Archived

This topic is now archived and is closed to further replies.

d000hg

Wheels skidding

Recommended Posts

d000hg    1199
I just added the start of tyre friction to my car model. I''m letting it be frictionless in the direction the wheel points for now. I''m currently applying a force (F=constant * suspension force up from ground) perpendicular to this direction to oppose this motion. But friction should stop it getting that velocity to start with not just oppose it once started (sticking friction as well as sliding). When I turn the wheels at more than slow speed, it turns but not as sharply as it should. However there are no spin-outs just the front wheels slide while causing a small turning effect. When I straighten the wheels the car quickly returns to a straight path with no over/understeer or sliding - I''m unable to get the car to slide. I''m a bit stuck how to improve this - any suggestions?

Share this post


Link to post
Share on other sites
Squirm    481
The ODE (Open Dynamics Engine) manual has a quite good explanation of friction on car tyres in it. Of course, if assumes you''re using ODE, but the methods involved are fairly clear - it would be worth having a look.

Share this post


Link to post
Share on other sites
d000hg    1199
quote:
Original post by Squirm
The ODE (Open Dynamics Engine) manual has a quite good explanation of friction on car tyres in it. Of course, if assumes you''re using ODE, but the methods involved are fairly clear - it would be worth having a look.
Happen to have a link?

Share this post


Link to post
Share on other sites
szinkopa    198
I suggest using a better model, like Pacejka.
With bad tyre model you will get bad carsim.
Goggle for it, Physics of racing series describes it.
It needs Fz, slip angle, camber angle for lateral force.
For longitudinal it needs Fz, and slip ratio.
You will have perhaps troubles when combining these 2 forces together, because there is a friction circle, which shows the maximum force at a given point of time.

http://phors.locost7.info/contents.htm
http://www.racer.nl/

Share this post


Link to post
Share on other sites
d000hg    1199
Well initially I am modelling a tyre as having sticking friction and sliding friction (in direction it skids not the way it wants to travel). But all I''ve actually implemented so far is sliding friction which is probably why it''s impossible to powerslide - it''s the interface between sticking and sliding on which this rides I believe. And the racing sim tuts just talk about F=uW until the final few chapters so I feel justified with that for now. This is all with the wheel treated as no friction in the direction it''s pointing, once I put braking in it''ll have to become more advanced!

Share this post


Link to post
Share on other sites
d000hg    1199
I''ve looked at this some more and still confused. It''s still very simple - if a wheel is moving perpendicular to the direction it''s pointing then a restoring force proportional to the vertical force on the wheel from the ground is applied to counteract this. While this makes the car turn properly at low speed, at high speed (>30mph) it displays what I believe is understeer - the car turns but not as tightly as it should hence the front wheels are skidding. Now this would be fine if I could generate oversteer but I can''t. If I have drive applied to the front or back tyres the effect is basically the same, also if I move the CofM of the car forwards. The aim is to get the car into a slide; I''ve just managed it but only by implementing 4-wheel steering which is ot what I want. While very simple my model should generate correct spin behaviour I believe. What''s wrong? I ight be able to stick up a demo...

Share this post


Link to post
Share on other sites
oliii    2196
maybe try...
- unbalance the torque between rear and front wheels. more torque to the rear wheels. they should slip earlier than the front wheel. also, a car with front wheel drive will understeer.

- shift the weight of the car slightly forward.
- unbalance the grip levels between the front and the rear. low grip at the erar should make the car more tail happy. although usually, the grip at the rear is higher, for rear wheel drive.
- use a limited slip differential. apply more torque at the wheels outside the corner than inside the corner.

Share this post


Link to post
Share on other sites
szinkopa    198
The fact that your car''s behaviour does not change when you change the configuration is due to the simple and not realistic tyre model. A tyre is quite more complicated than a simple static and dynamic friction. With a steel wheel this model would be suitable, but with a deformable tyre it''s not sufficent.
I suggest reading this page:
http://www.tut.fi/plastics/tyreschool/moduulit/moduuli_10/hypertext/index.html

Share this post


Link to post
Share on other sites
d000hg    1199
I don''t think that my model is that innacurate. Weight transfer is one of the most important processes for grip and this seems to be working very nicely.
Currently though I don''t even have sticking friction - I find this hard because by definition the restoring force is such that velocity is zero. But how do I work out what force to apply to the wheel so it''s sideways velocity is zero since it''s at the corner of an inertial body?

Share this post


Link to post
Share on other sites
szinkopa    198
Do I understand well? You want to calculate, how much force has to be applied in order to not slide the tyre?
If so, then I say it''s not the way it should be done.
You should have a static or dynamic tyre model. Static is simpler and is sufficent for such simulations. Then your tyre model has input data: Fz, slip angle, slip ratio, camber angle, and whatever you want. From this the model calculates the longitudinal and lateral force the ground exerts on the tyre at the contact patch. This is the output of the model. The calculation depends on the tyre properties (lateral, longitudinal stiffness, pressure, temperature,...), and the ground properties (for example DIV (difference in velocity) based friction coefficient))
Of course you may want to use simpler model, I think.
Don''t forget that even if your tyre is slipping, the frontal region of the contact patch may stick to the ground. The calculation of the area of this is the tyre model''s job. But actually it does not matter, you just have to apply the total lateral and longitudinal force.

The dynamic tyre model has state variables, so it can simulate transients, which is more realistic.

So you can choose: use static empirical model (Pacejka), or static analytical model (based on brush model). I don''t suggest using dynamic model yet.

Could you upload, what you have now? I am interested. Maybe I can help more.

Share this post


Link to post
Share on other sites