Degraded Su-27 aerodynamic lift

[draconus]

Idk if you heard but both ED and some 3rd parties made it clear that some aircraft, systems or weapon performance will not be simulated too close to RL counterparts, deliberately. Shocking for any sim enthusiast, I know. I appreciate when you try to made it better and look deepr but if it's not too far away from available data, obvious bug or if it's not defying physics it'll be hard to press for any changes.

[DummyCatz]

11 minutes ago, draconus said:

Idk if you heard but both ED and some 3rd parties made it clear that some aircraft, systems or weapon performance will not be simulated too close to RL counterparts, deliberately. Shocking for any sim enthusiast, I know. I appreciate when you try to made it better and look deepr but if it's not too far away from available data, obvious bug or if it's not defying physics it'll be hard to press for any changes.

Same question asked and I'll just link my answer, over at the F-18 forum where the FM is scheduled for a review and rework.

 

[DummyCatz]

21 hours ago, DummyCatz said:

Nevertheless, I can use another source and method to calculate the instantaneous performance, using the wind tunnel and flight tested Cz data, obtained by TsAGI performing a Cobra maneuver, from Figure 6 of https://arc.aiaa.org/doi/10.2514/6.1993-4737

Some basic equations:

Total Normal Force = 0.5 * ρ * V^2 * Sref * Cz

Nz (normal load factor, or Ny in Russian convention) = Total Normal Force / Weight

 

Note: Cz is not the same as CL in that Cz = - CL * cos(AOA) - CD * sin(AOA).

Now dial in the data:

Sref = 62m^2

ρ = 1.225 kg/m^3 at sea level, standard day, for example

V = 350 km/h = 97.2 m/s, at sea level for example

Weight = 20000 kg = 196200 N, for example

Cz = -2.0 at 60 deg AOA if only considering steady wind tunnel tests and not considering dynamic pitch rate effects like Cz_q. (See flight test curves)

 

So the resulting normal force = 0.5 × 1.225 × 97.2 × 97.2 × 62 × 2.0 = 717563 N

And Nz (russian Ny) = 717563 / 196200 = 3.66g, at 350km/h + 60 deg AOA + sea level, which should be the bare minimum as we're only considering steady flows. Now you're good to perform a Cobra to test it out.

Using this method I performed several tests checking if the G displayed at info bar agrees with my calculation. Firstly the track:

su27 cobra test.trk

Testing conditions are standard day, with aircraft weight at 20000kg.

And here's several screenshot explaining the test points:

1. calculated Nz = 0.5 * 1.225 * 93.7 * 93.7 * 62 * 2.0 / 196200 = 3.4g. Quite close, but this is using steady wind tunnel results and doesn't take pitch rate effects (Czq) and other dynamic effects into account.

If using the dynamic Cz from flight test result, which is around 2.8 at 60 deg AOA by counting the pixels in the image, then Nz should be 0.5 * 1.225 * 93.7 * 93.7 * 62 * 2.8 / 196200 = 4.76g, rendering DCS Su-27 significantly underperforming in dynamic pitch maneuvers.

2. calculated Nz (steady flow) = 0.5 * 1.225 * 91.55 * 91.55 * 62 * 2.0 / 196200 = 3.24g. Quite close too, but same issue as above.

calculated Nz (dynamic) = 0.5 * 1.225 * 91.55 * 91.55 * 62 * 2.8 / 196200 = 4.54g.

Edited December 18, 2024 by DummyCatz

[Xhonas]

12 hours ago, draconus said:

Idk if you heard but both ED and some 3rd parties made it clear that some aircraft, systems or weapon performance will not be simulated too close to RL counterparts, deliberately. Shocking for any sim enthusiast, I know.

That is not true, it has been proven many times that developers are using this as an excuse to not model certain things, either because it is not financially viable, lack of info or simply because some of them can't stand to be proven wrong. I remember that they used this excuse with the F/A-18 AoA, they said that they wouldn't simulate the Hornet aoa capabilities to its full extent because "classified", suddenly after some time they decided to implement its full aoa capability. (Old FM would reach 34-35° max, New FM: 55°). Similar things happened to other modules in DCS. Now i don't doubt that they might have some source that we dont know about, although very unlikely. To me it looks like they are just using that "restricted source" talk as an excuse because for them it isn't viable to change something to a FC3 product. I would love to be wrong tho.

