Use of infinity box in magnetic simmulstion software.

[alansimpson]

I am attampting to model in 3d, permanent magnetics using Cederat Flux3D. Computation takes a long time.
This is a very vague question as I am new to FEA.
Would there be any simple guidelines to sizing infinity box around magnets in order to optimise model performance?

 

[dgallup]

I use a different package (MagNet) but I worked with some people who used Flux2D & 3D. In the 2D package they had a neat transformation that very efficiently modeled the infinite boundary condition, don't know if the same is available in 3D.

If your model has a closed or nearly closed flux path there is no need to use a really big bounding box. just 3 times the model size has very little error. Also, your mesh can become very course as you get farther out, this will help reduce solve time.

With most FEA software, it is best to start with simple course mesh, solve, refine mesh (or increase bounding box) & compare results to see if you are sufficiently accurate.