Forum: General F-22A Raptor forum

shingen wrote:

That's why even if we get the wing loading it doesn't do us any good. We need the Ps diagrams which we're not going to get.

How close could someone get with a 3D model and MATLAB?

I've seen 3D aerodynamic modeling of design concepts, as well as existing jets. I still believe the F-22 has the best aerodynamics of any production fighter. I would be surprised if the T-50 or J-20 had less form drag than a F-22

The Su-27/30/35 partially compensate with more engine thrust.

F-16.net Sponsor

neurotech wrote:

shingen wrote:

That's why even if we get the wing loading it doesn't do us any good. We need the Ps diagrams which we're not going to get.

How close could someone get with a 3D model and MATLAB?

I've seen 3D aerodynamic modeling of design concepts, as well as existing jets. I still believe the F-22 has the best aerodynamics of any production fighter. I would be surprised if the T-50 or J-20 had less form drag than a F-22

The Su-27/30/35 partially compensate with more engine thrust.

It's not that simple, CFD is very difficult to get to work and even harder to get to work right (just ask the Super Hornet or F-35 development teams). The complex shapes of the aircraft and the strange flow condition during maneuvering (ie. high angles of attack) makes this a very difficult proposition.

It's like debating who has the best aircraft gun at this point. What makes these systems work is networking. The Boogeyman is alleged to be able to completely shut down US networking and then win the good ol' fashioned way.

If we want to have a discussion about T-50 aero we have some problems.

At this point we don't have:
Weight
Thrust
Dimensions

What we do have is:
Sweep angle
General layout
Size of stabs

What can we learn from these?

I find it interesting that they have widely spaced engines and small vertical stabs. Do they not care about safety or is there another way to deal with an engine out at high speed?

The thing I would like an answer on is how different the internal structure is from the T-10 and a good RCS estimate.

If the RCS is really .2m^2 they should have built the MiG 1.42.

Wasn't the wide engine spacing in order to give the trust vectoring some amount of roll control?

shingen wrote:

The thing I would like an answer on is how different the internal structure is from the T-10 and a good RCS estimate.

If the RCS is really .2m^2 they should have built the MiG 1.42.

The lowest RCS figure for the Super Hornet is 0.1m2, but other sources put it up to 1.0 m2.

The Su-35 apparently has an RCS of 1.0-3.0 m2. Some reports indicate the Su-35 uses a "ferromagetic RAM coating" on the compressor blades to reduce the turbofans returns. I'm not sure I'd believe that is nearly as effective compared to the ducting of the F-22 or the Super Hornet.

sirsapo wrote:

neurotech wrote:

shingen wrote:

That's why even if we get the wing loading it doesn't do us any good. We need the Ps diagrams which we're not going to get.

How close could someone get with a 3D model and MATLAB?

I've seen 3D aerodynamic modeling of design concepts, as well as existing jets. I still believe the F-22 has the best aerodynamics of any production fighter. I would be surprised if the T-50 or J-20 had less form drag than a F-22

The Su-27/30/35 partially compensate with more engine thrust.

It's not that simple, CFD is very difficult to get to work and even harder to get to work right (just ask the Super Hornet or F-35 development teams). The complex shapes of the aircraft and the strange flow condition during maneuvering (ie. high angles of attack) makes this a very difficult proposition.

The Super Hornet had problems with wing-drop early on. This was fixed with minor airframe changes and flight control software updates. I'd assume think CFD has improved since the mid-90s when the SH was being developed.

Does anyone have a moderately accurate 3D model of the T-50?

shingen wrote:

It's like debating who has the best aircraft gun at this point. What makes these systems work is networking. The Boogeyman is alleged to be able to completely shut down US networking and then win the good ol' fashioned way.

If we want to have a discussion about T-50 aero we have some problems.

At this point we don't have:
Weight
Thrust
Dimensions

What we do have is:
Sweep angle
General layout
Size of stabs

What can we learn from these?

I find it interesting that they have widely spaced engines and small vertical stabs. Do they not care about safety or is there another way to deal with an engine out at high speed?

The thing I would like an answer on is how different the internal structure is from the T-10 and a good RCS estimate.

If the RCS is really .2m^2 they should have built the MiG 1.42.

