calculate displacement with maxwell stress (mgdyn)

Numerical methods and mathematical models of Elmer
Post Reply
Bavragor
Posts: 32
Joined: 07 Nov 2013, 14:10
Antispam: Yes

calculate displacement with maxwell stress (mgdyn)

Post by Bavragor »

Hi!

How is it possible to get the displacement caused by the magnetic forces?
I have the maxwell stress tensor and the nodal forces. First problem: But I don't really know which element of the stress tensor in elmer (numbered from 1-6) belongs to which element in T_x,x ; T_x,y ...
Second: it is possible to set elmer a stress tensor as a body force, but when i try to set it equal as the maxwell stress tensor i get only a the trivial solution.


Thank you in advance
evgenpalam
Posts: 1
Joined: 25 Apr 2023, 22:29
Antispam: Yes

Re: calculate displacement with maxwell stress (mgdyn)

Post by evgenpalam »

The components seem to be (after calculating manually in Paraview):
1 -> xx
2 -> yy
3 -> zz
4 -> xy
5 -> yz
6 -> xz
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: calculate displacement with maxwell stress (mgdyn)

Post by kevinarden »

I do it by adding the stresssolver solver to calculate the displacement, then using the calculated nodal forces from magdyn for the stress solver as a load. Either as a body force or a boundary condition for the stress solver.

There is an example case here for a magnet pulling on a plate.
https://github.com/mrkearden/magnet

There are several sif files that go there various steps, but the displace.sif is the one that calculates the displacements

The nodal forces can also be used as a body force
Stress BodyForce 1 = Equals Nodal Force 2

just depends on what the model is trying to achieve
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: calculate displacement with maxwell stress (mgdyn)

Post by raback »

Hi

I would also warmly recommend using the nodal forces directly as in Kevin's example. This is the most accurate way of bringing the forces to the structural problem.

However, maybe Kevin there is a typo mapping "1" to "2"?

-Peter
Post Reply