Handing over voltage on boundary between Equation-Modules

General discussion about Elmer
Post Reply
Wil
Posts: 72
Joined: 10 Jan 2014, 13:26
Antispam: Yes
Location: Austria

Handing over voltage on boundary between Equation-Modules

Post by Wil »

Hallo,
I need to build a model with two domains one is included in the other. In the outer domain I need to compute the static E-field distribution which is given from the preset potential at the outer boundaries and in the inner one the carrier density has to be determined.
I have set up a basic model where I have activated the "Static Current Conduction" and "Electrostatics" in both domains and used iterative solvers. The Potential of the outer domain is influenced by the separated carriers and vice versa. But I do not see any coupling (see attachment). The outer field is completely unaffected by the conductor domain.

I need to know what I do have to fill into the Potential dialogs for both equations at their boundary. The dialogs are there but I do not know how the variables are called. The names form the "equation-> Edit Solver Settings -> Potential" dialog are not working

Thanks in advance,
Wil
Attachments
elmer_unaffeted_potential_in_two_intersection_domains.png
(52.5 KiB) Not downloaded yet
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Handing over voltage on boundary between Equation-Modules

Post by raback »

Hi

Perhaps you could write down the equations. It is not clear to me what kind of coupling you desire to have.

I'm afraid that the the XML definitions that define the GUI menus assume "Potential" for both equations. Therefore I don't see how you could properly treat the two equations with the GUI without touching at least the XML files. I would do the rest by hand.

-Peter
Wil
Posts: 72
Joined: 10 Jan 2014, 13:26
Antispam: Yes
Location: Austria

Re: Handing over voltage on boundary between Equation-Modules

Post by Wil »

Thank you peter.
I think I am a step further.
I now just use "Static Current Conduction" and set the conductivity of the surrounding air to a very small value. The resulting field distribution looks correct. This means I hopefully do not have to couple the both solvers any more.

Now I want to calculate the force on a boundary. I have found some information about what to insert into the .sif in 'Tutorial 6' 'Electrostatics' in the Elmer-non-GUI-Tutorials (http://www.nic.funet.fi/pub/sci/physics ... nonGUI.pdf).

The main thing here is that I want to/have to use the GUI and always changing the .sif does not work for me when I have changed something in the GUI. This is simply a error source.
There are free text input dialogs for almost all Equation and Boundary related things but no one for an additional solver like the "Electric Force" solver.
Do you have any suggestion how this could be solved? I mean how to use the GUI and insert anywhere the following solver related text for the electric Force solver.

Solver 2
Equation = Electric Force
Procedure = "ElectricForce" "StatElecForce"
End

An other question regarding the Electric Force solver is: How do I get the value of the force. I know that I have to specify a boundary where the force is then calculated but how do I get the overall value of this force (see the tutorial on page 30)

BR
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: Handing over voltage on boundary between Equation-Modules

Post by mzenker »

hi,

you can edit/build the xml file yourself as you need it - the syntax is not hard to understand, and is also documented in the ElmerGUI manual.

HTH,

Matthias
Wil
Posts: 72
Joined: 10 Jan 2014, 13:26
Antispam: Yes
Location: Austria

Re: Handing over voltage on boundary between Equation-Modules

Post by Wil »

Hallo,
I have built a small simple model to show the basic steps on how to build the above requested simulation. Since it was not really easy for me to figure all details out I hope others can benefit from this.

This model incorporates two touching domains, where each is solved from a different solver. The "air"-domian is solved by the "Electrostatic" (ES) solver and the "copper"-domain is solved by the "Static Current Condition" (SCC) solver. On the interface between these domains is the potential shared or "handed over" from the SCC to the ES. I introduced a new variable name for the "Static Current Condition" solver since both solvers do usually use the variabel "Potential" what would lead to false results (see also viewtopic.php?f=4&t=4094&start=0).

I have attached a picture of the model set-up and the results as well as the complete project as .zip. If you want to examine the project especially the variable names be aware of problems as described here viewtopic.php?f=16&t=4107&start=0.

If needed one can find the edf file for the electrostatic force here viewtopic.php?f=3&t=4100&start=0 .

Even when I was able to simulate the correct electric potential in the "Air" and the "Copper" domains there are some questions left:
- As can be seen in the attached picture is the electric energy density and the electric field are zero in the complete simulation. This is not correct but I do not know what is wrong.
- The solver for the electrostatic force yields the correct magnitude for the resulting force but the direction is wrong. To my knowledge should the force direct into the negative x-direction but it shows a positive sign. The magnitude of the force was analytically estimated with the formula F = epsilon0/2 * deltaV**2/dist**2 *Area (force acting on a parallel plate condensator) resulting in 1.77e-09N.

Does anyone has an idea what I have to do to resolve the two issues above?

BR
Wil
Attachments
Setup and results
Setup and results
setup and results.png (107.6 KiB) Viewed 6612 times
Elmer working with force.zip
Project files.
(124.16 KiB) Downloaded 386 times
Wil
Posts: 72
Joined: 10 Jan 2014, 13:26
Antispam: Yes
Location: Austria

Re: Handing over voltage on boundary between Equation-Modules

Post by Wil »

I have solved the described problem above by solving the model in Elmer and calculating the electrostatic force in paraview. After simulating the ES and SCC the flux solver provides the electrostatic flux and gradient for calculating the force in paraview by applying several filters such as threshold, calculator and integration. Threshold was used to filter the left surface using the Geometry Ids exported from Elmer (Equation->Result Output->Save Geometry Ids).
Post Reply