Board index FlightGear Development Aircraft

Stalls and spins

Questions and discussion about creating aircraft. Flight dynamics, 3d models, cockpits, systems, animation, textures.

Re: Stalls and spins

Postby sety » Wed Aug 06, 2008 10:37 am

HHS wrote:Depending what you mean with "stalling".


I would define a stall (for simplicity to make it easier to model) as when the AoA exceeds 15deg and the wing is no longer effective. (That's in a Cessna, I don't know if it's the same on all aircraft. Perhaps a function of wing chord, etc?)

A stall can happen with or without power and can even happen at speed. A "snap" is kind of a stall I think because you stall a control surface. I remember once my father lost his temper while mustering sheep (a musterer couldn't find the sheep he was diving on) and at about 100kts he slammed the yoke into the wall to dive on the mob again and the aircraft just suddenly "flipped" inverted and then nose down. It was quite violent and sent books, pens, tools, microphones etc all over the cabin.

Anyway, in a sim it shouldn't be too difficult to model stalls using AoA (is AoA on the wings and control surfaces monitored in FG?) If it is, I rekon it'd be a pretty easy way to model some of the more unpredictable behavior of the planes.

Cheers,
Stuart
sety
 
Posts: 24
Joined: Sat Aug 02, 2008 4:15 am

Re: Stalls and spins

Postby Natovr » Thu Aug 07, 2008 9:37 pm

xiii wrote:The A-10 has also nasty non symmetric stalls, very nasty indeed. I don't know how the real one goes, but you better watch your speed when approaching or when pulling lots of "G". I didn't succeed to put it in a spin.

The real F-14 is commonly subject to spin, our Flightgear Tomcat is not.
Both are YASim.

Alexis

Actually while practicing bombing rounds, I crashed quite a few times while turning. I went into a spin at low AGL and couldn't recover in time.
Callsign: ZS-NVR (mil/priv)IVA-003 (com)
Part of Island Virtual Airways, one of Flightgear's
Natovr
 
Posts: 958
Joined: Fri Feb 22, 2008 12:11 pm
Location: United Queendom

Re: Stalls and spins

Postby skypup » Mon Aug 25, 2008 10:00 pm

It seems to me the stall portion of this discussion is the more important for Flightgear as an accurate flight simulator. The spin portion maybe a little less so, as most GA flight instruction today focuses on avoiding spins in the first place. As I understand it, The GA accident rate involving spins dropped quite a bit when spin practice was dropped from the instruction. Most modern GA aircraft seem to be designed to reduce the likelyhood of entering a spin.
Stall, on the other hand, is a feature of almost all aircraft, and a major contributer to airplane shaped holes in the ground. Modeling accurate conditions that will induce stall for a particular aircraft in all phases of controled flight would seem to be a must for any serious flight simulator.
Modeling what will happen after the stall, spin, etc seems much more problematic, as you have gone from a controled state to a chaotic state. Making accurate prediction becomes a nightmare I would imagine.
However, the most important result of a stall still remains predictable, loss of altitude, and sometimes dramatic loss of altitude, depending on the aircraft. This remains predictable, or at least information should be avalable as to how much altitude will be lost. This should not be to difficult to include in the aircraft model. What the aircraft does during that loss of altitude would be nice to model, but would seem to be of secondary importance to the loss of altitude.
Most of the aircraft I have tried in Flight gear seem reasonable in the loss of altitude, though a few seem to get away with a lot less than would probably happen in the real world.
The part of the stall equation that seems to be missing though, is weight, or more specificly, G force. The given stall speed for an aircraft is at 1G. As G force goes up, so does stall speed. A large percentage of GA accidents have this as the root cause. Slow, high angle turns at low altitude. The extra Gs induced by the turn bring the stall speed above the air speed, and there isn't enough altitude to recover.
Search Youtube for "B52 crash" to see a classic example.
I haven't tried all the Flightgear planes, but the ones I have tried, the stall speed doesn't seem to change with G force. You can do 80 degree banks at 40 mph in the cub all day if you feel like it.
Flightgear regesters G force, as anyone who has used the HUD is probably aware, so it would be a useful feature to link it with stall speed. It would give the simulator added accuracy in a critical area, and who knows, it might just save the life of someone who continues on from Flightgear to real world aviation. Flightgear seems a much better place to make those fatal mistakes.
Is it nap time yet? And where are the doughnuts?
skypup
 
