Outerra forum
User mods, screenshots & videos => Aircraft => Flight Model Development => Topic started by: bomber on June 30, 2014, 02:08:38 pm
-
Well it is if you've 2 years to learn it....
Hi I'm a jsbsim flight modeller and that's a question I'd like to ask.
Does the modding community and flight modellers here wish this steep learning curve to exist into the future or is there a vision that flight modelling through the use of an interactive building tool be made if not easy, easier ?
Is there any drive towards that end ?
Simon
-
I wonder if there`s a program that used the 3D mesh to test aerodynamics and then write the basic JSB file.....
-
That's not what you need, it's really a poor way of flight modelling.. and as a professional 3D modeller I can tell you the flight models would only be 1/3 done, and not that good.
A flight model consists of 3 major elements.
Weight/Mass balance
Aerodynamics
Thrust/drag
There's also 'systems' such as fuel, hydraulics, pneumatics...etc
In my opinion these elements should be linked to damage reductions
-
Hmm it would be certainly a nice thing to have for modders and addon makers. Do you have a concrete idea how the tool would work and how it would be used? Could it be some kind of an addon in OT that would allow real-time testing of the airplane as well?
-
Well understanding that a tool should do it's job and no more, in the same way that a chissel isn't a screw driver, or a screw driver a hammer... I'd want a tool that's simple and does what it says on the label..
The first tool i'd want to be able to introduce to a fellow moder that wishes to do flight modelling would be mass/ballance tool.
In JSBsim this section can be filled in with a couple of numbers... Empty mass, CoG and moments of inertia.... And it's great if you've got them or are defining a new plane.. But try getting them for a 1917 bi-plane. I'd say 95% of planes built on JSBsim have these numbers guessed.
But here's the thing if someone's doing bi-planes, trawling through documents, over a period of time they've an idea as to how much 'generic' items weigh....
And after a period of study, from the overall weight of the plane minus these individual elements the mass of a planes 'structure' can be narrowed down.
Wing structure has it's own density, that being it's material plus air contained within it's shape.. the internal make-up of one wing to the next is pretty similar, the weight difference being it's plan form and thickness differences...
So the tool I'd want is one that allows me to enter the planes total weight plus on individual tabs add 'point masses' for the objects I know and point masses based on density, area and volume for wings and fuselage.. outputting these values as a complex mass ballance JSBsim section or a collated JSBsim mass ballance section.
Simon.
-
Well it is if you've 2 years to learn it....
Hi I'm a jsbsim flight modeller and that's a question I'd like to ask.
Does the modding community and flight modellers here wish this steep learning curve to exist into the future or is there a vision that flight modelling through the use of an interactive building tool be made if not easy, easier ?
Is there any drive towards that end ?
Simon
Greetings, Simon,
Yes, it would be wonderful if there was a GUI tool to help with development of JSBSim aircraft models. Over the past ten years I have started - and stopped - development of such a program, and there was at one time such a project being developed as an open source tool (JSBSim Commander). The more useful I see such a tool being, the less time I have available to develop one myself! :-( It may be that at some point I can develop tools that will help with specific parts of a JSBSim flight model, such as mass properties, systems, aerodynamics, etc. Better documentation and examples would be helpful, too, I am sure. But, myself and the other primary developers of JSBSim have been very busy lately with our day jobs. :-)
Jon
-
Oh I understand jon, we all have lives to live.
I'm not totally convinced that a single tool would be the best approach as this could very well 'straight jacket' fdm'ers into creating flight models in a single way.... As you know I don't create my flight models similar to how others do....I basically took your manual as a 'guide' to creating flight models within JSBsim with the understanding that I had a free hand if I thought to do it differently.
I think a series of small tools that link up even loosely, javafoil, javaprop, spreadsheets for weight ballance and the such I think would work better than an all enfolding tool.
Simon
-
While JSBSim is a well accepted and capable flight simulation dynamics application, it may be possible to create it's parameter tables using another well accepted and capable flight dynamics application that generates much of this data literally on the fly using blade element theory. One of these is X-Plane. I wonder if a plugin could be written for X-Plane to export JSBSim parameter files.
First, you model that aircraft in X-Plane's Planemaker application with a high level of accuracy.
