incompressible NS solver with multi-block structured grid
I recently wrote an incompressible Navier-Stokes solver code to deal with multi-block structured grid and immersed boundary method.
When I run sample cases with different Re conditions, I found some problem in my code.
When I run the case with Re of 100, it seems to work properly. But when I increase the Re to be 1000, then a certain source of error comes into play.
I attached the figures of results with different Re.
Why are problems? Does anybody have an idea on it?
I used preconditioned Bi-CG stab solver instead of LDU type solver to deal with asymmetric matrix due to block(refinement) interfaces.
The flow field behaves like checker board problem in some block interfaces.
I need your help orz
does not look like checker boarding to me. It looks like your convection term is creating unwanted sources. If this is the case first order upwind scheme should work fine. (for testing purposes).
|All times are GMT -4. The time now is 06:32.|