Predicting the effects of environmental parameters on the spatio-temporal distribution of the droplets carrying coronavirus in public transport - A machine learning approach

Mehrdad Mesgarpour; Javad Mohebbi Najm Abad; Rasool Alizadeh; Somchai Wongwises; Mohammad Hossein Doranehgard; Saeed Jowkar; Nader Karimi

doi:10.1016/j.cej.2021.132761

Predicting the effects of environmental parameters on the spatio-temporal distribution of the droplets carrying coronavirus in public transport - A machine learning approach

Chem Eng J. 2022 Feb 15:430:132761. doi: 10.1016/j.cej.2021.132761. Epub 2021 Oct 7.

Authors

Mehrdad Mesgarpour¹, Javad Mohebbi Najm Abad², Rasool Alizadeh³, Somchai Wongwises^{1

4}, Mohammad Hossein Doranehgard⁵, Saeed Jowkar⁶, Nader Karimi^{7

8}

Affiliations

¹ Fluid Mechanics, Thermal Engineering and Multiphase Flow Research Lab. (FUTURE), Department of Mechanical Engineering, Faculty of Engineering, King Mongkut's University of Technology Thonburi (KMUTT), Bangmod, Bangkok 10140, Thailand.
² Department of Computer Engineering, Quchan Branch, Islamic Azad University, Quchan, Iran.
³ Department of Mechanical Engineering, Quchan Branch, Islamic Azad University, Quchan, Iran.
⁴ National Science and Technology Development Agency (NSTDA), Pathum Thani 12120, Thailand.
⁵ Department of Civil and Environmental Engineering, School of Mining and Petroleum Engineering, University of Alberta, Edmonton, Alberta T6G 1H9, Canada.
⁶ Aerospace Engineering Department, Sharif University of Technology 14588-89694, Iran.
⁷ School of Engineering and Materials Science, Queen Mary University of London, London E1 4NS, United Kingdom.
⁸ James Watt School of Engineering, University of Glasgow, Glasgow G12 8QQ, United Kingdom.

Abstract

Human-generated droplets constitute the main route for the transmission of coronavirus. However, the details of such transmission in enclosed environments are yet to be understood. This is because geometrical and environmental parameters can immensely complicate the problem and turn the conventional analyses inefficient. As a remedy, this work develops a predictive tool based on computational fluid dynamics and machine learning to examine the distribution of sneezing droplets in realistic configurations. The time-dependent effects of environmental parameters, including temperature, humidity and ventilation rate, upon the droplets with diameters between 1 and 250 $μ m$ are investigated inside a bus. It is shown that humidity can profoundly affect the droplets distribution, such that 10% increase in relative humidity results in 30% increase in the droplets density at the farthest point from a sneezing passenger. Further, ventilation process is found to feature dual effects on the droplets distribution. Simple increases in the ventilation rate may accelerate the droplets transmission. However, carefully tailored injection of fresh air enhances deposition of droplets on the surfaces and thus reduces their concentration in the bus. Finally, the analysis identifies an optimal range of temperature, humidity and ventilation rate to maintain human comfort while minimising the transmission of droplets.

Keywords: COVID-19; Computational Fluid Dynamics; Droplet Suspension; Droplets distribution; Machine learning; Prediction Models.