corner treatment with fullway BB in a square obstacle

Hello to everyone,

I am having some trouble on the concept of BounceBack implementation in a particular case.

I am working with a square obstacle and a standard bounceback, 2DQ9, poiseuille.

I want to know if I need to consider the next suposition.

as I use BB, the real wall is exactly between the fluid/solid interface of these nodes, so If I set a square lenght of X as an obstacle, do I really need to sum 1 because of the increment of wall that comes from the bounceback when I calculate anything that requires the lenght of the square lenght (i.e. Drag/lift/strouhal, viscosity/Reynolds…)?

in addition, this is kind of difficult to explain:

if I study the forces on a corner of the square obstacle, let’s say the top left corner of the square and consider vector directions as:

v(0:8,x) = (/0,1,0,-1,0,1,-1,-1,1/)
v(0:8,y) = (/0,0,1,0,-1,1,1,-1,-1/)

I focus on the square corner node, so i set in as the (y,x) node of my image matrix and it is wall.

Now considering the fluid node(x-1,y), the program would identify solid adjacent nodes in these directions: v(1), v(5) because I am on (x-1,y). But if I take carefully a look on v(5), this one is a special case… it’s adjacent nose (x,y+1) is liquid. But it’s path between these two nodes comes across the REAL wall. Is there someone able to explain me if I am right and what do I have really to consider in these cases. This is relevant for me because I am evaluating the drag and lift forces with the Momentum Exchange method, and I do have some values of these coefficients that do not match perfectly with the bibliography that I have for this case. (S. Izquierdo et all, 2009, analysis of open boundary effects in unsteady lattice boltzmann simulations)


I place this image to exemplify it:

The shaded zone is wall between the 1/2 of the fluid solid nodes

So Do I have to take into account for the half way bounceback the orange direction from that point? In addition, do I have to take it into account for the force evaluation? If I do so, how do I apply the bounce back if it is only one direction for each point? It’s really confusing me.

I need response on this topic please

I appreciate the effort if just someone has read the thread.

Thanks in advance,

Albert P