Board index FlightGear Development Aircraft Flight dynamics model

Fundamental FDM design problems.

Good sims require good FDMs (the "thing" that makes an aircraft behave like an aircraft).

Re: Fundamental FDM design problems.

Postby swampthing » Wed Oct 09, 2019 11:36 am

@thorsten

I'm not sure if I'm addressing this correct for your last comment but from my experience more JSBsim systems and less nasal = beter FPS. I don't think for some things you can move away from nasal all together. It comes down to what your plane requires. I'm not aware of any way to make canvas displays with JSBsim code for instance. I'm not against nasal but I can see where people may want to do full JSBsim sytems and I think for the most part if you can get a good grasp of it, you can with JSBsim. I haven't done it but the Viggen is a good example of complex JSBsim systems. I'm just an old grease monkey still trying to learn. Sadly I wan't born knowing all of this and when I was born it didn't exist. ;) Another option is using property rules. @octal450 used property rules to make the Mirage-2000 fly more realistic. Something I forgot when I spoke about YASim earlier. This could be something Helija may want to try and impliment into some of his planes but with that many you just have to pick one and work away at it start using Flightgear again, test, test, test and move to the next. There are a lot of models out there. One thing that is a question for many is how many are no longer being developed by the author because life smacked them in the face and they just don't have time, lost interest or found a new hobby. Maybe the wife hit them over the head. I don't know. If there was a list of planes we know are never going to be touched by the original author maybe there could be further development. Who knows how many people look at a plane in FGaddon and would like to work on it. Maybe they already do but its hidden away on only their PC. Well that doesn't allow me to fly it. I know thats off topic but its a question @thorsten.
www.opredflag.com
I have sworn upon the altar of God, eternal hostility against every form of tyranny over the mind of man. -Thomas Jefferson-
swampthing
 
Posts: 534
Joined: Wed Oct 28, 2015 4:10 am
Location: Missouri
Callsign: swamp
Version: 2018.2
OS: multiple

Re: Fundamental FDM design problems.

Postby Thorsten » Wed Oct 09, 2019 11:56 am

I guess the relevant question is - at what interval do you need it?

Think of a high-level planning task - I want to steer the Shuttle to a runway - is my current energy sufficient, or should I change the approach? I don't need the answer to that every FDM iteration (it's comparatively expensive to get anyway) - I can live with getting it every few seconds and feed that information to the actual control loops which runs at a much higher rate. The weight of the task executed not at 1/120 of a second but at, say, 5 seconds is but a fraction of what a JSBSim implementation is.

(The extreme example is orbital targeting - computing a burn solution costs a few seconds in itself - trying to do this per JSBSim iteration would be ridiculous).

Think of a housekeeping task like fuel cell depletion - this happens over the course of 24 hours, so you absolutely do not need to do this more than once every few minutes.

Think of something as trivial as leakage simulation of tanks. You plainly don't need to check every FDM iteration whether the tank has a leak or not. You don't even need it every few seconds. You can catch it with a listener and only compute the changed flow pattern when the leak arises and then compute just once - then you write the leakage rate to the low-level simulation till another change occurs.

Doing tasks when you need them rather than when JSBSim makes an iteration is the key to much better framerate.

Of course, if you try to do a per-frame task in Nasal and factor plenty of badly designed property I/O into the equation, then you gain by doing everything in JSBSim.

The key really is to let JSBSim do the simulation and control tasks that need quick updates and let Nasal handle the high-level stuff and everything that needs to be touched just once in a while.
Thorsten
 
Posts: 10982
Joined: Mon Nov 02, 2009 8:33 am

Re: Fundamental FDM design problems.

Postby swampthing » Wed Oct 09, 2019 12:04 pm

Why did I have a feeling i was going to get shot to the moon? ;)
www.opredflag.com
I have sworn upon the altar of God, eternal hostility against every form of tyranny over the mind of man. -Thomas Jefferson-
swampthing
 
Posts: 534
Joined: Wed Oct 28, 2015 4:10 am
Location: Missouri
Callsign: swamp
Version: 2018.2
OS: multiple

Re: Fundamental FDM design problems.

Postby bugman » Wed Oct 09, 2019 12:46 pm

If you had the relevant skills (enough coding knowledge to put together a small c++ fragment), you could also create a flightgear submodule to perform the operations you are interested in and manually set its update interval dt. In some situations that might be a more elegant solution than a complex piece of Nasal code.

