3D simulation of a Motor without application of a load

[raback]

Hi Felix,

The torque in 2D and 3D is a little different. In 2D it must be multiplied by the length of the machine even when using 2.5D approach. In 3D not. In both cases the result must be multiplied with the number of symmetric sections.

Regarding BCs here are some that I've been using lately. Here only part of machine (1/4) is used and the BCs are detected automatically for a typical elecrtical machine. But you can use them even if you use the names or target boundaries.

It may be worth using conforming BCs in cases where the symmetry allows that. This reduced the size of the linear system to N-M instead of increasing it to N+M. Unfortunately the rotating interface can never be conforming.

There is a version of generic strong (edge-to-surface) projector that might be more robust. It would be nice to apply elimination for that.

Maybe I could have a look at the case at the elmag repo.

BR, Peter