A sound-driven digital twin for reducing passengers’ exposure to exhaled bioaerosols in an aircraft cabin

Abstract

Airborne transmission of exhaled bioaerosols poses a significant health risk in enclosed environments such as aircraft cabins, where traditional steady-state and fixed ventilation systems often fail to respond effectively to bioaerosol exhalation events by index passengers. This study introduces a digital twin control system based on real-time sound recognition of coughs to dynamically mitigate passenger-to-passenger bioaerosol transport in an aircraft cabin mockup. The system utilized acoustic sensors distributed throughout the cabin to detect cough events, which were considered as one of the indicators for potential bioaerosol exhalation. Machine learning models were employed to classify and localize these events, serving as input signals for the digital twin framework. To respond to the detected coughs, the system accessed a precomputed database of ventilation strategies derived from computational fluid dynamics (CFD) simulations. These ventilation strategies adjusted the supply air velocity and direction locally to accelerate the removal of bioaerosols exhaled by the index passenger. Experimental validation was conducted in a full-scale seven-row aircraft cabin mockup. The results demonstrated that the sound-driven digital twin dynamic ventilation control system achieved over 80 % reduction in particle concentration in the passengers’ breathing zones, without increasing the total ventilation rate or compromising thermal comfort. The proposed system represented a real-time and event-driven solution for effective infection control in aircraft cabin environments. Since the system does not distinguish between coughs produced by healthy and infected individuals, false-positive triggers of ventilation control are expected to occur in real applications. Future work should address this limitation by integrating multiple indicators.