Edited December 17, 2024 by Xhonas

[DummyCatz]

The 2nd issue is about pitching characteristics. Now referencing the Cm curve of Figure 7 from https://arc.aiaa.org/doi/10.2514/6.1993-4737

The aircraft also fails to achieve the expected maximum dynamic angle of attack (α_dyn_max) as depicted in Figure 7 when attempting the Cobra maneuver. Despite varying the center of gravity by adjusting fuel loads, the highest observed angle of attack was approximately 85°, falling short of the anticipated values for such a maneuver.

During several 1g stall tests with different center of gravity settings, the maximum sustained angle of attack (α_trim_max) stabilizes around 43-48°, contrary to ~70° in Figure 7. Above approximately 50°, the DCS aircraft demonstrates a clear nose-down pitching moment, with full aft stick. For comparison, a statically stable F-15C with CAS enabled would achieve a maximum sustained angle of attack of around 40-42° with full aft stick, raising questions about the DCS Su-27’s longitudinal static stability.

su27 1g stall test.trk

Aircraft with relaxed static stability typically exhibit marked instability in the 40-50° angle-of-attack range, driven by elevator stall and forward shift of aerodynamic center. For example the F-16, with Cm curve from NASA TP1538 (https://ntrs.nasa.gov/api/citations/19800005879/downloads/19800005879.pdf)

However, the DCS Su-27 does not show noticeable uncommanded pitch-up tendencies during 1g stalls. Releasing the stick promptly reduces the angle of attack, suggesting the presence of positive static stability, contrary to the expected behavior of a relaxed static stability aircraft. If you recall what @Yo-Yo said back then, the static margin of Su-27 should be neutral, rather than positive as of now.

What do you think?

Edited December 18, 2024 by DummyCatz

[DummyCatz]

The 3rd issue is about lateral-directional stability at 30-40° AOA where asymmetric flow separation and asymmetric vortex breakdown happens. Parameters such as Cn_beta_dyn should reflect this, according to the paper. The above track for 1g stall is sufficient to demonstrate this. And this is also a common phenomenon that is missing or ignored in other aircraft such as DCS F-18. The good side is that DCS Su-27 exhibits mild adverse yaw as well as low amplitude roll oscillation at 30-40° AOA, compared to nothing at DCS F-18, but still not enough amplitude, as explained in the paper:

I'm also not seeing asymmetric yawing moments that would cause yaw to diverge, so Cn0 and yaw stability (Cn_beta) might need checking too.

Edited December 18, 2024 by DummyCatz

[DummyCatz]

@NineLine @Yo-Yo Is this sufficient for bug reports, or do I need to create separate threads for the 3 issues above? In summary, the first is for lift/drag or Cz/Nz discrepancies during dynamic pitch maneuvers, the second is for pitch stability and third is for roll-yaw stability.

There's more to come with issues in the flight control system, but this is the first step.

[janitha2]

would this have an effect on the Su-33 also? or just the 27

[AeriaGloria]

41 minutes ago, janitha2 said:

would this have an effect on the Su-33 also? or just the 27

While there are many similarities, there is very little Su-33 documentation or information public, and there is no way to “test” its flight model except the few things that we know, that it’s draggier and then the Su-27 airframe and should have worse speed and sustained turn characteristics with positives in AOA control ability and slow speed handling 

[MA_VMF]

В 12/18/2024 в 4:50 PM, DummyCatz сказал:

The 3rd issue is about lateral-directional stability at 30-40° AOA where asymmetric flow separation and asymmetric vortex breakdown happens. Parameters such as Cn_beta_dyn should reflect this, according to the paper. The above track for 1g stall is sufficient to demonstrate this. And this is also a common phenomenon that is missing or ignored in other aircraft such as DCS F-18. The good side is that DCS Su-27 exhibits mild adverse yaw as well as low amplitude roll oscillation at 30-40° AOA, compared to nothing at DCS F-18, but still not enough amplitude, as explained in the paper:

I'm also not seeing asymmetric yawing moments that would cause yaw to diverge, so Cn0 and yaw stability (Cn_beta) might need checking too.

[DummyCatz]

15 minutes ago, MA_VMF said:

Thanks. Unfortunately DCS Su-27 is not able to attain such g and AOA. Agrees with my previous calculation.

