Su-27 Flight Model Discussion

[Esac_mirmidon]

Thanks to all. Glad to help.

 

Yes, the spanish version is mine but only for the first part of the manual. The second part is not translated yet because the work needed for the level of detail i want is SO huge that i get exhausted withe first.

 

Maybe one day i will start the second part.

[DarkFire]

Thanks to all. Glad to help.

 

Yes, the spanish version is mine but only for the first part of the manual. The second part is not translated yet because the work needed for the level of detail i want is SO huge that i get exhausted withe first.

 

Maybe one day i will start the second part.

 

Great job even doing that much :thumbup:

 

Ironhand - thanks for the clarification. Interesting that the AB fuel drain regime is different. This is pure supposition on my part but I would suggest that maybe the fuel drain is from tank 2 / feeder tank only, as that has the least probability of introducing any weight distribution imbalances by being rapidly drained. Similarly I would suggest that fuel dump may be sourced from tank 4 as it's the closest to the dump nozzle.

 

This calls in to question something else that I had believed: I'd always thought that the "standard" 60%-of-maximum fuel load for the Flanker was comprised of tanks 1 & 4, but I don't see how this could be true given the flow diagram, unless tanks 1 & 4 feed directly in to the cache tank and not in to tank 2, but this layout seems unnecessarily complicated to me in terms of plumbing. Unless filling tanks 2 & 4 would introduce an unacceptable shift in the CG maybe,

[Esac_mirmidon]

In the next paragraph of the manual the comsumption sequence is:

 

FOR A BASIC REFUELED AIRCRAFT:

 

5º.- Tank Nº3 ( both symetric ) completely.

6º.- 3210 kg from tank Nº2 ( except the content of the central feeder )

7º.- The rest of the Tank Nº2 completely.

 

If we take in count the Kg weight of every Tank:

 

№ 1 – 3180 Kg

№ 2 – 4160 Kg

№ 3 – 1060 Kg

№ 4 – 1000 Kg

TOTAL – 9400 Kg

 

Tanks Nº3 + Nº2 = 5160 Kg.

Total: 9400 Kg

 

Difference: 4240 Kg.

 

This 5160 Kg is almost 60% of max fuel load ( 55%), so i think for that 60% fuel load the fuel is used from Tanks 3 and 2, not 1 and 4.

[DarkFire]

In the next paragraph of the manual the comsumption sequence is:

 

FOR A BASIC REFUELED AIRCRAFT:

 

5º.- Tank Nº3 ( both symetric ) completely.

6º.- 3210 kg from tank Nº2 ( except the content of the central feeder )

7º.- The rest of the Tank Nº2 completely.

 

If we take in count the Kg weight of every Tank:

 

№ 1 – 3180 Kg

№ 2 – 4160 Kg

№ 3 – 1060 Kg

№ 4 – 1000 Kg

TOTAL – 9400 Kg

 

Tanks Nº3 + Nº2 = 5160 Kg.

Total: 9400 Kg

 

Difference: 4240 Kg.

 

This 5160 Kg is almost 60% of max fuel load ( 55%), so i think for that 60% fuel load the fuel is used from Tanks 3 and 2, not 1 and 4.

 

That makes complete sense. Wing tanks + centre tank means no chance in CG as the fuel is depleted.

[Ironhand]

That makes complete sense. Wing tanks + centre tank means no chance in CG as the fuel is depleted.

:) Maybe yes, maybe no. There are some changes in the airframe's behavior that I'm not willing to comment on yet. EDIT: This is not meant to suggest that something is wrong, though. Just that there are changes. :EDIT ENDS

 

Also seeing a definite change in behavior at about the time the drawdown of Tank 4 initiates which ends well before the drawdown completes. It seems to coincide with a distinct rise is the nose that trim can no longer accomodate when flying level at speed under full mil power with the FBW fully engaged. As soon as you are through that "burble" you again have sufficient nose down trim available. But I need to look at this stuff more closely.

Edited July 29, 2018 by Ironhand

[DarkFire]

:) Maybe yes, maybe no. There are some changes in the airframe's behavior that I'm not willing to comment on yet. EDIT: This is not meant to suggest that something is wrong, though. Just that there are changes. :EDIT ENDS

 

Also seeing a definite change in behavior at about the time the drawdown of Tank 4 initiates which ends well before the drawdown completes. It seems to coincide with a distinct rise is the nose that trim can no longer accomodate when flying level at speed under full mil power with the FBW fully engaged. As soon as you are through that "burble" you again have sufficient nose down trim available. But I need to look at this stuff more closely.

