Draft: BUG: fan: bc value not updated. See #3211
Added updating boundary values after relax.
Note: real issue is that the jump BC should be on p and not p_rgh
Plan:
- read p with user-defined jump
- read p_rgh to supply a special form of jump-cyclic that takes its jump from the p jump (and applies any corrections for the rgh varying across the disk). (or can we automatically convert pressure jump into p_rgh jump?)
- so when p_rgh get solved it applies the jump
- will still need the triggering of the correctLocalBoundaryConditions on p_rgh after relaxing it? (since original code worked on copy so would have calculated bcs)
- converting it back to p will automatically ('localConsistency') apply the jump bc?
The reading of p_rgh might be avoided if there is a way of specifying that the constraint type needs to be preserved instead of converting it to a 'calculated' type. But even then only p_rgh (or the top-level solver) knows how to convert a p jump into a p_rgh jump.
Edited by Mattijs Janssens