Magnetostatic 3D strange results

[noir-soleil]

Recently i have tried to calculate efficiency of different shields made of high permeability material (relative permeability = 20000).
Initial field was set using boundary condition (Ax = 0, Ay = 0, Az=y ==> Bx = 1, By = 0, Bz = 0).
In the center of area there was a cylinder (tube) (mu = 20000) with external radius = 0.5, wall thickness = 0.1 and length = 2.
The cylinder axis was directed along Z axis in the first case and along Y axis in the second.
While both cases seems identical, you can see on the images below that obtained results are differ significantly.
Along Z:

cylinder along Z.jpg (17.8 KiB) Viewed 2022 times

Along Y:

cylinder along Y.jpg (12.75 KiB) Viewed 2022 times

Both images shows Magnetic Flux Density distribution on the XY surface in the center.

Can anyone give me a clue of what i can possibly do wrong?

[noir-soleil]

Problem solved by changing boundary condition to: Ax = 0, Ay = -0.5*z, Az = 0.5*y. This gives the same initial Magnetic Flux and right results.

[karwan]

Hi,

I am trying to do similar simulations. Could you please post your sir file? It would be of great help as I have just started working with Elmer.

Thanks,

K