Can restaurants be made safe during the pandemic?

Non-regional discussions Restaurants meta
15

Much of metro Atlanta’s restaurant industry knew immediately that reopening would take time and that there wouldn’t be a rush back in the name of expediency. Hugh Acheson, a well-known Atlanta restaurateur, tweeted on April 21 that “No one tells me when to open.” Using the hashtag #GAHospitality, 50 restaurant owners representing 120 restaurants took out a full-page ad in the Atlanta Journal-Constitution giving their reasons for remaining closed for dine-in service despite being allowed to reopen. As we head into mid-May there have been no major updates on when Atlanta’s dining scene will reopen.

https://www.northwestgeorgianews.com/associated_press/business/conor-sen-georgia-s-fast-reopening-is-going-pretty-slowly/article_58b8b7b2-3f8a-5e5a-9f2d-be6bb2e24710.html

Which confirms this story from a few weeks ago:

16

Sure, there are lots of issues, but social distancing, seating only a portion of seats, servers washing hands even more than usual and wearing masks, and putting up barriers are just theater if recirculating air is spreading virus particles among customers, who obviously can’t wear masks while eating and drinking. And using disposables? That’s anti-scientific nonsense.

It’s also possible that eating outside presents little risk, and perhaps also that so long as there’s good ventilation with fresh air it’s equally safe to eat inside. Not that I’d bet my life on it without more research, but so far the patterns found through contact tracing suggest that virtually all infections have been through indoor air.

17

Duct cleaning and maintenance will put most of them out of business. It is not inexpensive.

18

Duct cleaning is unnecessary since the virus particles don’t stay viable very long. The Hunan bus study makes it clear that in recirculating air they can stay viable at least half an hour, but not much longer (since the last passenger to get infected got on half an hour after the one who was shedding particles got off).

How expensive it would be to eliminate the accumulation of virus particles in recirculated air will depend on the design of the HVAC system. Best case, there’s an option to choose between recirculated and fresh air. Worst case, there’s no exterior intake so it would require remodeling.

19

I’ve made some people seriously furious with me. It’s a cruise board and I’ve had the audacity to post info like the above. These people are willing to die, I suppose, to get back on a cruise ship. And here we are wondering if/when/how we’ll go to a freaking restaurant. Amazing.

20

In the Chang article that you posted here on May 7, people at table B were clearly infected by droplets from A1 being blown by the a/c fan over table A and then over table B. The people at table C may have been infected by droplets that managed to remain airborne on the return path to the a/c unit, or they may have been infected because the return airflow picked up new droplets from A1.

Air that is recirculated in airplanes passes through HEPA filters (I believe two stages) and is quite safe.

But: You don’t want to be downstream of a horizontal air current that is blowing air past an infected person who has coughed, sneezed, or talked loudly. Anywhere. Fans are just as bad as a/c units. Maybe worse. Note that no-one at tables D or E were infected, because there were no air currents from table A to those two tables (or anywhere else on that floor of the restaurant).

TY for having posted the referenced report by Lu, et. all here: Covid-19, and how to support the restaurant industry? - #191 by robert. In the report, one important finding was that there was no virus on the a/c unit. Virus-bearing droplets may have recirculated within the restaurant space, but neither they nor individual virus particles or clusters originated from the a/c unit itself. This is different in an important way from DD’s unfortunate experience with a mold-laden a/c in a test kitchen.

Also, there is a minimum number of virus particles necessary to infect. With bacteria, this is called the minimum number of “colony-forming units” (CFUs). One bacterial spore very rarely or never does it, although in principle it could, because a single spore by its lonely self is in fact a potential CFU. Similarly, while theoretically possible, it’s extremely unlikely to be infected by having inhaled fewer than some (unknown to me) minimum number of virus particles, alone, in solid clusters, or carried in droplets. Droplet/particulate density reduction through careful HVAC engineering, including filtering, are controls to limit CFU intake in an enclosed space.

As an aside with a suggestion, I used to do a great deal of business travel. If someone behind me or next to me on the airplane coughed or sneezed, even if there wasn’t anything “going around”, I would immediately hold my breath for as long as possible. That’s an easy defense to adopt, since coughing and sneezing are the primary mechanisms of covid transmission.

1 Like
21

It doesn’t matter if you’re downstream or upstream from the infected person, only whether you’re sharing recirculated air. Virus particles accumulate in the air over time, as shown by the person who was infected on the bus 30 minutes after the infected person got off. And as you said, the more virus particles, the more infectious the air is.

Asymptomatic carriers are a major source of infections, and they’re not coughing or sneezing.

https://www.nejm.org/doi/full/10.1056/NEJMe2009758

22

Here it is, direct from the State of California and CalOSHA, the guidelines for restaurants to reopen.

I’ve skimmed them and I’m not sure how doable they are for many restaurants. Sorry, Robert, no mention of A/C but there are 2 bullet points pertaining to fresh air and air circulation

1 Like
23

There’s been some mention of restaurant air circulation in restaurants with wall-mounted a/c units, vis-à-vis virus spread. (E.g., a video on CBS news today.) These small a/c units probably have dust filters, but those filters are only intended to trap dust that would interfere with the a/c efficiency, as in a home furnace or central a/c. I think.

