rec.autos.simulators

I've got a bit of a problem that I hope you guys can help with.

I've got a wheel and an engine and all the drivetrain.  In the wheel,
I calculate the following :

1) Calculate the slip ratio
  This is calculated from the following

    (wheelAngVel * wheelRadius - vehicleVelocity) / (wheelAngVel *
wheelRadius)

2) Calculate the traction force applied by the wheel on the road
  This is calculated by the following

     Traction = (axleTorque / wheelRadius) * slipRatio

3) Calculate the angular acceleration on the wheel

     angAcceleration = axleTorque / wheelInertia

4) Calculate the angular velocity of the wheel

     angVelocity += (angAcceleration * deltaTime)

The Traction is returned and is used to calculate the force applied on
the vehicle to move it forward.

First question - Is this right ?  Is this how you calculate the
traction force and the rotational acceleration of the wheel ?

Now, I take the force from each of the rear wheels and apply it to the
vehicle using f=ma and calculate the acceleration and thus the speed
of the vehicle.  The problem I have is that in first gear (ratio of
3.5, rear diff of 3.6, efficiency of 0.7, engine max torque = 190Nm),
I get up to a speed of approx. 70mph....  I'm not sure if I'm
calculating the traction force wrong, the resistive force, or
something else.  So, I just want to check this bit first.

Cheers,

Nick.

Actually, this is only right for small slip ratios. The slipRatio ->
tractionForce curve is a complicated curve (approximated by Pacejka
for example). There's a nice one in RCVD, and you could look at a
curve from my sim (sliptrac.crv) in Windows (it isn't even used
though; it uses Pacejka now).
The curve goes up almost linearly upto an optimum slipRatio (say 0.15
for example), then goes down again and settles a bit (at about 60% of
the maximum generally).
Same for braking, but the curve looks different (although, OMG! I use
the same curves for braking); it hasn't got the sharp peak that
traction curves have.

This really depends on how RPM you let the engine generate. You can
top it off hard (like a revlimiter) which would simulate the max
sparking rate, or (what I do) let the torque generation curve of the
engine go down to 0 at near maxRPM so the transition is more gradual.

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

Erm... you've missed two of the parameters we need to look at speed:
engine redline and drive wheel rolling radius.  70 is perfectly
feasible if you've got tractor tyres, or a bike engine...  Rolling
radius is also required before you can calculate acceleration.

Jonny

Cheers,

Nick.

Mike.

But that might take awhile to implement ;)

- Matt

the way to

Sounds fine to me too - the only major "improvement" would be to model
transmission wind-up, which is generally unnecessary unless you're
investigating "shuffle" or "rattle" behaviour, although it can be
useful to cushion gearchanges.

The trick then is to model large lumps of the powertrain (wheels,
gearbox, flywheel+crankshaft) as lumps of angular (rotary) inertia and
then connect them with springs.  IOW you apply engine torque to
accelerate the flywheel, but then subtract a back-torque obtained by
integrating the difference in speed between engine and gearbox (i.e.
how much the clutch has wound up); apply this back torque as the
forwards input to the gearbox, and repeat the process downstream with
propshaft, diff, driveshafts and wheels (if you really want!)

Jonny

  Probably overkill for a video game though.  You'd need a really high sampling
rate to keep everything stable, and not many people would notice the
difference.  I see a little play like this in my thing, but it's due to changes
in tire slip ratio with throttle/braking input and would undoubtedly mask the
effect of modeling drivetrain component flex/play, IMO.  
Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

Yeah, agreed.  I did include a single wind-up stage in a vehicle model
I developed during my final semester, partly 'cos the previous author
hadn't actually connected the engine to the wheels (it transferred
torque without any speed linkage!) and it was easier than completely
re-coding.

It was primarily a handling model in any case, and I ended up using a
sort of 'fluid flywheel' more often than the wind-up since this
allowed the car to change gears pretty smoothly without bothering with
a clutch/driver model, for things like constant-steer-angle skid pad
tests.

Jonny

Yes, you would actually have engine braking. If you draw the curve
engine torque = f(engine rotational speed, throttle), you would have a
point where it goes below zero.

--
Johann Deneux

...

Yes you have like Johann pointed out. It's just you have an
RPM->torque connection, but that doesn't mean that RPM can be
anything.
For starters, you can just take the average speed of all driven
wheels, multiply that by the gear ratios and there's your engine
speed. It's really not that bad. It's the clutch that makes things
different. But even there, what I now still have is 2 modes; clutch
engaged or disengaged. When engaged, you just lock all things together
like above. If disengaged, I let the engine run by itself using only
it's own inertia.

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

Nick.

Oh, one more thing.  Ruud - I tried downloading your racer, but the
files on High Gear don't have the sources.

Yep, some downloads are offline, and I thought the sources would hurt
less. Will be up again starting the beginning of next month.

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

  Good way to go about it and it accomplished what you needed.  Suppose it
worked like a loose, high stall drag racing torque converter :-)
Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com