Driver Resulting in Solver Failure

draucis

Hello,

Earlier I posted a question about the solver failing due to custom inertia values. Now that thing is sorted out, but the new thing is I am modeling with a detailed rider now, the solver is failing and sometimes not solving at all without error, so when I deactivate the driver, the solver works fine and when I activate it gives an error. I tried to simplify my model by adding a simple rider instead of the detailed one, but the error is still there. the error I get is the menitoned below. I have not edited the driver and have taken the default values for mass and inertia.

WARNING: Akima-interpolation -> Data out of bounds Spline id:0
         Future messages for Spline will be suppressed.
WARNING: Akima-interpolation -> Data out of bounds Spline id:0
         Future messages for Spline will be suppressed.
 Time=1.202E+00; Order=2; H=5.000E-04 [Max Phi=3.336E-07]
WARNING: [DASPK] The nonlinear system solver in the time integration could not converge
 .....
WARNING: EvaluateResidual failed (NaN encountered)! Future warnings will be suppressed.
         Residual idx=2214, GS/310001 X1
ERROR: [DASPK] IRES = -2 was encountered and control is being returned to the calling program
 .....

 DAE SOLVER STATISTICS (DASPK-Index3)
 ------------------------------------------------
 Number of solution steps                 = 1341
 Number of residue evaluations            = 445
 Number of jacobian computations          = 516
 Number of error test failures            = 4
 Number of nonlinear convergence failures = 31

WARNING: Repeated abnormal return from integrator!

ERROR: At time=1.254E+00 the integrator failed to proceed.

after this there is list of probable causes.

Regards 

Draucis

Jagadeesh R

hi Draucis,

The modified model is attached here.

The issue was not with inertia values. The powertrain solver variables and rear monoshock suspension properties needed changes to run events

1] Edited FMU vehicle speed and throttle inputs. Vehicle velocity needed in FMU control should be in m/s. throttle value is in %.

2] Propulsion torque which is coming form FMU need to be changed to model units so multiplied by 1000 in solver variable within Electric powertrain.

Note: You can refer help document and library/example models to verify the unit system followed in FMU control and within motionview model.

3] Observed rear monoshock suspension motion ratio, stiffness and preload value issues. I have moved back and up just to solve the model with edited powertrain parameters. Please consider correct stiffness, preload values for the monospring spring based on required motion ratio.

 

Jagadeesh R

hi Draucis,

In my opinion the issue may be still with mass, cg and inertia values. When you deactivate the driver and run , the ground to frame planar joint is still active which helps to run default transient  without any issue. Please refer MV example or library two wheeler model to compare the inertia values with your model.

In case the mass , CG, inertia entered values are good , then you can also look into (important) the "Lean PID" values available in the event editor "Driver" dialog page. The default  Lean PID values in the event editor are need to be tuned as the mass and inertia of each vehicles will be different to control the direction and lean angleS. The lean PID related help can be seen at  " ./hwdesktop/hwx/topics/motionview/tuning_controller_gains_to_improve_2wheeler_response_r.htm".

After all this, if you still see the problem,. you can share the model as Praful mentioned in the previous "custom inertia values" interaction so that we can study the model and help you.

thanks,

jagadeesh

draucis

Hello Jagadeesh,

Thanks for your reply. Yes, I will try both of these methods and try to troubleshoot my model. I'll update this post in a few days with my progress, if not successful I will share my model with you all.

Thanks 

Draucis

draucis

Hello Jagadeesh,

I have reviewed my model again with the points you mentioned, in the inertia values there are some components that contains inertia values larger than other components. As I don't have value for those components, I left it as default and I have done the PID tuning but that is not changing anything. I am attaching my model file here for your review.

Thanks a ton

Draucis 

draucis

Can you have a look.

Jagadeesh R

hi Draucis,

Yes I can open the model that you shared. I will take a look and let you know.

 

thanks,

jagadeesh

draucis

Jagadeesh R said:

hi Draucis,

Yes I can open the model that you shared. I will take a look and let you know.

 

thanks,

jagadeesh

Hello Jagadeesh,

Thank you so much.

Draucis 

Jagadeesh R

hi Draucis,

The modified model is attached here.

The issue was not with inertia values. The powertrain solver variables and rear monoshock suspension properties needed changes to run events

1] Edited FMU vehicle speed and throttle inputs. Vehicle velocity needed in FMU control should be in m/s. throttle value is in %.

2] Propulsion torque which is coming form FMU need to be changed to model units so multiplied by 1000 in solver variable within Electric powertrain.

Note: You can refer help document and library/example models to verify the unit system followed in FMU control and within motionview model.

