Direct numerical simulation of “cough/sneeze flows” to understand transmission dynamics of COVID-19 infections

Team Lead: Sourabh S. Diwan


The transmission dynamics of highly-contagious respiratory diseases like COVID-19 (through coughing/sneezing) is an open problem in the epidemiological studies of such diseases (Bourouiba 2020, JAMA). The primary cause of COVID-19 infections is believed to be droplet transmission from an infected person to a susceptible neighbour. WHO has recommended maintaining a distance of 1-2m from an infected person to minimize transmission to a neighbour. However, recent studies suggest that this could be an under-estimate and that the pathogen is likely to get transported over much longer distances, especially through sneezing (Bourouiba 2020). Thus, a better understanding of the transmission dynamics of the COVID-19 infection is the need of the hour.

The problem is basically the fluid dynamics of a transient turbulent jet/puff with buoyancy, laden with evaporating droplets carrying the pathogen. The objective of the present work is to develop a direct numerical simulation (DNS) code for studying “cough/sneeze flows” by a suitable combination of available DNS codes, and to generate useful data and physical understanding on these flows. We believe simulations of this kind can help devise more accurate guidelines for a safe separation distance between people and design better masks, towards minimizing the spread of respiratory diseases of the COVID-19 type.

There is a strong similarity between the dynamics of cough/sneeze flows and that of atmospheric clouds (especially cumulus clouds), both involving turbulent jet/plume, suspended droplets and their complex interaction including phase change and gravitational settling. We have extensive experience in studying cumulus-cloud flows (involving experiments, theory and computation). We are, therefore, in a position to leverage our expertise in cumulus-cloud computation towards investigating cough/sneeze flows through a DNS. For this purpose, we plan to use the existing DNS code called “MEGHA-5” (developed by S. Ravichandran and others to study cumulus clouds), with suitable modifications to include dynamics of liquid droplets of various sizes.

Current status

We have started making simulations on a “dry cough” (i.e. without including liquid droplets) as a first step towards making more realistic simulations involving droplets. The computational domain is shown schematically in the figure below. The cuboidal domain can be taken to simulate a room or a part of a room, with boundary conditions specified appropriately. The cough flow is initiated through an opening (simulating the mouth) on the left face, as if a person were standing close to the left wall in a room. We plan to make use of the typical parameters for the cough/sneeze flows available in the literature. The temperature difference between the cough fluid and ambient air, as well as the humidity of the ambient air, are treated as parameters.

The simulations are being run on the IISc CRAY XC40 (SahasraT) supercomputer. The plan is to make highly-resolved simulations, typically involving >2 billion grid points and 400,000 core hours for a single simulation run. The preliminary results on the dry cough should start coming in a month’s time. The development of the DNS code incorporating droplet dynamics (including evaporation and gravitational settling) should take about 6-8 months, with the first results expected within a year. The total duration of the project is envisaged to be 2 years. The computational resource anticipated is approximately 6-8 million core hours on CRAY per year.


  • Team lead: Sourabh S. Diwan, Department of Aerospace Engineering, Indian Institute of Science
  • Rohit Singhal, Department of Aerospace Engineering, Indian Institute of Science, Bangalore
  • S. Ravichandran, Nordita, KTH Royal Institute of Technology and Stockholm University, Stockholm
  • Rama Govindarajan, International Centre for Theoretical Sciences, Tata Institute of Fundamental Research, Bangalore
  • Roddam Narasimha, Engineering Mechanics Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore

Scroll Up