velocity_bfl Subroutine

public subroutine velocity_bfl(me, state, bcBuffer, globBC, levelDesc, tree, nSize, iLevel, sim_time, neigh, layout, fieldProp, varPos, nScalars, varSys, derVarPos, physics, iField, mixture)

Inlet Velocity BFL rule boundary condition

This is taken from the paper: "One-point boundary condition for the lattice Boltzmann method", by M. Junk and Z. Yang published in Physical Review E 72 (2005) and leads to the chapter 'B. BFL rule'. This boundary condition shall be an improvement of the already existing velocity bounce back condition. It introduces a term called $f_{i^{\star}}^{b}$. This one represents the particle distribution: $$\hat{f}_{i}(n+1,\mathbf{j}) = \hat{f}_{i^{\star}}^{c}(n,\mathbf{j}) + 6hf_{i}^{\star}\mathbf{c}_{i} \cdot \mathbf{\Phi}(t_{n},\mathbf{x}_{\mathbf{j}i})$$

This boundary condition has error of 1st order velocity and less 1st order pressure tested for wall is located exactly at q=1/2.

Like for velocity_bounceback_qval we compute the values at the boundary points.

Main differences to velocity_bounceback: - implemented "link-wise", only incoming links at the boundary are updated for each time step - uses q-Values in pre-processing (see mus_bc_header_module) - computes state values at the boundary nodes instead of barycentre position - can be an improvement of the bounce back rule for q-values ~= 0.5

This subroutine's interface must match the abstract interface definition boundaryRoutine in bc/mus_bc_header_module.f90 in order to be callable via fnct function pointer.

Arguments