Hi Hassen,
You could try and compile the model with "+d=failtrace", which may give
us a better idea of where OMC fails. If you are compiling a mos-file it
might be necessary for you to also add "+showErrorMessages".
It's a pity that the model is confidential, but if it's possible it
would be great if you could try and reduce the model to a smaller model
that still exhibits the same problems but is small enough that you don't
deem it confidential.
Best regards,
Per Östlund
2010-08-23 16:29, HADJ AMOR HASSEN skrev:
Thank you for answering.
Sorry, I can't send you the model, it is a confidential one.
The compiler failed to get the incidence matrix and I obtained an
error message like this one :
incidence_row failed
-DAELow.statesInEqns failed
-DAELow.statesInEqns failed
-DAELow.statesInEqns failed
...
...
...
I verified all the if-equations, when equations and all known omc bugs
via the Bug tracker..
The omc listed all the equations and variables and failed to get the
incidence matrix, this mean probably that an equation or a variable is
missing... But I am sure that there is no problem of singularity.
Any idea or suggestion is more than welcome.
Thank you in advance.
Best Regards,
Hassen
--- En date de : *Mer 18.8.10, Jens.Frenkel
/<[email protected]>/* a écrit :
De: Jens.Frenkel <[email protected]>
Objet: Re: Sorting equations
À: [email protected]
Date: Mercredi 18 août 2010, 19h49
Hi Hassen,
there is may be a bug in the compiler and Pantelides algorithm
would fail because of wrong input data.
The first step to check this is to compile the model with the
flags "+d=dumpdaelow,bltdump". Now you get the flat model and the
incidence matrix. Maybe you see some curious equations. Good
candidates are if equations or the equations the compiler print
after the error message "no states found in equations:".
You could also send the model via email, if it is possible for you.
--
Mit freundlichen Grüßen,
Kind regards,
Jens Frenkel
_________________________________________________________________
Dipl.-Ing.
Jens Frenkel
*****************************************************************
Technische Universität Dresden
Institut für Verarbeitungsmaschinen und Mobile Arbeitsmaschinen
D-01062 Dresden
Tel: 0351 463-39278
Fax: 0351 463-37731
eMail: [email protected]
</mc/[email protected]>
*****************************************************************
*****************************************************************
Dresden University of Technology
Institute of Processing Machines and Mobile Machinery
D-01062 Dresden
Germany
Tel: +49 351 463-39278
Fax: +49 351 463-37731
eMail: [email protected]
</mc/[email protected]>
*****************************************************************
Am 18.08.2010 17:06, schrieb HADJ AMOR HASSEN:
Hi,
When compiling a big model within OpenModelica, I obtained the
error message below :
" Error: Model is structurally singular, error found sorting
equations ............"
The error message listed also a number of equations and their
related variables. The model is not singular since it runs
correctly under Dymola.
Did this mean that the omc (DAELow module precisely) can't
perform BLT sorting or index reduction for this model due to its
complexity?
As indication, the model have 1200 variables, 385 Unknowns and 53
continuous time states. The same model runs correctly on Dymola
and it contains 3 numerical Jacobians.
Thank you in advance.
Regards,
Hassen
