For turbulent flows, this keyword may be used to invoke wall function boundary conditions on viscous walls, using the White-Christoph law of the wall. Wall function boundary conditions allow calculations to be performed with fewer grid points, and generally with higher CFL numbers.

A wall function boundary must lie on one of the boundaries of the selected zone, and may be an overlapped or internal boundary. The first grid point adjacent to the boundary must be within the log layer (roughly 15 < y⁺ < 100). The recommended y⁺ value at the first point off the wall is about 50. The boundary condition will revert to a no-slip condition when the first grid point off the wall falls below y⁺ = 15, indicating that it is within the laminar sublayer.

Note that the wall function boundary condition is used to eliminate the grid points that would otherwise be required to resolve the laminar sublayer and the logarithmic layer. The grid beyond the first point off the wall should be the same as for a run without wall functions, in order to properly resolve the rest of the boundary layer. For structured grids, the cfsubset utility may be useful when modifying an existing grid for use with wall functions.

The implementation of wall functions in Wind-US involves only the modification of the wall flux, and does not reset the values of any flow quantities.

Viscous forces on wall function boundaries calculated using the LOADS keyword should only be used for determining convergence. For structured grids, correct values can be calculated from the .cfl file using the integrate force command in CFPOST.

Based on limited test cases, the use of wall functions is not recommended if accurate predictions of wall heat transfer are required, or for flows with strong shock waves. The default setting is wall function off.

See Also: DEBUG 35