Convergence of the Poisson solver compromised because refine/unrefine

[jose.lopez...@gmail.com]

Hi all,

I am computing the electric potential distribution in a relatively complex domain with embed solids and an spatial variable permittivity. (attached the code)

Using uniform grid the solver works faultlessly (see attached) but if I want to refine (or unrefine) a portion of the computational domain, the solver does not converge doing the solver very weird things.

Has anyone experienced this problem and knows what I am doing wrong? It is a bug?

Cheers

Jose

[j.a.v...@gmail.com]

Hallo Jose,

You should call `boundary ((scalar *){epsilonv});` after you have computed its values

An more thorough explanation can be found here:

https://groups.google.com/g/basilisk-fr/c/oYn8FHpdhIU/m/JWED6iusCQAJ

Antoon

Op dinsdag 2 november 2021 om 14:52:34 UTC+1 schreef jose.lopez...@gmail.com:

[jmlopez...@gmail.com]

Hi antoon,

You are absolutely right! Thanks for saving me! I thought it was unnecesary to apply BC to epsilonv because it is a face vector (obviously I was wrong...).

Cheers

Pepe