rec.autos.simulators

by **J. Todd Wass** » Sun, 08 Sep 2002 22:02:12

Is tire load sensitivity, or the variation in effective friction coefficient
with load, generally approximately the same in both the longitudinal and
lateral directions? My thinking is that if you look at the strain pattern in
the contact patch of a tire in the longitudinal direction, there is a
predictable curve from slip ratio, and there is another similar distortion that
would exist even with 0 slip ratio that would take away from the lateral force
that's available. (Even at 0 slip ratio/angle, the front of the tread is being
bent forwards at the road, and rearwards at the rear of the tread, resulting in
essentially 0 longitudinal force if rolling resistance was neglected, but would
cause a loss of available lateral force dependent mostly on load, I'm
thinking.)

However, thinking this way and rotating this system 90 degrees, the
longitudinal force would not decrease in the same manner. There is a somewhat
equivalent loss laterally I think because the center of the tread is being
compressed, which would cause a lateral strain pattern resulting in a loss of
available longitudinal force (perhaps the left/right edges might tend to slide
outwards a bit during free rolling, similar to how the front/mostly rear might
work), but my thinking is that this would not be as severe in this direction
because the strain here would not be as great....maybe...

So.... This would indicate that as load is increased, a tire would tend to
lose more lateral force than longitudinal force capacity, causing the friction
circle to become not only larger, but more elongated along the longitudinal
direction. I haven't read anywhere even a hint that this is indeed what
happens though, so I'm questioning my thinking here. Can anybody shed some
light on this?

Of course, I'm assuming constant temperature, pressure, camber, etc..

Thanks!

Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://www.racesimcentral.net/

My little car sim screenshots:
http://www.racesimcentral.net/

by **Jim Seamu** » Mon, 09 Sep 2002 14:01:27

I've got some pacejka coefficients for a road tyre (you've got them too
Todd, I forwarded them to you last month!). At zero load this tyre has an
effective friction coefficient of 1.201 (lateral) and 1.571 (longitudinal)
but then these coefficients are reduced by 0.038 per kN load (lateral) or
0.080 per kN load (longitudinal). So if your tyre had a 2 kN load you'd have
an effective friction coefficient of 1.125 (lat) and 1.411 (long).

So the tyres appear to be twice as sensitive to load in the longitudinal
direction.

by **Tim Epstei** » Mon, 09 Sep 2002 22:03:52

Hi, I'm no physics guru, but am curious if you've ever come across research
on modelling of tire dynamics on gravel surfaces. I'm really keen to
experience a rally simulator that emulates dirt/gravel surfaces. Whilst LFS
is brilliant on tarmac, and good on gravel, the experience of sliding the
car on dirt can't be described as realistic.

On tarmac surfaces, modelling is much simpler, as the surface is effectively
2D, with only camber angles to worry about (and of course wet surfaces).
However, gravel is a 3D surface, with a tire digging in. As such, the
physics are quite different. For instance, a car on a dirt surface will stop
fastest when the front wheels are completely locked up, as they build a
"wall" of gravel up in front of them. Conversely, a car on tarmac will stop
fastest when the wheels are still rotating, but at slower speed than which
the car is travelling on the road.

One of the biggest differences between tire dynamics on tarmac and unsealed
surfaces is when the car has attitude (oversteer or 4 wheel drift). Rally
cars have the best grip when going sideways, under either acceleration or
braking. This allows for the controlled power slides that are the art of
rally driving, and which I've never seen realistically moddeled on a Sim.

I'd be very interested into any research into this..

Tim

by **Dave Pollatse** » Tue, 10 Sep 2002 01:19:53

Try searching the internet for Bekker's soil model. People seem to be using
this model for things like ATVs and humvees. The last time I searched, I
also found some tire data for humvee's on dirt--from the graphs it looked
like load sensitivity was "backwards" for lateral grip (e.g., higher
coefficient with more load). Which actually seems plausible, because you
can imagine the more vertical sides of the tire digging in better than the
sloping front face of the tire, which would probably cause the tire to float
up.
-Dave

> Hi, I'm no physics guru, but am curious if you've ever come across
research
> on modelling of tire dynamics on gravel surfaces. I'm really keen to
> experience a rally simulator that emulates dirt/gravel surfaces. Whilst
LFS
> is brilliant on tarmac, and good on gravel, the experience of sliding the
> car on dirt can't be described as realistic.