All one would then need to do is to fly a series of development flights with this model that hit all of the regimes that would be need to be measured. Naturally this sort of data could then be "massaged" with reference to real world performance values for the aircraft in question, if it actually exists.
-
Steve. I think you've misunderstood how JSBsim works in that it also generates the data 'on the fly'
The problem people have is that they've heard that JSBsim uses 'tables' of data and come to the conclusion that JSBsim is somehow not working things out..
Blade theory is simply splitting of the wing into smaller sections and then calculating the forces generated by these sections and then integrating these forces back together to come up with the total forces generated by the wing.
I genuinely think that's the correct way of doing it, especially if the wing has different airfoils, twist, dihedral, sweep along it's length, simply averaging out the wing and generating a co-efficient table for the whole wing just to my mind doesn't cut it.
In JSBsim you can split the wing up into sections, you can calculate the forces for these sections and you can use these values to calculate the overall lift and moments of the whole....
Simomn
-
One guy of the 3rd party dev's on the DCS-Forum has made a nice tool to debug flight models see ->http://forums.eagle.ru/showpost.php?p=2159719&postcount=26 (http://forums.eagle.ru/showpost.php?p=2159719&postcount=26)
(http://i.imgur.com/hB9ZXGC.png?1)
https://www.youtube.com/watch?v=4Ub0Ru96xeQ (https://www.youtube.com/watch?v=4Ub0Ru96xeQ)
-
Any chance of opening a dialogue with these guys ?
Gota be careful as there’s normally a mine field of politics to avoid in flight sim cross community com's.
But I'd welcome that kinda of tool.
-
Any chance of opening a dialogue with these guys ?
I don't know if they are willing to release it to the public, cause they made there business indirect with these kind of tools.
You can contact Cory Parks he is the developer, currently i see 2 way to get in contact with him:
DCS-Forum -> User: SilentEagle (http://forums.eagle.ru/member.php?u=3929)
YouTube -> Channel: Virtual Patriots (https://www.youtube.com/watch?v=4Ub0Ru96xeQ)
Maybe there is a way to motivate him to open his business for a future Outerra Flight Sim. It would be nice to see a proove of concept to see if it is possible to port an DCS aircraft to Outerra.
-
No harm in asking....
-
Cory is also present here: http://forum.outerra.com/index.php?topic=1172.msg25244#msg25244 (http://forum.outerra.com/index.php?topic=1172.msg25244#msg25244)
-
well ... just rendering the KA-50 in OT would be a crazy thing -would fly that piece of beauty all around the world !
-
I have before expressed interest in developing tools and aircraft on the Outerra engine, but it does not yet have an SDK or multiplayer support. When those are available, it will be a no-brainer to support Outerra.
-
I totally understand the need for multi-player support and as my goal is WWII heavy bombing crewing it's a necessity...
However Chicken or Egg.... which comes first.
Flight modelling and producing the best, and being known for having the best will drive development both by other flight modellers, 3d and 2d artist but also by Outerra developers to match it..
Raising the developers priorities of multiplayer online support is a must, and producing the best FDM's is the only way I can see of doing it...
I respect your opinion to sit it out until such a time as they exist, because I know others are doing the same, but I don't agree with it...
And I hope in the future we can work together.
Regards
Simon
-
I'm not holding off because it doesn't have multiplayer, I'm unable because it doesn't have an SDK. The multiplayer is necessary for a commercial product, but won't stop the development.
-
It's more about the tool for me..
I use jsbsim flighmodels, although not conventional, yet looking at the video it seems to me you have 4 normal forces coming off of each wing so maybe your flight models aren't conventional either.
Adapting that tool such that it would allow it to display the properties I use would greatly further the understanding of the small amount of flight modellers we have here.
That's all I'm asking.
Anyway thanks for the reply
Simon
-
Right, I don't use JSBSIM to model forces, only to integrate the aircraft position and orientation in that program. Tools such as that could be adapted for use with any flight model type. I also have ideas for aircraft development tools for things such as lift curves, aircraft sectioning (mass distribution, differing incidence, dihedral, camber, twist, control surfaces etc.), gear and suspension geometry, etc. All of this is done by hand in the flight model you see in the video, but if I built a standardized flight model interface, making those tools would be necessary to accelerating development and making it easier for other devs.
-
I can only dream....
