Discussion Forum

Posted: 2 decades ago Aug 18, 2009, 3:15 a.m. EDT

Hi I assume you are in 2D mode, then the "thickness" is the extend of your model in the "out of plane" direction, this is related to certain variables such as a density given in kg/m^3 and that need a thickness value (3rd dimension) to calculate correctly, or more related to electrostatics alone : a space charge C defined in C/m^3. Normally COMSOl is working coherently in 3D but gives you the possibility to work in 2D by fixing one direction through the "thickness" value, In the same way, in 2D axisymmtry, to calcualte a volume you must multiply by 2*pi*r the corresponding "thickness" for an aximsymmetric case hope this helps (and that my interpretation is fully correct) good luck Ivar

Posted: 2 decades ago Aug 18, 2009, 5:14 a.m. EDT

Yes I agree with that, The thickness is activated only in 2D mode. There is no thickness parameter for 3D or 2D axial symmetry. You can notice that the thickness (d) is multiplied to the volume charge density (rho), this gives you somehow a surface charge density in your model. I hope it helps, Cheers

Posted: 1 decade ago Dec 15, 2010, 7:07 p.m. EST

Hi: I am using 4.0, and I could not specify differnet regions with different thickness(though I do know how to perform this task with version 3.5.. ) Could some please help me with this simple question? Thanks! Jason

Posted: 1 decade ago Dec 16, 2010, 2:22 a.m. EST

Hii 1) thickness in 2D axi is th "loop length" that is 2*pi*r and should not be forgotten, there is a check box to pass from 2D integration to 3D via the loop length within COMSOL (but it has not been implementd for 2D thickmness as in 3.5 yet). Note that in v4.0a there was a convention change about the 2*pi, but I cannot rember if it was maintained in v4.1. I ALWAYS check such things by integrating the surface to volume and check my results with my CADsurfaces/volumes dump to be sure I apply the correct formula, it's too easy to get that wrong ;) For the change in thickness, there are subnodes to be aplied , at least in v4.1, so get the la version and add the latest patch then you will be back in work -- Good luck Ivar

Posted: 1 decade ago Dec 16, 2010, 4:56 p.m. EST

Hi Ivar: I am using 4.1 now, but still do not have a clue to specify different thickness in different domains. When using 3.5, I can easily do that by selecting different domains and set its own physical subdomain property(Like Jason Zhang post ), but with 4.1, I could not do it.

Attachments:

• thickness.png

Posted: 1 decade ago Dec 17, 2010, 1:20 a.m. EST

Hi have you tried to right click your nodes and sub-nodes (all the way to the bottom leaf ? then some nice extra items/subnodes appear -- Good luck Ivar

Posted: 1 decade ago Dec 17, 2010, 11:04 a.m. EST

Hi Ivar: I have right-clicked every subnode/subleaf multiple times, spending two days but still couldn't find this trivial function, and this 4.1 is worse than 3.5 in terms of interfacing, Could you please teach in more detail? And is there any technical support I could consult by telephone? Thank you very much. Jason

Posted: 1 decade ago Dec 21, 2010, 3:16 a.m. EST

Hi in V4 you must also check that you turn on see more settings (there is one to select in the preferences side by side with the equation view, and another in the advanced under the main physics. As I never remember what is hidden and what is always there I turn always on all options equations more settings hidden features etc, but its true it becomes crowded in there ;) -- Good luck Ivar

Posted: 1 decade ago Dec 21, 2010, 7:38 a.m. EST

[QUOTE] Hi: I am using 4.0, and I could not specify differnet regions with different thickness(though I do know how to perform this task with version 3.5.. ) Could some please help me with this simple question? Thanks! Jason [/QUOTE] Hi, In version 4, it is no longer possible to set the thickness per subdomain in 2D electrostatics and electric currents. You can set a thickness parameter for the entire physics inter face on the physics interface node. A possible workaround is to multiply the permittivity by a domain specific thickness factor. It makes sense to use a domain specific thickness in stationary electric currents as the current may well be confined within the thickness of a slab (surrounded by a non conducting medium). For electrostatics and high frequency electric currents modeling it is more questionable as the electric displacement is never that well confined. Dielectric contrast is rarely more than an order of magnitude. The main reason it has not been implemented in version 4 is that we thought it may encourage the less experienced user to apply a questionable approximation. Best regards, Magnus Olsson COMSOL

Posted: 1 decade ago Dec 21, 2010, 9:19 a.m. EST

Hi Magnus Thanks for the explanations, indeed several of us missed the subtilities of the change, but you have some good arguments there How many of us can really tell what the fringe effects are in absolute/relative values for a 2D model w.r.t a full 3D one ? My comment applies to EM too, by the way -- Good luck Ivar

Posted: 1 decade ago Dec 21, 2010, 5:51 p.m. EST

Thanks for the answer, and I sort of get around with it using simple scaling technique. However, I have a question regarding post-processing: How could I plot the intensity of norm electric field along a streamline in"electrostatics" module? Thanks. Jason