> On tarmac surfaces, modelling is much simpler, as the surface is
effectively
> 2D, with only camber angles to worry about (and of course wet surfaces).
> However, gravel is a 3D surface, with a tire digging in. As such, the
> physics are quite different. For instance, a car on a dirt surface will
stop
> fastest when the front wheels are completely locked up, as they build a
> "wall" of gravel up in front of them. Conversely, a car on tarmac will
stop
> fastest when the wheels are still rotating, but at slower speed than which
> the car is travelling on the road.

> One of the biggest differences between tire dynamics on tarmac and
unsealed
> surfaces is when the car has attitude (oversteer or 4 wheel drift). Rally
> cars have the best grip when going sideways, under either acceleration or
> braking. This allows for the controlled power slides that are the art of
> rally driving, and which I've never seen realistically moddeled on a Sim.

> I'd be very interested into any research into this..

> Tim

> > Is tire load sensitivity, or the variation in effective friction
> coefficient
> > with load, generally approximately the same in both the longitudinal and
> > lateral directions? My thinking is that if you look at the strain
pattern
> in
> > the contact patch of a tire in the longitudinal direction, there is a
> > predictable curve from slip ratio, and there is another similar
distortion
> that
> > would exist even with 0 slip ratio that would take away from the lateral
> force
> > that's available. (Even at 0 slip ratio/angle, the front of the tread
is
> being
> > bent forwards at the road, and rearwards at the rear of the tread,
> resulting in
> > essentially 0 longitudinal force if rolling resistance was neglected,
but
> would
> > cause a loss of available lateral force dependent mostly on load, I'm
> > thinking.)

> > However, thinking this way and rotating this system 90 degrees, the
> > longitudinal force would not decrease in the same manner. There is a
> somewhat
> > equivalent loss laterally I think because the center of the tread is
being
> > compressed, which would cause a lateral strain pattern resulting in a
loss
> of
> > available longitudinal force (perhaps the left/right edges might tend to
> slide
> > outwards a bit during free rolling, similar to how the front/mostly rear
> might
> > work), but my thinking is that this would not be as severe in this
> direction
> > because the strain here would not be as great....maybe...

> > So.... This would indicate that as load is increased, a tire would tend
> to
> > lose more lateral force than longitudinal force capacity, causing the
> friction
> > circle to become not only larger, but more elongated along the
> longitudinal
> > direction. I haven't read anywhere even a hint that this is indeed what
> > happens though, so I'm questioning my thinking here. Can anybody shed
> some
> > light on this?

> > Of course, I'm assuming constant temperature, pressure, camber, etc..

> > Thanks!

> > Todd Wasson
> > ---
> > Performance Simulations
> > Drag Racing and Top Speed Prediction
> > Software
> > http://PerformanceSimulations.Com

> > My little car sim screenshots:
> > http://performancesimulations.com/scnshot4.htm

by **Tim Epstei** » Tue, 10 Sep 2002 04:04:16

Thanks very much for the lead.

I just found this:
http://arc.engin.umich.edu/arc/research/T2RealTimeTire.htm

Its only abstract, but means that this area is being seriously looked at.

Cheers,
Tim

> Try searching the internet for Bekker's soil model. People seem to be
using
> this model for things like ATVs and humvees. The last time I searched, I
> also found some tire data for humvee's on dirt--from the graphs it looked
> like load sensitivity was "backwards" for lateral grip (e.g., higher
> coefficient with more load). Which actually seems plausible, because you
> can imagine the more vertical sides of the tire digging in better than the
> sloping front face of the tire, which would probably cause the tire to
float
> up.
> -Dave

> > Hi, I'm no physics guru, but am curious if you've ever come across
> research
> > on modelling of tire dynamics on gravel surfaces. I'm really keen to
> > experience a rally simulator that emulates dirt/gravel surfaces. Whilst
> LFS
> > is brilliant on tarmac, and good on gravel, the experience of sliding
the
> > car on dirt can't be described as realistic.

> > On tarmac surfaces, modelling is much simpler, as the surface is
> effectively
> > 2D, with only camber angles to worry about (and of course wet surfaces).
> > However, gravel is a 3D surface, with a tire digging in. As such, the
> > physics are quite different. For instance, a car on a dirt surface will
> stop
> > fastest when the front wheels are completely locked up, as they build a
> > "wall" of gravel up in front of them. Conversely, a car on tarmac will
> stop
> > fastest when the wheels are still rotating, but at slower speed than
which
> > the car is travelling on the road.

