Ventilation efficiency of street canyons with intersections

When looking at domains with several building densities, the resulting average concentrations caused by traffic when unknown, are often linearised between known building densities. To verify if this assumption is correct, several cases of street canyons with intersections, are simulated using the DALESURBAN model. The traffic is represented by a constant emission line source spanning the entire width of the periodic domain. The building density is increased, by increasing the span wise width of the obstacles, while keeping the obstacle height and stream wise obstacle width constant. The aspect ratio of the street canyons is kept constant for all cases and thereby remains in the skimming flow regime. Four span wise obstacle widths are compared in this study and tested under four wind angles ranging from perpendicular to the street canyon to almost parallel to the street canyon. To test the effects of obstacle placement on the spreading of pollutant, the obstacles are placed in an constant and non-constant pattern. For the constant spreading the stream wise streets are all of equal width, and for the non-constant pattern, a wide street is followed by a small street resulting in an alternating pattern. The resulting mean volume-averaged concentrations for the cases above complemented with, mean retention time, ACH and PCH at roof level of the volume show a linear trend for almost all cases. The exceptions in this linear trend are found at the non-perpendicular wind angles. Where at the two lowest wind angles and the highest building density, the pollutant is recirculated in the domain through the span wise periodic boundaries, resulting in a steep increase in mean volume-averaged concentration and mean retention time. Another exception is there for the wind angle just below the perpendicular one, where a large concentration is contained in the cavity at high building density, which is not present at lower building densities.