For those who have read about HOLT’s CFD model and watched our simulations, we have put together an FAQ below that provides answers to the most commonly asked questions.
Why did HOLT Architects choose to do this study of their office?
To reopen our physical office, we wanted to better understand how possible “infected” droplets could be transported within our office, and then consider mitigations for a safer reopening.
Why would an architect do this study and not an epidemiologist, doctor or scientist?
It is common practice for architects and engineers to utilize CFD (Computational Fluid Dynamic method) to design, analyze, and predict the flow of air or wind in or around a building. Our firm always engages a consulting specialist (in this case M/E Engineering’s Syracuse Office) who runs the simulations from virtual building models provided by HOLT. See M/E Engineering’s CFD experience here.
Does the HOLT CFD study relate to other buildings?
This model is specific to our building and our HVAC system. HOLT Architects’ Ithaca office is designed as a Net-Zero facility. The building envelope is very “tight” and a high quantity of the indoor air is recirculated for energy efficiency. Many of the same concepts will apply to other buildings but every building will perform differently.
What methodology was used to perform this study? What assumptions?
Much is yet to be known about the novel corona virus. This simulation has been based on current literature and studies from within the scientific and medical community. Our assumptions are as follows:
- Sneezes, coughs and talking produce different quantities of droplets/aerosols into the air.
- Each droplet distribution follows a Rosin-Rammler distribution (Like log normal). The Rosin-Rammler expression is a mathematical representation of the droplet size for liquid sprays.
- The aerodynamics of the aerosol droplets follow three primary modes of travel: Heavy droplets fall quickly, medium droplets remained suspended in the air for relatively short durations and very small droplets remain suspended for hours.
- In all cases, the CFD model accounts for evaporation, droplet adhesion if a surface is encountered and gravitational forces.
- There is a huge variation between various studies of virus particle concentration in saliva, infectivity, TCID50 (Tissue Culture Infectious Dose for 50%) and HID50.
- The TCID50 that we used is equivalent to approximately 650 RNA copies (virus particles or virions). The HID50 = 3 x TCID50 = 1,950 copies.
- Between 1,000 and 10,000,000 virions fit into aerosolized droplets with diameters between 1 and 10 microns. 200,000,000 can fit into all ranges of droplets from sub-micron up to 1mm in size.
- Somewhere around 30% of individuals are believed to be “speech-superemitters” where they naturally produce more aerosolized droplets than the rest of the population. Studies have shown how about 20% of the population could be responsible for about 80% of the transmission and therefore the modeled assumes a “speech-superemitter”.
- Talking and breathing is assumed to produce an average of around 1,000 droplets/breath at slow velocity, a cough is assumed to produce an average of 3,000 droplets at 50mph, and a sneeze will produce an average of 30,000 droplets at 200mph.
- Breathing produces only about 50 RNA copies of the virus per minute, talking produces about 200 copies per minute, coughing will produce up to 20,000,000 copies per cough and a sneeze will produce up to 200,000,000 virions. Not all virus particles will be infectious and not all droplets will travel more than a few feet from the infected person. Once this is accounted for, the volume fraction of airborne infectious droplets is diminished by 10 to 100 fold.
- The actual HID-50 for this novel corona virus is yet to be determined. The study uses some data from SARS 2003 (a Corona Virus), but mostly Influenza A to approximate our values for COVID19. The assumption is that COVID19 is at least infectious as Flu A. It’s important to note that the values in the video are Percent of HID50; so, a value of 100% is actually the Human Infectious Dose for 50% of the people exposed. A value of 500% would indicate that the exposure is 5X the HID50.
- All of the information and assumptions incorporated into the CFD model are consistent with other works in literature today.
Does the model assume cloth masks or N95 masks?
For the masked simulations we modeled double or triple layer cloth masks that cover both the mouth and nose and do not have a check valve; pseudo-surgical masks like what can be purchased at Home Depot or convenience shops fall into this category also. Depending on the design, they provide anywhere from 50% to 85% effectiveness. Bandanas, single layer cloth or any masks that do not also cover the nose are not very good – 0% to 25% effective, depending on the design. We did not model N-95 masks as most non-medical individuals do not wear N-95 masks.
How about air purifiers? Are they being used?
The model does not assume air purifiers. Unless the air purifier is placed between the sick person and the healthy, in our model it would have little effect in the 10-minute period. The same is true of filters in the HVAC ductwork. The droplets shown in this specific model mostly never reach our office’s return air distribution system. Our third video will look at mitigations and we will discuss strategies like air purifiers.
We have begun to list some articles and other resources that may be of interest:
Controlling COVID-19 Transference through Heating, Ventilation and Air Conditioning Systems
By: Christopher J. Colburn, MEng, CSP, CHMM, Grad IOSH; April 30, 2020
Experts Agree on Aerosol Transmission of COVID-19: How Can Schools Manage Indoor Air to Stop the Spread? https://syracusecoe.syr.edu/stop-the-spread-in-schools/; Webinar- July 22, 2020