The Bizarre Way We Track Fake Virus Particles That Shouldn't Work
ฝัง
- เผยแพร่เมื่อ 5 ก.ย. 2024
- The Lattice Boltzmann Method of modelling fluid flow (computational fluid dynamics) sounds like it couldn't possibly work but it has amazing predictive power. Coupled with Lagrangian Particle Tracking, it's a clever way to figure out how virus particles move around a building.
This video is paid for by Dassault Systèmes who are experts in computer modelling and have helped hospitals to reduce the spread of coronavirus using these methods. Find out more here:
go.3ds.com/Dass...
You can buy my books here:
stevemould.com...
You can support me on Patreon here:
/ stevemould
just like these amazing people:
Matthew Cocke
Glenn Watson
Joseph Rocca
Joël van der Loo
Doug Peterson
Rashid Al M
Paul Warelis
Will Ackerly
Heather Liu
Twitter: / moulds
Instagram: / stevemouldscience
Facebook: / stevemouldscience
Buy nerdy maths things: mathsgear.co.uk
I don't know who you are. I don't know what you want...
Chills mate.. chills
You've Taken my breath away
I just want to watch this video...
... so bugger off,,, whoops sorry!
I first understood Emmanuel Bernoulli and just thoght .. ofcourse he knows with fluid dynamics
Before I saw his name printed on screen, I heard "Emmanuel Bernoulli", and I thought, "well, no wonder he's good at this stuff!!" :))
Me too!
en.m.wikipedia.org/wiki/Nominative_determinism
Haha yeah
well we know that at a point the flow goes over to a turbulence. Very interesting indeed.
If only! Love a good niche joke.
"Imagine it's a spreadsheet. It probably isn't..." Um, you've known Matt Parker for _how_ long?
Matt appears in the example spreadsheet from the video 5:43
"Matt Parker Loves Tau" XD
@@tarkanal-kazily4950 So it seems. And spreadsheets don’t lie...
@@tarkanal-kazily4950 The Letter of the Greek alphabet or the Movie though?
@@Thermalions The measure of 2π(the mathematical constant) is specifically being referred to here.
I first heard you say "Vanoli" as "Bernoulli" and thought that was a suspiciously serendipitious name for an expert in fluid dynamics.
Same.
Same!
@Mark Smileer bro fucking lmao. There is entirely too much wrong with your run on sentence for me to even attempt to break it all down for you. All i can say is stop getting your news from your crazy aunt rita on facebook, and start getting your news from legit news sources, including international news sources if you can even receive the signals for them with that tin foil hat on ;p
@Mark Smileer bro, no one cares!
@Mark Smileer wow, that's a lot of words.
"All models are wrong. Some models are useful."
-My professor on biological modeling
"Assume the cow is a sphere."
Not your biological modeling professor: en.wikipedia.org/wiki/All_models_are_wrong
@Mark Smileer Are you trying to go for a lie/bullshit/misinformation speed run or something?
There is so many factually incorrect things in there that I can only sum up the response as "lol no".
@Mark Smileer Whoa buddy, I think you need to take a break from the internet. From the sketchy corners at the very least!
@Mark Smileer I genuinely feel sorry for you. You're either completely gullible or you've been hurt so severely by something that you cling to anything that seems to give you a definite answer in an inherently scary and imprecise world.
That is a worryingly good Liam Neeson impression.
That's just because he's really just Liam Neeson wearing a mask. Steve Mould doesn't exist.
Actually, it is terrible, really terrible. Really don't do it again, smh. You're friends obviously like you too much to tell you it sucks.
th-cam.com/video/jZOywn1qArI/w-d-xo.html
@Mark Smileer [CITATION NEEDED]
If this TH-cam thing doesn't work out for him, he could fall back on acting in action movies.
So did anyone measure the actual airflow at sample points in the hospital to check the accuracy of the model?
You said that way more succinctly than me. Had the same question! And, to be clear, I'm rooting for the model to be useful. Is important to check though.
It's amazing how often the experts forget this simple point
At least in science, yes, you would test your model and calibrate the model with the results. Not a fluid modeler myself, but worked with them on several occasions.
They don't mention it, but they probably validated the model with smoke or something.
They may have but the results were so abysmal that they cut it out. Unsteady aero is practically uncharted territory in CFD if strouhal number isn't high enough to cancel the convective term your model will probably suck.
I think the lattice model needs to use hexagons, because they are the bestagons.
icosahedra in 3D ftw!
The 3D lattice model uses hexahedra, close enough?
@@mm1979dk
Nah, we need some kind of plesiohedron.
Grey
Indeed, hexagons ARE the bestagons.
