Regarding Stat Elec Solver and Boundary conditions

Numerical methods and mathematical models of Elmer
Post Reply
ajay0705
Posts: 17
Joined: 13 Nov 2018, 01:17
Antispam: Yes

Regarding Stat Elec Solver and Boundary conditions

Post by ajay0705 »

Hi,

I am trying to model 2 cuboids in gmsh and find out the capacitance between them and of each metal with ground.
I have attached my .sif and .geo file to this query.

I am using the Stat Elec solver to get the values of the capacitance. I have defined boundary conditions for each of the 2 metals as given in the .sif file.
My question was, what should be the potential values in each of the boundary conditions and would it be based on the circuit that i make or do they have some specified constant values?

Also just for the sake of comparison i changed the potential of one of the metals from 10 to 3 and kept the other metal potential at 0, these were the capacitance values that i got.

Potential = 10
StatElecSolve: 1 1 -7.21090E-16
StatElecSolve: 1 2 7.15129E-08
StatElecSolve: 2 2 6.32850E-15

Potential = 3
StatElecSolve: 1 1 -6.47565E-17
StatElecSolve: 1 2 7.15129E-08
StatElecSolve: 2 2 9.48448E-16

Also, could you help me if there is anything missing in my .sif file that you could spot out?

Hope to hear from you soon.

Thank You,
Ajay Balaji
Attachments
test.sif
(1.99 KiB) Downloaded 287 times
test.geo
(2.54 KiB) Downloaded 273 times
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Regarding Stat Elec Solver and Boundary conditions

Post by raback »

Hi

The "Capacitance Body" keyword will permutate appropriate combinations of zero and one potentials. Don't specify "Potential" as the two are conflicting. You could specify potential for the farfield condition to give you self-capacity (or use the infinity condition). To summarize if you want to have 2x2 capacitance matrix you need 3 BCs. You're missing the reference.

-Peter
ajay0705
Posts: 17
Joined: 13 Nov 2018, 01:17
Antispam: Yes

Re: Regarding Stat Elec Solver and Boundary conditions

Post by ajay0705 »

Hi,

Oh okay. Thanks a lot for clarifying my doubt.
And by reference do you mean the ground surface?
If so then will I have to define it as potential = 0 or not?

Thanks
Ajay
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Regarding Stat Elec Solver and Boundary conditions

Post by raback »

Yes, reference would be ground given either by zero potential or the farfield condiction (assumes spherical potential from origin). -Peter
ajay0705
Posts: 17
Joined: 13 Nov 2018, 01:17
Antispam: Yes

Re: Regarding Stat Elec Solver and Boundary conditions

Post by ajay0705 »

Hi,

Thank you so much. So for adding reference I should add another capacitance body? Would i have to specify potential as 0 for that or is there another way to make the solver understand that it is ground?
Also, could you suggest some ways to get a more accurate value for capacitance of certain test cases? As in, are there any specific boundary conditions or body force conditions to be applied to a model?

Thank you.
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Regarding Stat Elec Solver and Boundary conditions

Post by raback »

Hi

Ground can be either "Capacitance Body = 0" or "Potential = 0".

If you look for accuracy then refine the mesh and study "grid convergence" of your results.

-Peter
ajay0705
Posts: 17
Joined: 13 Nov 2018, 01:17
Antispam: Yes

Re: Regarding Stat Elec Solver and Boundary conditions

Post by ajay0705 »

Hi,

Thank you for helping me out earlier.
I have another question.
In the case of calculating capacitance between 2 3D metals, inside a tall dielectric stack-up, how do you control the amount of fringing capacitance?
I observed that the final capacitance values change drastically when I truncate some part of the dielectric stack-up.
Is there an ideal situation that should be considered for this kind of a scenario?

Thanks
Ajay
Post Reply