Tire force combining review

[NaN]

Here I am back again playing around with tire physics

Some time ago I've added a tire size correction factor function to the Pacejka tire model. The calculation is derived from the physical model, scales max friction based on tire dimensions.

The results are pretty good. Looking at the 360, with its 215mm front and 275mm rear tires, the maximum rear tire force is scaled so that the yaw torque on the car when sliding is almost zero (it is a tiny bit larger on the rear, like 1%). This explains the neutral handling, and also the reason why the car has those tire dimensions!

To test this observation I've looked at the F1-02, another car that is hard to handle. I replaced the custom tires with the default racing variant to enable scaling, increased rear tire width from 355mm to the 380mm per F1 rules, adjusted ballast position to get front weight from 43% to 44% as stated in the specs. And again I am getting zero yaw torque, which is pretty cool.

I'll push the F1-02 changes soon, maybe also add yaw torque calculation to cartest, as it allows to see how well balanced a car is.

[colin78]

I just read this awesome post!! great job!!
any explanation on the parameters of the tire model?
any info about the temperature/pressure influence to the grip?

[NaN]

I just read this awesome post!! great job!!
any explanation on the parameters of the tire model?
any info about the temperature/pressure influence to the grip?

The tire models I am writing about are contained in vdrift source code github.com/VDrift/vdrift and also here (friction curve editors) github.com/logzero/vdrift-tools

Tire pressure is a parameter in the physical model affecting contact patch size and pressure. It should also have some effect on carcass stiffness, but I have no data on that. Currently a fixed pressure value is used for vdrift tires.

Temperature is a very interesting parameter. Some might argue the most important one, responsible for the nonlinear friction behavior. Interesting for me is probably how it affects peak friction and cornering stiffness.

I haven't looked into implementing it yet. Here is a paper that might be a good starting point: publications.lib.chalmers.se/records/fulltext/143233.pdf

[colin78]

great things...
yes, 'cause the temperature affect 2 things:
the core temperature just the pressure (linear changes) and the tread temperature (that affect the grip)

the process should be like this:

- the slip of the wheel heat the tread, that changes the grip (I think that for the other sim the optimal grip is between a range of temperature, if is less than the min or more than the max, the grip is reduced.
- heating the tread, the air inside the tire will be heated
- heating the air inside, the pressure will be increased

if the car run faster, wiith a small amount of slip (straights) the tread will be cooled and after also the core

the road temp influence the tread heat, like the air temp

[2spitfire2]

@NaN
Did you ever give the implementation of that paper a shot? I had trouble implementing it because of the missing coefficients.

I found this a while ago, it's from 2017 and also compares a couple models to the Chalmers paper you linked. The required coefficients are mentioned below the graphs in 2.4. Very cool stuff, but I'm struggling a lot to convert the equation into code.

https://www.researchgate.net/publication...modynamics

https://www.academia.edu/36799804/Optimi...modynamics

[NaN]

Sory for the late reply. RL kept me busy.

From a quick look their temperature model has 3 coeffs: H = 350J/K, k 1 = 5W/K, k 2 = 1s/m

The peak friction coeff is a parabola with a peak at 1.05 at 80°C and about 0.96 at 0°C.