I did notice this (brief) comment in the CalOSHA guidelines:

“Consider installing portable high-efficiency air cleaners, upgrading the building’s air filters to the highest efficiency possible, and making other modifications to increase the quantity of outside air and ventilation in all working areas.”

24

Interesting to compare the CA with the WA guidelines:

Especially this:
9: “If the establishment offers table service, create a daily log of all customers and maintain that daily log for 30 days, including telephone/email contact information, and time in. This will facilitate any contact tracing that might need to occur.”

More, here:
https://www.seattletimes.com/life/food-drink/to-reopen-washington-state-restaurants-will-have-to-keep-log-of-customers-to-aid-in-contact-tracing/

25

In the Guangzhou restaurant, only two of the seven sitting at the table in front of the air conditioner got infected, vs. three of the four people at the table at the opposite end of the room.

1583×2048 273 KB

California at the state level has been doing a pretty crappy job so far compared with the SF Bay Area health officers.

26

You might want to take a look at the CBS video, or the report by the researchers that it references (at UC Davis and Univ. Oregon). The report title in part is …“Built Environment Considerations…” The 3D simulation is that of a small 4-table restaurant, not the one in Guangzhou where A1 was the original source and people at three tables were infected. It’s a different situation and therefore an interesting and relevant addition to this kind of information.

The CBS video has the title “How a restaurant’s ventilation system can affect the spread of the coronavirus”.

P.S. My apologies, the infected person in the CBS video was not directly under the a/c unit (although that was the case on the bus).

27

Simulations are only as good as their assumptions, and experts have been making a lot of bad ones. A four-seat restaurant’s not a very common real-world scenario.

The Hunan bus had 13 seats. The infected person was sitting in row 12, and infected passengers in rows 5, 6, 8, 9, 11, and 13.

Do you have a cite for where the air conditioner vents were on the bus? I haven’t seen that.

28

Here’s a preprint of a second study of the Guangzhou restaurant. Excellent 3-D simulation on page 10.

1 Like
29

Very good point, Robert. No I don’t. I looked at the diagram of the bus on the link you posted (repeated here) and assumed that the a/c was in the very back of the bus, which would be pretty much right above the infected person in row 12.

30

Outstanding! This, in my opinion, is an excellent example of a truly scientific study, incorporating both 3D CFD (including a widely-accepted turbulence model), and experimental data collection in a very carefully-constructed full-scale physical simulation; in conjunction with thorough analysis and comparison of both. Their concluding remarks and discussion are very clearly written and direct.

Transmission of the covid due to the airflow patterns in restaurants is something we very much need to fully understand right now. IMO, this should be required reading by all restaurant owners and designers of their ventilation systems. Retrofitting the HVAC systems in restaurants as needed should be at least as high a priority as many of the other 10,000 steps that have been government-mandated to limit transmission in CA restaurants so that they can re-open.

31

Return vents are often at the back of the bus, but supply vents are usually toward the front or spaced throughout the bus. Sometimes they’re over every seat, like on a plane.

32

OK, so if someone who’s infected is seated near the return vent in the back of a bus, particles/droplets could be drawn into the return and then spread throughout the bus, or deposit residue on the surfaces of air ducts for later spread.

That was originally my thinking when I said, “don’t sit near an air return” (in a restaurant). If you’re infected, you could infect others.

33

It doesn’t matter where in the cloud of recirculating virus particles you sit.

34

I can’t agree with that, but I’m willing to agree to disagree.

If you’re infected and sneeze directly under a wall-mounted horizontal a/c return vent, I think the probability of transmitting particles through the ventilation system is much greater than if you’re further away, because in the latter case, turbulent diffusion (i.e., decreased concentration) will have diluted the cloud as it expands, before reaching the a/c return.

Turbulence plays a major role in dispersing/diffusing droplets and particles – i.e., spreading them out and reducing their overall volumetric concentration. The large-scale (room-size) source-generated vortices in Guangzhou established the largest-scale, primary vortex structures, but those have to have induced secondary, smaller vortices due to shear between them, and then on and so forth, until the vortical eddy structures finally got so small (near-microscopic) that they were kinematically stopped and dissipated by aerodynamic drag, a process referred to as cascading. This is an ongoing process in a “steady-state” flow; that is, one in which a general “picture” of the flowfield doesn’t change with passing time.

In steady state, turbulence production and dissipation balance, and all scales of hydrodynamic mixing proceed at established, constant rates, progressing in scale and kinetic energy content through the turbulence cascade process. The Guangzhou restaurant airflow wasn’t in a true steady state condition, but largely so over time.

Below some concentration level (and exposure time), even if you’re right inside the cloud you won’t get infected. Which is why the servers didn’t get infected in the Guangzhou restaurant. The amount of virus that gets into your lungs is a function of the concentration of particles in the air you breath; your rate of breathing; and the amount of time you’re exposed; i.e., the overall dose of infectant. The servers weren’t exposed to as much nor for as long as those who were dining in the area of TA, TB, and TC.