Posts: 37
Joined: Mon Mar 10, 2008 3:55 am
Location: western Washington

Re: Stalls and spins

Postby MD-Terp » Mon Aug 25, 2008 10:21 pm

Stalls aren't induced by airspeed or G force -- they're induced by the angle of attack of the lifting surface against the airflow, and the result is very predictable indeed -- dramatic loss of lift and increase in drag on that particular lifting surface.

However, I agree that it seems like some of the modeled aircraft do very well in simulating stall conditions, and others are severely more forgiving than their real-world counterparts. With open-source projects of this scope there will always be some quality control issues, but maybe we can support the diversity of the project by documenting which models behave realistically and which are too "permissive" of out-of-envelope flight parameters.

I wonder if this discussion would best be moved to the "aircraft modeling" forum?
Cheers,
-Rob.

"Retired" from FlightGear involvement as of July 2010.
viewtopic.php?f=3&t=8809
User avatar
MD-Terp
 
Posts: 2410
Joined: Wed Jan 23, 2008 7:37 am
Location: Baltimore, Maryland, USA
Callsign: N531MD, AVA0025

Re: Stalls and spins

Postby skypup » Tue Aug 26, 2008 1:05 am

MD-Terp wrote:Stalls aren't induced by airspeed or G force -- they're induced by the angle of attack of the lifting surface against the airflow, and the result is very predictable indeed -- dramatic loss of lift and increase in drag on that particular lifting surface.

However, I agree that it seems like some of the modeled aircraft do very well in simulating stall conditions, and others are severely more forgiving than their real-world counterparts. With open-source projects of this scope there will always be some quality control issues, but maybe we can support the diversity of the project by documenting which models behave realistically and which are too "permissive" of out-of-envelope flight parameters.

I wonder if this discussion would best be moved to the "aircraft modeling" forum?
Cheers,


I agree with you that airspeed or G force in themselves don't induce a stall, But both are major factors in causing an aircraft to assume an angle of attack where it will stall. A more heavily loaded airplane will fly at a greater angle of attack, and close to stall, for a given airspeed than if is lightly loaded. It also requires more power due to the added resistance of the greater angle of attack.
An airplane entering a steep turn has become heavier for the duration of the turn due to the G force. It will assume the greater angle of attack, and also the greater resistance that goes with it.
Going into a turn at a low speed can start a deadly cycle, the angle of attack increases due to the increased G force, the resistance goes up due to the greater angle of attack, speed drops due to the increased resistance, and the cycle continues until stall is induced. It can all happen very quickly at a low enough speed.
In Alaska my sister says they are called "Moose Stalls". People get distracted looking at wildlife and don't notice they are getting into trouble until it is too late.
It is a factor that is very important in real world aircraft, and kills or injures quite a few people each year. I think it should be a factor in a serious flight simulator.

I think you are right that it would be best in "modeling forum, and I had been thinking for some time of posting about it there, when I found the stall thread here. I don't know how to go about getting it moved. Does anyone else reading know how to do this?
Is it nap time yet? And where are the doughnuts?
skypup
 
Posts: 37
Joined: Mon Mar 10, 2008 3:55 am
Location: western Washington

Re: Stalls and spins

Postby skypup » Tue Aug 26, 2008 3:49 am

Thanks to whoever moved this to Aircraft Modeling, I think it is a better place for this thread.
Is it nap time yet? And where are the doughnuts?
skypup
 
