The problem with using Tyre Curves.

[quickquestion]

Ok, here is an example of a Tyre Curve:

This is totally missing a dimension: speed.

If a car is going 1 mile per hour...shouldn't it have 100% grip and no slip at all times?

Similarly, if a car is going 200 mph sideways, shouldn't it have much less grip than the diagram implies?

 

So I'm basically unsure how to approach this in terms of using this for realistic car physics.

[imatfaal]

Where is the graph taken from ? Might help to be able to read axes and context.

 

In a simple system maximum frictional force is only related to the materials of the two surfaces and the normal force. I know from extensive testing that for a bicycle this is still the case and width of tyre etc has very little bearing on the matter - all you need to know is the tyre type, road surface, and the mass of the rider (and not even the mass of the rider for flat roads). Car tyres might deform slightly more (my bike tyres are at 7.5 bar which is far higher than any car - but then the tyres are only 25mm wide) but I see no reason that their maximum frictional force should vary much from the simple model which holds for bicycles.

 

Remember that friction is a balancing force and most of the time will be in equilibrium - in most cases it will be below its maximum possible value that is predicted by F_fric = N * mu_s. When braking at the absolute limit (but not skidding) or when accelerating at the limit (but not wheel spinning) then you closely approximate the maximum value.

 

If you are talking about locked wheel skidding then you use the kinetic coefficient of friction F_fric= N * mu_k

[quickquestion]

Where is the graph taken from ? Might help to be able to read axes and context.

 

In a simple system maximum frictional force is only related to the materials of the two surfaces and the normal force. I know from extensive testing that for a bicycle this is still the case and width of tyre etc has very little bearing on the matter - all you need to know is the tyre type, road surface, and the mass of the rider (and not even the mass of the rider for flat roads). Car tyres might deform slightly more (my bike tyres are at 7.5 bar which is far higher than any car - but then the tyres are only 25mm wide) but I see no reason that their maximum frictional force should vary much from the simple model which holds for bicycles.

 

Remember that friction is a balancing force and most of the time will be in equilibrium - in most cases it will be below its maximum possible value that is predicted by F_fric = N * mu_s. When braking at the absolute limit (but not skidding) or when accelerating at the limit (but not wheel spinning) then you closely approximate the maximum value.

 

If you are talking about locked wheel skidding then you use the kinetic coefficient of friction F_fric= N * mu_k

Its just a generic graph found in most car physics lectures and documents.

X axis is slip angle, and Y axis is Normalized Lateral Force.

 

Knowing the Normalized Lateral Force is somewhat useless because obviously, if car is going 1 mile an hour, the Normalized Lateral Force is not going to be "6000 pounds at a slip angle of 4" like many charts say.

 

Now, the deeper you get into car physics the more confused you get (At least I did.)

 

For locked wheel skidding it implies that the equation of Fmax=uK*weight is inaccurate, since the friction coefficient changes based on slip angle. For example this chart shows it as such.

Let me explain the reason for this. Newtonian friction models are an inaccurate representation of friction modelling...friction is not a single-direction thrust vector like the Newton diagrams imply.

In reality, a wheel behaves like a cog on rails longitudinally, and laterally it behaves similar to a damping field.

Edited April 12, 2017 by quickquestion

[Bender]

In reality, a wheel behaves like a cog on rails longitudinally, and laterally it behaves similar to a damping field.

Longitudinaly, a belt on a pulley is a better analogy, since there is always slip. Laterally, it is sliding friction and the curve does not apply.

 

The normalised force is used because it depends less on other variables such as load and temperature. Often you only need the slip at maximum force anyway (except in eg model predictive control).

The curve is used for slip control to maximise the force during acceleration. The left part can be used to maximise breaking (or to explain eg ABS).

[quickquestion]

Longitudinaly, a belt on a pulley is a better analogy, since there is always slip. Laterally, it is sliding friction and the curve does not apply.

 

The normalised force is used because it depends less on other variables such as load and temperature. Often you only need the slip at maximum force anyway (except in eg model predictive control).

The curve is used for slip control to maximise the force during acceleration. The left part can be used to maximise breaking (or to explain eg ABS).

Yes, belt and pulley is a better analogy for my purposes (large scale). Gearcog is what I meant on a microscopic level.

 

Anyway, I would like to explain what exactly the problem is, more clearly.

The slip charts are fine for large velocities, where the car is sliding at 200 mph. At these times you can just apply the maximum amount of lateral force for each tire.

But my problem comes in when figuring out what amount of lateral force to apply at low levels.

For instance, say a car is going at 3 or so units per hour, and the car is sliding a bit and Each tire is at a random angle (This is a hypothetical test car for simulation accuracy, that has tires that can point every which way.)

What amount of force do I apply to each tire, such that the car's grip behaves realistically?

To make it easier on ourselves, let's just simplify things and say these are very sticky tires that should totally grip completely at low speeds.

 

The problem with these charts is that most likely these are done with Single Tires on a belt in a lab...

It does not tell you how each tire behaves once it is connected with 3 other tires on a chassis.

I would pay actual money to find a chart or diagram that shows me 4 tires with different rotations, and how much lateral force is exerted by the tires to keep the car gripping.

Because with 1 tire it is much more easy, I simply put the lateral force to tiremass*tirelateralvelocity and apply that as a 1 frame impulse, to cancel the velocity of the tire.

But with four tires all with different rotations, I can't simply use the carmass (because each tire is not part of the total car body itself, it is only part of a fraction of the carbody, and the fraction of the carbody is also dependent on the other tires - for instance, if only the front tires were going against the grain (because they had a steering angle of 90 degrees), but the backwheels were totally rolling with the car, then the fronttires would use the total carmass to stop the car. Or, if it was a tricycle, and only the front wheel was 90, then the whole frontwheel would have a carmass*tirelateralvelocity applied to it to stop the car.)

Edited April 14, 2017 by quickquestion