Simulazione e modellazione della dinamica dello spray turbolento

The present dissertation investigates and models the dispersion/evaporation behaviours of droplets within turbulent jet-spray conditions, employing both high-fidelity simulations as well as low-order modelling approaches. Turbulent sprays are complex multi-phase flows where two or more distinguished phases move together, mutually exchanging mass, momentum, and energy. These complex flows play a crucial role in many industrial applications and in a large variety of natural and environmental processes, whose significance has been further emphasized during the outbreak of the COVID-19 pandemic since respiratory events like sneezing, coughing and speaking are kinds of turbulent sprays laden with pathogen-bearing droplets. Therefore, it is vital to achieve a satisfactory comprehension of the mechanisms governing the process and enhance the model capabilities for applications. To this purpose, the present research was firstly concentrated on the effect of bulk Reynolds number on the evaporation process and clustering of dispersed droplets within a turbulent jet by employing a solver based on a low-Mach number Navier-Stokes equations and the point-droplet approximation for the Lagrangian phase under the Direct Numerical Simulation (DNS) approach. A detailed and systematic analysis was reported in Paper I. Using the DNS data as reference, the model capabilities for Lagrangian droplets evolution under the Large-eddy Simulation (LES) framework by considering the parcel concept has been analyzed (Paper II). Then, the focus was moved to a more practical topic, i.e. virus transmission via respiratory droplets due to the sudden outbreak of the COVID-19 pandemic. Actually, the SARS-CoV-2 virus mainly spreads from an infected individual's mouth or nose when they speak, cough and sneeze, ejecting pathogen-bearing droplets of different sizes, from drops, O(1mm), to small aerosols, O(1μm). To better evaluate the exposure risk related to these respiratory droplets, accurate and computationally intensive LES of turbulent puffs emitted during sneezes in different environmental conditions were performed. The simulations showed how different could be droplet evaporation and virus exposure as a function of environmental conditions (Paper III). An existing numerical parallel in-house code (CYCLON), written in FORTRAN90, has been modified to allow simulations for different scenarios. Inspired by previous findings showing an extended droplet evaporation time with respect to the widely used D2-law model, a revision of the D2-law has been proposed and tested against reference data from DNSs (Paper IV). On the footing of this revised D2-law and an effective correction to the classical Stokes drag to account for droplet inertia as well as considering the latest research on turbulent jets and puffs, an integrated framework able to describe the evaporation-falling-travelling dynamics of respiratory droplets for different environmental conditions and respiratory activities are being put forward and used to assess the effectiveness of physical distancing and face coverings, which was elucidated in Paper V.