On the estimation of wind speed in urban canyons for ventilation purposes—Part 2: Using of data driven techniques to calculate the more probable wind speed in urban canyons for low ambient wind speeds

For low ambient wind speeds, airflow in deep urban canyons is characterized by a high scatter and important fluctuation as no coupling is established between the undisturbed wind flow and the flow inside the canyon. Thus, thermal and mechanical forces determine the wind speed characteristics. Existing studies based on experimental comparison have shown that under the above boundary conditions, deterministic models may predict with sufficient accuracy the mean wind speed but not the fluctuation caused by the thermal phenomena. In the present paper, data have been collected through an extensive experimental campaign in seven canyons. Then, data driven techniques, to predict the more probable wind speed in deep urban canyons as a function of the prevailing thermal and inertia phenomena, have been developed. The proposed methodologies are strictly valid inside the limits of the experimental data, i.e. aspect ratios between 1.7 and 3.25, and can be used to estimate the more probable wind speed close to the facades of urban canyons.