The research institute i worked at had 2 corridors with rooms in between in a building, the clean corridor was slightly positive pressure, the room was neutral, the dirty corridor was negative, this meant air could only go from clean to dirty
(clean and dirty related to diseases)
@Mark Smileer where'd you get that copypasta?
@Mark Smileer I used to work with someone like you
I’m happier now
@Mark Smileer no, he was a religious and conspiratorial nut, who (also) hid behind a Facebook page to publicly harass his former coworkers
@Mark Smileer and FYI, I still have my job
@Mark Smileer
Supposedly, China isolated the virus when they released the information on its DNA in the early months of last year.
The PCR test can theoretically detect anything with DNA, at the minutest level, I don’t know what other tests are available for detecting a virus but it’s entirely plausible that’s the only available test in many area
PCR is common when reading DNA
Finding a vaccine? The vaccines are all genetically modified. If they can read the DNA using PCR, then they can design an mRNA vaccine, or in Oxford, modify a monkey adenovirus vaccine.
And yeah, running more cycles means more false positives, but at a certain threshold, like containing a new virus that we barely understand and barely know how to treat, it’s better to err on the side of caution. At least in some countries, erring on the side of caution allowed authorities to conduct contact tracing and get the virus under control
In college I did some work with modeling wildlife populations. The rule that all models follow: you don't need to be accurate, just accurate enough to be useful. The useful metric is determined by running the model and comparing the result to historic/real world data. The more dynamic the system, the worse all models are at peering into the future.
In this case, just seeing what its like in real world would have been a lot quicker and correct(er). I dont even see what the model is supposed to be for in this case. It might not even be correct, no word about validation was lost and thus I doubt they validated it. People move around, structures are not perfect as in the very simple model (all sorts of things are in the rooms, not just walls).
All in all, this seems like "ah this looks good" but it has either no real value or only by chance happens to work out in the real world.
Im saying this as someone modelling system like complex heat exchangers, chemical reactions or whatnot. I have seen models that were not validated (correctly) - and they are just trash.
@@leocurious9919 hardest thing to model is the validity of your model 🤔
@@leocurious9919 So, you're saying opening these two windows, which our best effort at the moment say would be helpful in limiting spread, is too much hassle in the end? [Yes, it is a rhetorical question]
@Mark Smileer wow thered be a lot to unpack here if any of it actually made a lick of sense lmao
@Mark Smileer dont bring poor mr freeman into this. Just cause hes famous and beloved doesnt mean hes right or that intelligence/education has anything to do with delusion.
Like comedian Pete Holmes says, the best way to do a good impression is to insist that it was good
Or a gun
@@tedarcher9120 Or a dumb haha lol
Now if only Pete Holmes could do an impression of a good comedian.
@@Flabulo that's like asking Seinfeld to do an impression of himself
@@doggfite wHaTs ThE dEaL
Hey I've seen a lot of interesting videos from you.
As a German with only basic English knowledge i can say:
You are one of the best international artists on TH-cam. I love scientific content from a lot of TH-cam producers, but you manage to explain that I understand every word without a lot of pictures. Even though it's a foreign language for me. Because of your videos, I first dared to watch English content and now, after years, I no longer shy away from seeing every video in English.
Thank you for the good content and the great pronunciation so that people who speak other languages understand it well.
Hey Guten abend Volker Yes i understand your problem, i have that problem but the other way around! i am an English guy living in Germany, what i use to do is to watch a film on a DVD in German but with the untertitle turned on in German! so i could read and listen!
if i just listen it was always to quick to hear and understand, maybe i got 40% but if i read and listen ! i understand up to 90% or more.
Schönes abend
While it's true that my sample size is pretty small, I've never met a German who doesn't have a good grasp of the English language - and better than they believe they have in almost every case. As a monolingual American I'm always impressed by this. Kudos to your modern culture for bringing into being an education system that creates this result. Love from Texas 🤠👍🇩🇪
@@Jesse__H Bretter than they believe, but not good enough!!! Writing and speaking is so much more difficult than listening and understanding! Without a translator, I have no chance of saying what I want to say.....
I take your text as a compliment, thank you very much.
@@mirage4014 hey Julian.
my Real Name is also Julian! :)
Try hear in German and read the undertitle in English.
If you want to speak in German, you can’t do anything with the letters, you have to hear how it is spoken.
That's just a tip from me.
I wish you every success and all the best.
Schönen Abend noch ;)
That is a very lovely and motivating comment! Have a wonderful day! ☺️
We used a teen's vape to visualise airflows around an external supermarket queue with a 90 degree bend in it - proving 50% of all possible wind directions can cause particulate transfer, and you're never too old to be barred from Waitrose. Old uk duffer here :)
Yes I could see vape being excellent for visualizing airflow.
