Abstract:
Bounce-back boundary (BB), the popular wall boundary in lattice Boltzmann method (LBM), corresponds to the no-slip boundary and does not provide an accurate shear drag on walls in some cases. This study discusses a new wall boundary's effectiveness with a wall function in LBM-based large-eddy simulation (LBM-LES) to predict indoor and outdoor flows in the built environment. “Wall-function bounce” boundary (WFB, using Spalding's law) is conducted and compared with BB. Two validation cases of indoor convection in a 9 m × 3 m × 3 m room and of turbulent flow around a single 1:1:2 building are employed. Results show that BB provided a lower shear drag accuracy on walls than Spalding's law in both indoor and outdoor cases, particularly when using coarse grids. WFB compensated for this and yielded a more accurate shear drag. WFB yielded an overall simulation accuracy similar to BB with half-length grids in indoor and outdoor cases and achieved grid independence using a coarser grid resolution. In indoor case, WFB improved the accuracies of both time-averaged and fluctuating velocities in both near-wall and off-wall regions. In outdoor case, LBM-LES with WFB obtained similar time-averaged flow structures to those with finer-grid BB. WFB also improved the accuracies of the time-averaged velocity and turbulent kinetic energy near roof and ground. This study indicates that a boundary with a wall function (e.g., WFB) is important for LBM-LES in built environment flows because it can yield a better simulation accuracy utilizing coarser grids and reduce the demands of computation.
