Abstract: The correct prediction of air pollutants dispersed in urban areas is of paramount importance to safety, public health and a sustainable environment. Vehicular traffic is one of the main sources of nitrogen oxides (NO ) and particulate matter (PM), strongly related to human morbidity and mortality. In this study, the pollutant level and distribution in a section of one of the main road arteries of Antwerp (Belgium, Europe) are analyzed. The assessment is performed through computational fluid dynamics (CFD), acknowledged as a powerful tool to predict and study dispersion phenomena in complex atmospheric environments. The two main traffic lanes are modeled as emitting sources and the surrounding area is explicitly depicted. A Reynolds-averaged Navier–Stokes (RANS) approach specific for Atmospheric Boundary Layer (ABL) simulations is employed. After a validation on a wind tunnel urban canyon test case, the dispersion within the canopy of two relevant urban pollutants, nitrogen dioxide (NO) and particulate matter with an aerodynamic diameter smaller than 10 m (PM10), is studied. An experimental field campaign led to the availability of wind velocity and direction data, as well as PM10 concentrations in some key locations within the urban canyon. To accurately predict the concentration field, a relevant dispersion parameter, the turbulent Schmidt number, , is prescribed as a locally variable quantity. The pollutant distributions in the area of interest – exhibiting strong heterogeneity – are finally demonstrated, considering one of the most frequent and concerning wind directions. Possible local remedial measures are conceptualized, investigated and implemented and their outcomes are directly compared. A major goal is, by realistically reproducing the district of interest, to identify the locations inside this intricate urban canyon where the pollutants are stagnating and to analyze which solution acts as best mitigation measure. It is demonstrated that removal by electrostatic precipitation (ESP), an active measure, and by enhancing the dilution process through wind catchers, a passive measure, are effective for local pollutant removal in a realistic urban canyon. It is also demonstrated that the applied ABL methodology resolves some well known problems in ABL dispersion modeling.

Abstract: The main purpose of this research is to manage simultaneous measurement of velocity and concentration in large cross-sections by recording and processing images of cloud structures to provide more detailed information for e.g. validation of CFD simulations. Dispersion from an isolated stack in an Atmospheric Boundary Layer (ABL) was chosen as the test case and investigated both experimentally and numerically in a wind tunnel. Large Scale-Particle Image Velocimetry (LS-PIV), which records cloud structures instead of individual particles, was used to obtain the velocity field in a vertical plane. The concentration field was determined by two methods: Aspiration Probe (AP) measurements and Light Scattering Technique (LST). In the latter approach, the same set of images used in the LS-PIV was employed. The test case was also simulated using the CFD solver FLUENT 6.3. Comparison between AP measurements and CFD revealed that there is good agreement when using a turbulent Schmidt number of 0.4. For the LST measurements, a non-linear relation between concentration and light intensity was observed and a hyperbolic-based function is proposed as correction function. After applying this correction function, a close agreement between CFD and LST measurements is obtained. (C) 2009 Elsevier Ltd. All rights reserved.