@Heather Petersen I love how 9/11 is completely off topic, its an art to be so... disconnected
"Emmanuel Vanoli has a very particular set of skills... skills that he acquired using today's sponsor, Skillshare!"
And he practices them every day by solving challenges on Brilliant.
"And when he's done learning, he'll play Raid Shadow Legends!"
@@mayabartolabac and once he inevitably deletes that he'll want to get his friends to talk to him again, and for that, he'll need new clothes, which he can get a discount on simply by using Honey!
And if that someday happens for real I probably will immediately get on with it :D
When someone tells you marketing doesn't work, show him this thread
Air curtains are underutilized.
One of the easiest ways to isolate an area with very little construction.
Starbucks loves them.
@@x--. glad they do it helps prevent insects from entering the store.
most retailers use them, besides countering insects, using positive air pressure in a vestibule also helps lessen heat loss from doors that open automatically, because the pressure of the air in the vestibule is higher than both the warm air in the store itself and the colder air outside (or vice versa during the hot months)
Very useful, including in saving energy in commercial temperature regulations
I suspect for hospitals they want better isolation than that. Like, they go for full-on negative pressure, or nothing at all. But anything is better than nothing!
Come to think of it... I've never seen Steve and Liam Neeson pictured together. Are we sure they're two different people?
You make a good point
If they enter the same room they both exit via the two sides
@Mark Smileer @Mark Smileer I'm sorry, could you please use punctuation in your comment, so that your run-on sentence doesn't read like the ramblings of a madman?
Actually, don't bother. I don't think the world needs to know how 9/11 somehow correlates to Covid-19.
@@autumnhd he has copied and pasted that on EVERY COMMENT.
when working on the design of a hospital many years ago. there is a lot of thought put into which spaces are positively and negatively presurised. it's not as simple as spaces only having extraction. sany sensitive spaces like an operating theatre, ICU etc are positively presurised, so they dont get any nasties into them, things will tend to flow outwards, quarantine wards would be the reverse. the challenge is often getting spaces next to each other that work well like this, you cant have every room negatively pressurised, so it works really well to alternate them. The surprising thing here is having opening windows, you wont see that on many newer hospitals, precisely because it can screw a lot of this up.
Yeah I was wondering about that. The HVAC system would definitely have fans bringing in fresh air and recirculating conditioned air
Yee. This video is definitely not about a quality, modern hospital, built within the last 30 years.
As a game developer who makes a lot of particle systems, I learnt some cool new stuff watching this, so really appreciate the content.
Steve you can really explain complex topics to the common people. Thank you for making every topic interesting.
Thank you!
@@SteveMould are you a physicist by training?
@@ericl8743 I have a degree in physics
@@SteveMould I love your videos. They're always presented brilliantly!
So... fluid flow is basically like a big game of chess... where it's just Kings... but they can occupy the same square? sometimes.
it's nothing like chess. This is a dreadful analogy
Hikaru be like staring at the ceiling, drawing thirty arrows and then going "Yeah but if I don't open the window then the virus is winning because after takes takes takes takes then I get checked and virus h1 is a clearly winning position... Yeah blunder"
Agadmator be like "And it was in this position that humanity resigned... The virus is simply that good."
sometimes.. most of the times... once in a blue moon it does.
D9n't forget that they all move at the same time...
@@Moinsdeuxcat hahaha this made me spit out my drink
@@Moinsdeuxcat "Beth Harmon, Hikaru Nakamura, and Antonio Radić walk into a hospital ward..." is such a bizarre premise, and nobody asked for such a specific amalgam of chess references, but you somehow made it work
Hahaha man 0:41 exactly what I was thinking when you said "particular set of skills"
And then in a Linus voice: "speaking of particular set of skills, Skillshare!"
skills to pay the bills
Particle-lar set of skills
@Mark Smileer ok mate
@Mark Smileer stop posting the same reply to every comment, no one wants to read it
My dad was in a neg-pressure room in hospital which were positioned at the end of hallways. They replaced one window in each room with plywood and a circular vent connection and placed the portable A/C unit's exhaust air hose but used it only in fan mode to pump air outside so that room air would not seep back out into the hallway. It was a cheap and easy fix to protect the staff and limit spread.
Fun fact: The lattice at 8:55 works like pixels in an image or monitor. This allows developers to abuse the existing drawing functions in graphics cards to simulate fluids & gases using the same technology used for Textures in 2D/3D graphics. Each new image you draw is essentially a new time step in the simulation. The colors of the R G B pixels can be (ab-)used to represent 3d coordinates for position or speed
Steve: "and they have a chance to be convected upwards, away from the people around you"
Overhead fan: are you sure that's a good thing?