> > One of the biggest differences between tire dynamics on tarmac and
> unsealed
> > surfaces is when the car has attitude (oversteer or 4 wheel drift).
Rally
> > cars have the best grip when going sideways, under either acceleration
or
> > braking. This allows for the controlled power slides that are the art of
> > rally driving, and which I've never seen realistically moddeled on a
Sim.

> > I'd be very interested into any research into this..

> > Tim

> > > Is tire load sensitivity, or the variation in effective friction
> > coefficient
> > > with load, generally approximately the same in both the longitudinal
and
> > > lateral directions? My thinking is that if you look at the strain
> pattern
> > in
> > > the contact patch of a tire in the longitudinal direction, there is a
> > > predictable curve from slip ratio, and there is another similar
> distortion
> > that
> > > would exist even with 0 slip ratio that would take away from the
lateral
> > force
> > > that's available. (Even at 0 slip ratio/angle, the front of the tread
> is
> > being
> > > bent forwards at the road, and rearwards at the rear of the tread,
> > resulting in
> > > essentially 0 longitudinal force if rolling resistance was neglected,
> but
> > would
> > > cause a loss of available lateral force dependent mostly on load, I'm
> > > thinking.)

> > > However, thinking this way and rotating this system 90 degrees, the
> > > longitudinal force would not decrease in the same manner. There is a
> > somewhat
> > > equivalent loss laterally I think because the center of the tread is
> being
> > > compressed, which would cause a lateral strain pattern resulting in a
> loss
> > of
> > > available longitudinal force (perhaps the left/right edges might tend
to
> > slide
> > > outwards a bit during free rolling, similar to how the front/mostly
rear
> > might
> > > work), but my thinking is that this would not be as severe in this
> > direction
> > > because the strain here would not be as great....maybe...

> > > So.... This would indicate that as load is increased, a tire would
tend
> > to
> > > lose more lateral force than longitudinal force capacity, causing the
> > friction
> > > circle to become not only larger, but more elongated along the
> > longitudinal
> > > direction. I haven't read anywhere even a hint that this is indeed
what
> > > happens though, so I'm questioning my thinking here. Can anybody shed
> > some
> > > light on this?

> > > Of course, I'm assuming constant temperature, pressure, camber, etc..

> > > Thanks!

> > > Todd Wasson
> > > ---
> > > Performance Simulations
> > > Drag Racing and Top Speed Prediction
> > > Software
> > > http://PerformanceSimulations.Com

> > > My little car sim screenshots:
> > > http://performancesimulations.com/scnshot4.htm

by **Tim Epstei** » Tue, 10 Sep 2002 04:07:40

Bullseye!
http://www.***net.co.jp/dads/seminar/userconf1999/Tire-Soil_Interac...
f

This includes all the equations, including lateral force slip angle, based
on how much the surface compresses.
Only problem is the CPU performance hit trying to model this in real time -
imagine with 8 x AI cars as well!

> Try searching the internet for Bekker's soil model. People seem to be
using
> this model for things like ATVs and humvees. The last time I searched, I
> also found some tire data for humvee's on dirt--from the graphs it looked
> like load sensitivity was "backwards" for lateral grip (e.g., higher
> coefficient with more load). Which actually seems plausible, because you
> can imagine the more vertical sides of the tire digging in better than the
> sloping front face of the tire, which would probably cause the tire to
float
> up.
> -Dave

> > Hi, I'm no physics guru, but am curious if you've ever come across
> research
> > on modelling of tire dynamics on gravel surfaces. I'm really keen to
> > experience a rally simulator that emulates dirt/gravel surfaces. Whilst
> LFS
> > is brilliant on tarmac, and good on gravel, the experience of sliding
the
> > car on dirt can't be described as realistic.

> > On tarmac surfaces, modelling is much simpler, as the surface is
> effectively
> > 2D, with only camber angles to worry about (and of course wet surfaces).
> > However, gravel is a 3D surface, with a tire digging in. As such, the
> > physics are quite different. For instance, a car on a dirt surface will
> stop
> > fastest when the front wheels are completely locked up, as they build a
> > "wall" of gravel up in front of them. Conversely, a car on tarmac will
> stop
> > fastest when the wheels are still rotating, but at slower speed than
which
> > the car is travelling on the road.