The vertical stabs do appear to be all moving. As a result they won't suffer from moments ordinarily generated by discrete control flaps generating shock waves and imparting a twist to lower the effective alpha of the entire stab surface. The YF-23 used a similar tail configuration in that regard. Still, those tails are still quite small and I will agree that making them all-moving won't mean much when they're too small to really do anything useful during a flame out or inlet unstart.

I'm with neurotech in the sense that I can't imagine many fighters out there that have the F-22 beaten in terms of aerodynamics.

neurotech wrote:

The Super Hornet had problems with wing-drop early on. This was fixed with minor airframe changes and flight control software updates. I'd assume think CFD has improved since the mid-90s when the SH was being developed.

Does anyone have a moderately accurate 3D model of the T-50?

I was refering to CFD failing to properly predict stores separation on the Super Hornet, resulting in those draggy outward canted inner pylons. But you bring up the transonic roll-off issue, which highlights how hard this stuff really is, since even the F-35 has been having some trouble with it, and they have the most modern setup so far.

The problem is that during any maneuvering, the angle of attack gets high enough that all the simple methods of fluid analysis don't work any more, and you end up having to use very complex techniques (and therefore super computers) to get any useful data. The USAF has an entire program dedicated to doing CFD for stuff like stores separation and whatnot called Seek Eagle, and even they get this stuff wrong 50% of the time.

Bottom line: if you want to do enough analysis on a whole maneuvering airplane to build a Ps chart, you're going to need a super computer and a PhD, or else your numbers wont mean a thing.

count_to_10 wrote:

Wasn't the wide engine spacing in order to give the trust vectoring some amount of roll control?

I asked johnwill about that once. He said it was unlikely because TVC can only control role along the aircraft axis, not flight axis. His guess was that they went with that arrangement because it was what they were used to already (see Su-27 and Mig-29).

1st503rdsgt wrote:

count_to_10 wrote:

Wasn't the wide engine spacing in order to give the trust vectoring some amount of roll control?

I asked johnwill about that once. He said it was unlikely because TVC can only control role along the aircraft axis, not flight axis. His guess was that they went with that arrangement because it was what they were used to already (see Su-27 and Mig-29).

I think he's right on that as well. The T-50 is not a modified Su-27 in the same way an Su-35 is just a modified Su-27. But it's clearly still based on the Su-27's design in many ways, the most obvious being the engines/intakes/center tunnel layout. I think they just couldn't afford to start over from scratch.

Roll control is also not useless by any means, but it's not something that really needs thrust vectoring either. Rolling about your own axis won't get your plane out of someone else's gunsight; pitching the plane nose-up and generating more lift relative to your direction of flight will. Plus you'll get more efficient rolling by moving control surfaces way out on the tips of the wings, rather than trading thrust for torque much closer to the centerline of the plane. Thrust-vectored roll control is something Sukhoi realized they had in the Su-3x due to the existing Su-27 design, and therefore hype as a marketing point, not something they designed into the Su-27 anticipating they'd eventually put thrust vectoring nozzles on it.

TVC roll control can be useful to a limited extend at speeds where your control surfaces are less or even ineffective or to minimise control surface deflections in order to reduce trim drag.

sirsapo wrote:

neurotech wrote:

The Super Hornet had problems with wing-drop early on. This was fixed with minor airframe changes and flight control software updates. I'd assume think CFD has improved since the mid-90s when the SH was being developed.

Does anyone have a moderately accurate 3D model of the T-50?

I was refering to CFD failing to properly predict stores separation on the Super Hornet, resulting in those draggy outward canted inner pylons. But you bring up the transonic roll-off issue, which highlights how hard this stuff really is, since even the F-35 has been having some trouble with it, and they have the most modern setup so far.

There is plans by Boeing aerodynamic improvements to the Super Hornet to reduce drag and improve stores separation. The Navy has not ordered this version, but is offered to Brazil or possibly Canada.

sirsapo wrote:

The problem is that during any maneuvering, the angle of attack gets high enough that all the simple methods of fluid analysis don't work any more, and you end up having to use very complex techniques (and therefore super computers) to get any useful data. The USAF has an entire program dedicated to doing CFD for stuff like stores separation and whatnot called Seek Eagle, and even they get this stuff wrong 50% of the time.