[MA_VMF]

В 12/23/2024 в 8:35 PM, AeriaGloria сказал:

While there are many similarities, there is very little Su-33 documentation or information public, and there is no way to “test” its flight model except the few things that we know, that it’s draggier and then the Su-27 airframe and should have worse speed and sustained turn characteristics with positives in AOA control ability and slow speed handling

null

null

Edited December 26, 2024 by MA_VMF

[AeriaGloria]

10 hours ago, MA_VMF said:

null

null

Still means you end up with an aircraft that has worse turn rate sustained and instantaneous, worse acceleration, worse climb, speed, G limits. I’m sure a lot of that is due to the weight. 
 

Would be interesting to test them both in DCS at equal weight 

Edit: according to CL’s chart here, Su-27 with about 25% fuel (2350 kg, Su-33 is 2200 kg heavier empty), still beats Su-33 by a good margin 

However perhaps it is tested without emergency power. In DCS.silver.ru the Su-27 still has advantage with normal AB Su-33 but with emergency power added the Su-33 actually slightly beats it. 
https://dcs.silver.ru/165-7175,34,168,turnrate

 

 

Edited December 26, 2024 by AeriaGloria

[MA_VMF]

11 часов назад, AeriaGloria сказал:

Still means you end up with an aircraft that has worse turn rate sustained and instantaneous, worse acceleration, worse climb, speed, G limits. I’m sure a lot of that is due to the weight. 
 

Would be interesting to test them both in DCS at equal weight 

Edit: according to CL’s chart here, Su-27 with about 25% fuel (2350 kg, Su-33 is 2200 kg heavier empty), still beats Su-33 by a good margin 

However perhaps it is tested without emergency power. In DCS.silver.ru the Su-27 still has advantage with normal AB Su-33 but with emergency power added the Su-33 actually slightly beats it. 
https://dcs.silver.ru/165-7175,34,168,turnrate

 

 

It doesn't say anything about the best maneuverability characteristics.

In DCS, they are literally the same.

Edited December 27, 2024 by MA_VMF

[MA_VMF]

nullWith the same mass, the Su-33 should be superior to the Su-27 at low speeds.

[Yo-Yo]

On 12/17/2024 at 5:27 PM, DummyCatz said:

Using this method I performed several tests checking if the G displayed at info bar agrees with my calculation. Firstly the track:

su27 cobra test.trk 320.51 kB · 3 downloads

Testing conditions are standard day, with aircraft weight at 20000kg.

And here's several screenshot explaining the test points:

1. calculated Nz = 0.5 * 1.225 * 93.7 * 93.7 * 62 * 2.0 / 196200 = 3.4g. Quite close, but this is using steady wind tunnel results and doesn't take pitch rate effects (Czq) and other dynamic effects into account.

If using the dynamic Cz from flight test result, which is around 2.8 at 60 deg AOA by counting the pixels in the image, then Nz should be 0.5 * 1.225 * 93.7 * 93.7 * 62 * 2.8 / 196200 = 4.76g, rendering DCS Su-27 significantly underperforming in dynamic pitch maneuvers.

2. calculated Nz (steady flow) = 0.5 * 1.225 * 91.55 * 91.55 * 62 * 2.0 / 196200 = 3.24g. Quite close too, but same issue as above.

calculated Nz (dynamic) = 0.5 * 1.225 * 91.55 * 91.55 * 62 * 2.8 / 196200 = 4.54g.

Status bar shows acceleration in local body-fixed coordinate system. Cy in all articles and documents is given for velocity coordinate system. It is almost correct to neglect this difference for low AoA, but it is not correct to mix them in the common formula in the case of high AoA. Thus, this calculations operating with only Cy value has no sense in this case. 

[Yo-Yo]

On 12/23/2024 at 6:18 AM, DummyCatz said:

@NineLine @Yo-Yo Is this sufficient for bug reports, or do I need to create separate threads for the 3 issues above? In summary, the first is for lift/drag or Cz/Nz discrepancies during dynamic pitch maneuvers, the second is for pitch stability and third is for roll-yaw stability.

There's more to come with issues in the flight control system, but this is the first step.

I think, I answered the question in the previous post.

