Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
Transient electric field analyses
Posted Jan 18, 2011, 1:02 a.m. EST Low-Frequency Electromagnetics, Modeling Tools & Definitions, Parameters, Variables, & Functions Version 4.1 18 Replies
Please login with a confirmed email address before reporting spam
Hi,
I am trying to build a 3D model of induced currents due to application of time dependent potential as a boundary condition. I have two electrodes immersed in medium, first is grounded and the exponential growing electric potential is applied to the second electrode. Will be thankful for assistance with defining the boundary conditions for Comsol 4.1. Can this be done in AC model?
I am trying to build a 3D model of induced currents due to application of time dependent potential as a boundary condition. I have two electrodes immersed in medium, first is grounded and the exponential growing electric potential is applied to the second electrode. Will be thankful for assistance with defining the boundary conditions for Comsol 4.1. Can this be done in AC model?
18 Replies Last Post Jan 28, 2011, 12:56 p.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
if you want the Voltage of one BC to evolve with time vrite the value as
(V0[V])+(V1[V/s])*t
You might want to use the lumped port BC's too if you want to link to a SPICE or other
--
Good luck
Ivar
if you want the Voltage of one BC to evolve with time vrite the value as
(V0[V])+(V1[V/s])*t
You might want to use the lumped port BC's too if you want to link to a SPICE or other
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
Thanks,
The problem is that I need to calculate the induces currents. It appears that voltage as boundary condition is not available in Magnetic Field or Electric and Magnetic Module in 4.1. Where can I define it?
Thanks,
Alex
Thanks,
The problem is that I need to calculate the induces currents. It appears that voltage as boundary condition is not available in Magnetic Field or Electric and Magnetic Module in 4.1. Where can I define it?
Thanks,
Alex
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Here is the model
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I suppose that an EC should do. But I'm not sure exactly how to mix the physics with the current density from the EC and input it correctly to the MF to get the field out. But then begins the mix: as the field should also induce Eddy currrents in the conductors these will interact with the current, and the field should add some Lorentz term, no ?
Is'n' this the task of the electrochemistry module to have these mixes pre-cooked ? (one I do not have)
--
Good luck
Ivar
I suppose that an EC should do. But I'm not sure exactly how to mix the physics with the current density from the EC and input it correctly to the MF to get the field out. But then begins the mix: as the field should also induce Eddy currrents in the conductors these will interact with the current, and the field should add some Lorentz term, no ?
Is'n' this the task of the electrochemistry module to have these mixes pre-cooked ? (one I do not have)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi ,
I tried EC. Is solves perfectly the voltage as function of time. However when I combine this solution of MF module it does not work. Seems there is a need for additional boundary conditions.
I tried EC. Is solves perfectly the voltage as function of time. However when I combine this solution of MF module it does not work. Seems there is a need for additional boundary conditions.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
with EC it solves V, and gives you "J" the current density, I believe the issue with MF then is due to the fact you need a full current loop otherwise the calculations of A are not coherent. And in your case the current appears and disappears in the system, which is not strictly physical. You could perhaps manage better by having your two electrodes going all to the external boundary, and then have the current naturally flowing in the external boundary. In this way you respect the physics.
Interesting issue by the way.
These are typically things I ask "support". If you get a good reply pls inform us too ;)
--
Good luck
Ivar
with EC it solves V, and gives you "J" the current density, I believe the issue with MF then is due to the fact you need a full current loop otherwise the calculations of A are not coherent. And in your case the current appears and disappears in the system, which is not strictly physical. You could perhaps manage better by having your two electrodes going all to the external boundary, and then have the current naturally flowing in the external boundary. In this way you respect the physics.
Interesting issue by the way.
These are typically things I ask "support". If you get a good reply pls inform us too ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Yes, the problems is that there is a closed loop part of current goes through the matter.
This is what I recently got from "support". I do not really understand this solution. It seems to lead me away from physics and real model.
If it is absolutely necessary to model this electromagnetic problem in
time domain, then you would only need to use the Magnetic Fields interface
in COMSOL. In that case you cannot specify voltage and ground as boundary
conditions. You would need to assume the input time varying voltage and
resistance of the closed current flowing path and calculate a current
density which can be entered as a model input. In order to model the
time-varying induced current, you would then need to integrate the time
varying electric field and find the back emf. This back emf would then need
to be added to your original input voltage to adjust the actual total
voltage.
This is what I recently got from "support". I do not really understand this solution. It seems to lead me away from physics and real model.
If it is absolutely necessary to model this electromagnetic problem in
time domain, then you would only need to use the Magnetic Fields interface
in COMSOL. In that case you cannot specify voltage and ground as boundary
conditions. You would need to assume the input time varying voltage and
resistance of the closed current flowing path and calculate a current
density which can be entered as a model input. In order to model the
time-varying induced current, you would then need to integrate the time
varying electric field and find the back emf. This back emf would then need
to be added to your original input voltage to adjust the actual total
voltage.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm not sure anybody here has fully understood your model.
For me it looks like electric flow in some liquid with air around, via two rod electrodes, is it so ?.
If you want the magnetic field from this, I really suggest that you get what looks like two electrodes up to the external boundary. the you can define, even only in MEF I believe (to be checked again) a current flow into one electrode, that will loop back into the other and around on the external air surface (the latter to be set to conducting surface)
--
Good luck
Ivar
I'm not sure anybody here has fully understood your model.
For me it looks like electric flow in some liquid with air around, via two rod electrodes, is it so ?.
If you want the magnetic field from this, I really suggest that you get what looks like two electrodes up to the external boundary. the you can define, even only in MEF I believe (to be checked again) a current flow into one electrode, that will loop back into the other and around on the external air surface (the latter to be set to conducting surface)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Yes,
I have two electrodes immersed in the conductive liquid. Air is around.
