Tyre forces vs speed

[Jersey Tom]

@JT: I still believe you've got your terminology confused, "physical" is by no means a synonym to "analytical".

I don't think I'm confused at all. In the world I work in - the world of vehicle dynamics, tire data, etc - they are used synonymously. That's just how it is. Don't know what else to say. You can have an empirical / curve fitted / test data driven model... or a physical / theoretical / analytical model. Applies to vehicle characterization as much as it applies to tire characterization. You can have a suspension which is defined by curve-fitted K&C data, or you can have one that's computed on-the-go by a multibody simulation. Each have their own strengths and weaknesses.

Beyond that, this is so past the point of this thread that it borders on asinine. But this seems to be the case every time you bring something up - it's not to further the value of the thread through meaningful discussion - it's prodding at trivialities and unproductive BS. Like I say, this is the world I work in.. and I like to shed some light on things from time to time using what public domain data as I can. If you're not interested in in it, then go do something else rather than derail discussion.

On to the matter at hand...

So unless, we define the term "speed" a bit more precisely in the OT quoted statement, I would be careful, to dismiss the statement out of hand or call it BS.

Trimming this quote down quite a bit, but to me... it effectively makes no difference whether you're thinking of speed with sliding velocity or whatever.

For one, if you're using something like an industry standard Magic Formula model... the post-saturation "sliding" characteristic of the tread is already implicitly defined in the model by the lateral and longitudinal force shape factors. Don't even have to worry about it - it's already built in. With regard to physical models - I don't worry about them because for the vast, vast majority of end users you're just not going to have pressure / velocity / temperature sensitive data so it's totally impractical. Besides if you're way post saturation in your simulation (i.e. spinning out or brutally beating up on the fronts) you have bigger issues to worry about. I leave that realm to the video gamers.

Then with regard to tire rotational speed - again, there's a reason why even in the latest and greatest empirical models after 15+ years of development, there's not even a single parameter given to it (turnslip and parking maneuver behavior excepted - but that's not exactly of concern in racing). Just completely drowned out by the other state variables. Maybe that's just my own working experience.. after going through gigs and gigs of data, some things are more on my radar than others.

More to the point, even if you DO add it in - it doesn't mean you're going to get any value out of it. In my earlier post - your vehicle speed at any point on the racetrack is what it is. Not going to change by a dramatic amount. I'd rather focus on the things that ARE in my control, namely load, camber, slip, and inflation.

It's very easy to get into the mindset that just because something is there that you have to model it. Or that racing is all about "all the little things" and that's what's most useful to track down. You can spend the time to make your model more complex and then still arrive at the same end answer; you've added cost and complexity but it hasn't been value-added work. Given that my experience in racing has been that you are dramatically time limited in how much you can get done for a race, I see MUCH more value in getting a "close enough" answer today, rather than a marginally more precise or complete answer several days or a week from now.

My MO is to be as effective and impactful as you can, use the simplest models or predictive tools as you can - not the most complicated or involved.

[xpensive]

I beg to differ, a mathematical model is either analytical or empirical, the xpression "physical" obviously refers to something completely different, why don't you look it up in a dictionary?

In all honesty believe that you made it up as you typed it, in that "world you live in", didn't you?

[munks]

I can confirm that tire models are usually categorized as empirical or physical. Regardless of what the dictionary says, this is how people use it.

EDIT: I might add, though, that each can have elements of the others. My understanding is that the simplest form of Pacejka's curve is equivalent to a simple brush model, although I haven't confirmed that mathematically. And likewise with a physical model, at some point you need curves to define, for example, the rubber grip vs. sliding speed & temperature on a given surface. At least until we have the computing power to model every molecular chain.

[Jersey Tom]

Like it or not, that's the terminology that I see / hear used. How precise it is to Merriam Webster's standards is irrelevant. This whole discussion of semantics is irrelevant and unproductive. If you want to continue to be stubborn about it, that's your problem. I'm trying to discuss the topic of the thread.

[xpensive]

Like it or not, having a technical discussion with someone confusing basic relations with out of context terminology is pretty useless, just like when you some time ago argued for retaining the Imperial unitst agains SI-units, by referring to "Planck's constant", as if that had anything to do with the kind of engineering or science that was being discussed?

I think your repeated use of the term "physical" is a tell-tale of the true status of "the world you work in".

Once in a while my doctor gives me a physical exam and I don't mind getting physical with my mistress however!

