ODE and mu contact parameter

Started by
4 comments, last by Eddycharly 15 years, 5 months ago
Can someone explain to me what the mu contact parameter is used for in ODE ? I understand it's used to calculate friction but how can i determine the value i must set it to ? I am trying to implement a car with ODE. For now, i want my wheels to never slip so i put mu to dInfinity. I also put slip1 and slip2 to 0. As long as i drive in a straight line, that works fine. As soon as i turn my front wheels, the car stops instantly. Now if i put mu to 5 (slip1 and slip2 to 0), i can drive in straight line and steering works. Why does steering works with mu = 5 and not with mu = infinity ? What do slip1 and slip2 really mean ? How all these parameters are used together by ODE ? All in all, how can i determine the parameters i need to set and their value in order to produce a given behaviour ? Thanks in advance.
Advertisement
Quote:Original post by Eddycharly
Can someone explain to me what the mu contact parameter is used for in ODE ?

mu is the maximum amount of friction force that a surface can exert per unit of normal force (pushing against the ground).

Quote:I understand it's used to calculate friction but how can i determine the value i must set it to ?
Some googling will turn up plenty of tables of mu values. Search for "friction coefficient" or such.
Quote:For now, i want my wheels to never slip so i put mu to dInfinity.
I also put slip1 and slip2 to 0.
The two values you should rarely use in physics simulations are infinity and zero. Neither one occurs very often in nature, and physics simulations are intended to mimic nature and do not do well with unrealistic situations.
Quote:
mu is the maximum amount of friction force that a surface can exert per unit of normal force (pushing against the ground).


Thanks, that makes a lot of sense, that's why it's harder to make something heavy slide than something light.

Quote:
For now, i want my wheels to never slip so i put mu to dInfinity.
I also put slip1 and slip2 to 0.


Well, now i understand mu, can you explain what slip1 and slip2 are used for please ?
What is this force-dependent-slip they talk about in the ode user guide ? is it something that exists in real life too ? or is it an ode concept ?

For now, i think that if my car stops when i rotate the direction wheels is normal. The forces involved, with mu at infinity, create a set of constraints that cannot be solved.
At least, one thing i'm almost sure is that i need a differential, as on real cars.

Thank you.
Force-dependent slip appears to be an attempt to model a non-coloumb-based sliding which is caused by small-scale deformation of the tire surface. It's an exclusively ODE thing, and TBH, I don't really know much about it.
Thanks, i'm going to play with mu for the moment. I'll keep slip1 and slip2 for next week.

I made a quick test to experiment with mu. I hope it's correct.

I put my car in front of a wall, so that it is blocked.
My car is 200 kilograms, gravity is 9.81, so the weight force should be 1962 newtons.
Now, i accelerate, and my wheels start to slide when rear axle torque is around 340 Newton/meter.
I set mu to 1.

I think it's not correct. Weight force is 1962, and i have four wheels. It gives me 1962 / 4 = 490.5 newton on each wheel.
As mu is 1, my wheel should start sliding when i put more than 490.5 newton at tire contact, right ?

My wheel radius is 0.7 meter, so the minimal torque needed on the axle to make the wheel slide should be 490.5 / 0.7 = 700.71 newton/meter, isn't it ?

Why does it start sliding at 340 N.m-1 of torque ? did i miss something ?

Another question, do you think that ODE will take weight transfer into account ? Will it transfer more weight from the car body to the rear wheels when i accelerate ? or is it something i need to code myself ?

Thanks a lot.
Hmmm, it seems like i was wrong.

torque = force * radius

and NOT

force = torque * radius

With the first formula, the minimal torque needed on the axle to make the wheel slide should be 490.5 * 0.7 = 343.35.

Now it seems to be correct.

This topic is closed to new replies.

Advertisement