Reducing simulation time of Covid spread events from weeks to hours
Reducing simulation time of Covid spread events from weeks to hours lead image
The Covid-19 pandemic taught many people the importance of covering their faces and maintaining distance from others to reduce the spread of infection. However, these measures only help against short-range transmission avenues, which are not the only ways viruses can reach new victims. Small, virus-laden droplets can linger in the air for hours, slowly drifting from one side of a room to another upon currents of air.
Determining where these droplets end up requires simulating their movement throughout their lifetimes, a computationally difficult task. Xiang et al. employed a reduced order modeling (ROM) approach to cut down on processing time.
“A simulation that would otherwise take weeks can be done in an hour or so,” said author Cheol Lee. “And this weeklong simulation was done on a high-end workstation with multicore processors while our ROM approach was simulated on a laptop.”
The team employed an Arnoldi-type algorithm to project a computational fluid dynamics solution onto a Krylov subspace, reducing computational complexity without sacrificing accuracy. The result allows simulations to be quickly modified to explore the effects of conditions like ventilation, room layout, and differences between coughing and sneezing.
While this technique was developed to model virus spread, its general nature makes it widely relevant.
“The approach pursued in this paper can be applied to other fields, like thermal management or heat transfer problems,” said author Oleg Zikanov. “Anything that can be described by linear transport equations can be approached by this model.”
The authors hope to expand their technique to incorporate nonlinear effects and more complex phenomena involving turbulent flow.
Source: “Reduced order modeling of transport of infectious aerosols in ventilated rooms,” by Linyan Xiang, Cheol W. Lee, Oleg Zikanov, Mohamed Abuhegazy, and Svetlana V. Poroseva, Physics of Fluids (2023). The article can be accessed at https://doi.org/10.1063/5.0158941 .
This paper is part of the Flow and the Virus Collection, learn more here .