[GSpeedR]

Like it or not, having a technical discussion with someone confusing basic relations with out of context terminology is pretty useless, just like when you some time ago argued for retaining the Imperial units by referring to "Planck's constant", as if that had anything to do with the kind of engineering or science that was being discussed?

I think your repeated use of the term "physical" is a tell-tale of the true status of "the world you work in".

Once in a while my doctor gives me a physical exam and I don't mind get physical with my mistress however!

Physical -> based in physics.
Empirical -> based in observation.

As mentioned above, these are standard terms for tire models.

[xpensive]

Must be something lost in translation there, Imperials perhaps?

[Jersey Tom]

I think your repeated use of the term "physical" is a tell-tale of the true status of "the world you work in".

The world I (and others here) work in revolves around winning races on Sundays. It's been my profession for the past 5 years. Whether or not all the terminology we use is academically 100% precise or meets your expectations or is equivalent to any other industry is inconsequential IMO. The rest of us seem to be in agreement. You can dig through SAE papers on the topic which use the same terminology as well. It is what it is, we're just calling it like we see it.

The other aspect here - I don't want it to come across like I'm arbitrarily dismissing speed sensitivities as "BS." Does running a tire at varying speeds have an effect? Sure. Everything does something. I just can't think of a situation where taking into account direct speed dependency on in-plane forces is a mission critical item... or more to the point, would affect the decisions made from simulation results. That being the case, I keep it off my radar. The less complicated and convoluted things can be the better.

That industry standard tire models follow suit in not having speed dependency on in-plane forces, even after decades of use and development, I think backs up that I'm not alone in this opinion. Incidentally, I'd argue that a Magic Formula model is to an extent better at broad application / to-the-limit use than it is at low slip angle on-center behavior. But that's a separate topic.

[Jersey Tom]

My understanding is that the simplest form of Pacejka's curve is equivalent to a simple brush model, although I haven't confirmed that mathematically.

Almost missed this. From my recollection they are not equivalent, at least the closed-form example solution for the brush model given in Dr. HP's book - which I believe is dependent on the footprint pressure distribution assumed.

Even the brush model which one could say tries to have some physical meaning to its parameters... still requires an empirical definition of that footprint pressure distribution. I think Pacejka's example just uses a parabola.

[gato azul]

all fair and good JT, I see where you are coming from, and I don't disagree with most of what you are saying.
Nevertheless, I do think, that you are a bit "in haste" , calling things BS, just because they do not fit your experience, opinion or are related to your field of expertise/work.

From the OP, it is not clear, that his question is related to racing or cornering, as you mainly imply in your answer(s).
For all we know, he may needs to design an ABS or TC controller, where these effects are perhaps more dominant, than in cornering.

Furthermore, do we assume "a perfect tyre" or "a real tyre" which may (or may not) have any amount of non-uniformity to it, and all the effects which come with it?
I'm not trying to nitpick or have a go at you, but your statement

Says who? I call BS. Or if it were to, it would be by a very small amount. Load, camber, slip angle, and slip ratio sensitivities are the big ones. There's a reason why tire models have those as variables with generally no direct speed parameter.

sounds like a bit of "hand waving" to me.
Just because, it is not important for you, does not mean that the effect(s) don't exist, nor does it make the statement quoted in the OT BS.

as for "industrial standard models", well which model do you refer to?
I assume MF, because it does not has a direct speed dependency, so what, there are a couple of others, which have:
among them:
- LuGre
- Bliman
– Kinematic
– TameTire
- TreadSim

Wonder why this is, when it is totally useless according to your assessment.
As I said, it depends on what you try to achieve, and what you define as "mission critical", that can mean that for you it can be perfectly o.k. to use a MF model (which is mainly based on "static tyre data" curve fits), but is fast and easy to integrate into simulations, but for others it may is not.

some test data for lateral forces @ different velocities

and the fitting in the various tyre models


some test data for longitudinal forces @ different velocities

and the fitting in the various tyre models


I will leave it to you, to decide if the variation in longitudinal forces vs. slip ratio for different velocities in the posted data is significant or not.

[Jersey Tom]

Valid points. To be fair though, many of those charts are model data rather than test data... and even most of those test curves have been pre-processed. That the four different model responses and test data are so different makes a point in itself... and I'm wary that the LuGre model for example predicts such an immense change in peak forces over a relatively small range of velocity (up to 60 mph).

But yes, I'd call MF industry standard, moreso than other models. MF is just very prevalent. Things like Tametire I almost view as academic. It's cool and all but the sheer volume of extremely specialized test data you need to populate it (and that you'd likely only get it from a tire supplier) makes it impractical for most applications IMO. That's been my view of things.

