Electrostatic: mesh with wires

Numerical methods and mathematical models of Elmer
LowLow
Posts: 9
Joined: 15 Jan 2018, 20:16
Antispam: Yes

Electrostatic: mesh with wires

Post by LowLow »

Hello,
I am new in Elmer and I am struggling with ElmerSolver (I generate the mesh with gmsh).
I try to get a electric potential for the following configuration:
- a cathode at a given potential CATHODE_V=-300 V
- an anode at a given potential ANODE_V= +800V
- an in-between electrode (copper wires) at a potential MESH_V=-150 V. The in-between electrode has the particularity to be a set of intersecting wires (see paraview screenshot visu).
- a gas in the volume space

I was expected to get an almost constant potential at the Z-coordinate of this electrode, which is not the case accordingly to the result (see paraview screenshot in attached file). The solver does not seem to ensure the continuity between the electrode boundary condition and the space volume: there is a linear decrease between the cathode and the anode. ElmerSolver seems to not take into account the electrode boundary condition to calculate the potential in the gas volume.

I put my .sif file in the attachment. Firstly I tried just using the "StatElecSolver" and since it doesn't work, secondly I tried to couple it with "StatCurrentSolver" as it is advised in another post, but this configuration makes the program crashes (this is the commented configuration in my .sif file).
The boundary conditions that I have set aim to reproduce this potential field infinitely.

Please, if someone has any idea what I am doing wrong? It must be silly but I can't figure it out...
Thanks in advance!
Attachments
In between electrode visualization
In between electrode visualization
potentialConfig_inBetweenElectrode.png (17.7 KiB) Viewed 4333 times
result_potential_config_problem
result_potential_config_problem
potentialConfig.png (8.87 KiB) Viewed 4333 times
inputElmerSolver.sif
sif file
(4.44 KiB) Downloaded 248 times
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: Electrostatic: mesh with wires

Post by mzenker »

Hi,

I don't understand the definition of the potential in your boundary conditions. The variable names you use have to be assigned numerical values somewhere.

Matthias
LowLow
Posts: 9
Joined: 15 Jan 2018, 20:16
Antispam: Yes

Re: Electrostatic: mesh with wires

Post by LowLow »

Hello Matthias,
Thanks for the reply. Basically, I replace the variable CATHODE_V, ANODE_V and MESH_V with a Python script which assign the potential number (I have just posted my template). After the Python script finished, the final .sif file is as you can see in the attached file. Is that answer to your remark?
Since yesterday, I have played with the solver options but I have not figured out a solution yet...

Loïc
Attachments
inputElmerSolver.sif
input .sif with assigned potential
(4.46 KiB) Downloaded 239 times
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: Electrostatic: mesh with wires

Post by mzenker »

Hmmm...

Just to be sure that I understand correctly:
Your cathode and anode planes are in the xy plane, at differenz z positions?

Then I would expect the result shown in your screenshots, especially the second one:A linear decrease between anode and cathode. For the wire, I cannot distinguish the potential distribution. One would need to zoom in to see if there is a linear increase/decrease towards the wire.

What is the error message for the StatCurrent solver?

Matthias
LowLow
Posts: 9
Joined: 15 Jan 2018, 20:16
Antispam: Yes

Re: Electrostatic: mesh with wires

Post by LowLow »

Yes, the cathode and the anode are in the XY plane at a given Z position: anode_Z=0. cm and cathode_Z=1.5 cm. The "in between" electrode (the mesh) is at mesh_Z=0.5 cm.
I expect a linear decrease between the cathode and the anode but with: one linear decrease with a given slope A between the cathode and the mesh AND a linear decrease with another slope B between the mesh and the anode.
In fact, for an "in between" electrode made just of a set of parallel wires placed periodically, I found with the electrostatic code Garfield++ (which only calculates an electric field having an analytical solution) the behavior that I have just described see the attached file (Garfield_analytical_solution_forParallelWires). Thus I was expected such a behavior when basically I made this previous set of parallel wires (X direction, in the XY plan) intersects perpendicularly with another set of parallel wires (Y direction, in the XY plan). In both cases I take the same wire pitch and wire diameter.
But with Elmer, it is clearly not the case see file: Potential_along_Z-axis, representing a plot of the potential along the Z-axis at the center of the geometry X=0. and Y=0..
However, Elmer does take into account the boundary conditions as you can see on the figure: potential-ElectricField_AlongTheWire_Mesh-Z.png
With the "StatCurrent solver" I manage to get ride of the errors that I have had. I set in attached file the output and my current .sif file.
I am a little bite lost... Could it be a scaling problem?
Thanks for your help!

Loïc
Attachments
Potential along the X-axis at the coordinate Z=mesh_Z
Potential along the X-axis at the coordinate Z=mesh_Z
potential-ElectricField_AlongTheWire_Mesh-Z.png (14.1 KiB) Viewed 4291 times
Potential_along_Z-axis
Potential_along_Z-axis
potential-ElectricField_AlongThe_Z-axis.png (6.29 KiB) Viewed 4291 times
PotentialProfile_alongZaxis_GarfieldCode.pdf
Garfield_analytical_solution_forParallelWires
(19.99 KiB) Downloaded 217 times
LowLow
Posts: 9
Joined: 15 Jan 2018, 20:16
Antispam: Yes

Re: Electrostatic: mesh with wires

Post by LowLow »

Here are my .sif file and the code output.
Attachments
outputElmer.txt
Elmer output
(49.62 KiB) Downloaded 215 times
inputElmerSolver.sif
sif file
(4.7 KiB) Downloaded 228 times
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: Electrostatic: mesh with wires

Post by mzenker »

Hi,

I see your point. Could you post your mesh (if it is too big, then try to make a coarser one, that should not make a big difference here)? I will try to have a look on the case later today.

Matthias
LowLow
Posts: 9
Joined: 15 Jan 2018, 20:16
Antispam: Yes

Re: Electrostatic: mesh with wires

Post by LowLow »

I forgot to put the potential curves assigning to the previous geometry screenshots, sorry
Attachments
Potential along the Z-axis at X=0. and Y=0.
Potential along the Z-axis at X=0. and Y=0.
potential-ElectricField_AlongThe_Z-axis_Potential.png (10.76 KiB) Viewed 4289 times
Potential along the X-axis of the mesh at Z=Mesh_Z
Potential along the X-axis of the mesh at Z=Mesh_Z
potential-ElectricField_AlongTheWire_Mesh-Z_Potential.png (8.02 KiB) Viewed 4289 times
LowLow
Posts: 9
Joined: 15 Jan 2018, 20:16
Antispam: Yes

Re: Electrostatic: mesh with wires

Post by LowLow »

You do not arrive to have a .msh file less than 1MiB... However, I upload the file on WeTransfer and you can download in following the link:https://wetransfer.com/downloads/1362cd ... 026/ac590a
Is it ok for you?

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

Re: Electrostatic: mesh with wires

Post by mzenker »

I will try later.
You can also zip the mesh file, and/or you can post the gmsh geo file so that I can simply rebuild the mesh.

Matthias
Post Reply