Depends on which way it's spinning.
"Fallout"
Fluid dynamics is the dark sorcery of engineering.
Thank you for a peek into the grimoire!
*Steve:* _"How many lives could you save....."_
*My idiot brain for no reason:* _"Twelve."_
"Three, take it or leave it."
@Mark Smileer Excuse me WHAT?????
5:43 Matt Parker Loves Tau lol
"The artist formerly known as particle"
Is that what they call that guy with the peculiar skill set?
His calling sign is just a point to represent the infinitesimal size of atomic and subatomic particles.
Hi Steve, I just wanted to thank you. I received and still receving a lot of kind messages from your community since this video!
Thank you so much! If someone is interested, we published some papers in Scientic Report & Environment International on this very topic!
Happy new year to you and your community & stay safe
Before covid outbreak: How to put spoilers on my car.
During covid outbreak: How to put spoilers on my mask.
"That will be the final word of the video; You saying that.", he says, stealing the final word.
with a knowing smirk
Have been watching and liking your channel for years. This is the first time you are touching fields that are surprisingly close to my heart: I did tons of Lattice Boltzmann simulations coupled with Lagrangian Particle Tracking in my PhD thesis, and I now actually work for Dassault System (even though it is in the Electromagnetism department). You did an amazing job explaining these issues, and that is high praise from someone who worked "in the field"...
"It's actually really good."
Classic phrase of an expert mimic.
I worked for a year as a tech for a company that installed special filters for clean rooms. After installation, we would balance the air thus so when entering an air lock between the dirty and clean spaces, the air would pull in and not escape from the dirty spaces. Fascinating physics and fun models
The man's a genius.
He got a hospital to pay him for the expert advice of "Open the window"
the true question was to open the right ones. Are you sure you watched the video? :)
Your both correct guys
@@cockatieltime2259 Well, they are both right, but only one was funny...
Steve has interest in all fields n thats what makes him my favourite.
How robust is the solution though? If someone has a fever, or turns an extra computer on, will it break everything?
the particular behavior of the particular particle is going to be different everytime in every single simulation.
However, there are hundreds of thousands of particles and thus altogether they will exhibit the average behavior which is quite certain, one can even derive timing distribution to predict how likely those particles will hit certain location and when.
I do expect they've varied the parameters, such as the number of patients, heat sources, etc. in order to find general trends
This kind of modeling has very little to do with things that actually happen in the real world.
@IUIUI You're talking about uncertainty from repeating exactly the simulation but with different stochastic factors.
I'm talking about the stability of the solution when changing the operating settings. In my experience that's the difficult thing to get from a numeric simulation, and it isn't addressed at all in the video. I hope that they did what @EcceJack suggested, but there's no sign that they did so.
Those factors wouldnt change pressure at room scale so the pressure differentials would still work. Here the modelling is used for visualisation imho the same thing could have been done writing out the pressures. Having the computer visualisation probably allows faster visualisation and helped the engineers see what factors are important, what assumptions to make. This is probably accompanied by rough calculations on the back of an envelope. The computer models are easier to communicate and seem more convincing.
I found it particularly amazing how probability in the fluid simulation represents real world fluid physics so well because it seems to reinforce the notion that comes from quantum physics that everything is at its core based on probability. We are random chance. Crazy to think about.
Well in this case the actual underlying physics are not governed by inherent randomness, just high nonlinearity and complexity. Therefore it is usually easier to just average out some effects, such as in RANS and LBM.
That simulation reminded me of the Ph D my old Industrial Placement Tutor was doing in the 80s, but he was trying to control fires in copper mines, to allow some attempt at evacuations. I don't know if he ever completed it. He was using a BBC B to preprocess the matrix work, then loading the data into the mainframe
The great thing about Steve Mould is the very diverse range of topics he covers and the fact that he somehow manages to explain complex science and engineering in a way that non-specialists can understand. My own field is electronics and I've watched a few of his videos on electronics and they're always spot on.
Steve, thanks for making fascinating content that I can use to engage my 8-year-old son in the sciences. Thanks for being a science communicator.
When I saw the lattice and the particle with the 8 arrows, I immediately thought of Mine Sweeper.
Just a note: the most common CFD applications use the Finite Volume Method for discretization (and I'd say the great majority of simulations aren't even dealing with transient problems). The Lattice Bolzmann Method, though cool and powerful (and currently a little hyped) has quite a few limitations.
At the end of the day CFD is still a decent guess, at best. It is quite useful and, if you use the correctly, can have a huge impact in several areas (assuming proper verification and validation of your models isn't skipped, which I'd say is a problem in most cases you see). And I make a living out of it, so I guess I can talk sh*t about it every now and then.