"(mass distribution, differing incidence, dihedral, camber, twist, control surfaces etc.),"
I do all those things by hand...
-
The SDK is the biggest thing that we need right now, so that we can take the flight model Cory is working on and integrate it into Outerra, and be able to create a complete interactive cockpit with full systems simulation, and build a complete aircraft.
At the moment we can import models and use JSB sim, but that is far short of what Outerra is capable of. Outerra is in a good position, but we need an SDK to really show how great it can be.
-
JSBSim can do all those things now..
So the question I have is why split a flight modelling community in two ?
That's a mistake Flightgear made.
-
JSBSim can do all those things now..
So the question I have is why split a flight modelling community in two ?
That's a mistake Flightgear made.
Not just flightgear, have you seen the way the IL2 community behaves? not to mention the wider flight sim community and their bizarre intra sim wars, the sad part is that when Outerra starts gaining ground in the sim arena it will start to attract some of this flak.
-
JSBSim can do all those things now..
So the question I have is why split a flight modelling community in two ?
That's a mistake Flightgear made.
Not just flightgear, have you seen the way the IL2 community behaves? not to mention the wider flight sim community and their bizarre intra sim wars, the sad part is that when Outerra starts gaining ground in the sim arena it will start to attract some of this flak.
I'm kinda dreading that......
-
Why is it a bad thing if a couple of different methods are available to do something?
AFAIK these different techniques (JSBSim, or SE's methods) don't require specialized support in the core. Surely it is better to have a development environment whereby you can create a dynamics package as complex as you want?
-
Been around flight sims long ?
I started in warbirds circa 2000, then Targetware, then Flight gear, now here...
Working on flightsim attracts some very passionate people, some would say they have an evangelical attitude to the sim, developers, tools and techniques they use.
And it causes politics and flamewars on forums ad nauseum. I like this place it has potential and having seen how these flamewars develop I'd rather it was 'cut it off at the pass'
In my mind that's not a bad thing.
I'd ask, if the JSBSim flight model I use is everything you'd ever want, accuracy wise, why would you want an alternative one ?
If the flight model isn't up to your requirements, isn't it better to have the community focused on a solution rather than bickery over perceived insults about the accuracy of one flight model over another... And that will happen.
I build flight models for accuracy, not to hit the numbers in a table but to be the result of the sum of all it's parts, and as a combat sim enthusiasts damage and its effects on this summation is very important...
If you check out the Ask-13, you might be surprised by it. Its a glider, flight modelling wise there's nowhere to hide as its either aerodynamically accurate or its not, there is no engine to make it hit the numbers. Its 2 years in the making, over 9 rewrites from top to bottom, but if yours is a better flight model that incorporates damage, systems, prop, piston, superchargers, all the range needed from the Wright brothers to modern helicopters and beyond designs then I'll drop it like a stone...
Would you do the same if it turned out the other way round ?..... Or should we get the choir warmed up
Simon
-
Well ... having a different approach to the simulation must not be a bad thing. Just need to compare them to the reality and see what has witch better cuts of it. ... as a science interested, ill say, there is no point of argue before a hefty research paper is written there. So, tell them arguers to rent the Boeing research labs for a week to proof their point. :D :D
And off course a power-point presentation .... always a power-point presentation ! Do not ewer take an statement of theirs whiteout one ! ;D
(http://www.generalcomics.com/14.gif)
Also - true story, better to focus on a solution whyte constructive criticism and open mind, than the beer-filled Bar sitting one brother of the criticism guy. Than its just politics .... we dont need politics in our labs, dont we ? :D
-
I believe developers should have choices when it comes to how complex they want their aircraft to be. Not everyone is capable of developing to the same level of fidelity. The current situation in several flight sims is that devs are forced into a tiny box of accepted methods and this severely limits their potential. Many FSX devs have decided to bypass as much of the base SDK as possible because of it.