That is why I cannot take them out of the medium. Part of the electrode is in the air and part in the medium.
So you propose to make a closed loop?
Thanks,
Alex
I have two electrodes immersed in the conductive liquid. Air is around.
That is why I cannot take them out of the medium. Part of the electrode is in the air and part in the medium.
So you propose to make a closed loop?
Thanks,
Alex
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I was not proposing to remove the electrodes.
But to make them longer to go all the way to the limit of your external "air" boundary. Then you can force a current flow in one and have the return via the external boundary "shell". In this way you respect the current loop for the MEF calculations
--
Good luck
Ivar
I was not proposing to remove the electrodes.
But to make them longer to go all the way to the limit of your external "air" boundary. Then you can force a current flow in one and have the return via the external boundary "shell". In this way you respect the current loop for the MEF calculations
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
Thanks. Indeed making the electrodes up to the shell solved part of the problem. However in order they work there is a need to out current in and the same current out.
Currently the main problem is that the formulation of the task with current as boundary condition is problematic. I think you proposal about solving first the EC with variable voltage as boundary condition and then submitting this solution to MF to get induces currents may work.
Thanks. Indeed making the electrodes up to the shell solved part of the problem. However in order they work there is a need to out current in and the same current out.
Currently the main problem is that the formulation of the task with current as boundary condition is problematic. I think you proposal about solving first the EC with variable voltage as boundary condition and then submitting this solution to MF to get induces currents may work.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
what is so complicated ?
you seelct a short section in air and impose a current (only in one of the two electrodes), then the current flow will buildup by itself (in my view) if you use MEF alone. Try it, I believe it should work I do not have access to my COMSOL just now ;)
--
Good luck
Ivar
what is so complicated ?
you seelct a short section in air and impose a current (only in one of the two electrodes), then the current flow will buildup by itself (in my view) if you use MEF alone. Try it, I believe it should work I do not have access to my COMSOL just now ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi ,
I decided to go on a much simple mode ; however I want to solve it by a time dependent study.
I have an electrode inside cylinder as attached. There in a exponentially increased increased current and I want to calculate the eddy currents in a specific time. Any computer I try is hung at 23% of time-dependent solver (BDF).Any ideas why? I never worked with time dependent problems in COMSOL
I decided to go on a much simple mode ; however I want to solve it by a time dependent study.
I have an electrode inside cylinder as attached. There in a exponentially increased increased current and I want to calculate the eddy currents in a specific time. Any computer I try is hung at 23% of time-dependent solver (BDF).Any ideas why? I never worked with time dependent problems in COMSOL
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
when you do transient studies, normally you give a "range" (hit the square button to the right of the time field) of time to allow the solver to step over several time sequences, as often it needs even the 2-3 first to stabilise correctly. In this way you get the responce over time, because the way you are doing is a stationary case at the conditions of t=something fixed.
Another way to go even faster is to start to work in 2D (which means working in 3 with a depth of by default 1[m] (Z direction out of paper). Then you can learn how to set upthe physics and BC's before you port it to 3D
There are many details to lern when you use COMSOL that are easy once you know them but I find not trivial to learn if you start fully from scratch. Therefore I can only suggest to mange to follow some of the COSMOL organised courses. I learned a lot by following them the first yeasr, it really gives you a kick start
--
Good luck
Ivar
when you do transient studies, normally you give a "range" (hit the square button to the right of the time field) of time to allow the solver to step over several time sequences, as often it needs even the 2-3 first to stabilise correctly. In this way you get the responce over time, because the way you are doing is a stationary case at the conditions of t=something fixed.
Another way to go even faster is to start to work in 2D (which means working in 3 with a depth of by default 1[m] (Z direction out of paper). Then you can learn how to set upthe physics and BC's before you port it to 3D
There are many details to lern when you use COMSOL that are easy once you know them but I find not trivial to learn if you start fully from scratch. Therefore I can only suggest to mange to follow some of the COSMOL organised courses. I learned a lot by following them the first yeasr, it really gives you a kick start
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
here you have an example of MF, MEF, and EC for 2 electrodes in a liquide chamber (in v4.1.0.112)
The trick is to use MEF (alone) and to get the correct voltage/current setup that is "hiding" under the "magnetic isolation" sub-node (it took me some time to find them back) so you can use only MEF to drive your current in your electrodes, but you must get them all out to the external boundary to get the current flow in a closed loop, otherwise you violate the physics
the MF and EC are there just as illustration since MEF is a combination of both
have fun comsoling ;)
--
Good luck
Ivar
here you have an example of MF, MEF, and EC for 2 electrodes in a liquide chamber (in v4.1.0.112)
The trick is to use MEF (alone) and to get the correct voltage/current setup that is "hiding" under the "magnetic isolation" sub-node (it took me some time to find them back) so you can use only MEF to drive your current in your electrodes, but you must get them all out to the external boundary to get the current flow in a closed loop, otherwise you violate the physics
the MF and EC are there just as illustration since MEF is a combination of both
have fun comsoling ;)
--
Good luck
Ivar
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
I have the same problem. I am not able to see any of the results after loading ur attachment. I dont know what the problem is. Can you resend ur .mph file.
I have the same problem. I am not able to see any of the results after loading ur attachment. I dont know what the problem is. Can you resend ur .mph file.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
my file is a 4.1 version (patched to 4.1.0.112 see main COMSOl site), you might not be able to load it if you have the "early " v4.0 or "a"
--
Good luck
Ivar
my file is a 4.1 version (patched to 4.1.0.112 see main COMSOl site), you might not be able to load it if you have the "early " v4.0 or "a"
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
Thanks.. I can see it now.. The file just took a long time to load.
Thanks.. I can see it now.. The file just took a long time to load.