Still the biggest and most important question is - even if there is a real effect there (and I'm not denying there is), is it going to affect the end result of your car design / setup? THAT is the question of significance - not at the component level of each individual tire but of the package as a whole I'd argue that most if not all the time, it won't. Just my opinion. Like I say, you can add model complexity and take into account more and more effects of things but it's not necessarily value added work if there's no way you can exploit it.

So yes, we can add these extra model factors but if it doesn't get you to a better system design... has it been practical or just academic?

Back to the original question of this thread, I still call BS that speed globally reduces forces at any given tire state. If anything, the rest of these models and charts have backed that up.

[Greg Locock]

gato_azul, yes that is useful data (I'm not so fussed about models). Which book/paper is it from?

[munks]

JT- Thanks for the reply on the Pacejka vs. brush model, I read that somewhere (maybe even on a forum!?) but it must have been mistaken.

Gato Azul - I'm also curious where that data came from, particularly that labeled Figure 6.2. I don't recall seeing such data before, so that was cool.

[gato azul]

the data are out of an thesis, and where collected using the TNO Tyre Test Trailer

you can find some more/similar data here:
Implementing inflation pressure and velocity effects into the magic formula tyremodel


There a couple of papers from Guo et al. out there, which deal with the subject as well.
for example:
- Effect of Tire Rolling Speed on Cornering Characteristics of Vehicle
- Affecting Factors Experiment in Tire Lateral Force
- Influence of dynamic friction property on the combined side and longitudinal slip properties of tire
- Experimental Research on Friction of Vehicle Tire Rubber

K. Guo, Y. Zhuang, D. Lu, S. Chen, andW. Lin. A study on speed-dependent tyre-road friction and
its effect on the force and the moment. Vehicle System Dynamics, Vol.43(Supplement):329–340,
2005.

So, I will leave it at this, that should be enough information for anyone, to make up their own mind, if they feel like ignoring the effect, or considering it - it will mainly depend on your application.
The effect is stronger in longitudinal direction, then it is in lateral, so in TC, ABS and ESC controller design it becomes more of an consideration in newer designs.
For a lapsim or game model or to investigate pure lateral performance of an free rolling tyre, you can probably ignore it.
There are a couple of other papers out there, mainly about aircraft tyres, with some test data.

Just for giggles, here some historic data, dealing with the subject (1956)

[gato azul]

But honestly i was not able to find any reference explaining why and how much this phenomenon is influencing tyre force at contact patch.

@Silente
in parting, just an (perhaps lame) attempt to provide an direction to look into for an answer to your original question.
I'm not sure how deep you want to go into this, if you have a concrete application in mind, or if it was just and
question out of curiosity and general interest, nor do I know your background.
There is no cut and dry, simple answer to it, I suppose, at least non I'm aware of.

The way I see it (which as always can be wrong&subject to change), one of the parameters in this phenomena, is the interaction of the tyre with the surface macro/micro texture of the ground/road/track.
For a given texture (wavelength etc.) the frequency with which the tyre (rubber) gets excited will change with velocity (over ground), especially in situation, where large parts of the tyre patch are in "slip" condition.
The energy dissipated, due to hysteresis in the rubber, & heat generated by this, and it's effects play a role in this.
It's a multidimensional problem, as almost anything tyre related, and there are different theories and explanations "floating around" in "academia land".

If you have an deeper interest into this, you may look for some papers/publications by B.N.J. Persson, he (some co-authored by others) has published quite a few over the last years on the subject of tyre road surface interaction.
But they are mostly academic papers, so it's no "easy read", a list of some publications, you can find here

and an example, to get you started would be this one.

to summarize from the paper, a possible shortish answer for your question, or at least some food for thought:

.........
The energy dissipation will result in local heating of the rubber.
Since the viscoelastic properties of rubber-like materials are extremely strongly temperature dependent, it is necessary to include the local temperature increase in the analysis.
In this paper I have developed a theory which describes the influence of the flash temperature on rubber friction.
At very low sliding velocity the temperature increase is negligible because of heat diffusion, but already for velocities of order 10−2 m s−1 the local heating may be very important, and I have shown that in a typical case the temperature increase
results in a decrease in rubber friction with increasing sliding velocity for v > 0.01 m s−1.
This may result in stick–slip instabilities, and is of crucial importance in many practical applications, e.g. for tyre–road friction, and in particular for ABS braking systems.
.......

Have fun & good luck