A regression

[massimo ceraolo]

The attached "EV_DAEissue.mo" contains a model named EVbasic.
I could run this model (and a lot of more complex models based on this one) with OM.
This model runs well under OM 1.9.3 (release)
On the latest nightly build it does not run, and emits the following error:

[3] 09:53:16 Translation Error
[CodegenCFunctions.tpl: 5009:16-5009:16]: Template error: daeExpRelation {{0.0, 0.0}} <> {{0.0, 0.0}}

[Lennart Ochel]

I cannot reproduce the issue. For me the simulation just works fine. This is the script that I used for testing:

loadModel(Modelica); getErrorString();
loadFile("EV_DAEissue.mo"); getErrorString();
simulate(EV_DAEissue.EVbasic); getErrorString();

[Adrian Pop]

Same here, I can simulate the model fine with OpenModelica v1.9.4-dev.beta2-10-g4d534f7.
The error seems to be some comparison of tuples but there are none in the flattened code.

[massimo ceraolo]

Maibe omeditcommunication.log can help. I enclose it.
I used the exe downloaded today: OpenModelica-v1.9.4-dev.beta2-15-g2dbe6e3.exe

[Adrian Pop]

This is strange, I just installed your version (the latest nightly build) loaded your model and simulated it and it still works fine for me:

stdout | OMEditInfo | <p>C:/Users/adrpo/AppData/Local/Temp/OpenModelica/OMEdit/EV_DAEissue.EVbasic.exe -port=39743 -logFormat=xml -override=startTime=0,stopTime=80,stepSize=0.1,tolerance=0.0001,solver=dassl,outputFormat=mat,variableFilter=.* -r=EV_DAEissue.EVbasic_res.mat -dasslJacobian=coloredNumerical -w -lv=LOG_STATS</p>
LOG_STATS | info | <p>### STATISTICS ###</p>
stdout | OMEditInfo | <p>Simulation process finished successfully.</p>

[Adrian Pop]

Did you restart OMEdit and tried again?

[massimo ceraolo]

Here are my steps to reproduce the issue:
1) load OMEdit
2) load (from OMEdit) "EV_DAEissue.mo"
3) select EVBasic
4) click on the "Simulate" button

A screenshot of the result is on the file "ScreeenShot.png"

[Adrian Pop]

Replying to ceraolo:

Here are my steps to reproduce the issue:
1) load OMEdit
2) load (from OMEdit) "EV_DAEissue.mo"
3) select EVBasic
4) click on the "Simulate" button

A screenshot of the result is on the file "ScreeenShot.png"

I do basically the same and the model builds fine. Do you have some special settings in OMEdit?
What Modelica library do you use?

[Adrian Pop]

Can you please do right click Instantiate Model on the model and attach the result here?

[Adrian Pop]

OK, if found the issue. You have the special flag +gDynOpt which causes the problem.

[Adrian Pop]

With this flag an equation of this form is generated:

3/3 (1): dragF.f.$pDERA.dummyVarA + -2.3907 * (mass.v ^ 2.0 * dragF.Sign.$pDERA.dummyVarA + 2.0 * mass.v * mass_vSeedA * dragF.Sign) = if dragF.locked then dragF.sa.$pDERA.dummyVarA else 2039.7832 * (if dragF.startForward then (if mass_vSeedA <> 0.0 then mass_vSeedA * $DER$$PModelica$PMath$PtempInterpol1(mass.v, {{0.0, 1.0}}, 2, 1.0, {{0.0, 0.0}}) else 0.0) + (if {{0.0, 0.0}} <> {{0.0, 0.0}} then {{0.0, 0.0}} else {{0.0, 0.0}}) else if dragF.startBackward then (-(if (-mass_vSeedA) <> 0.0 then (-mass_vSeedA) * $DER$$PModelica$PMath$PtempInterpol1(-mass.v, {{0.0, 1.0}}, 2, 1.0, {{0.0, 0.0}}) else 0.0)) - (if {{0.0, 0.0}} <> {{0.0, 0.0}} then {{0.0, 0.0}} else {{0.0, 0.0}}) else if pre(dragF.mode) == 1 then (if mass_vSeedA <> 0.0 then mass_vSeedA * $DER$$PModelica$PMath$PtempInterpol1(mass.v, {{0.0, 1.0}}, 2, 1.0, {{0.0, 0.0}}) else 0.0) + (if {{0.0, 0.0}} <> {{0.0, 0.0}} then {{0.0, 0.0}} else {{0.0, 0.0}}) else (-(if (-mass_vSeedA) <> 0.0 then (-mass_vSeedA) * $DER$$PModelica$PMath$PtempInterpol1(-mass.v, {{0.0, 1.0}}, 2, 1.0, {{0.0, 0.0}}) else 0.0)) - (if {{0.0, 0.0}} <> {{0.0, 0.0}} then {{0.0, 0.0}} else {{0.0, 0.0}}))   [dynamic]

Which contains the problematic relation:

if {{0.0, 0.0}} <> {{0.0, 0.0}} then {{0.0, 0.0}} else {{0.0, 0.0}}.

[massimo ceraolo]

Ah, thank you.
It was very difficult to find this (I did not at all).

Is there a way to warn the user somehow to avoid this?
Or maybe it is possible to avoid this conflict? I don't have this issue with 1.9.3 release even while having +gDynOpt active (on a Mac).
I believe (but I'm not absolutely sure) that I didn' have this issues on a Win PC with +gDynOpt.

[Adrian Pop]

This for sure is a bug but I don't know to which person should I assign it, maybe Vitalij?
Lennart, what do you think?

[Willi Braun]

I will check that, since it seems to be an expression of the symbolic jacobian.

[Adrian Pop]

Replying to wbraun:

I will check that, since it seems to be an expression of the symbolic jacobian.

The bad expression seems to come from Differentiate.createPartialSum while differentiating Modelica.Math.tempInterpol1. It seems not to find Modelica.Math.tempInterpol1_der?

[Adrian Pop]

Replying to adrpo:

Replying to wbraun:

I will check that, since it seems to be an expression of the symbolic jacobian.

The bad expression seems to come from Differentiate.createPartialSum while differentiating Modelica.Math.tempInterpol1. It seems not to find Modelica.Math.tempInterpol1_der?

Ah, yes, is the jacobian calculation, not wrt time.

[Willi Braun]

fixed in d6e73f0/OMCompiler.