Traction Control Acceleration Performance

[Gytakemoto]

Hi,

I’ve been trying to implement a traction control system in Simulink by using the CM4SL resource in my PTControl, but I have some questions regarding acceleration performance.

For some context, I am currently using a parametrized car from IPG for Formula Student teams, named FSEAllWheelDrive with MF_205_60R15_V91 tire. I tried using a parametrized vehicle from my team but it was just giving me problems. Anyways, I modified the Powertrain section, placing 2 motors and an Universal Driveline connecting the Motor + Gearbox directly into each rear Wheel. Besides that, I’ve made a custom hypothetical RPM x Torque curve in order to provide sufficient Torque to make the wheels spin.

Then I proceeded to configure my IPGDriver. My goal was to simulate the effects of Traction Control on a straight line, with an accelerating Maneuver. So I configured my driver to request full throttle unitl 150m.

But after simulating this scenario with and without my Traction Control, I found that using it was slowing accelerating time, which is odd for me. Then I repeated simulations by using Carmaker provided Traction Control in Driver’s configuration. The results were the same: slower accelerating times with Traction Control.

The results comparing them are below:

Analyzing the results, with TCS on, the slip ratio, and therefore, wheelspin were much lower, which proves that Traction Control is indeed working as expected.

But, and my question resides here, when TCS is off, the slip ratio is absurdly high, which, in my understanding, means that the wheels are spinning and, therefore, it should have had a negative impact on acceleration performance, which didn’t happen, as the vehicle in fact ACCELERATED faster with TCS off. I’ve been trying to come up with a solution, but without success.

Am I doing something wrong? I wanted to develop a TCS system which makes my vehicle go faster during acceleration, by minimizing slip ratios to an optimal range.

Any guidelines?

[tim.sherrington]

Hi Guilherme,

Thank you for your question with the detailed information behind it.

The reason for the behaviour you are seeing is that you are using the tyre model outside of its validated parameters, therefore the traction control off results are not realistic.

The .tir file for that tyre model states the longitudinal slip range for which the model is valid (I copied that below).

----------------------------------------------------------------long_slip_range [LONG_SLIP_RANGE] KPUMIN = -0.30 Minimum longitudinal slip
KPUMAX = 0.30 $ Maximum longitudinal slip

Do you have access to any tyre models valid at high slip ratios?

Thanks,

Tim

[Gytakemoto]

Oh, I see, it really makes sense.

Indeed, upon examining these tires with ModelCheck, I’ve noticed the Fx values were saturated when reaching these KPUMAX. I modified the .tir file in order to set really high slip ratios (-1,1) and also manipulated scalling factors to produce a more abrupt downfall, as the image shows:

It worked as intended, and when the tire reached high slip ratios the Fx suddently dropped, which meant the spin was, in fact, slowing vehicle down, contrary than before.

This approach really helped me to examine my TCS better, but I recognise it is far from ideal, because I simply developed an imaginarium tire. To your question, no, I don’t have access to any model nor data for such high slip ratios. I tried to model my team’s tire with Tire Test Consortium data but in this case it didn’t work, as TTC just had data within [-0.18;0.18] range.

Would you have such models to share? I guess it would be a much better approach than the one I’ve come up with.

Thanks in advance,

Guilherme.

[tim.sherrington]

Hi Guilherme,

That approach looks much better and the Fx vs long slip curves look a lot more realistic.

I am not able to share any high slip models with you, but there is lots of literature available showing Fx vs long slip curves, so perhaps tune a model to look like a plot you can find.

Thanks,

Tim

[Gytakemoto]

Oh, I see.

Anyways, thank you very much Tim, I’ll take this approach. I’ve been stuck on this for a long time .

Guilherme.

[Farzad]

Hi everyone,

Does anyone know how CarMaker calculates slip ratio when Vx is close to zero?
I tried to calculate it as s=(-Vsx)/max(1,Vx) but the value is different from the one that Carmaker is showing me?

Thanks