Wheel roational inertia mod. Volunteers are welcome!

[NaN]

The wheel rotational inertia on a lot of cars are wrong. I am looking for volunteers with vdrift-data svn access to fix them.
You may use the calculator on this website: http://stephenmason.com/cars/rotationali...html#calcs

[joevenzon]

So it sounds like a realistic value is around 0.5 or 0.6? At values that low, the simulation seems to have some issues....

[NaN]

A car tire weights about 10kg. A wheel is between 6-20kg.

For the MI(i think it has the smallest wheels in the game):
Wheel: 10"
wheel radius = 0.5 * 10 * 0.0254 = 0.127m
wheel weight = 10kg
wheel inertia(all weight in rim) = 0.16129kgm2
Tire: 165/70-10
tire radius = rim radius + 0.165 * 0.7 = 0.243m
tire weigth = 10kg
tire inertia(all weight in tread) = 0.59049kgm2
total inertia = 0.75178kgm2

For the 360:
Wheel: 18"
wheel radius = 0.5 * 18 * 0.0254 = 0.229m
wheel weight = 10kg
wheel inertia(all weight in rim) = 0.52258kgm2
Tire: 215/45-18
tire radius = rim radius + 0.215 * 0.45 = 0.325m
tire weigth = 10kg
tire inertia(all weight in tread) = 1.0563kgm2
total inertia = 1.57888kgm2

So realistic values would be somewhere between: 0.75 - 2.0
Will the simulation be able to handle it?

[joevenzon]

With the 0.75 value on the MI, in first gear it will peel out from a stand still. This is because the lower inertia is allowing the engine to move the wheel quickly enough that the slip ratio is high enough to cause the pacejka equation to lower the force that the tire generates. I kind of doubt that the MI will really do this, so I tweaked the tire's longitudinal force coefficient b0 from 1.6 to 1.3, which lessens the falloff in grip at higher slip ratios (note that changing other parameters such as b3, b4, b5, etc will do this as well) and now the MI drives normally again.

So after looking at it a bit, yes, I think it's fine and will increase realism, although the tires may need some adjustments afterward.

[NaN]

I see.

Alternatively we state in the wheel parameters description that the inertia has to be n times the real value.

Or we could add a correction factor(like n times the real inertia) into wheel inertia load code and comment it explicitly as a numerical integration workaround. Just for the case someone(like me) looks into the car parameters and thinks they are wrong, tries to fix them.

I will look into the pacejka coefficients. I found a paper online containing coefficients for a Goodyear FSAE tire. :lol:

edit:
The motivation to fix the inertia came from the desire to get a correct wheel lock up torque, as it's proportional to inertia and angular velocity. On the other side the lock up torque also depends on tire grip/friction. So I am not sure if adjusting the tire coefficients is going to help.

[joevenzon]

I think it's worthwhile trying the real numbers and see how things go. The tire coefficients for the MI were probably very incorrect anyway since tiny little tires behave a lot differently than big thick tires with stiff sidewalls.

I checked in a pacejka editor recently with little fanfare, you can get it here:
http://svn.vdrift.net/repos/pacejka/trunk

The best guide I've found to understanding how real tire properties affect the coefficients (the website seems down so here's the wayback archive):
http://web.archive.org/web/2005091305222.../epcjk.htm

Brian Beckman has some info on his site:
http://www.miata.net/sport/Physics/Part21.html

[stan.distortion]

Quick note, brakes and transmission also need to be included so the weights will be quite high, especially for the transmission as drive shafts, gearbox internals etc. are often spinning a lot faster than the wheels.