Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Surface charge density / COMSOL tomograph example
Posted Apr 20, 2011, 10:17 a.m. EDT Low-Frequency Electromagnetics Version 5.2a 9 Replies
Please login with a confirmed email address before reporting spam
Hi, I have a question regarding surface charge density calculation in the electrostatic module.
In the example of the tomograph [1] the surface charge density at the two electrodes (boundaries) having a set voltage is found using the equation:
es.nD = -es.nx*es.Dx-es.ny*es.Dy-es.nz*es.Dz
This equations yields a surface charge density distribution on the outer boundaries.
My question is: can this method of surface charge density calculation be used if a boundary has an electric field on both sides? A boundary inside a model can for example be made by creating a 2D workplane in a 3D model. Or is this method illigal.
If not, are there ways to extract surface charge density in such a boundary accurately?
Thanks in advance for your help!
[1] Electric Sensor; Model 211; www.comsol.fi/showroom/gallery/211/
In the example of the tomograph [1] the surface charge density at the two electrodes (boundaries) having a set voltage is found using the equation:
es.nD = -es.nx*es.Dx-es.ny*es.Dy-es.nz*es.Dz
This equations yields a surface charge density distribution on the outer boundaries.
My question is: can this method of surface charge density calculation be used if a boundary has an electric field on both sides? A boundary inside a model can for example be made by creating a 2D workplane in a 3D model. Or is this method illigal.
If not, are there ways to extract surface charge density in such a boundary accurately?
Thanks in advance for your help!
[1] Electric Sensor; Model 211; www.comsol.fi/showroom/gallery/211/
9 Replies Last Post Apr 26, 2011, 4:52 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Apparently 4.1 does support charge density for interior boundaries... too bad it was so darn hard to find!
4.0a:
es.nD = -es.nx*es.Dx-es.ny*es.Dy-es.nz*es.Dz
4.1:
es.nD = unx*(down(es.Dx)-up(es.Dx))+uny*(down(es.Dy)-up(es.Dy))+unz*(down(es.Dz)-up(es.Dz))
I hope anyone else having questions about this finds this thread in the future....
4.0a:
es.nD = -es.nx*es.Dx-es.ny*es.Dy-es.nz*es.Dz
4.1:
es.nD = unx*(down(es.Dx)-up(es.Dx))+uny*(down(es.Dy)-up(es.Dy))+unz*(down(es.Dz)-up(es.Dz))
I hope anyone else having questions about this finds this thread in the future....
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I cannot fully follow you there. Your equations looks like if you are changing from "union" mode to "assembly mode" so you cannot be telling the full story for us to understand it.
You can define a surface charge on an internal boundary, in both "union" and "assembly" mode (the latter as a sub node to the pair continuity note), so why cant you measure it ?
--
Good luck
Ivar
I cannot fully follow you there. Your equations looks like if you are changing from "union" mode to "assembly mode" so you cannot be telling the full story for us to understand it.
You can define a surface charge on an internal boundary, in both "union" and "assembly" mode (the latter as a sub node to the pair continuity note), so why cant you measure it ?
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar!
Ah, I'm only working in union mode here. If I may ask, what makes you think I'm not?
In an electrostatics model, I can assign an electric potential to a boundary, and in the equation view it shows the variables generated. One of these variables is the surface charge density (es.nD). In 4.0a the equation is noticeably different to 4.1 as can be seen in my above post.
I did a little test with a simple 2D model built in 4.0a with a couple of boundaries set at a fixed potential. In this model the exact same result is generated for es.nD for an exterior boundary as in 4.1, but using an interior boundary the calculated surface charge densities differ quite a lot.
This results in what I believe is incorrect when using the es.nD values calculated in 4.0a for surface charge density of an interior boundary.
It could of course be that I'm making an error here, but you would have to point that out for me ;) Thanks for your response in any case! (as I understand you are not a COMSOL employee?! Your discussion contributions are deeply appreciated, which I'm sure I can say for all forum members)
Ah, I'm only working in union mode here. If I may ask, what makes you think I'm not?
In an electrostatics model, I can assign an electric potential to a boundary, and in the equation view it shows the variables generated. One of these variables is the surface charge density (es.nD). In 4.0a the equation is noticeably different to 4.1 as can be seen in my above post.
I did a little test with a simple 2D model built in 4.0a with a couple of boundaries set at a fixed potential. In this model the exact same result is generated for es.nD for an exterior boundary as in 4.1, but using an interior boundary the calculated surface charge densities differ quite a lot.
