Hi,
I'm currently trying to model the stator of a motor but I have some trouble to apply the current density in the conducting part.
To simplify the situation as much as possible, I only consider one phase in static configuration. Everything is linear.
I'm using the StatCurrentSolver to simulate the current injection in the conductor based on the different models I could find related to this topic on the Forum.
When using Dirichlet BC (imposing the potential difference through the conducting domain), everything seems to work (At least I get a decent solution). However, I would like to obtain a somehow uniform current density by imposing the value of J instead of the potential (Neumann BC). But, in this case, I can't get any convergence.
As I'm very new on Elmer, I'm probably missing something quite basic.
(Here is the link to the mesh file in case you need it as the file is too big. Sorry for that...)
https://we.tl/tTtYtygjKes
Thanks a lot,
Sebastien
Neumann BC with Static Current Solver

 Posts: 4
 Joined: 14 Nov 2023, 04:39
 Antispam: Yes
Neumann BC with Static Current Solver
 Attachments

 case.sif
 (3.34 KiB) Downloaded 3 times

 Site Admin
 Posts: 4698
 Joined: 22 Aug 2009, 11:57
 Antispam: Yes
 Location: Espoo, Finland
 Contact:
Re: Neumann BC with Static Current Solver
Hi
You need at least one node where potential is fixed. Otherwise there is mathematically no unique solution. Just a quick guess...
Peter
You need at least one node where potential is fixed. Otherwise there is mathematically no unique solution. Just a quick guess...
Peter
Re: Neumann BC with Static Current Solver
I'd say that is a very good guess...
See attached modified sif file.
Rich.
See attached modified sif file.
Rich.
 Attachments

 case.sif
 (3.42 KiB) Downloaded 4 times

 Posts: 4
 Joined: 14 Nov 2023, 04:39
 Antispam: Yes
Re: Neumann BC with Static Current Solver
Indeed,
Quick, but good guess !
The results look decent by imposing the potential value on one of the boundaries only.
Thanks a lot for the quick reply by the way.
Sebastien
Quick, but good guess !
The results look decent by imposing the potential value on one of the boundaries only.
Thanks a lot for the quick reply by the way.
Sebastien