mil aircrafts there is no standard mission. At design generally a mix of missions is used to give equivalent representative mission. But typically for life debit counting purpose one use of full throttle is considered one cycle. I have seen variations based on OEM to OEM on this one. But all of them give life numbers in hours. For ex Su30MKI's engine Al31FP has total life of 2000hrs and MTBO of 1000rs. GE F414 has something like 2000hrs life for its turbine module while its EDE version has 6000hrs life. And so on. But in actual deployment The Life tracking system counts equivalent hours of life debit for every actual mission flown using railflow counting model. Same as you mentioned. (Kaveri does have LTS using rainflow counting which is similar to what RM1w and GE F404 (and possibly F414 also) uses.

Since typically Mil jets OEMs talk in terms of no of hrs for life and you mentioned no of cycles, I asked what is the conversion from Cy to hr. And what is the definition of cycle you use.

Above all is for LCF life. Additional factors like Creep would come into picture for mil jets based on what is the exposure to highest TET. But we can let that aside for now.

When you mentioned 4300cy beforeN I thought you are talking of specifically Kaveri. I have seen one old reference, a paper from someone from GTRE which says the targeted life for Kaveri was 8000hrs. Which is very high. I have been trying to find out what are actual life numbers for Kaveri but I see no references at all. I forgot to ask this q in AI2017, and even now to exGTRE director through you, when you offered earlier.

[SaiK]

okay..(^^for my questions) I got some answers but still not clear on the exact problems with K9.

High turbulence and fast mixing would tend to produce greater magnitudes of acoustical pressure oscillations, this level of thermo-acoustic stability might cause the fluid and combustion instabilities.

or is this the issue? [**might**]

[JayS]

I had posted this previously. Read here, some generic details of the physics.

http://etd.iisc.ernet.in/handle/2005/581

We know that Kaveri has screech problem. GTRE folks claimed during AI that they have solved it.

Its nigh impossible to know exact details of issue with Kaveri unless you talk to some Aero guy in GTRE. But above thesis is as close as you can get to it since it looks like it is actually very closely related to Kaveri. The work is done in GTRE.

But All we need to know here is there are combustion instability issues ( screech to be specific) which doesnt allow AB to work at target efficiency, resulting in lower than expected wet thrust.