Optimal ventilation strategy for multi-bed hospital inpatient wards: CFD simulations using a genetic algorithm

Abstract

Optimising ventilation strategy for an indoor environment necessitates systematically evaluating the influence of a diverse combination of physical and operational parameters in the design space. This study proposes a methodology that couples an evolutionary algorithm (genetic algorithm) with an evaluation mechanism (computational fluid dynamics) to determine the optimal ventilation strategy for an inpatient ward. The traditional approach would exhaustively simulate numerous scenarios to identify the optimal combination of parameters meeting the design objective. The proposed methodology would iteratively evaluate diverse design solutions with fewer CFD simulations than the traditional approach. The results of design space exploration suggest that design parameters, namely, location of the infected patient; air change rate; flow rate through local exhaust grille; and number, location and size of supply air diffuser and local air exhaust grille, are critical in minimising the risk of cross-infection caused through contact transmission in a ward.