Posts: 37
Joined: Mon Mar 10, 2008 3:55 am
Location: western Washington

Re: Stalls and spins

Postby skypup » Tue Aug 26, 2008 8:00 am

Ok, the plot thickens. xiii mentioned how the A10 stalls. I had not flown it, so I gave it a try. yes, quite a nasty stall. It wants to make that airplane shaped hole and take you with it. I lost 5000 feet in one stall before recovery.
It has a pretty consistant stall at about 95 knots at about 1G. In a 2G turn the stall speed goes up to about 155 knots. So in this airplane stall speed seems to be linked to G force.
Next I tried the piper cub. With a lot of work I finally got a stall at 20 knots. No chance of a stall in turns, no matter how tight. The stall speed seemed unreasonably low, so I did some reaserch.
A cub's stall speed with the 65 hp engine is around 33 Knots. Quite a difference.
Next, I took a look in the cub XML files and played around changing a few things. Got the stall in the neighborhood of 33 knots, but still no stall in tight slow turns.
Anyway, it would be intersting to hear from folks who are developing the models. Is stall speed always linked to G force? Sometimes? Only by choice of the developer? Or something else all together?
Is it nap time yet? And where are the doughnuts?
skypup
 
Posts: 37
Joined: Mon Mar 10, 2008 3:55 am
Location: western Washington

Re: Stalls and spins

Postby ajm » Tue Aug 26, 2008 8:32 am

The complete flying behaviour of any FG aircraft (from whether it sits stationary and level at rest to more complex issues such as discussed here) depends entirely on the FDM configuration created by the modeller responsible.

There are three different FDMs in use; the two main ones are YASim and JSBSim. The configuration file for JSBSim lets you use pretty much any aerodynamic data you can lay hands on to model the flying behaviour - lift, drag, alpha - and another whole raft of forces. The "problem" is almost always in obtaining the data in the first place as it's a time consuming and therefore expensive business, the results of which are unlikely to be shared around by manufacturers for free.

With YASim, the focus is on providing the geometry and weight distribution of the aircraft and control surfaces, and several known "performance points". There are quite a few things you can tweak to obtain the behaviour you want (understanding how the FDM works is an advantage, obviously!) including tunable parameters to define stalling behaviour for wings.

You should use the readme.yasim file supplied in the docs directory of FG to understand that a bit better, but it should be borne in mind that people will be generally not too keen to accept FDM config modifications which don't come from someone with practical hands on experience of a particular aircraft unless you can provide sufficient documentary evidence for the behaviour you've achieved.

I'm still waiting hopefully for someone with flying time on the Lightning to give me some pointers on how the config might be tweaked - without much hope, since nobody has yet come forward to help with something as common and popular as the Bocian!
ajm
 
Posts: 258
Joined: Wed Dec 20, 2006 6:05 pm

Re: Stalls and spins

Postby DFaber » Tue Aug 26, 2008 9:12 am

skypup wrote:Anyway, it would be intersting to hear from folks who are developing the models. Is stall speed always linked to G force? Sometimes? Only by choice of the developer? Or something else all together?


In YASim, Stall speed and behavior for a wing or surface is specified via stall speed and Angle of Attack. A common pitfall here (which appears to be with the pub too) is an aircraft being unable to reach the AOA during slow flight (18 degrees here). In this case the Aircraft doesn't stall at all, but glides with fully pulled elevator, at low speed slowly downwards.
Usually it is sufficient to decrease the stall-aoa for the Wing.

Greetings
Detlef Faber
FlightGear Development:
http://flightgear-de.net

my 3D-Art:
https://www.sol2500.net
DFaber
 
Posts: 709
Joined: Fri Dec 01, 2006 8:51 pm
Location: Aachen, Germany
Version: GIT
OS: Linux

Re: Stalls and spins

Postby Galloglas » Tue Aug 26, 2008 12:35 pm

skypup wrote:Thanks to whoever moved this to Aircraft Modeling, I think it is a better place for this thread.


