Pressure de-coupling with fractional step
Hi, this is a fairly specific question so, sorry...
I have been writing an incompressibile 2D unstructured, unsteady NS solver using a fractional step approach. It uses non-staggered 2nd order upwind FV's on the median dual mesh (vertex centred). 1st order pressure projections, (like Chorin) work fine, but 2nd order projections (pressure corrections) lead to a grid scale pressure de-coupling. Some people claim to be able to use the 2nd order projection without de-coupling on collocated structured grids. Does it seem reasonable to be getting pressure de-coupling?? Can anyone suggest a fix?? |
All times are GMT -4. The time now is 07:27. |