Whitney AV Solver:: Boundary tree edges error

Numerical methods and mathematical models of Elmer
Post Reply
Xan
Posts: 8
Joined: 25 Aug 2015, 12:18
Antispam: Yes

Whitney AV Solver:: Boundary tree edges error

Post by Xan »

helm.sif
(5.39 KiB) Downloaded 305 times
Hello,
I am trying to simulate a Helmholtz-like coils in the earth magnetic field, for its compensation. I created a model of the coils, which are powered by the potential difference on its edges.
Flowing current is calculated by Static Current Solver and computes properly. The resulting volume current in each node is transcript to MGDynamics solver as a body force.
When the coils are placed in a sphere, and boundary condition (Potential=0) is set on the sphere surface, solver computes magnetic field properly.

Now I am trying to model uniform magnetic field, which will simulate the Earth magnetic field. I put coils in the cylinder, and applied following BCs:

Code: Select all

Boundary Condition 5
   	Name = "BCn Flux Parallel"
   	Target Boundaries(1) = 25
    Magnetic Flux Density {n} = real 0;
	
End

Boundary Condition 6
   	Name = "l_storona"
   	Target Boundaries(1) = 26
   	Magnetic Flux Density {n} = real -1;
	
End

Boundary Condition 7
   	Name = "p_storona"
   	Target Boundaries(1) = 27
   	Magnetic Flux Density {n}= real 1;
	
End
When i run this simulation, according to attached .sif, when coming to WhitneyAVSolver I get:
Whitney AV Solver:: Boundary tree edges: 11881 of total 35640
and then Elmer crashes. I don't think it's a problem with computer (tested on 3 different machines) and generally run out of idea's. Thus I'll kindly appreciate any help.
Mesh files can be downloaded from here: https://www.dropbox.com/s/c32ubpfscaujlof/helm.rar?dl=0
With best regards,
Paweł
kataja
Posts: 74
Joined: 09 May 2014, 16:06
Antispam: Yes

Re: Whitney AV Solver:: Boundary tree edges error

Post by kataja »

Have you tried as a workaround to set boundary condition for vector potential so that its curl is just constant vector (eg, A_x = -y/2, A_y = x/2 for B=+- u_z)? I'm not sure, however, that this is a physically proper workaround to your problem. It might be..

Anyways, on a sif level it'd look something as follows:

Code: Select all

Boundary Condition 5
  Name = "BCn Flux Parallel"
  Target Boundaries(1) = 25
  A {e} 3 = REAL 0
  A {e} 2 = Variable Coordinate 1
  REAL MATC "-tx(0)/2"
  A {e} 1 = Variable Coordinate 2
  REAL MATC "tx(0)/2"
End
Cheers,
Juhani
Post Reply