You're welcome!! :D

G
Galloglas
 

Re: Stalls and spins

Postby skypup » Wed Aug 27, 2008 6:02 am

A number of intersting points have been discussed in this thread, but my question still goes unanswered;

What, if anything, can be done to link stall speed to G force?

Or is it already, as it seems to be with the A10? But not with the cub? I seem to get conflicting results with different aircraft.
I have looked through the FDMs, and the explanation of the FDMs, but could find nothing that really indicated a link to me. Granted, I am not too well versed in programs and computer models though.

However, if Flightgear is still pursuing the aim of being a highly accurate flight simulator, this is an area of the flight model that should be addressed.

Thanks
Is it nap time yet? And where are the doughnuts?
skypup
 
Posts: 37
Joined: Mon Mar 10, 2008 3:55 am
Location: western Washington

Re: Stalls and spins

Postby DFaber » Wed Aug 27, 2008 6:54 am

skypup wrote:What, if anything, can be done to link stall speed to G force?

Or is it already, as it seems to be with the A10? But not with the cub? I seem to get conflicting results with different aircraft.


A properly set up YaSim FDM will have a higher stall speed at higher g-loads. This is already the case. Try the Bf109, Fw190, F4U, these Aircraft were trimmed to stall properly (and they can be put into a spin too).
It is the Aircraft Modellers, (or willing contributors) job to adjust stall characteristics.

Open Aircraft/j3cub/j3cub.xml in your favorite text editor and try for yourself:

Change line 35 to:
Code: Select all
<approach speed="35" aoa="15">


line 64 to:
Code: Select all
  <stall aoa="11" width="4" peak="1.5"/>


and finally line 77 to:
Code: Select all
  <flap0 start="0" end="1" lift="1.5" drag="1.2"/>


The Cub will now stall at approximately 33 kts and in tight turns at about 40 kts.

This is only a quick example, which can surely be extended and improved.

Greetings
Detlef Faber
FlightGear Development:
http://flightgear-de.net

my 3D-Art:
https://www.sol2500.net
DFaber
 
Posts: 709
Joined: Fri Dec 01, 2006 8:51 pm
Location: Aachen, Germany
Version: GIT
OS: Linux

Re: Stalls and spins

Postby skypup » Wed Aug 27, 2008 8:05 am

Thanks much, I'll give it a try tomorrow. :D
Is it nap time yet? And where are the doughnuts?
skypup
 
Posts: 37
Joined: Mon Mar 10, 2008 3:55 am
Location: western Washington

Re: Stalls and spins

Postby HHS » Thu Aug 28, 2008 2:54 pm

Hi,

I had an unintended try yesterday with the PC 6 Porter. It really shows that a lot of aircrafts seems not to be realistic in stalling behavior and low speed range.

I made a steep curve short to final, when suddenly the aircrafts stalled and I coulden't react quick enough....I was really surprised, but positiv! :D

Fine work Detlef!

Cheers
HHS
Up, up and away
User avatar
HHS
 
Posts: 3625
Joined: Thu Jul 19, 2007 9:09 am
Version: GIT

Re: Stalls and spins

Postby syd_cyyf » Thu Aug 28, 2008 8:26 pm

While we're on the subject , every one of my aircraft definitely needs re-tuning , though I'm not comfortable with adding unrealistic wing stall AOA's , I guess the end result is what's more important .
My experience has been getting a proper approach speed , but then unable to get to stall speed, or a proper stall speed , but appraoch speed too high.
So .... another round of updates coming up.
syd_cyyf
 
Posts: 168
Joined: Fri Dec 01, 2006 1:58 am
Location: Lethbridge AB
Callsign: C-CYYF
IRC name: syd
Version: GIT
OS: Linux

PreviousNext

Return to Aircraft

Who is online

Users browsing this forum: No registered users and 16 guests