I’m coding multi component single phase D3Q19 shan-chen LBM. In my simulation fluid2 occupied entire the channel except 10 initial lattice that occupied by fluid1. In y and z direction I have bounce-back and in x I have Zou/He pressure boundary. When the simulation starts, fluid1 force fluid 2 in to the outlet and progressing in the channel but after a while everything’s blowup because fluid2 is entered to the channel from inlet corner nodes!! Where is my mistakes, please help me.
As you’ve almost said yourself, there’s probably an inconsistently in the boundary conditions at the corners. At an inlet there are 5 unknown f_i for a D2Q9 model, more for D3Q19, and you probably have something wrong here. Also, as a thought , Zou and He finds populations moving along the boundaries, while bounce-back doesn’t allow the streaming and collision of particles along the wall, so you have to be very careful (especially if you are using half-way bounce back, because then the two methods don’t necessarily meet at the corners). Why not try something simpler first, like a D2Q9 implementation (perhaps even using just one method) to see what you get? It’s much easier to understand something if you start with the simplest possible situation.