Anyway, I think that YASim produces a decent result if you are in the situation of having no real aero data at your disposal. If you had a decent set of aero data, then JSBSim would be better. I know that Helijah has an interest in obscure, historical, unusual, and even fictional aircraft. And that his focus is on large quantities of aircraft rather than a few of top quality, while maintaining a decent and functional quality level not found in many FGAddon aircraft. So, in his situation, YASim is probably the best. If you were to tell me that he should use JSBSim for the Macchi M33 from the Porco Rosso film from Ghibli Studios - based solely on your personal FDM preferences - I'd probably think that you are nuts ;) (and the same if you told me that YASim should be used for NASA aircraft).

Regards,
Edward
bugman
Moderator
 
Posts: 1702
Joined: Thu Mar 19, 2015 9:01 am
Version: next

Re: Fundamental FDM design problems.

Postby xcvb » Wed Oct 09, 2019 1:53 pm

If you put the same amount of energy into an XFLR5, OpenVSP or Datcom model that you would put into a Yasim FDM, you will get at least the same quality result. The big difference is that JsbSim allows you to check every little detail, whereas Yasim is a kind of VooDoo magic and you can be happy if a little change doesn't break everything.
xcvb
 
Posts: 86
Joined: Sat Mar 14, 2015 2:08 pm
Version: 2018.1.1
OS: Fedora 27

Re: Fundamental FDM design problems.

Postby legoboyvdlp » Wed Oct 09, 2019 3:23 pm

Bugman, I think you're missing the point slightly.
It's not so much that you should always use jsbsim no matter what.

The problem here is that what I can only call fabrications are being posted about JSBSIM. I highly object to this, as it can lead people astray into avoiding JSBSIM.
User avatar
legoboyvdlp
 
Posts: 7007
Joined: Sat Jul 26, 2014 1:28 am
Callsign: YV-LEGO
Version: 2018.3.1
OS: Windows 10 HP

Re: Fundamental FDM design problems.

Postby bugman » Wed Oct 09, 2019 3:45 pm

These arguments go back for many years. This is a double-edged sword! I think Helijah's dislike of JSBSim and some of his arguments against it are misplaced. I also know that there are a few on this thread who dislike YASim so much that they would not touch it with a barge pole. So the discussions on the subject tend to be non-logical and revolve around the concept of who hates the other FDM more. So my point is countering those who say you should use JSBSim no matter what (there are quite a few who argue this way). Note that there are also more YASim haters on these threads than JSBSim haters.

My opinion is quite different. I think there is a place for both YASim and JSBSim - as well as UIUC. If you have no or little aero data, use YASim. Otherwise use JSBSim. If you are interested in properly simulated stalls or bizarre flight models such as the ornithopter, use UIUC!

Regards,
Edward
bugman
Moderator
 
Posts: 1702
Joined: Thu Mar 19, 2015 9:01 am
Version: next

Re: Fundamental FDM design problems.

Postby legoboyvdlp » Wed Oct 09, 2019 3:47 pm

I quite agree with you about there being a place for both - regardless, I do feel that little is accomplished for instance by falsely claiming that the Cessna 172 can land upside down, bounce, turn round on its own, and fly away - it simply can't -- unless you use the menu options to reset it, of course.
User avatar
legoboyvdlp
 
Posts: 7007
Joined: Sat Jul 26, 2014 1:28 am
Callsign: YV-LEGO
Version: 2018.3.1
OS: Windows 10 HP

Re: Fundamental FDM design problems.

Postby Thorsten » Wed Oct 09, 2019 4:16 pm

Much of the discussion is actually a meta-discussion about the style of a certain person to make contradictory claims over the course of a few posts, spiced with entertaining examples for that habit.

But as far as the FDM discussion goes, there is a specific claim which was made:

JSBSIm makes him bounce planes on the water ridiculously.


The claim is - as written - very general. And it is just plain wrong and needs to be countered, because that wrongness has nothing to do with whether one likes JSBSim or not or whether one likes YaSim or not. People are entitled to whatever opinion they may want to cultivate about the usefulness of the various FDMs, but not to their own facts.

It's demonstrated by the existence of JSBSim ships and water planes that JSBSim can do very good hydrodynamics interactions - and anyone who claims that it makes 'plane bounce on the water' plainly has no clue about that.

