Error when whole grid rotates - why?
 

Hi,

I'm trying to simulate a 2d airfoil heaving/pitching using fractional step and ALE. I used 2 ways to achieve that.

BC is inlet fixed at u=1, top/bottom v=0, du/dy=0, outlet du/dx=0,dv/dx=0

1. whole grid rotates/moves. no deformation

2. external boundary of grid fixed. mesh deform to accomodate motion. GCL used.

The strange thing is that mtd 2 works fine, except when the grid quality deteriorates. However, mtd 1 usually works for abt 8s (~50000 time step) and then starts to go haywire.

u velocity on top/bottom/outlet boundary can get 1.5~2. outlet may also get reverse flow. I wonder what wrong...

Is there mass in/out which I must handle? I thot the poisson eqn ensures global mass conservation. Or should I change the BC. I've used convective BC for outlet but it makes not much difference.

what should I do or check?

Thank you.

Re: Error when whole grid rotates - why?
 

Dear Zonexo,

You need to carefully check your outlet BC. Unlike (2) where the outer boundary is held fixed, in (1), the boundary also moves,so that the grid motion has to be accomodated in the BC. If you are sure that there is no major differences between your codes using (1) and (2), except at the BCs then this has to be the only cause of failure.

Hope this helps

Regards,

Ganesh