 

There's another trim oddity which normally manifests at low altitude. The trim behaviour normal up to about 780 Km/h. As IAS increases through 800 the trim effect inverts a little and the nose starts to dip but then the effect reverses again to normal behaviour as IAS increases through 890 Km/h.

 

What sort of altitude are you testing at? I always put the pitch trim anomaly at 800-875 Km/h IAS down to an anomaly in the control laws applied to pitch trim, but It could conceivably be be that the two effects are in fact linked, depending on what IAS full military thrust produces at a given altitude.

 

Edited to add: just gave this a try. Standard DCS day. Flew at 5,000 m altitude. 100% dry RPM gave me 1150 Km/h TAS and 890 Km/h IAS. If I've understood the drain sequence correctly tank 4 should start to drain at 7,680 Kg total fuel. I didn't observe any anomalous changes in trim, at least within the limits of my ability to manually maintain zero trim.

Edited July 30, 2018 by DarkFire

[Ironhand]

...

What sort of altitude are you testing at? ...

 

Edited to add: just gave this a try. Standard DCS day. Flew at 5,000 m altitude. 100% dry RPM gave me 1150 Km/h TAS and 890 Km/h IAS. If I've understood the drain sequence correctly tank 4 should start to drain at 7,680 Kg total fuel. I didn't observe any anomalous changes in trim, at least within the limits of my ability to manually maintain zero trim.

I was actually just doing some flying at the time checking something else but using the mission I had made for testing. So I would have been somewhere between 2500-3000 meters and was flying at full mil and had nudged up to about 1040 kph, when it happened. I thought: That's odd. Then I checked my fuel quantity and saw a familiar number 7400. After a little more than 100 kg or so, everything was back to normal. I'd say the entire sequence spanned roughly 300 kg of fuel.

[Esac_mirmidon]

Well if we take in count that the first tank to be used is Nº1 in the forward fuselage section, and 9400 Kg total capacity - 1720 kg consumed first from Tank Nº1 is = 7680 Kg, this behaviour makes a lot of sense.

[Ironhand]

Well if we take in count that the first tank to be used is Nº1 in the forward fuselage section, and 9400 Kg total capacity - 1720 kg consumed first from Tank Nº1 is = 7680 Kg, this behaviour makes a lot of sense.

Not sure it does. The data I have suggest a shift to the rear, while the No 4 tank is being emptied which shouldn't be happening.

 

I'll try to put what I have together over the next few days and post it for you folks to look at and experiment with.

 

I also just completed a l-o-n-g 3-hour direct control flight trimmed for level flight. Fortunately, once I was at the right airspeed for neutral trim, I could leave the computer and go on with life while the simulation ran it's course. I was surprised to see that I was still airborn all that time later, when I checked in. Watched the fuel run out and, in the end, had my only bit of fun with a deadstick landing.

 

I'll have to run through it to see what there is to see. I suspect that, since nothing upset the flight in my absence, there'll be nothing to jump up and down about.

[Esac_mirmidon]

Thanks for the time invested mate.

[DarkFire]

Hmm, I tried this again and I think that my manual trim control isn't sufficiently accurate to be able to differentiate between divergent vertical velocity changes due to non-neutral trim and any changes that are occurring due to shifts in CG due to fuel depletion. However small a nudge I give the trim adjustment it eventually enters a runaway divergence from neutral.

 

There were very subtle hints at possible trim changes at roughly between 7900 - 7750 Kg of fuel loaded and maybe another at some point between 7,500 and 7,300 Kg fuel but again I can't trim accurately enough to tell if these are simply symptomatic of non-neutral trim or subtle CG changes.

 

I would imagine that the way to do it would be to use alt hold and then watch for changes in the trim position commanded by the ACS, but without having the sort of access to the FM that YoYo would have, I'm not sure there's any way for us to get that data. I tried watching the control monitor graphic but the changes in ACS-commanded trim position are just too small to see on that small graphic.

 

Anyway, the relevant track is attached.

 

Edited to add: when testing, are we using a full gunpad? IIRC that would be an additional 147Kg of ammunition that should be forwards of the CG. Not sure if it makes a difference here in the Su-27 under these test conditions, but given that the potential trim anomaly happens during consumption of a similarly small amount of fuel it makes sense to standardise our tests. I had a full ammunition load for the gun in the attached test track.

