Setting thermals around the plane - implement on asw20

[D-ABBA]

Hello developers of asw20,

I've become a fan of the thermal gliding in FlightGear. But I must note that the automatically generated thermals (advanced weather - generated thermals) are very difficult to use. Have the function ctrl-t setting Thermals arround the plane built by the DG-101G. These thermals can be used well with the asw20. Very good indeed (vertical speed indicator shows the rise and fall quietly and evenly).

The automatically generated thermals of FlightGear are difficult to use. The variometer (vertical speed indicator) wriggles wildly.

Who can help me with the bug fix.

D-ABBA (D-ABAB)

[legoboyvdlp]

Um what if there are no 'developers' Maybe they aren't on the forums. Look, do it yourself if you want it.

[Thorsten]

The automatically generated thermals of FlightGear are difficult to use.

The author of the automatically generated thermals (=me) has the pleasure of having been a real glider pilot and having a few hundred hours of experience flying thermals and doing cross-country flights (in the ASK-21 mostly, but also in an ASK-13). And comparing my experiences in the ASK-13 in FG and reality, I can assure you they do what they should.

I have also noticed that e.g. the DG-101G had a problem with inertia - which didn't show up in still air but was very prominent in both entering thermals and doing aerotow. So - are you reasonably sure the problem is not with the glider?

[galvedro]

D-ABBA, not sure what you have in mind when you say they are difficult to use.

Since you mention the asw20, I think you might be bumping into an inconsistency we have between the two flight models (yasim vs jsbsim). If there have been no changes to this functionality since the last time I took FG for a spin, turbulence in advanced weather thermals affect jsbsim gliders like the asw20 out of proportion, making them difficult to control and not that realistic. However, yasim models (e.g. the ask13) do actually feel quite natural.

Overall, I would say the best thermaling experience you can get out of FG right now is with advanced weather together with good yasim model like the ask13.

EDIT:

Since you mention you are interested in improving this, and assuming the problem you see is related to the problem I mentioned, what I would suggest is to modify yasim turbulence model to behave closer to what jsbsim does, and then tune the weather system/s to produce the same level of turbulence effects they are now producing for yasim. Why touch yasim if it's the one that I'm saying feels more natural? Because jsbsim is a reference flight model also used outside FG while yasim scope is probably limited to internel use only. (My opinion, and I might be wrong).

[Thorsten]

Since you mention the asw20, I think you might be bumping into an inconsistency we have between the two flight models (yasim vs jsbsim). If there have been no changes to this functionality since the last time I took FG for a spin, turbulence in advanced weather thermals affect jsbsim gliders like the asw20 out of proportion, making them difficult to control and not that realistic.

Actually, I am no longer sure of that one.

I've tested a number of JSBSim planes since.

* The DG-101G indeed had severe issues with turbulence - but also aerotow was way off. I have flown aerotow in real life, and basically the plane doesn't bounce on the rope - it's not easy to slacken the rope at all. Whereas the DG-101G would slacken the rope easily as soon as a force spike came through. Which is consistent with an overall problem with the inertia (tensor) of the plane which makes the response to all external forces wrong, not necessarily with a flaw in weather modeling.

* The SenecaII performed just fine with turbulence and thermals, I haven't seen anything grossly unrealistic

* The F-16 funnily enough had a clear response to thermals, again turbulence didn't seem to affect it out of proportion.

* I'm not sure what the glider was but it was also a JSBSIm model - I got winch launch altitudes of 600-800m easily (real numbers range from 250 to 350 meters) - again, indicating that the plane doesn't have the correct response to a given external force.

So, thing is, lots of aerodynamical behavior is just dimensionless ratios - like lift/drag (the gliding behavior) - you can get this right without having the overall scale of the force right. Which is to say, you can get plausible behavior of a plane in the absence of external forces, but once you apply a force with given magnitude, the response of the plane is off. There is no doubt in my mind that some planes have a wrong response to external forces.

So the first thing to investigate is - how much of what we're seeing is just improper response coded into the planes and missed in still air, and how much is actually the difference in turbulence modeling? That'd need some attention from someone who is familiar with aerodynamics and can do some standardized tests - unfortunately I have no time to delve into this given my other FG projects. I think technically adjusting the strength of turbulence in the code is easy once we have the data telling us how it should be set.

(Side note: You know that the amount of turbulence in thermals is customizable using the convective stability slider, right?)

[galvedro]

Interesting. I agree, the first thing would be to measure the forces each fdm imposes on the model at a given turbulence setting. Both should produce comparable magnitudes.

It may be handy to have a glider featuring carefully crafted yasim and jsbsim models. I happen to have one at hand in this branch. It is a work in progress ASK13 where I am integrating a jsbsim model by Simon (Bomber) together with the existing yasim model. I have flown the new one (not measured it) and it feels quite plausible, although it showed the same problem with turbulence if I remember correctly, it's been some time since I worked on this,

[D-ABBA]

Hello Thorsten,
Hello galvedro,

Thank you for your response. I think that it probably to YASim and jsbsim
Models is what the observed effect leads.

As I have discovered that? Firstly, I have my own thermal_demo
(folder fgdata\ AI) created using thermals places near ED03,
EDFQ and EDGS. Changed thermal_cap.xml (folder fgdata\Models\Geometry)
like this: boole operant from false in true...

<PropertyList>
<path>thermalcap.ac</path>

<animation>
<type>billboard</type>
<spherical>true</spherical>
</animation>

</PropertyList>

After that I see generic clouds (and in the map, Traffic activated) marking my thermals.

Now I had to compare with the thermals from the advanced settings.
With ASK13 (YASim model) Was the thermal_demo thermals well and evenly
centered. Equally, the thermals from the advanced setting be used.

The aircraft asw20 and DG-101G (jsbsim) is the thermal_demo thermals as the
ASK13 usable. Centered and uniform Variometer (VSI) rashes.
The automatically generated thermals of FlightGear are difficult to use.
The Variometer wriggles wildly.

What can you do about it? I'm not able change the FGFS source to make.

I have installed tone to the DG-101G e-variometer (vsi).
In the asw20 I installed with ctrl-t generating thermals arround asw20 (from DG-101G copied).
With both aircrafts I have many successfully flights done in the past month.

I will upload my changes to

https://svn.code.sf.net/p/flightgear/fg ... /Aircraft/

asw20 and DG-101G with my changes using SVN-commit to the /fgaddon/trunk/Aircraft.
I use Tortoise SVN on my computer (Win7 OS). (But) Succeeded, it is not me.
Please give me the idea how I can do that.

Must for asw20 and DG-101G upload my changes to fgaddon/trunk/Aircraft ?

D-ABBA