Re: [Flightgear-devel] Abnormal surface friction factor

[Wayne B. <kil...@co...>]

Dany, thanks for taking the time to dig into this code and explain the logic. That make sense. 

Couple of things here, static VS rolling. Typically you would always think static should be higher than rolling, it’s not. In a very short test I saw ranges from .7 to 1 for static values. I need to test more terrain types. I’m not sure exactly what relationship comparison these exact properties have in the sim though.

Typically
Static, how much force does it take to make the tire slip from a standing position.
Rolling, how much force does it take to make the tire roll.

When I see a static friction factor of 1 and a rolling friction factor of 50, I automatically think something is wrong. 50 to 1, rolling to static is wrong in every case in my opinion, again context might make a difference here.

I did a ton of work on the aircraft side to gain the ability to distinguish between the various gear and undercarriage changeouts. I went back to grass in the Cub and the c172p, both aircraft I did this work to. They still perform as I expect. On grass the standard tires need a bit of throttle to start rolling, with tundra tires, it rolls much easier. Grass is rolling = 5. So I think I maybe jumped the gun on something being different. The high end may have always been 50.

But that doesn’t dissuade me from my argument that 50 is way to high for any terrain other than mud (this need qualified).

A couple points.

1) static values VS rolling values, makes me think somewhere along the line things got mucked up somehow.
2) Town and city with all the buildings and potential cover values is rolling = 1. So the argument that it is supposed to be forest with trees and therefore we should have a high rolling value, I think is not a valid argument. 
These two cover types are at a completely different ends of the spectrum, yet in reality both could have valid arguments for sitting on either end of the scale.
  
It should be based on the ground type and not necessarily what is growing or sitting on it. We should figure out a robust collision system though. We allow landing on urban cover, we should allow for landing on leaf litter. I wouldn’t have issue if it were high enough that most default tires couldn’t roll on it. But it needs to be low enough that a larger tire could accomplish a landing. Please try either the dev version of the cub or the c172p on grass (I don’t know if the 2020.3 FG version has the work included in it yet) and switch between the default tires and the largest tundra tires available and you will start to get why I am so adamant on making this more realistic. It’s a really cool feature. Erik has a chance to extend this unique feature by adding weather elements to it. Is it wet or dry, frozen or not. Both easy to adjust for with his season system.

David’s slant on more robust scenery is not without merit. If we had large sections of forest that you could only land on a road or clearing, that would be great. Same for a river bottom, you need a sand bar or a flat, uncluttered piece of ground. Now what is the best way to accomplish this world wide? I don’t think cutting it into the scenery is going to be viable in some cases. Specifically forests, tree collision I think would be the answer here. Partly because there are many who dial down the physical tree count and the land looks mostly bare. I guess a large fire or loggers worked all that land. Also, technically a tundra tire could possible handle a lot of that ground cover where there is access to the actual ground.

I think am getting hung up on the millions of acres I can’t land my bush aircraft in at all. Even when I see a completely open area, Same with river bottoms, depending on the land class.

I need to go through this code completely in order to make a valid argument, I just haven’t had the time lately. I think this will prove to be an easy fix though. To Erik and Stuart, I hope you will at least keep this in mind as you go through your efforts of improving the scenery systems. Again, I offer to help you any way I can, just ask.

Wayne

From: Erik Hofman
Sent: Thursday, June 15, 2023 4:12 AM
To: fli...@li...
Subject: Re: [Flightgear-devel] Abnormal surface friction factor

On 6/15/23 10:49, Dany wrote:
> I think I progressively understand the logic (including the 50 factor).
> 
> Everything starts from the (geodinfo) rolling_friction.
> On asphalt, this (geodinfo) rolling_friction is given the value 0.02. 
> Which, multiplied by 50, gives rolling_friction_factor = 1. In this 
> case, no correction, the final rolling friction coefficient is the value 
> given in the aircraft FDM:
> (in the FDM, e.g. <rolling_friction> 0.03 </rolling_friction>)
>   /fdm/jsbsim/gear/unit[]/rolling_friction_coeff
> 
> On another terrain, the (geodinfo) rolling_friction has another value. 
> Let's take "grass": 0.1. Which, multiplied by 50, gives 
> rolling_friction_factor = 5. Thus, the final rolling friction 
> coefficient = 5 x the value given in the aircraft FDM:
> 5 x /fdm/jsbsim/gear/unit[]/rolling_friction_coeff
> If the aircraft (FDM) initial rolling_friction_coeff = 0.03, it gives 
> 0.03 x 5 = 0.15 for the final rolling friction coefficient.
> 
> Example: C172P, 180hp. Full thrust at rest: 549 lbs.
> If the aircraft weight is 2000 lbs, it will not move if the final 
> rolling friction coefficient is greater than 549 / 2000 = 0.275.
> 
> If the aircraft will not move on a terrain, the first (and easier) thing 
> is to check if the (geodinfo) rolling_friction is not the culprit.
> 
> Is it true? Please fix me otherwise, thank you.

Yes, that's the idea.
This would allow a different friction factor for different terrains 
while preserving the possibility to set the friction coefficients in the 
JSBSim configuration file.

Erik

-- 
http://aeonwave.xyz
High performance audio scenegraph and sound simulation.



_______________________________________________
Flightgear-devel mailing list
Fli...@li...
https://lists.sourceforge.net/lists/listinfo/flightgear-devel