Stores separation CFD is significantly more complex than calculating the Ps diagram of a clean jet.

sirsapo wrote:

Bottom line: if you want to do enough analysis on a whole maneuvering airplane to build a Ps chart, you're going to need a super computer and a PhD, or else your numbers wont mean a thing.

GPGPU technlogy is improving so that a high-end desktop can do what would require super computers with CPUs only. PhD...I'd have to ask around.

The NAVAIR & NASA LRC teams actually do a lot of aerodynamic research with minimal compute capability.

neurotech wrote:

There is plans by Boeing aerodynamic improvements to the Super Hornet to reduce drag and improve stores separation. The Navy has not ordered this version, but is offered to Brazil or possibly Canada.

I'm not saying that the Super Hornet is a bad airplane or that it needs improvement, I'm just using it as a case where CFD done by very smart people with tons of resources failed to properly predict "real life"

neurotech wrote:

Stores separation CFD is significantly more complex than calculating the Ps diagram of a clean jet.

The problem is that from a fluid dynamics standpoint, a clean jet is not clean at all. You have both laminar and turbulent flow separating in weird places from the airplane. CFD is very good at predicting lift and drag under cruise (ie low AoA) conditions, but for maneuvering flight, it is orders of magnitude more complex. To get accurate results you must also add onto that the fact that your 3D model (which is called a grid in CFD) must move the flight controls to the right deflections for each test point and flight condition. Plus if you want a full Ps diagram, you probably want to include the transonic (mach 0.8-1.2) region, within which the flow behavior is nearly impossible to predict with even the best CFD software.

neurotech wrote:

GPGPU technlogy is improving so that a high-end desktop can do what would require super computers with CPUs only. PhD...I'd have to ask around.

The NAVAIR & NASA LRC teams actually do a lot of aerodynamic research with minimal compute capability.

I don't doubt that organizations use small clusters and PC's to do this kind of research (I did some myself when I was getting my degree), but those are all very small scale experiments, that don't compare to the complexity of modeling a real airplane throughout its entire flight regime.

I'm not saying that you couldn't throw together a PAK-FA grid and run it through some public domain CFD code, but the results you get will not be close to the real thing. Getting CFD to run properly and give you proper results is an art that takes a long time and lots of experience to get proficient at.

Again, this isn't a personal attack on you, but I think you're underestimating how daunting a task something like this would be.

sirsapo wrote:

I'm not saying that you couldn't throw together a PAK-FA grid and run it through some public domain CFD code, but the results you get will not be close to the real thing. Getting CFD to run properly and give you proper results is an art that takes a long time and lots of experience to get proficient at.

Again, this isn't a personal attack on you, but I think you're underestimating how daunting a task something like this would be.

I don't buy the whole PAK-FA advantage over an F-22. If its anything like previous intelligence blunders (MiG-25 Foxbat turning performance?). It's likely the PAK-FA is heavier, and has less maneuverability than the F-22.

Wasn't it Jon Beasley(test pilot from Lockheed) who criticized the Su-30 demonstration as being good for an air show, not so good in an actual dogfight. The likely hood is that the T-50 wont be significantly better than the Su-30MKI in maneuverability. I'm going to rewatch some of the Su-30MKI dogfight videos.

The whole MiG-25 thing was due to the West's inability to get a good look at the plane (especially what was not on the outside) and understand its intended role. The Foxbat could have been a superfighter for all they knew at the time. Remember, the F-15 doesn't look terribly different.

It seems like the PAK-FA is very much intended to compete with the F-22, or at worst dominate the F-15. I don't see it turning into a MiG-25. That said I don't think it will beat the F-22, but at this point it's basically speculation.

On CFD, I'd gladly do it if someone provided me with a model and $3,000 to buy the new computer I was looking at. Being serious though, trying to get a full flight envelope for a plane using CFD on one computer could be as time consuming as actually flying the plane to generate the envelope. Without (or maybe even with) serious hardware, the time it takes you to run CFD on the plane in some cruise state could be enough to get the real think to take off, do a bunch of maneuvers, and land. And as has been stated, high AoA and transonic flight can be very difficult to do. Those subjects are tricky even off the computer.

http://www.youtube.com/watch?v=cenIBqda8zA

According to that video, every sub second time step of simulation took 20 seconds of real time.

Forum: General F-22A Raptor forum

PAK FA will "remove the advantage held for decades"