Ivar KJELBERG
COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
18.04.2010, 16:53 GMT-4
Hi
well if you have only got a few hours of Comsoling and want to do such nice examples I can well understand some frustration, COMSOL is very powerfull, but that means also quite some conventions and tricks that require time to get used to, even in only one physics.
First, I would say that whatever you try to do in FEM you should have first a good clue of how it's supposed to behaive by analytical simple analysis. FEM calucations will give you the details, but you MUST always check the results against a global qualitative and quantitative response obtained by other means. It's so easy to get something wrong.
This said, if you want to work in a fixed (or sweeping) frequency mode it's normally the "(Time-)harmonic" mode to use, and the frequency in "emqvw" application mode is by default "nu_emqvw" it is defined in the "Scalar variables", you can edit it there. But often it is overwritten i.e. when you use the "Solver Parameter" time harmonic mode it is the "parametric" solver that comes up and "nu_emqvw" is usually set as the parameter for the sweep, and it will overwrite the value you define in the scalar variables (if you change mode the scalar value takes effect again).
Now, I see that you used the nu_emqvw for your harmonic sweep in the other model, this time you can set it only to 50 and do a harminic analysis at only one frequency = 50[Hz] (no ?). By changing the application mode (stationary, Time harmonic ...) COMSOL adapts the formulas it is using (see the doc and check the equations how they change in "Physics - Equation systems - subdomain settings, or boundary settings).
I would set the "scalar variable" nu_emqvw to 50, and in the "parametric solver" set V0 as "sweep parameter name "and define a "range" in volts to scan (i.e.1-100kV or something like that). Then suppress your "Constants" "V0", "frequency" and "w" (as they are misleading but it is not required, the V0 is simply replaced by the parametric value anyhow and your "w" and "frequency" (if not used elsewhere) are then just ignored.
Then plot the electric field in V/m, use the manual "range" scaling and set a minimum to 2500000[V/m] and a maximum say twice or 4 times higher. Then anything below or above the min max will be "white" and the rainbow will start at blue for anything above your 2.5kV/mm.
there is another way to have COMSOL calculating the V0 value for you if you now the region where the field should not exceed your value, you then define V0 as a global equation variable such that the field in this region is optimised to your value, but this demands some in depth understanding of COMSOL and would be rather long to explain here.
Hope this helps, have fun Comsoling
Ivar
and then solve it to see
Hi
well if you have only got a few hours of Comsoling and want to do such nice examples I can well understand some frustration, COMSOL is very powerfull, but that means also quite some conventions and tricks that require time to get used to, even in only one physics.
First, I would say that whatever you try to do in FEM you should have first a good clue of how it's supposed to behaive by analytical simple analysis. FEM calucations will give you the details, but you MUST always check the results against a global qualitative and quantitative response obtained by other means. It's so easy to get something wrong.
This said, if you want to work in a fixed (or sweeping) frequency mode it's normally the "(Time-)harmonic" mode to use, and the frequency in "emqvw" application mode is by default "nu_emqvw" it is defined in the "Scalar variables", you can edit it there. But often it is overwritten i.e. when you use the "Solver Parameter" time harmonic mode it is the "parametric" solver that comes up and "nu_emqvw" is usually set as the parameter for the sweep, and it will overwrite the value you define in the scalar variables (if you change mode the scalar value takes effect again).
Now, I see that you used the nu_emqvw for your harmonic sweep in the other model, this time you can set it only to 50 and do a harminic analysis at only one frequency = 50[Hz] (no ?). By changing the application mode (stationary, Time harmonic ...) COMSOL adapts the formulas it is using (see the doc and check the equations how they change in "Physics - Equation systems - subdomain settings, or boundary settings).
I would set the "scalar variable" nu_emqvw to 50, and in the "parametric solver" set V0 as "sweep parameter name "and define a "range" in volts to scan (i.e.1-100kV or something like that). Then suppress your "Constants" "V0", "frequency" and "w" (as they are misleading but it is not required, the V0 is simply replaced by the parametric value anyhow and your "w" and "frequency" (if not used elsewhere) are then just ignored.
Then plot the electric field in V/m, use the manual "range" scaling and set a minimum to 2500000[V/m] and a maximum say twice or 4 times higher. Then anything below or above the min max will be "white" and the rainbow will start at blue for anything above your 2.5kV/mm.
there is another way to have COMSOL calculating the V0 value for you if you now the region where the field should not exceed your value, you then define V0 as a global equation variable such that the field in this region is optimised to your value, but this demands some in depth understanding of COMSOL and would be rather long to explain here.
Hope this helps, have fun Comsoling
Ivar
and then solve it to see