How to simulate the compression of the dielectric in a capacitor?

[tfsilva]

Hi,
I am currently working on the simulation of the compression of the dielectric in a capacitor (sort of).
I am really confuse on how to set the interaction between the different solvers (electric potential, electric force, deformation of the dielectric, and the mesh update).
I am a new user of Elmer and many things are still confused.
Please, see attached files.

First thing I do not understand is why I can not see the electric force results in Paraview.
The second is, when I added the deformation solver, the warning message "WARNING:: ListFind: Requested property: [Poisson Ratio], not found" appears insistently.

If anyone can give me some light on this, I would appreciate.

[kevinarden]

Regarding the warning, the stress solver is applied to all bodies using equation 1. Therefore all bodies have to have solid mechanical properties defined in their material such as Poisson ratio. This can be avoided by using 2 equations

Equation 1
Active Solvers(2) = 1 2
Calculate Electric Energy = True
End

Equation 2
Active Solvers(1) = 3
End

Where equation 2 is the stress solver. The other bodies would reference equation 1, and the deformable body would reference equation 2.

[kevinarden]

To use the stress solver you need to specify boundary conditions for the stress solver. Restrained displacement on the lower copper plate, and added the electric force as an applied load on the upper copper plate. Changed steady state iterations to 1. Changed to 2 equations.

gemcell.sif

(4.01 KiB) Downloaded 133 times

displacement.png

(43.78 KiB) Not downloaded yet

[tfsilva]

Dear @kevinardem,
thanks for the help!
It shows Elmer can simulate what I need.
However, I try to impose a different boundary condition.
Instead of restraining the displacement of the lower copper plate, I would like to restrain only the X and Y displacement on boundaries 5,6, 9, and 10.
I tried. Elmer finishes calculation but with no convergence on the displacement computation. I also applied electric force on the lower copper plate.
Modified sif file in the attachment.
Should I impose the restrain in the displacement of the middle plan of the dielectric? How to do that?
Any suggestion?

[kevinarden]

StressSolver needs to be restrained from rigid body motion in a static solution. Meaning in 3D there has to be a restraint on Displacement 3 somewhere.

[tfsilva]

Hi kevinarden,
Thanks again!
I finally managed to fix the boundary conditions.
Results seems reasonable, but there is one thing still not working properly.
The upper copper plane is stretching while compressing the dielectric. Please, see attached figure. Isn't it weird?
How can I impose the thickness to not to change?
Thanks in advance!

[kevinarden]

Deflection in an orthogonal direction is a function of Poisson ratio. However I believe that there is meshing issues for good analysis of StressSolver. The mesh is tetrahedrons and should be higher order for better results. However the mesh size in way of the dielectric is large.

displacement.png

(453.89 KiB) Not downloaded yet

[tfsilva]

Hi kevinarden,
Thanks! You are right! I have to fix that.
But my question was about the grey part. It was initially as thin as the red one. When compressing it get thicker and thicker...
Is this also a mesh problem? Apparently there is a mistake defining the boundary condition or the force in this copper electrode.
Thanks in advance.

[kevinarden]

Looks like the force is applied in a positive number which is extension not compression.

[tfsilva]

Well, it looks like a Mesh Update solver is missing to compute for the movement of the upper copper plate.
The rough mesh is a consequence of the stretching of the mesh since the body does not move.
I was unable to make it work properly. Could you please take a look and give a hint on what is wrong?
Thanks again!