But like all computer simulations EVER, people tend to think all colorful images are automatically correct, and oh boy they are not.
This type of air management becomes far more critical when viral load is factored into the equation. Managing airflow in a pandemic saves lives by reducing new infectios but more so in the severely infected by reducing viral load. Putting sick people together just increases deaths. No govts consider these dynamics & few humans think this deep so thanks for surfacing this content.
The Lattice Boltzmann Method bares some striking resemblance to quantum field theory.
I was thinking the exact same thing!
You mean how it's exactly the same? Yeah I noticed that too.
Well, Boltzmanns work was foundational to a lot of modern physics
Steve mate, isn't it Das- oh as in Ren-oh? Delt a lot with Das-oh Falcon 20s in me time. Great vid.
Thought it was too!
I came looking for this exact comment.
Yes you're right!
Est-ce que tu peux parler français ? J'ai vu que tu as travaillé avec cet entreprise
Yep, "daso" / "das-oh" / "da-sow", etc.. Definitely not "deh-salt".
Steve, your content just keeps getting better and better.
Thank you!
@@SteveMould Thank you for the effort. Not only does it show in the quality, but it helps me learn effectively, engages me in deeper ways, and makes my brain wrinklier.
I worked using heat introduction into structures for pest control for years and became quite good at determining beforehand what windows to open, which to close, which rooms or areas should have fans placed and in which direction, how to get the fastest and most even proliferation and distribution of the heat really. It was important to reach temperature in a relatively simultaneous time to avoid giving the insect time to leave the areas or hunker down beneath something that provided them higher insulation. It was not an intuitive process and I found it was remarkably difficult to teach. It's one thing to react to the temperature readouts throughout the structure and then, through trial and error, make adjustments to see what works, and another to do so beforehand. This video was an excellent explanation as to why and also what to look (or feel) for. Thank you!
What can we say about that Liam Neeson impression?
Do you have any validation studies for this work? What factors were considered, and which models were used for the viral particles? Were they just a simple scalar? Was the effect of outside wind/pressure variations considered?
Was any of this validated against experiment? How do you know that you haven't fallen into the usual CFD trap of garbage in, garbage out (and also colourful pictures to wave at the uninformed out)?
I just found your channel. Really enjoying it. I appreciate the way that you don't talk down your audience like some of the more saucy channels out there. One thing I notice that other channels do to help keep attention that you don't do id to keep changing the set throughout the video. Your background gets boring to look at after a while. Just filming in other rooms or your garden would help I think. Just my two cents
Particle McParticle Face, yes of course one of my favourite airborne particles 👍
I dont know why I'm here and I have no idea what your talking about but, I'm completely invested in this video.
I've found the last part with the convection through your body heat very interesting. Is there less convection in the winter, when we are thermally isolated (clothing) then in the summer? Or is the difference in temperature high enough? And could it provide to the higher spreading of the common cold during winter times?
How cold would your clothes be? The rate of convection depends, among many many many other things, the difference in temperature as you said. So if the clothes are similar temperature to the air then convection would be lower. And if it's a warm day (like 30°C) and you have light clothing on, then the difference is very low too and convection would be low.
5:42 I LOVE that particles 18 & 19 are called "Particle McParticle Face" & "The artist formally known as particle"
18 is "Matt Parker loves Tau"
Wouldn't "closing the wrong window" mean it should have been open?
Yes.
It's important to open the right windows, but also to keep them open ;)
I really enjoyed this little slice of expertise from a field I am unlikely to ever be involved in. But even if I'm never directly involved in modeling airflow through a building, it could very well prove useful to understand how such complex dynamics are simplified into models that can be simulated, and therefore provide utility.
This is amazing! There’s just something about the physics of flow, whether it’s in the form of heat, liquids or gases etc. that’s just so satisfying. Is this reason enough to finally decide what region of physics I wish to pursue as a career? I think so
I just realized that code I wrote at ~16 is exactly like the particle simulator shown here. Nothing like these videos was available to me then, so I _never_ put together that they were the same...and today those kinds of connections are _easy_. I hope this gets in front of more eyes.
Very Cool! And a good description of how to make a simulation... must have taken hours of compute time to create those animations... This is such a complex thing to model, because you have so many confounders such as people moving in and around the building, they will open doors/windows at various times, outdoor weather patterns, maybe barometric pressure and temperature changing atmospheric density... all those things perturbing the air flow, but don't get me wrong... it's really good work, just super hard to nail down the real-world! and finally what if Air flow is not the only vector this virus spreads? The simulation can only account for what we assume are the relevant facts.