[DummyCatz]

On 12/29/2024 at 9:19 AM, Yo-Yo said:

Status bar shows acceleration in local body-fixed coordinate system. Cy in all articles and documents is given for velocity coordinate system. It is almost correct to neglect this difference for low AoA, but it is not correct to mix them in the common formula in the case of high AoA. Thus, this calculations operating with only Cy value has no sense in this case. 

It is clearly Cz (normal force coefficient) in the TsAGI paper, which IS defined in the body Z axis, not wind axes frame.

It is not CL (lift coefficient), which is in the wind Z axis. As I already said, Cz = - CL * cos(AOA) - CD * sin(AOA). The paper is using western conventions of flight dynamic symbols when published at AIAA.

 

CL and Cz can be easily differentiable by looking at whether the curve drops at AOA higher than 35 degrees (usually where CLmax happens). For example the NASA F-18 data (https://ntrs.nasa.gov/api/citations/19900019262/downloads/19900019262.pdf)

In comparison, the Su-27 (TsAGI data):

So my calculation is correct as is. But the first issue depicted in my post is really about whether DCS has implemented dynamic, non steady flow effects such as pitch rate (q) or AOA rate (alpha_dot) contributions to Cz, so that the total Cz is higher than steady wind tunnel data.

 

18 hours ago, Yo-Yo said:

I think, I answered the question in the previous post.

It’s too long a thread. Would you elaborate?

Edited December 30, 2024 by DummyCatz

[Yo-Yo]

I agree that almost constant C coeff up to 90-90 degree of AoA is a clear sign of body-fixed system, so your calculations were correct. I have no access to the article you refer to, and I would like to have a look at the whole  article, because discussing the thing using only excerpts can not be very fruitful. For example, CoG position of the plane that was tested for this article (earlier model) can be different (more aft) from the plane that was modelled in DCS (later variants). So, any discussion of issue related to trim are useless without knowing the exact CoG position. And, by the way, I can not admit ~70 of steady trim angle shown in the article because of trim  data I have. Referring to quite different F-16 aerodynamics is not very useful in this case. So, 40-50 degrees regarding CoG is correct value.
When developing the model, the primary objective was to achieve the most accurate representation of steady-state flight regimes, as these regimes, being the most prolonged in duration, have the greatest impact on the flight trajectory and the maneuvering characteristics of the aircraft.  Dynamic changes in lift coefficient during rapid changes in angle of attack are a phenomenon well-known at least since the 1940s. This behavior occurs universally and is not limited to aircraft such as the Su-27 or MiG-29, as might be inferred from the article by G.I. Zagaynov. We could theoretically use this effect in flight modeling (FM), the mechanism is no mystery to us. However, before demanding this from developers, consider the following questions:

1. To what extent would a trajectory or a trajectory parameters differ from calculations using a static lift coefficient if a very brief 30% increase in load factor occurs during an increase in angle of attack, followed by a subsequent decrease when the angle reduces, given that velocity is an integral and displacement is the double integral of acceleration?
2. Given that reliable data on the hysteresis loop shape for the lift coefficient (Cy) is available for only a handful of aircraft, do you believe applying unsteady aerodynamics exclusively to these aircraft is fair to others? Alternatively, would it be valid to use data obtained, for example, for the Su-27 to model the F-18?

[DummyCatz]

3 hours ago, Yo-Yo said:

To what extent would a trajectory or a trajectory parameters differ from calculations using a static lift coefficient if a very brief 30% increase in load factor occurs during an increase in angle of attack, followed by a subsequent decrease when the angle reduces, given that velocity is an integral and displacement is the double integral of acceleration?

Thank you for your prompt response and Happy New Year. I appreciated the detailed points you’ve raised.

While your point about the cumulative impact being minor for prolonged flight path is valid, dynamic pitch pointing (and unloading) capabilities would certainly benefit from the 30% increase/decrease in Ny. This would also explain why the data graph of performing a Cobra posted by MA_VMF above doesn't agree with the Ny and/or AOA data seen in DCS, which could be a real question posed by other users who would love to compare such data and diagrams.

3 hours ago, Yo-Yo said:

Given that reliable data on the hysteresis loop shape for the lift coefficient (Cy) is available for only a handful of aircraft, do you believe applying unsteady aerodynamics exclusively to these aircraft is fair to others? Alternatively, would it be valid to use data obtained, for example, for the Su-27 to model the F-18?