Su-27 Fuelk Test 1 Track.zip

Edited July 30, 2018 by DarkFire

[Ironhand]

DF,

 

I thought I had posted this but I guess I didn’t. I have not been able to reproduce the untrimmable pitch up @ 7400 kg. So far I’ve only been able to reproduce it during high AoA flight. I’m curious what the trimmed direct control flight will show. I can tell you, however, I will not be sampling data every 30 seconds like I did with the high alpha flight(s).

[DarkFire]

DF,

 

I thought I had posted this but I guess I didn’t. I have not been able to reproduce the untrimmable pitch up @ 7400 kg. So far I’ve only been able to reproduce it during high AoA flight. I’m curious what the trimmed direct control flight will show. I can tell you, however, I will not be sampling data every 30 seconds like I did with the high alpha flight(s).

 

30 seconds? Damn, that must have resulted in the world's biggest spreadsheet :shocking:

 

I need to have a look at the in-game lua engine documentation to see if it's possible to script data output to a log file.

 

In terms of relatively high AOA flight, I remember there being a region of stability up at around 13-14,000m at what used to be cruise throttle (87% RPM I think) which would settle at IIRC about 240 Km/h IAS and around 8 degrees AOA. The problem will be getting up there before tank 4 starts to empty. It'll be interesting to try though. Might give this a bash later.

 

Edited to add: Of course it'll be possible. Air start... :doh:

Edited July 30, 2018 by DarkFire

[Ironhand]

DF,

 

:) When I said high AoA flight, I was talking about AoA in the 40’s.

 

Do look at that LUA. I’d love to simplify the process.

Edited July 30, 2018 by Ironhand

[Ironhand]

...

 

Edited to add: when testing, are we using a full gunpad? IIRC that would be an additional 147Kg of ammunition that should be forwards of the CG. Not sure if it makes a difference here in the Su-27 under these test conditions, but given that the potential trim anomaly happens during consumption of a similarly small amount of fuel it makes sense to standardise our tests. I had a full ammunition load for the gun in the attached test track.

Yes. I've changed nothing internally except the fuel.

[DarkFire]

DF,

 

:) When I said high AoA flight, I was talking about AoA in the 40’s.

 

Do look at that LUA. I’d love to simplify the process.

 

Ah I see! Slightly different flight regime to what I was testing :)

 

Reading up about the data export abilities of the in-game LUA engine, I'm not sure that what we want to achieve will be possible. It's obviously a very powerful tool but there's no obvious functions for accessing the guts of the flight models. In fairness this may be an entirely deliberate step by ED to help prevent others from simply copying their R&D work on flight models.

 

It looks like it is possible to output information such as pitch angle and what I think is the total velocity vector of a given unit but I'm not sure that data would help to asses potential anomalies in CG / CP behaviour. Unless I've missed something or there are undocumented lua functions I think getting that sort of data is probably only possible at the development / code testing level.

[Ironhand]

Ah I see! Slightly different flight regime to what I was testing :)

 

Reading up about the data export abilities of the in-game LUA engine, I'm not sure that what we want to achieve will be possible. It's obviously a very powerful tool but there's no obvious functions for accessing the guts of the flight models. In fairness this may be an entirely deliberate step by ED to help prevent others from simply copying their R&D work on flight models.

 

It looks like it is possible to output information such as pitch angle and what I think is the total velocity vector of a given unit but I'm not sure that data would help to asses potential anomalies in CG / CP behaviour. Unless I've missed something or there are undocumented lua functions I think getting that sort of data is probably only possible at the development / code testing level.

:) Just slightly different...

 

No problem on the LUA stuff. I've decided to "cheat" and use Tacview for some of this. I don't fully trust it but it seems to mirror what I'm seeing accurately so far. With some Tacview results and a spreadsheet for one of the flights, you'll be able to get a good sense of what's going on. I'll try to have something for you folks to look at and verify (or not) over the weekend.

 

...

 

Edited to add: when testing, are we using a full gunpad? IIRC that would be an additional 147Kg of ammunition that should be forwards of the CG. Not sure if it makes a difference here in the Su-27 under these test conditions, but given that the potential trim anomaly happens during consumption of a similarly small amount of fuel it makes sense to standardise our tests. I had a full ammunition load for the gun in the attached test track.

Thinking about it, this might help explain an oddity I've noticed. Once you're down to the final 1200 kg or so of fuel, the AoA reduces. Think I'll try to make another high AoA flight without ammo and see what happens.

Edited August 1, 2018 by Ironhand

[DarkFire]

I actually like Tacview for this sort of thing. I think the readings are relatively accurate. The only issue is the relatively low (for this sort of testing) sample rate, but I guess it has to be otherwise .acmi files would be gigantic for missions of any sort of real duration. Hmm, I might suggest a "test" mode for TacView which could be 2x or even 5x sample rate but limited in track length.