I used to work with a team that made a radiation modeling program and in training they talked about figuring out which details were relevant to what you're modeling. They joked that if they added a helix shape to the geometry library users would insist on putting the threads on the screws.
@@adamsbja LOL, Yep that sounds eerily similar... there's always one of those "users" in every crowd... mostly I see that kind of stuff coming from the marketing department, and so I'm working in the engineering team and we see the shifting project requirements: the dreaded feature creep!!! I'd say to the team: Buckle up guys... the ride's going to get a bit bumpy!
@Mark Smileer A couple of things:
1) the point my instructors were making was that the screws were provably irrelevant but people would insist on every detail just in case.
2) use of past tense does not necessarily mean recent past. "It only took 20 years" is wrong. I'm old.
I'm skeptical on the accuracy of a lot of these particle simulations. This video goes into a lot of specifics, but it really just scratches the surface. For example it doesn't appear that this simulation considers that people are breathing in. It seems like it's assuming that these sick patients are introducing more air (more fluid) into the system. However they are not, they are taking in just as much fluid as they are (forcefully) expelling. It's also not considering how doctors and nurses and patients are "stirring" this fluid as they walk around the hospital. It's not considering curtains or the central air system either.
That said, I really like the closing of the video though where it explains wearing a mask keeps your particles close to you which limits their spread. (Kind of reflecting my skepticism of the macro model.) Something else I wonder about mask usage is even though much of the air your breath out is ejected through the sides of the mask, the path of the particles out the sides is quite turbulent and chaotic. Much of these particles will come in contact with the surface of the mask as well as your face (and beard if you have one).
While there's a lot of experiments on the effectiveness of forcefully coughing through mask material, there doesn't seem to be much information on the more practical every day activity of casually breathing into a mask and the turbulent flow of particles around it. Considering how effective masks are in the aggregate I have to assume this is more effective than what most people assume it is; even with materials that are rated poorly when forcefully coughed through.
Good point, well made. We used a teen's vape - showing well fitted masks were best, tho good quality visors were helpful. Masks for all, visors for the mask intolerant - to see cashiers with no masks and a badge saying they're exempt is just nuts :)
I agree with you about the model. It feels like there’s a step missing - is there ever a measurement in the real world that tests whether the model’s recommended intervention actually has the predicted effect? I presume it ought to be possible to measure air flow at the boundary between the COVID ward and the non-COVID ward, to make sure the prediction was correct.
those particles are not air, those are the tracked viral particles exhaled from patients.
The Liam Neeson impression was sick!
My partner pointed out another corona virus airflow issue that she came across when converting a theatre in an area for covid patients (needed because they are running out of space in the normal places).
Operating theatres run a positive-pressure air-conditioning system (air con pushes in clean air, dirty air goes out the door) which is the exact opposite of what you want in an infection control area as all it'll mean it's that,as soon as anyone opens a door, there's a risk that covid particles get pushed out the door - so they had to get all the air-conditioning turned off
I'm jealous, your Liam is spotless, I was really Taken by it.
I just realised why youtubers put their face on the thumbnail, as I've avoided this video for two weeks thinking it was something clickbaity or boring, until I realised it was from this channel and then realised it would be interesting and clicked immediately.
Except, most hospitals that I have ever been to have a very slight negative pressure in all of the rooms, only the hallways and entries provide the air for the hospital....which is bio filtered.
Hospitals that are both newer and have a big enough budget will do that. An older building, or an underfunded hospital can't afford to update the systems, or even keep them properly running all over.
Thank you Steve Mould for actaully disclousing you were paid by this company. Veratasium rubbed me the wrong way a few years ago because he did not mention (or did so at the very end) he was sponsored by a company.
Glad you have some integrity.
« Coronavirus » & « fake virus » in the same video. Finally someone is intelligent there...
I’m interested by this last point about body heat convection. In enclosed spaces doesn’t this greatly increase the virus in air time? Also at what distance would gravity vs convection be predominant/ should I be standing or sitting on the subway?
That’s some serious Testing Adjusting and Balancing. Did he consult with T.A.B.B. to figure out all of the airflow specifics?
Interesting video. Its amazing how Florence Nightingale came to the same conclusion, with regard to
ventilating wards, without a computer :-) Her 'Notes on Hospitals' is well worth reading.
The artist formally known as particle 😂
I did particle simulation in the 1970s for electron plasma flow for CTR. It used 1.5D 2D and 3D modeling.
5:41 Fenton!!!!
Great info! And as an FYI from a licensed building engineer, in the US, code bodies such as the International Mechanical Code set requirements for hospital HVAC design, such as pressurized buildings, all air returned from rooms in ducts, and patient rooms specifically being positively or negatively pressurized, amung many many other requirements.
