M.Tech (Res): Direct Numerical Simulation of Turbulent Aerosol Transport in Human Respiratory Flows

Airborne transmission of respiratory diseases is governed by the coupled dynamics of the exhaled turbulent flow and the transport of pathogen-laden aerosols and droplets. This thesis studies two representative respiratory activities, speech and cough, and quantifies the infection probability using Direct Numerical Simulations. The incompressible Boussinesq Navier-Stokes equations are solved using Megha-5 solver, including the required scalar and dispersed-phase transport equations. Speech flow is studied for a repeated plosive-rich utterance at soft, normal and loud levels. Aerosol exposure is quantified from the instantaneous and time-integrated flux through a face-sized region representing a silent listener at different streamwise separations. Normal and loud speech develop from individual puffs into a jet-like mean flow, whereas soft speech remains comparatively underdeveloped. The centreline exposure increases in the approximate ratio 1:5.4:18.4, governed mainly by the source conditions. At a separation of 1 m, the infection probabilities are approximately 12%, 28% and 63% for soft, normal and loud speech, respectively. For cough flow, the existing variable-timescale model is compared with a variable-number-density closure based on the Direct Quadrature Method of Moments. The latter transports droplet number density and total surface area as Eulerian fields and captures strong spatial non-uniformity, particularly within the toroidal vortex at the puff head. It increases the centreline infection probability by 10-30%, while a laterally displaced breathing zone produces a non-monotonic variation with distance. The results show that infection risk cannot be described by separation distance alone. The framework provides distance-resolved inputs for room-scale and epidemiological transmission models.

Speaker : SUBBA REDDY GARI SHASHANK REDDY

Research Supervisor : Sourabh Suhas Diwan