Experimental validation of a computational fluid dynamics code to predict the wind speed in street canyons for passive cooling purposes

Natural ventilation is one of the most efficient passive cooling techniques for buildings. Knowledge of the wind speed in street canyons is the necessary condition for the application of such a technique in dense urban configurations. Thus, prediction techniques to evaluate the microclimate and dispersion parameters in street canyons, has become a subject of intense scientific research in recent years. In most of cases, wind flow and pollutant dispersion characteristics have been studied numerically and experimentally with a view to provide an insight in urban dispersion. The microscale model MIMO was employed in order to perform a three-dimensional modelling of the wind field within three typical deep street canyons, located in the centre of Athens. Computational results were compared to field data collected during consecutive three-day experimental campaigns that took place in the summer period. Results from the computations have shown that the wind field in urban areas is quite complex, presenting areas of very low wind speeds and convergence of vortices. The model underestimated the measured wind speed intensities, which may be partly explained by the uncertainty of specific input parameters, the necessary simplifications for the application of such models and finally the geometrical complexity of the area modelled. Finally, computations were performed for a reference velocity of 2 m/s, which is considered to be a threshold value. The wind field developed departed from the one observed in the previous cases. 2005 Elsevier Ltd. All rights reserved.