If I had any say in the SDK and flight modeling methods, I would allow the developer some choices. Maybe one method can produce asymmetric lift, while another cannot, but the developer's knowledge level, data available, and time commitment should factor into his decision on which method to use. JSBSim certainly won't get dropped from Outerra support unless it is found incompatible with the physics integrator and collision detection with other simulator cores (which I believe is currently the case). Besides, the accuracy of a flight model is only as good as the data and estimations the creator used. One creator can make a much more accurate JSBSim representation of flight than someone who doesn't know what they are doing and is attempting to use blade element or lifting-line method. It is less about the accuracies of each method and more about the secondary effects and finer details that one method captures over another. Coefficient buildup method is great when you have a 100+ page NASA research document of all possible lift, drag, and moment coefficient to in every scenario, but may not be the best when attempting to model an aircraft with no data or flight in the post stall regime.
Edit: Of course, I would always allow the developer to plug in their own completely custom method through dll.
-
I use 8 per wing sections to determine asymmetrical lift and post stall behaviour using JSBsim.
If someone wants to create a simple flight model its easy to dumb down a complex flight model using simplified data.. But if the techniques and methods of creating a flight model are poor to start with then even hundreds of pages of data will still produce a poor flight model across all situations.
That visualisation tool could be used for JSBsim and would make flight modelling easier, producing more accurate flight models..
I'd rather not see two distinct methods of doing flight models in outerra as it very rapidly polarises the community and that's a very bad thing.
-
I'd rather not see two distinct methods of doing flight models in outerra as it very rapidly polarises the community and that's a very bad thing.
It wouldn't be two distinct methods...it would be numerous methods, because it would depend upon the developer as to how they chose to do it.
If you want to use JSBSim, then do that! There is no point limiting the sim to "only" JSBSim, or any other method, because of a perceived "split" that might arise.
It makes no sense to limit what is possible based upon someone's biases towards a particular method. It should be open so anyone can use any method they like (within certain technical constraints). If you're not careful, you get into the trap that numerous simulators of the past have faced - that they force the developer to using one method, and as a direct consequence limit what developers are able to do.
That visualisation tool could be used for JSBsim and would make flight modelling easier, producing more accurate flight models
Yes, and not only JSBSim, but other techniques as well. As SE explained elsewhere, the visualizer only uses JSBSim for integration methods (essentially, to emulate the missing half of the simulator) so the flight model code can run complete for analysis; it doesn't have constraints as to how the flight model is computed - it doesn't know or care if you use XYZ method or JSBSim for the calculations.
-
Agree to disagree then. That flight model visualizer is just the tip of the iceberg of what I'm planning. This series of tools I'm working on should allow people to define their aircraft however they wish and be able to write all the data to a standardized file format for the sim to read.
-
SilentEagle....fly on. Sounds interesting. The JSBSim and Outerra combination is no doubt terrific, as I have seen, but it pales in comparison to the ease with which one can create an aircraft in X-Plane. I like the sound of your endeavors. Really, it's all about the tools.
-
Well said Steve..... The point of this thread.
Do we want to make flight modelling easier and how do we go about it ?
-
Great thread. It's true that it can be tedious to create a detailed JSBSim flight model. If you have the data, it's a bit easier to make simple models. Getting the data (or creating it) for specific aircraft can be hard - but that's true for any flight sim. I've wanted for a long time to create a very capable editor for building models, but that's a very serious effort. There is one open source effort that was begun years ago, but that has not been updated for a while. Someone could pick that up and turn it into a nice application, with some work. You can find it here: http://sourceforge.net/projects/jsbsimcommander/ (http://sourceforge.net/projects/jsbsimcommander/) and here: http://jsbsimcommander.sourceforge.net/. (http://jsbsimcommander.sourceforge.net/.) There is also a product called DATCOM+ that can help build up an aerodynamic mode: http://www.holycows.net/datcom/ (http://www.holycows.net/datcom/)
-
Have small question because of JSBSim and what he is able to do.
So, I have Flight dynamism in wind tunnel(Wiederstandsdaten-Cw...) as well as polar of airplane X (original),
how would be exact ones I able to do with JSBSim FM for airplane X programming? In other words, is one
able to do with JSBSim exact FM to create... faithful?
-
Faithfully?
Jsbsim can do anything you want.....set the bar as high as you want.
Simon
-
Faithfully?
Jsbsim can do anything you want.....set the bar as high as you want.
Simon
Hehe... Simon, this was only one question, because I do not know 100% what is able to do JSBSim. Hence, no panic on the Titanik!.... and Thanks! :P :D ;)
Merry Christmas and Happy Holidays!
Gruß,
Stefan