Wall functions in Edge
again questions for developers:
I would like to use wall functions for rotating case (some turbomachinery, IROT=1, etc.). However, first tests crashed and I suspect wall function has something to do with this. Is there any problem in principle or is it just implementation?
Some issues I'd like to solve:
1) Wall functions are available only for strong conditions (let's take 'strong adiabatic'). In hprokz_m.f there is however some turbulence production modification for 'weak adiabatic'. Isn't that some relic from the older version where the wall canditions were called 'weak' even if many things were set strongly and the wall function could be used for this bc? I.e., shouldn't the 'strong adiabatic' be included here to really do some modification?
2) Was the case of moving walls considered in the construction of wall functions?
Why I ask: in fbcwal_m.f90 the array BCUU (and BCVV,BCWW) is set to velocity of the wall, then for instance for coarser grids (strong condition & wall function)
BCUU(I) = BCUU(I) + UU(IBCN(I))
... i. e. ABSOLUTE velocity is added to wall velocity and then
UU(IBCN(1:NB)) = BCUU(1:NB)
... for this strong conditions.
For me it would make some sense if the 'first' UU was relative velocity.
I can add that the case worked somehow if one grid was used and blew up for multigrid and also for full multigrid when going from the coarsest to the next finer multigrid stage.
Can anyone give me some reference to the theory this implementation of wall function is based on?
Thanks for help!
|All times are GMT -4. The time now is 00:12.|