01:00
You should do a video on Warren Weaver's essay 'Science and Complexity' in which he (somewhat accurately) predicted our ability to model such systems as our computational power increased far beyond what they could imagine in the mid-20th century.
The hospitals have positive pressure systems to control air flow.
positive pressure and negative pressure basically each wing? or area can be controlled independently
outside > Low
hospitals > high
Covid 19 rooms/wing >low
this way the Covid can't escape to rest of the hospital since the air pressure is greater then the Covid 19 rooms, but the hospital since it has high pressure keeps germs and dust outside of the hospital
a whole book on that ashrae.iwrapper.com/ASHRAE_PREVIEW_ONLY_STANDARDS/STD_52.2_2017
ASHRAE
The American Society of Heating, Refrigerating and Air-Conditioning Engineers is an American professional association seeking to advance heating, ventilation, air conditioning and refrigeration systems design and construction.
in America they kind of do all the building and there ones that put standards and rules on buildings and such
like low rise buildings homes.lbl.gov/ventilate-right/ashrae-standard-622
hospitals will usually have UV lights and high Merv rating filter depending location
www.lakeair.com/merv-rating-explanation/
big building are not like a house, they usually have a few hvac systems throughout the building
@@knightwolf3511 HVAC's were assumed as a problem for some meat processing plants in Germany.
Nice video, Steve. There is certainly a whole world of information that was not covered in this video.
There was no mention of boundary conditions in the simulation. As it looked like the hospital simulation didn't account for external building flow, it would be nice to know the boundary conditions used at all of the doors and windows. External building flow would have a significant impact on the low pressure areas around the building, and therefore which windows and doors pull/push air.
Another significant aspect to this kind of science/engineer is validation, which I heard no mention of. Was there any validation work performed at the hospital? Without this, there's really no good way to know if the model is accurate. Garbage in = garbage out.
What would be immensely useful would be a system, from Dassault Systems, which allows people to roughly model their own building's (layout, people positions, and windows) with primitive shapes and locally run a simulation (on a gaming pc GPU) of how covid could spread and how it could be limited.
The trouble is how large an effect small inaccuracies can have. Even the CFD built into CAD packages for non-CFD specialist engineers can lead to suspect results, whether from limitations of the simplified software, or the engineer making the wrong assumptions. At an even lower level of fidelity, you're better off just following rules of thumb and general guidelines that have been published for decades. That said, an average gaming PC these days can easily run modestly-sized CFD sims.
doesn't solidworks allow for fluid simulation ?
@@shigekax Yes, but as C N above mentioned; it would be a case of garbage in garbage out if everyone started running studies. Simulation methods have huge caveats the users must be aware of.
@@csn583 the wrong assumptions that most people would make are the things that I would hope would be pre-set in the software solution. So all the user has to input is a rough floor plan of the building including the location of people, and out spits info of which windows to open/keep closed and where people should not sit to reduce the spread of the virus.
I.e. If the breath of people in a corridor are shown to travel around the intire building they should not sit there.
This is the sort of things that aren't laid out in guidelines, but could have a huge accumulative impact.
@@dfgaJK part of the risk is that if the software produces inaccurate results in some cases, people may change their behaviour due to these results and actually end up worse off
Thanks for this one, Steve!
Lord it reminds me of statistical mechanics lectures when I was a chemistry under grad in Manchester in the 70s. We used to keep a score of how many blakcboards this guy filled in a lecture. the record was, I think, 11, closely packed with partial differential equations. And that was just ideal gases in closed containers. All to derive the gas equations. We did a bit on diffusion, which is toward the topic of this talk, which gave a feel for the subject and hinted at the complexity of making more simple models that adequately reflect reality.
PS it might have been Dr Bullock (He pronounced it Boo 'lock)
The Liam Neeson impression was epic🤣
I love your videos, Steve. Your way of explaining things makes things so incredibly easy to understand to me. Thank you so much for doing this!
When you are early you Don't know what to write...; )
I did my master thesis using the LBM to model flow through a nasal cavity, after some work the results were also really good. Nice to see the method used and gaining momentum, it is easier to use then standard computational fluid dynamics because the grid is easier to create.
Fenton!
Content of this video I didn’t find as elegant and thrilling as Steve's other videos. The reason obviously is the constraints due to sponsorship. Still, I appreciate how STEVE cleverly MOULD it to maintain the quality. Limited numbers of sponsored videos are necessary for the financial stability of an independent youtuber and It's absolutely justified. I am glad Steve did not became like Johnny Harris (in his China video sponsored by WEF).
