Discussion Forum

error message in english

Segregated group X2
Failed to evaluate variable.
- Variable: mat.epsilonr33
- Geometry: 1
- Domain: 8 10 11 14 17 18
Failed to evaluate variable.
- Variable: mod2.es.epsilonrzz
- Defined as: mat.epsilonr33
Failed to evaluate variable.
- Variable: mod2.es.Dz
- Defined as: (mod2.es.Ez*(mod2.es.epsilonrzz*epsilon0_const))
Failed to evaluate expression.
- Expression: d((-mod2.es.Dx*test(mod2.V3x)-mod2.es.Dy*test(mod2.V3y)-mod2.es.Dz*test(mod2.V3z))*(dvol_spatial),{test@2})

Hi

I suspect you have not defined correctly all material variables, or that you have given istropic values, but set the mterial to anisotropic, this canges the material parameters needed, check your material node

--
Good luck
Ivar

thanks ivar
i've checked the model and i find that the materials variables are correct, but i didn't respect the continuity of the charge when i've defined the eletric potential, however thanks to your hints i'm able to run the simulation but i've encountered another problem. The simulation ended between 4% and 8% and return no error but either no results( displacement =0 elettric field =0.
do you have any suggestion?
thanks
andrea

PS:the model is about mems variable capacitor, i used structural mechanics module+ elettrostatic module+ ale module

Ps:i'm apologize for mine english

Hi

You should not expect that my Italian is better than your English ;) Anyhow one has to jump into it, luckily, for me, we are supposed to stick to English such that "everyone can follow"

Well no I cannot think of something obvious. But you should perform a few checks:
1) always recheck the BC's after having rechecked that all material are OK, because when you tweak a model physics, sometimes you change some settings that ask for other (typically anisotropic) material properties and these are not always defined. Then BC's are the most important to set up, just enough not to many nor too few, this is easily done wrong.

Then you should load your initial values and take a look, the method is:

if a new solver case, not yet solved, right click the Study - Solver Configuration node and right click + Show default solver
if the model has already run skip the point above

Select the Study ... Solver - Dependent Variables node and right click "Compute to selected"
This will load your matrices, prepare the results, interpret most variables and do a first consistency check. If no errors appear, take a look at the different "initial values" you have in the result tab

Ask yourself if the values make sense, does the fields appear coherent ? Nothing have been solved so its NOT the solution it's your starting point. perhaps some initial values appear funny, or you see immediately that the solution would look like something easy to get closer to, by some additions to the initial conditions for the full study, or for a 2nd initial condition node for only part of the model ...

Then solve and see if it goes better. Now you have also the Definition - probe plots to show you how some values changes with the solver process, use these, just as you can turn on the "plot while Solving" in the Study - Study step node" (for each saved value") or deeper down i.e. "Study ... ... Fully coupled" node for each solver iteration

By the way the early 4.0 was not really "ironed out", you should really check and get hands on a later version, if possible a 4.2 or newer, you will be less frustrated ;)
--
Good luck
Ivar