Learning Forum

[Yusuf Akcay]

Hello all,

I have a 2-pole section of a motor (symmetry applied) with boundaries on the sides (dependent and independent). I usually perform an adaptive mesh in the static solver and export it to the transient solver (the same geometry). However, the accuacy drops as soon as rotor rotates. The adaptive mesh is only created at the initial rotor position (assuming no shift and sides of the rotor and stator are aligned). 

Is it possible to make an adaptive mesh at each rotor position? Based on my experice, I cannot move the rotor in the static solver as it clashes with the boundary. 

Thanks,

[GLUO]

Hi, 

In transient solver, no. If it is a PM motors, you can simulate it in magnetostatic solver with a "mask" to handle the boundary. Although the matching boundary enforces the periodicity of fields at the boundary edges (or faces in 3D), it does not move the geometry out of the boundary back to the solution domain in the magnetostatic solution. To create a valid partial model for a magnetostatic analysis where the position/angle is going to be swept, the geometry must be fully contained in the periodic sector modeled.

I would still recommend using transient solver and segmented bands (in each time step, rotor should rotate integer number of segments) + mesh operations to improve the accuracy.

GL