There's so many variables, though... How do they know that the models correctly predict reality when there's so much abstraction? Do they track air flow experimentally to validate the results?
Yeah, it's always helpful to have an empirical bases to compare your model results to. More experimental results/data allow better and better refinement of models. It's sometimes difficult to anticipate what variables matter and what are the unknown unknowns that aren't built in. Some model systems (like Rosetta protein folding/modeling software suite) are not actually based on a physics based understanding as much as being based on experimental results from protein crystal structures. They started with what researchers observe happening and worked 'backwards' to create general rules that don't really make sense physically, but have been very good at prediction for some interactions/cases. I'm pretty sure google's AlphaFold works similarly with machine learning based on the massive & growing database of protein structures.
@@bradywells1293
Exactly!
And in here they are basically turning a pdf into a complex model for fluid dynamics and just assuming there are no large flaws?
I don't think they would necessarily need to track air flow itself for validation since the exact details of air flow aren't the end goal. They could instead take swabs around the "clean" area before and after the changes the model suggested were made and run PCR on them to see if the amount of virus detected changed. If it didn't change or increased, that tells them the model does not sufficiently reflect reality and needs updated. They definitely would have had to go through multiple iterations and Steve even mentioned the first hospital simulations went with the "full lagrangian" approach. So something must have given them feedback that such an approach wasn't a good choice here.
I see no reason to think the people working on this assume the model is without flaws. Anyone who works in modeling should have come across the aphorism, "All models are wrong, but some are useful".
Edit: found this from the link in the description: www.3ds.com/newsroom/media-alerts/air-flow-simulation-used-3dexperience-lab-open-covid-19-community-reduce-virus-propagation-risk-french-hospital
"Hospital technicians are testing with sensor equipment from local industrial companies to validate the different simulations."
Unfortunately vague, but it does show they were validating their simulations with observations.
@@IanGrams:
Thx! That was helpful info that should've been in the video. =)
Steve kept his word with Emmanuel. Those comments at the end, were indeed, the last words in the video. ;)
As usual, both fascinating and informative... and applying science and models to real life.
"Matt Parker loves Tau" that very brave of you to say
Some theatres do model their airflow! The West End and Broadway versions of Wicked monitor their temperatures in different parts of the theatre to make sure their smoke effects and drapery move correctly.
Sound engineers have to consider this too, as the heat produced by a concert audience can cause sound lensing. I think Dave Rat (sound engineer for red hot chilli peppers and coachella, among other things) has a video on it.
th-cam.com/video/4dU3RlYQIVM/w-d-xo.htmlm10s
th-cam.com/video/_lB8e9OY0-U/w-d-xo.html
Wicked Cool
Thank you Tony the Taiwanese Turtle!
Alternative video title: How this hospital saved lives by opening two windows.
Absolutely mind blowing 🤯 great video!
5:40 Matt Parker Loves Tau
I see what you did there 😂
A circular reference, how appropriate!
@@rednammoc also appropriately in a spreadsheet.
On November 8, 2020 I suffered a STEMI heart attack (the worst kind) and survived (obviously) and was admitted to ICU during a peek in the Corona crisis. I have no idea what other patinas were on my ward, but I was alone in my room. The following day I was transferred to IMCU and the day after I was sent home.
If memory serves, when my grandmother suffered a heart attack in 1978 I think she spent a week in ICU and 2 weeks after that in the hospital (when we kids could visit, we were not allowed to visit her in ICU).
I do remember quite clearly thinking in the ambulance, "Oh no, they are taking me to where all the sickest Covid people are!"
Luckily, it's a very good hospital, and I was not infected, but they were sure anxious to get me out of there. I do wonder if I would have spent more time if it were not for Covid.
"Buoyancy", what a beautiful word !
Military Industrial Complex-sponsored videos ARE THE BEST!
It took way too much scrolling to find this. Another channel I watch is Smarter Every Day, but I get a bit uncomfortable with the number of videos he releases that are essentially advertisements for the US armed forces.
"matt parker loves tau"
What an incredible thumbnail. I have never been more intrigued by a video thumbnail
Ok, but the airflow of a cow still is more interesting
It depends. is it an idealized spherical cow in a frictionless vacuum?
@@temseti0 nope, that's the actual aerodynamic of an accurate model of a cow...
"www.reddit.com/r/interestingasfuck/comments/bnxjkc/aerodynamics_of_a_cow"
@@emanuelescarsella3124 Interesting, but I still prefer Pink Floyd's aerodynamic study of a pig
@@sofa-lofa4241 agree😂👍🏻
Those animations of a person just SPEWING diseased air at another person is unintentionally hilarious.