3] Observed rear monoshock suspension motion ratio, stiffness and preload value issues. I have moved back and up just to solve the model with edited powertrain parameters. Please consider correct stiffness, preload values for the monospring spring based on required motion ratio.

 

draucis

Hi Jagadeesh,

Thanks for going through my model, I was stuck with this issue for a while now but thanks a lot for helping me out. I will definitely keep those points you mentioned in my mind. Earlier I had not correlated the powertrain with my suspension parameters, I left it at its default values. 

Thanks

Draucis

draucis

Hello Jagadeesh,

When I am opening the model, I am struck with a message about the FMU, but the solver is now running fine, it's just that in the FMU I am getting this message, is it of some concern? Do i have to point it towards a FMU file or something ?

Regards 

Draucis

Jagadeesh R

hi Draucis,

Browse the the  FMU file from the installation:  ( \hwdesktop\hw\mdl\mdllib\Common\FMU_Library\Motor\FMU_source\FMUs\win64).

note: The fmu example file path can be seen when you assemble the library scooter or motorcycle. I suggest you to use 2022.2v so that FMU parameters can be edited (for custom values) easily within in motionview.

 

thanks,

jagadeesh

 

draucis

Hi Jagadeesh, 

Oh alright, got it. Sorry for bothering you so much. 

Thanks 

Draucis 

draucis

Hello Jagadeesh,

I tried your points for correcting the preload with reference to the motion ratio, spring stiffness. So, I took one of the examples in motionview as reference and my values for monoshock are based mostly on that. Still after tweaking the parameters you the model failed to solve, the model which you provided earlier, even that is failing to solve. I am out of ideas, is there anything else I might be missing out on. 

 

This is the error the solver is giving:

X value out of range in AKISPL (spline id=-1)
Current x value 4555.71 > xend 4000, extrapolated y value : -6.49789
ERROR: Failure evaluating expression : "rtod*asin(VARVAL(30401300))"!
       ASIN argument must be between -1 and +1!

And 30401300 is the id for a solver variable which is for Pitch. 

 

Regards

Draucis

 

 

 

Jagadeesh R

draucis said:

Hello Jagadeesh,

I tried your points for correcting the preload with reference to the motion ratio, spring stiffness. So, I took one of the examples in motionview as reference and my values for monoshock are based mostly on that. Still after tweaking the parameters you the model failed to solve, the model which you provided earlier, even that is failing to solve. I am out of ideas, is there anything else I might be missing out on. 

 

This is the error the solver is giving:

X value out of range in AKISPL (spline id=-1)
Current x value 4555.71 > xend 4000, extrapolated y value : -6.49789
ERROR: Failure evaluating expression : "rtod*asin(VARVAL(30401300))"!
       ASIN argument must be between -1 and +1!

And 30401300 is the id for a solver variable which is for Pitch. 

 

Regards

Draucis

 

 

 

hi Drucis,

Please share the latest model which is failing after update.

 

thanks,

jagadeesh

Jagadeesh R

Hi draucis,

Please share the latest model which is failing after update.

 

thanks,

jagadeesh

draucis

Sure,

Session 2 is the model where I have updated it, 

Scooter_v1 is the model which you shared earlier on.

Thanks

Draucis

Jagadeesh R

Hi draucis,

The model which I shared to you after correction belongs to 2022.1v which will not run in 2022v (version you are using) because of fmu inputs units difference. If you refer the help , fmu inputs and outputs are clearly mentioned in the table with units used. 2022.1v/2022.2v is recommended to use for electric powertrain as the powertrain inputs and outputs are standardised along with limited fmu  parameters ( users to edit for new motor specs).

I have attached the mdl files, one belongs to 2022v and second model belongs to 2022.1v. between these versions fmu file from installation, fmu inputs and propulsion torque units are the only difference, please use these models and let me know if you face any new problems. 

It is good to move forward with 2022.1v attached model running in 2022.1v/ 2022.2v MV . Also please refer 2022.1v/2022.2v  library scooter model  and help document for fmu file and units.

thanks,

jagadeesh

 

 

draucis

Hello Jagadeesh,

As of now everything looks smooth appears to be sorted and is running fine. Yes, I will refer to those documents you mentioned. Thanks again! for your help, I appreciate it a lot. 

Regards 

Draucis

Jagadeesh R

draucis said:

Hello Jagadeesh,

As of now everything looks smooth appears to be sorted and is running fine. Yes, I will refer to those documents you mentioned. Thanks again! for your help, I appreciate it a lot. 

Regards 

Draucis

hi draucis,

good to know, it's working for you.

 

regards,

jagadeesh