I think this is more of a question as of to what extent the FM realism should achieve in individual aircraft. Modules like the F-16 and F-18 also benefit from extensive public data (e.g. NASA wind tunnel data for both the F-16 and F-18), which can create a disparity in FM fidelity. The NASA public data also include aerodynamic coefficient derivatives such as Czq and Cz_alpha_dot (for example in NASA TM 107601 for the F-18), which inherently places them at an advantage in terms of potential FM realism, if you consider fairness. I would personally prefer realism to fairness as long as the price are justified. But it's your decision afterall.

3 hours ago, Yo-Yo said:

I agree that almost constant C coeff up to 90-90 degree of AoA is a clear sign of body-fixed system, so your calculations were correct. I have no access to the article you refer to, and I would like to have a look at the whole  article, because discussing the thing using only excerpts can not be very fruitful. For example, CoG position of the plane that was tested for this article (earlier model) can be different (more aft) from the plane that was modelled in DCS (later variants). So, any discussion of issue related to trim are useless without knowing the exact CoG position. And, by the way, I can not admit ~70 of steady trim angle shown in the article because of trim  data I have. Referring to quite different F-16 aerodynamics is not very useful in this case. So, 40-50 degrees regarding CoG is correct value.

You mentioned discrepancies in the trim data (~70° α_trim_max from the article vs. ~40-50° in DCS). As a user, I cannot independently verify the trim data used in the Su-27 FM. While the TsAGI’s article shows higher trim angles, which of course can be indicative of a very aft CoG, as there's no CoG data shown in the article. But there's other points that I based my speculations on:

There's uncommanded pitch-down behavior at negative AOA in DCS Su-27, which is a very noticeable unstability. I'm using this as a basic feeling to the negative static margin.
Observations from 1g stall tests in DCS suggest the Su-27 exhibits behavior closer to a statically stable aircraft, if compared to a F-15, as there's no noticeable uncommanded pitch motions.This behavior aligns with positive static stability characteristics rather than the neutral or slightly negative stability expected from a relaxed static stability design like the Su-27.

I mentioned the F-16 is because as a typical behavior, it exhibits uncommanded pitch-up with a neutral elevator at 40-50° AOA (and the F-18 too as there's a Cm curve out there), due to the elevator stall, if kept at neutral position, and would cause the aerodynamic center to shift forward, resulting in a decrease in static margin. This phenomenon is pretty common to those aircraft, but I won't know the behavior of Su-27 without any data for sure.

 

Appendix. Cm curve of F-18C from https://www.scribd.com/document/721840808/F-18-HORNET-S-NEST

[MA_VMF]

В 12/31/2024 в 4:08 AM, Yo-Yo сказал:

Например, положение центра тяжести самолёта, который тестировался для этой статьи (более ранняя модель), может отличаться (находиться дальше от хвостовой части) от самолёта, смоделированного в DCS (более поздние варианты

В ДКС смоделирован франкенштейн из самолетов разных серий 

[MA_VMF]

null

[MA_VMF]

null

[GT400]

I have a few questions about Su-27, and my apologies in that I cannot understand Cyrillic.  But in TsAGI Su-27S paper, what does 1.6, 1.08, 1.4 numbers represent for Su-27, F-15, F-16? 

 

Also, so once Su-27S and SK are are equal or less than 19,000kg they can both pull 9G?  171000/9=19000kg and as long as it's less than mach .85?

Lastly,  I get it that in the TsAGI report it lists 18920kg as design flight mass at 50% fuel reserve,  but does any one know the exact amount of fuel that is and the Su-27S operating weight?  (easier to write it as operating weight?+fuel?= 18920kg)

 

Thank you.

 

Edited January 3 by GT400

[AeriaGloria]

3 hours ago, GT400 said:

I have a few questions about Su-27, and my apologies in that I cannot understand Cyrillic.  But in TsAGI Su-27S paper, what does 1.6, 1.08, 1.4 numbers represent for Su-27, F-15, F-16? 

 

Also, so once Su-27S and SK are are equal or less than 19,000kg they can both pull 9G?  171000/9=19000kg and as long as it's less than mach .85?

Lastly,  I get it that in the TsAGI report it lists 18920kg as design flight mass at 50% fuel reserve,  but does any one know the exact amount of fuel that is and the Su-27S operating weight?  (easier to write it as operating weight?+fuel?= 18920kg)

 

Thank you.

 

4700 kg Is 50%