[Ironhand]

I actually like Tacview for this sort of thing. I think the readings are relatively accurate. The only issue is the relatively low (for this sort of testing) sample rate, but I guess it has to be otherwise .acmi files would be gigantic for missions of any sort of real duration. Hmm, I might suggest a "test" mode for TacView which could be 2x or even 5x sample rate but limited in track length.

 

Tacview has its uses. A higher sampling option might be good. OTOH, the tracks I’m working with are 45 minutes to 3 hours long. So...

 

Anyway, I want to make another high-AoA flight starting with full fuel and running it down to zero but no ammo. It’ll be a nice double check on the data.

[Ironhand]

Tacview has its uses. A higher sampling option might be good. OTOH, the tracks I’m working with are 45 minutes to 3 hours long. So...

 

Anyway, I want to make another high-AoA flight starting with full fuel and running it down to zero but no ammo. It’ll be a nice double check on the data.

And after this flight, I can say that the internal stores have little, if any effect. I'm still seeing an abrupt drop in AoA once you have less than roughly 1200 kg of fuel remaining--the drop is small-- roughly 3° but takes place consistently. There are, in fact, several consistent abrupt changes taking place.

[DarkFire]

Interesting! Looking forwards to having a look at your track if it's not too huge to upload.

 

I've done a couple of tests now in level flight at a couple of different altitudes (~2,750m, 5,000m & 7,500m) and so far I've not come across anything that I can definitively say was outside of the experimental uncertainty produced by my lack of ability to maintain dead neutral pitch trim.

 

I'll see if I can put some time in over the weekend to test very high altitudes where any variance will hopefully be amplified by the very low IAS.

[Ironhand]

Interesting! Looking forwards to having a look at your track if it's not too huge to upload.

 

I've done a couple of tests now in level flight at a couple of different altitudes (~2,750m, 5,000m & 7,500m) and so far I've not come across anything that I can definitively say was outside of the experimental uncertainty produced by my lack of ability to maintain dead neutral pitch trim.

 

I'll see if I can put some time in over the weekend to test very high altitudes where any variance will hopefully be amplified by the very low IAS.

Don't go crazy with this just yet. Wait until I post the data I've collected along with my conclusions. Then you'll have something more definitive to go on. Some things might bear looking into. Others may not be.

 

If you want to have some fun, start with 9400 kg of fuel at 3000 meters with a TAS of 350. Take direct control of the aircraft, pull up into a stall while throttling up to full mil. Keep the stick in your lap for the next 52+ minutes standing on your tail, while using the rudders to hold your heading as the airframe shakes and rattles around you. Somewhere around 300 meters altitude--about the time you really start to worry that you'll run out of altitude--your thrust to mass ratio will become positive and you'll start climbing again. :)

[DarkFire]

Don't go crazy with this just yet. Wait until I post the data I've collected along with my conclusions. Then you'll have something more definitive to go on. Some things might bear looking into. Others may not be.

 

If you want to have some fun, start with 9400 kg of fuel at 3000 meters with a TAS of 350. Take direct control of the aircraft, pull up into a stall while throttling up to full mil. Keep the stick in your lap for the next 52+ minutes standing on your tail, while using the rudders to hold your heading as the airframe shakes and rattles around you. Somewhere around 300 meters altitude--about the time you really start to worry that you'll run out of altitude--your thrust to mass ratio will become positive and you'll start climbing again. :)

 

Well that's damned hard work to keep it there for the entire fuel load duration, but good fun! I'm going to play around with this over the weekend & see what the permissible ambient temperature limits are that make this possible.

[Ironhand]

Well that's damned hard work to keep it there for the entire fuel load duration, but good fun! I'm going to play around with this over the weekend & see what the permissible ambient temperature limits are that make this possible.

It does require a lot of concentration and wrist strength. ????

 

My guess, unless you increase the starting altitude will be about 30C. Maybe less if you reset the trim to neutral before pulling up into the near stall.

[Iriya]

Not sure about IRL but the flight model of Su-27 makes me feel controlling a brick. I mean, Su-27 loses air speed so quick when I reduce fuel flow which makes easier to stay in formation comparing to other aircraft. If it's correct as it be I'm totally happy with it. On the other hand, the flight model of MiG-29 is excellent thanks to update around 2018. I understand Su-27 is much heavier than MiG-29 and totally different aircraft by any aspect but still.