Problems with periodicity by rotation

[Erwan Le Coupanec]

Hello,

The problem is that the joining does not perform properly. Periodicity of rotation is based on a non conforming joining between the boundary faces of your mesh that are images of one another by the rotation you set. In your case, with your settings, a hundredth or so of boundary faces belonging to periodic_1 and periodic_2 are not joined, and they remain boundary faces after the joining. Then in your boundary conditions settings, these leftover boundary faces are not provided with any type of BC, so that the computation can't run.

I made it work by setting an angle of rotation of 270° (the other way around) and by increasing the fraction from 0.1 to 0.2 (tolerance distance ratio for the joining).

But I think your mesh could be nicer considering that the geometry is rather basic. At least, the periodic_1 and periodic_2 could be conform with one another.

Erwan.

[quarian_engineer]

Dear Erwan,

thanks a lot! Your solution could be appropriate.

But where do you set the tolerance distance of joining?
The one thing I can match with this in code saturne is the partitioning, which is actually set to "default" in my case.

best regards,
Selora.

[Yvan Fournier]

Hello,

As Erwan explained, periodicity is handled as an extension of joining. The joining tolerance in the periodicity definition is defined by the "Fraction" (and "Plane") parameters in the periodicity selection definition.

Partitioning is not related (it should impact only performance).

Regards,

Yvan

[quarian_engineer]

Finally, it works (with fraction 0,2 and angle of rotation 270°).
Thanks a lot!