Development and validation of a computational fluid dynamics modelling methodology for isolated and urban street canyon configurations using wind tunnel measurements

Abstract

Precise prediction of air quality in a street canyon under diverse conditions could be established through the comprehensive validation of velocity of wind profiles and the concentration distribu- tion of pollutants. In this study, a two-step approach was developed using Computational Fluid Dynamics simulations. The first step involved the validation of wind velocity profiles obtained using wind tunnel experimental measurements of an isolated street canyon discussed in ref. [1], while the second step focused on the validation of dispersion of pollutants from wind tunnel measurements discussed in ref. [2] conducted on isolated and urban street canyons. The wind velocity profiles obtained at five distinct vertical planes between the leeward and windward walls in the wind tunnel study [1] were validated by simulating the 2D cross-section of the entire wind tunnel domain with high accuracies; R 2 values of 0.931–0.986 were obtained across the canyon depth. The concentration distribution of the pollutant in the wind tunnel study [2] were validated for a range of velocities (0.5, 1, 2 and 4 m/s) using both 2D and 3D models. A verification of the Reynolds independent nature of the flow was performed by comparing the wind tunnel and street scale models and suitability of employing K- ε turbulence model with Enhanced Wall Treatment and K- ε Low Reynolds Number Model for the wind tunnel scale, and Standard Wall Functions for the street scale were observed. A 2D simulation of urban street canyon flow representing the whole wind tunnel cross-section in the flow direction was also studied to observe repetitive flow nature and thereby a potential to employ fully developed flow conditions for the same. The urban street canyon flow is established through the means of fully developed periodic flow profiles, which inherently restricts the additional mass sources in the flow domain. The emission scenario in the fully developed flow was captured by means of flow profile mapping at the upwind edge of the leeward building. To estimate the minimum number of downwind canyons required to keep up the fully developed flow profile at the target street canyon, a parameterization of the same was per- formed. Finally, the validation of the concentration profiles was obtained with parameterization of the Schmidt number, and an optimal Schmidt number was obtained in the case of using Realizable K- ε turbulence model. The developed and validated methodology provides a robust and efficient means of modelling air pollution dispersion in the isolated and urban street canyons for future research investigations.