[quarkz]

Hi,

I'm trying to use ghost cells to implement boundary condition for the 2d lid driven cavity problem. I'm using staggered grid and I wonder if my implementation is correct because the ans diverged after 20+ time steps.

My BC are:

u(1,:)=-u(-1,:) where u/v(-1,:) is the ghost cell for left wall
v(1,:)=-v(-1,:)
p(1,:)=-p(-1,:)

u(size_x,:)=-u(size_x+1,:) where u/v(size_x+1,:) is the ghost cell for right wall
v(size_x,:)=-v(size_x+1,:)
p(size_x,:)=-p(size_x+1,:)

u(:,1)=-u(:,-1) where u/v(:,-1) is the ghost cell for bottom wall
v(:,1)=-v(:,-1)
p(:,1)=-p(:,-1)

u(:,size_y)=-u(:,size_y+1)+2. where u/v(:,size_y+1) is the ghost cell for top wall
v(:,size_y)=-v(:,size_y+1)
p(:,size_y)=-p(:,size_y+1)

There is a "+2" becos at the top wall, u=1. At other walls, dp/dn=0,

For left/right wall

du/dx=0,dv/dx=0

For bottom wall

du/dy=0,dv/dy=0

Is my implementation correcto? tks!