Predicting the airborne microbial transmission via human breath particles using a gated recurrent units neural network

Purpose The purpose of this paper is to devise a tool based on computational fluid dynamics (CFD) and machine learning (ML), for the assessment potential airborne microbial transmission in enclosed spaces. A gated recurrent units neural network (GRU-NN) presented learn predict behaviour droplets expelled through breaths via particle tracking data sets. Design/methodology/approach methodology used investigating how infectious particles that originated one location are transported by air spread throughout room. High-fidelity prediction indoor airflow obtained means an in-house parallel CFD solver, which uses equation Spalart–Allmaras turbulence model. Several flow scenarios considered varying different ventilation conditions source locations. model computing trajectories emitted human breath. numerical results ML training. Findings In work, it shown developed model, GRU-NN, can accurately movement across environment vent operation prove able provide accurate predictions time evolution distribution at locations space. Originality/value This study paves way development efficient reliable tools predicting virus under positions within environment, potentially leading new design. parametric carried out evaluate impact system settings variation breath space considered.