This results in what I believe is incorrect when using the es.nD values calculated in 4.0a for surface charge density of an interior boundary.
It could of course be that I'm making an error here, but you would have to point that out for me ;) Thanks for your response in any case! (as I understand you are not a COMSOL employee?! Your discussion contributions are deeply appreciated, which I'm sure I can say for all forum members)
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
My advice:
forget v4.0, do not waist time on it, and concentrate on v4.1 with latest patch
--
Good luck
Ivar
My advice:
forget v4.0, do not waist time on it, and concentrate on v4.1 with latest patch
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Okay. I hope you understood my post though.
My university currently only has 4.0a available for the network license, but I will try and use 4.1 as we have a stand alone licence on more powerful simulation computers.
Still kind of erroneous the help file in 4.0a says it does support surface charge density for interior boundaries, while it apparently does not. I can of course always enter the equation manually and it works like a charm then, but besides this 4.1 does have a lot of features that are missing in 4.0a.
Thanks for replying. Have a nice day ;)
My university currently only has 4.0a available for the network license, but I will try and use 4.1 as we have a stand alone licence on more powerful simulation computers.
Still kind of erroneous the help file in 4.0a says it does support surface charge density for interior boundaries, while it apparently does not. I can of course always enter the equation manually and it works like a charm then, but besides this 4.1 does have a lot of features that are missing in 4.0a.
Thanks for replying. Have a nice day ;)
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm afraid I didn't really, here is a simple file where I apply a charge density on an internal boundary with ES, but I calculate the charge via D or Vy (model v4.1.0185)
--
Good luck
Ivar
I'm afraid I didn't really, here is a simple file where I apply a charge density on an internal boundary with ES, but I calculate the charge via D or Vy (model v4.1.0185)
--
Good luck
Ivar
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Louis,
you are correct. Since 4.1 the definition for nD (and nJ in the Electric Currents interface) has been changed to its correct expression nD = n dot (down(D) - up(D)), where down(D) and up(D) indicate the values of D on the opposite sides of a boundary. Adding a surface charge density feature to an interior boundary sets nD equal to the surface charge density.
--
Andrea Ferrario
Electromagnetics Group
COMSOL AB
you are correct. Since 4.1 the definition for nD (and nJ in the Electric Currents interface) has been changed to its correct expression nD = n dot (down(D) - up(D)), where down(D) and up(D) indicate the values of D on the opposite sides of a boundary. Adding a surface charge density feature to an interior boundary sets nD equal to the surface charge density.
--
Andrea Ferrario
Electromagnetics Group
COMSOL AB
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar and Andrea,
First of all thanks for your model Ivar. I think I see our misunderstand as I'm trying to extract a surface charge density from an interior electrode where I apply a voltage (like on your outer boundaries), and I'm not applying a charge density myself. The nD in your model shows the correct value since it is done in 4.1, but in 4.0a it would show a different incorrect nD for the interior boundary.
And thank you Andrea for the conformation! I'm now confident enough to proceed working with the 4.1 function in 4.0a. Maybe you should include it in an update for 4.0a, if there is going to be one anytime soon ;)
Best regards!
First of all thanks for your model Ivar. I think I see our misunderstand as I'm trying to extract a surface charge density from an interior electrode where I apply a voltage (like on your outer boundaries), and I'm not applying a charge density myself. The nD in your model shows the correct value since it is done in 4.1, but in 4.0a it would show a different incorrect nD for the interior boundary.
And thank you Andrea for the conformation! I'm now confident enough to proceed working with the 4.1 function in 4.0a. Maybe you should include it in an update for 4.0a, if there is going to be one anytime soon ;)
Best regards!
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
forget 4.0, as the new 4.2 will soon show up I have heard, but already 4.1 patch 3 is very good (4.0 was the first issue after the (r)evolution of the new graphical approach and change of application modes to general physics).
My experience is that first major releases of any software are just good to learn the new things, then to forget as soon as the first or second serious minor release comes out.
I like better the new v4 approach as it fits my understanding of physics
--
Good luck
Ivar
forget 4.0, as the new 4.2 will soon show up I have heard, but already 4.1 patch 3 is very good (4.0 was the first issue after the (r)evolution of the new graphical approach and change of application modes to general physics).
My experience is that first major releases of any software are just good to learn the new things, then to forget as soon as the first or second serious minor release comes out.
I like better the new v4 approach as it fits my understanding of physics
--
Good luck
Ivar