Although ventilation effectiveness varies in different sections of indoor spaces owing to imperfect mixing condition, previous studies on cross-ventilation in a model sheltered by urban-like arrays were limited to the computation of boundary-based volume flow rates and air exchange rate. To evaluate the local ventilation efficiency in the cross-ventilation model, we conducted large-eddy simulations (LESs) to generate the velocity and concentration fields of an indoor space surrounded by blocks arranged in a staggered manner to compute local ventilation using the age of air, purging flow rate, and net escape velocity (NEV*). We solved the distributions of seven scalars with corresponding sources positioned at the inlet of the indoor space (flow fluctuation is intermittent), indoor corners (flow stagnation is present), and downwind location, for two window-opening cases: parallel and perpendicular to the flow direction. The results indicated that the location of scalar sources significantly affects the ventilation of the scalar, resulting in different concentration fields, even when the sources are near each other. Additionally, we pioneered the use of NEV* for an unsteady simulation using LES and the decomposition of NEV* into two components: advection (NEVA) and turbulent diffusion (NEVT). The NEVA field appeared to be the velocity field, whereas the NEVT field appeared to be the induced or equivalent velocity field owing to turbulent diffusion. NEVT significantly affected local ventilation and even surpassed the effect of advection in regions with low mean velocities. This study reinforced the use of ventilation indices over boundary-based ventilation rates when analyzing the sheltering effect.
!!!All Science Journal Classification (ASJC) codes