> > One of the biggest differences between tire dynamics on tarmac and
> unsealed
> > surfaces is when the car has attitude (oversteer or 4 wheel drift).
Rally
> > cars have the best grip when going sideways, under either acceleration
or
> > braking. This allows for the controlled power slides that are the art of
> > rally driving, and which I've never seen realistically moddeled on a
Sim.

> > I'd be very interested into any research into this..

> > Tim

> > > Is tire load sensitivity, or the variation in effective friction
> > coefficient
> > > with load, generally approximately the same in both the longitudinal
and
> > > lateral directions? My thinking is that if you look at the strain
> pattern
> > in
> > > the contact patch of a tire in the longitudinal direction, there is a
> > > predictable curve from slip ratio, and there is another similar
> distortion
> > that
> > > would exist even with 0 slip ratio that would take away from the
lateral
> > force
> > > that's available. (Even at 0 slip ratio/angle, the front of the tread
> is
> > being
> > > bent forwards at the road, and rearwards at the rear of the tread,
> > resulting in
> > > essentially 0 longitudinal force if rolling resistance was neglected,
> but
> > would
> > > cause a loss of available lateral force dependent mostly on load, I'm
> > > thinking.)

> > > However, thinking this way and rotating this system 90 degrees, the
> > > longitudinal force would not decrease in the same manner. There is a
> > somewhat
> > > equivalent loss laterally I think because the center of the tread is
> being
> > > compressed, which would cause a lateral strain pattern resulting in a
> loss
> > of
> > > available longitudinal force (perhaps the left/right edges might tend
to
> > slide
> > > outwards a bit during free rolling, similar to how the front/mostly
rear
> > might
> > > work), but my thinking is that this would not be as severe in this
> > direction
> > > because the strain here would not be as great....maybe...

> > > So.... This would indicate that as load is increased, a tire would
tend
> > to
> > > lose more lateral force than longitudinal force capacity, causing the
> > friction
> > > circle to become not only larger, but more elongated along the
> > longitudinal
> > > direction. I haven't read anywhere even a hint that this is indeed
what
> > > happens though, so I'm questioning my thinking here. Can anybody shed
> > some
> > > light on this?

> > > Of course, I'm assuming constant temperature, pressure, camber, etc..

> > > Thanks!

> > > Todd Wasson
> > > ---
> > > Performance Simulations
> > > Drag Racing and Top Speed Prediction
> > > Software
> > > http://www.racesimcentral.net/

> > > My little car sim screenshots:
> > > http://www.racesimcentral.net/

by **Jim Seamu** » Tue, 10 Sep 2002 08:23:33

Hey great find Tim!

Ruud, are you reading this? :0)

http://www.***net.co.jp/dads/seminar/userconf1999/Tire-Soil_Interac...

by **J. Todd Wass** » Tue, 10 Sep 2002 15:49:37

Thanks, Jim :-)

Once the 2kN load is applied, the longitudinal capacity is 89.9% of what it was
at 0 load, while the lateral is 93.7% of the 0 load value, so is the
longitudinal really twice as sensitive?

But then again, if you went to 4 kN load, you'd have 1.049--87% lateral and
1.091--77% longitudinal.

Will have to look into this a bit further, but it looks like I need to go back
to the old drawing board on a thing or two :-)

Thanks again

Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

My little car sim screenshots:
http://performancesimulations.com/scnshot4.htm

by **J. Todd Wass** » Tue, 10 Sep 2002 16:08:06

I haven't really come across anything that gets into the areas you're looking
for (this is the first time I've heard of Bekker's model, thanks Dave). I've
seen some limited stuff in SAE papers that looks at pure slip ratio on snow and
gravel. Basically, the curves looked the same but occured at a much lower
level, and the force didn't drop off with increasing slip ratio (the curves
stayed flat after the peak).

The main paper that comes to mind didn't get in to the slip angle aspect,
though. I just downloaded the Bekker model, but must be using an old Acrobat
Reader because I can't view the thing! :-)

I think with the type of model I'm developing this would be fairly
straightforward to approximate, as I'm already calculating this "digging in"
area. It should be pretty simple to calculate a theoretical mass of
dirt/whatever moved by this area and take the slip angle into account to get
the necessary forces. I haven't done it yet though, and things don't work the
way I think they will quite often, so I won't say I've "solved" this yet in my
own work :-)

Someone (maybe you?) once said that with LFS's knobbly tires the off road
behavior was much better. What do you think?

Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

My little car sim screenshots:
http://performancesimulations.com/scnshot4.htm

rec.autos.simulators

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited

Tire load sensitivity revisited