OptimumTire MATLAB Troubles

[Adam Farabaugh]

Has anyone had luck using the OptimumTire COM addin with MATLAB to evaluate OptimumTire models in MATLAB?
I am getting different results (clearly incorrect) in matlab with the same units and inputs as in OptimumTire. See this screenshot, I would expect OptimumT graphs to be identical to the MATLAB ones. The add-in is giving results that are not even close, for two of the models they are completely backwards. For both OT and MATLAB I am using Fz=-1kN, IA=0, P=.55bar. This is lateral data/models only for now.
http://imgur.com/BozHUD8.jpg

Anyone had a similar problem? So far no luck getting support from OptimumG from this, frankly I'm quite disappointed with the cavalier way they have responded to my inquiries for support (suggesting I use radians instead of degrees when documentation clearly says degrees are the input unit, saying that no one is doing any development on OptimumT anymore, etc).

Thanks for any help anyone can provide. I'll pay you back in beers at Lincoln 2016.

[BillCobb]

Go back to the raw data at your input conditions for the tire you have been working with. If you'd care to post the ID of this tire, maybe some of us would take a stab at helping you with this problem.

1) If you have the Pacejka coefficients, then independently produce results from a Matlab Pacejka function that you write yourself.

2) Focus on just one input data group (1 deg slip, |load|, inclination and pressure, and one FY and one MZ). Then try alternative signs, units and input parameter order and see if you can make sense out of the Matlab results. to discover whether its a translation issue or a computational issue.

3) Fit your own Pacejka function to the raw data and compare it to the OptimumT results.

4) Properties of COM objects can be displayed to the Matlab command window using the GET function, and are displayed graphically using the property inspector. Take a look at the Saved elements.

5) Properties of a COM object can be changed using the SET function.

[BillCobb]

OT indicates 8 psi as the test pressure. Is the registered .COM input unit psi or bar?

[Adam Farabaugh]

Thanks for the suggestions Bill.

1) I'm going to do this after class this afternoon. Currently OptimumT has fit a pac2002 model w/pressure effects. Green line is EDIT: REMOVED , I doubt giving away that much is detrimental to the TTC. Didn't realize the equations themselves were so easy to come by.

2) Great idea, I had been stuck permuting units and the like for the full slip sweeps, should have thought to isolate even more.

3) This leads to not needing OptimumT at all, which likely won't help me for this design cycle but would be great for the future

4) and 5) unfortunately no properties are visible. I assume this is because the addin is meant to be as closed and inaccessible as possible :D

[Adam Farabaugh]

OT indicates 8 psi as the test pressure. Is the registered .COM input unit psi or bar?

Per the documentation is bar, but it's relatively insensitive to pressure (I tried 8.0 instead of .55 with only small changes)

[DougMilliken]

... Green line is R25B 20.5x7.0-13 Hoosier, I doubt giving away that much is detrimental to the TTC. ...

Per the TTC data license, please don't publish data publicly with the tire identified. You can have an open discussion on the TTC forum including identified plots.

[Adam Farabaugh]

With the coefficients listed in OptimumT I am unable to reproduce the model by evaluating the coefficients myself, probably because I am not normalizing the inputs correctly or I don't have the same scaling factors used by OptimumT.

[BillCobb]

Since a .COM element is a Windows subroutine, you should also be able to register it with Excel and call for results back to a cell there. Can you get any better results using that method ?

[Adam Farabaugh]

http://sae.wsu.edu/ttc/viewtopic.php?f=5&t=155

[Adam Farabaugh]

Turns out that pressure inputs must be in Pascals instead of bar. OptimumG support came through in the end :)

[BillCobb]

The industry generally uses psi or kPa units.