| Non-structural damage to the facade of tall buildings, particularly those exposed to high wind speeds, is often caused by the impact of wind-borne debris. This dissertation describes the development of a probability-based framework for the analysis of debris trajectory in simulated boundary layer winds and for the prediction of the probability of impact against the vertical facade of these buildings. The work focuses on compact debris, i.e., objects of small dimensions and negligible mass moment of inertia (e.g., roof ballast elements, gravel, "bluff" shingles, etc.) The trajectories were computed for three different wind fields: (i) uniform wind field with constant horizontal velocity and no turbulence, (ii) "sudden" vertical gust superimposed to the uniform wind field and (iii) fully turbulent wind field.;Debris elevations at takeoff, drag coefficient and Tachikawa number were modeled as random parameters to estimate the trajectory and to derive "Iso-probability Impact Contours". These contours describe the impact probability associated with "randomly flying" objects, as they hit against a cladding element located on the vertical facade of a building. Monte Carlo methods were used to estimate the probabilities. Object momentum and angle at impact were also calculated.;This study also proposed and investigated the use of a "Universal Probability Curve", describing the probability of impact for objects against the facade of a benchmark tall building as a function of distance. It was proven that this curve is "universal" and can be used independently of wind velocity.;Moreover, turbulence effects on the trajectory of wind-borne debris were investigated. First, a "Kussner-like" sudden vertical gust model was proposed to account for the influence of local recirculation in the flow around a building. Second, trajectories were estimated in fully turbulent winds in 2D; it accounts for atmospheric turbulence and wind shear effect, by modeling a variable mean velocity with elevation. Synthetically-generated and partially-coherent turbulence time histories were digitally simulated by the wave superposition method.;Simulated trajectory results were compared to experimental data derived both from literature and wind tunnel experiments, conducted at Northeastern University. |
