Hi Matthias!
I have made some experiments during the night.
I used working example: mgdyn_anisotropic_rel , I substituted .F90 file by permfile.F90:
Code: Select all
!-------------------------------------------------------------------------------
SUBROUTINE permf(Model, n, X, Y)
!-------------------------------------------------------------------------------
USE DefUtils
IMPLICIT NONE
TYPE(Model_t) :: Model
INTEGER :: n
REAL(KIND=dp) :: X(*)
REAL(KIND=dp), POINTER CONTIG :: Y(:,:)
!-------------------------------------------------------------------------------
Y = 0._dp
Y(1,1) = 50000.1_dp
Y(2,2) = 2.1_dp
Y(3,3) = 2.1_dp
!-------------------------------------------------------------------------------
END SUBROUTINE permf
!-------------------------------------------------------------------------------
and I made simple magnet-anisotropic plate test case in NETGEN:
Code: Select all
#
## aniso test
#
# Netgen 6.2 required
#
algebraic3d
solid core = cylinder (0, 0, -1.5; 0, 0, 1.5; 3)
and plane (0, 0, -1; 0, 0, -1)
and plane (0, 0, 0; 0, 0, 1) -maxh = 0.3;
solid magnet = cylinder (0, 0, 0.1; 0, 0, 3.9; 1)
and plane (0, 0, 0.2; 0, 0, -1)
and plane (0, 0, 3.5; 0, 0, 1) -maxh = 0.3;
solid range = sphere (0, 0, 0; 25)
and not magnet
and not core;
tlo core -col=[0,0,1];
tlo magnet -col=[0,1,0];
tlo range -col=[1,0,0] -transparent;
aniso_test.sif is in the attachment.
Please see the results:
It works, solves, but unfortunately, there is NO ANISOTROPY.
I agree, we need help. Fortran sources are to sophisticated for me. I give up for this moment.
Thank you for your help and discussion!
With best regards,
Roman