what is the difference between Pac02 tire and Pacjka tire model in chrono?

[stefan]

Hello,

The pacjka tire models expressed in the manual  are Pac89 and Pac02,  But I found pacjak tire model in HMMWV vehicle, and the simulation results derived from Pac02  and Pacjka model are obviously different, I don't know which model is better for vehicle simulation and which derived simulation result is closer to the real one. Otherwise, could you tell me the difference between Pacjka model and Pac02 model from the implementation principle.

[dr.ratz...@gmail.com]

Hi Stefan,

the Pacejka tire model in chrono is just outdated. While it is nearly feature complete, it lacks stability. Another problem is the data input routines. It pretends to read tiem orbit files (Adams/Car), but expects the data entries to be in a certain order. It is very likely that nonsense input is read from a valid Adams/Car tire file, as long as you don't reorder the data lines, I guess nobody is eager to do that.

If you want to use a Pacejka tire model in chrono, use Pac02. Pac02 hasn't all features possible, but it works and doesn't crash chrono vehicle simulations. For a benchmark project we needed a Pacejka89 (Adams/Car) model. This model has a very clear implementation structure, this structure was the also taken for Pac02. The slip definitons differ of course, as do the formulae.

Pac02 and Pac89 are comparable in simulation accuracy. Pac02 can be easily rescaled to different tire load levels due to its skaling factors, so it is more flexible to use.

[stefan]

Hello Ratzel,  

 Thanks for your replying, but I still have some questions about that. Recently I tested all the tire models in HMMWV vehicle with the conditoin that a constant motor torque was applied to the vehicle. It is true that the pacejka tire model is 

less robust than others, but it can give reasonable results when the simulation step size 1e-3 was given.  The pictures I attached can explain what I said(The first picture is for PAC02). The acceleration shouldn't increase with time under the constant motor torque. And I also found that acceleration was related to the vehicle speed, and acceleration was not zero when the vehicle had high speed and engine power was cut off. That is obviously wrong according to the vehicle dynamics.

I want to know why it happened,  and it can be described as a bug to some extent? or I miss something? 

And I also want to solve the problem with your guys and I am checking the codes of tire parts in chrono. I would be very appreciated if you have more advice for me.

Thanks again!

Stefan

[dr.ratz...@gmail.com]

Hi Stefan,

your observation is correct. Actually we know this happens only with handling tires. It is better to make a test with a vehicle with a real powertrain. Applying a const torque means that the accelerating power increases with the speed/angular speed which is not realistic.

We had some discussions about this issue, most likely it is related to limits in the integrator. There is some work in progress about it, but it will take a while to get a working solution.

[stefan]

Hi Ratzel,

Thanks for your answer again. I am doing a pure electrical vehicle simulation, the torque applied to the vehicle is constant when the electric motor start.  The process of constant torque will maintain for a time. 

Can it be concluded that the pacejka model has a different integrator from other tire models if it is due to the integrator?  Since reasonable results can be got from pacejka tire model.

[dr.ratz...@gmail.com]

Hi Stefan,

Pacejka uses its own integrator for calculation of tire relaxation, a possible source of instabilities. Rolling straight forward without crashing is not good enough for practical use. So don't waste your time with it.

Btw. even electric motors cannot deliver constant torque with increasing angular speed, since Power = Torque*omega. It is only possible, as long the max. power of the engine is not reached.

[stefan]

Hello, Ratzel

I have modified the Tmeasy tire code and used the RK integration method to calculate tire relaxation which is the same method as in pacejka, but it still can not give the right results.  

So it can be concluded that it is not the reason of integrator.

I am very confused now, could you give me more clues about how to solve the problem?