Likewise, while YaSim can be 'as good as it gets' with the available data - if the focus is actually on realism, that's not good enough. At some point we have to realistically say that with 1000 pages of wind tunnel data we can make a very realistic FDM in JSBSim - but with just a grainy photograph of some obscure historic plane, we can not possibly hope to get this to the same level in YaSim.

So the point that YaSim intrinsically limits realism - you can only ever get so far no matter how much data and work you put in - is entirely valid and also deserves to be made.
Thorsten
 
Posts: 10982
Joined: Mon Nov 02, 2009 8:33 am

Re: Fundamental FDM design problems.

Postby xcvb » Wed Oct 09, 2019 5:40 pm

In addition there are also some things you can't simulate with Yasim at all, like supersonic flight.
xcvb
 
Posts: 86
Joined: Sat Mar 14, 2015 2:08 pm
Version: 2018.1.1
OS: Fedora 27

Re: Next-generation scenery generating?

Postby helijah » Thu Oct 10, 2019 12:53 am

wkitty42 wrote in Tue Oct 08, 2019 10:23 am:i don't know which c172p you are testing but the one that is distributed with FG surely doesn't do that... not unless you specifically click on the repair button which is specifically included so you don't have to exit the sim and restart it just to go flying some more... perhaps you can/should provide a video of this situation you describe... if it is a real problem, the dev team for the craft can then fix it...


Indeed it was a version that was not up to date. Please forgive me. The latest version is indeed much better when it comes to its reactions with the soil.
Some planes (and other) for FlightGear
http://helijah.free.fr
and
http://embaranger.free.fr
User avatar
helijah
 
Posts: 1029
Joined: Wed Dec 27, 2006 12:35 pm
Location: Chartres (France)
Callsign: helijah
IRC name: helijah
Version: GIT
OS: GNU/Linux

Re: Fundamental FDM design problems.

Postby Octal450 » Thu Oct 10, 2019 1:42 am

xcvb wrote in Wed Oct 09, 2019 5:40 pm:In addition there are also some things you can't simulate with Yasim at all, like supersonic flight.


Or realistic flight dyanmics :mrgreen:

Kind Regards,
Josh
Don't ever contact me about FlightGear again.
Octal450
 
Posts: 4393
Joined: Tue Oct 06, 2015 12:51 pm

Re: Fundamental FDM design problems.

Postby V12 » Thu Oct 10, 2019 4:43 am

xcvb wrote in Wed Oct 09, 2019 5:40 pm:In addition there are also some things you can't simulate with Yasim at all, like supersonic flight.

Hmmm, there are YASIM XB70 and probably SR71 too. Or nice Mirage 2000.
Last edited by V12 on Thu Oct 10, 2019 5:08 am, edited 1 time in total.
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 1498
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Next-generation scenery generating?

Postby wkitty42 » Thu Oct 10, 2019 4:58 am

helijah wrote in Thu Oct 10, 2019 12:53 am:
wkitty42 wrote in Tue Oct 08, 2019 10:23 am:i don't know which c172p you are testing but the one that is distributed with FG surely doesn't do that... not unless you specifically click on the repair button which is specifically included so you don't have to exit the sim and restart it just to go flying some more... perhaps you can/should provide a video of this situation you describe... if it is a real problem, the dev team for the craft can then fix it...

Indeed it was a version that was not up to date. Please forgive me. The latest version is indeed much better when it comes to its reactions with the soil.

thank you for that admission... the current one even lands and takes of on water if you switch the gear to pontoons or amphibious... make sure you have the wheels up if you try to land on water with the amphibious or you'll flip and endo and go for a swim ;)
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 5644
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Fundamental FDM design problems.

Postby abassign » Thu Oct 10, 2019 6:02 am

Thorsten wrote in Wed Oct 09, 2019 8:58 am:
I'd like it to be more flexible and complete to be able to replace NASAL more, as JSBSim is much more suitable for real-time applications.


Why would you want your screwdriver heavier so that you can use it better to hammer in nails when you could... simply use a hammer for the job :?: :?: :?:


Because the world is much more varied than your eyes are able to see and your mind is capable of understanding, so it is never certain that the best solution is the one you think, if you continue to live in the prison of your unlimited presumption.
abassign
 
Posts: 805
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

PreviousNext

Return to Flight dynamics model

Who is online

Users browsing this forum: BLEXBot [Bot] and 4 guests