114 - Ventilation and fire flow paths with Craig Weinschenk

And all that based on an enormous collection of fire experiments full-scale fire experiments conducted by the Fire Safety Research Institute .

My guest today is Dr Craig Weissung , who you may recall from a previous episode about the mission of FSRI , and today I talk with Craig about his experience in running multiple full-scale fire experiments and especially on how the ventilation conditions in those experiments have influenced the outcomes of these fires .

And this impact of ventilation includes openings , doors , windows , establishing flow paths in your rooms , having stack effect and also having wind effects and other flow phenomena that can happen if your structure is of great volume . We also delve into the experiments with e-scooters , e-bikes as targets and as sources of fire .

So if you're interested in this body , perhaps need to listen almost till the end of the episode , but I promise it's good , it's worth it . This is as much fire science as it can be and I really hope you will enjoy this one . So let's spin the intro and jump into the episode . Welcome to the Fire Science Show .

My name is Vojtche Vykhsiński and I will be your host . This podcast is brought to you in collaboration with all of our consultants , a multi-award-winning independent consultancy dedicated to addressing fire safety challenges . Ofr is the UK's leading fire risk consultancy .

Its globally established team has developed a reputation for preeminent fire engineering expertise , with colleagues working across the world to help protect people , property and planet .

In the UK that includes the redevelopment of the Printworks building in Canada Water , one of the tallest residential buildings in Birmingham , as well as historic structures like the National Gallery , national History Museum and National Portrait Gallery in London .

Internationally , its work ranges from Antarctic to the Atacama Desert in Chile and the number of projects in across Africa . In 2023 , ofr will grow its team and is keen to hear from industry professionals who want to collaborate on fire safety features this year . Get in touch at OFRConsultantscom . And now back to the episode about ventilation and fires .

Hello everybody , welcome to the Fire Science Show . I'm today with Craig Weissink from Fire Safety Research Institute . Hello , craig , great to see you . How's it going ? Thanks for having me back . You're very welcome .

You've invited me to a very secret stream of extremely interesting fire experiments , and if you think that secretfully inviting me to witness a fire experiment is what gets you into the show , then you're gotten right . It actually does , but you've built my curiosity and we're going to discuss those experiments , those recent experiments as well , in the show .

But I've known you for years and I've studied your work for years , and you've been researching something that's very interesting for me , which is smoke flows and how the smoke behaves in the fire , and as a smoke control engineer , this is something very close to me , though you deal mostly with different types of structures than I do .

So maybe let's start with why , for so many years , at your time at NIST , now at ULF-SRI what still gives you the idea to research how flows behave in compartment settings ? Where's the merit for this type of research ?

We have an advisory board made up of active and regional retired fire service professionals to make sure that the work we're doing is relevant to them . And as the fire environment continues to evolve with new construction , with new fuels , recognizing how their actions interact with that fire environment are important .

So we've been doing this work going on a little over a decade now , at least for me . In terms of how do we understand what's going on ? And if you look back kind of the history of , we'll say , just with FSRI's work , we've been stacking kind of complexity of experiments and so originally it was let's just look at ventilation .

It's as much as we can isolate ventilation from everything else . What we'll do is we'll do a series of ventilation . The first is let's ventilate on plane , so it'll be horizontal ventilation , so same plane as the fire . And how does changing those ventilation tactics change the flows within the structure ?

They were things that we could repeatedly open or close it and in a fixed time they were non combustible . So it allowed us to just look at the true effect of changing openings with structure . Then we got a little more complex and we said well , now what if we vent off plane ? So what if we start taking ceilings and roofs on single and multi-story structures ?

How does that start to change the fire behavior ? And so we looked at all the different ventilation tactics when it comes to , now , horizontal and vertical ventilation , and we could see that as we change that vertical ventilation we're called timing we can critical , we can open that vertical vent .

You know it does get a little bit of relief , so to speak , for the firefighters because you drop the pressure locally , because it's going to expand and travel vertically .

However , you've created a really efficient flow path and by this you've got , you know , mostly unidirectional exhaust through that vertical vent and then mostly unidirectional inlet through that corresponding horizontal ventilation . And so if you don't have suppression closely tied you're going to result in larger fire growth because basically made a chimney .

But if we take advantage of that initial unidirectional intake , firefighters now have a cleaner path towards the fire . They can leverage some of that to their advantage . But that time when it was fixed can be very , very short .

In our engineering , if you use design tools such as these simulations or zone models , you very often prescribe a very specific environment in your building in which the fire , the action , is happening , and very rarely you would connect the size , how big the fire is , with your ventilation , because you would go back to the concept of a design fire which is related

to occupancy , not related to ventilation conditions . In that case , usually , and even if you did try to merge them , join them a bit , you would very rarely simulate any transient changes . Perhaps you would if you simulate , let's say , a jet-fan system because you have to start the system at some point . But that's your transient change .

I start the system , that's it . But here the window can break , the doors can be opened . You've used the phrase that there are actions in the fire response that change it completely and , understanding those actions , can you give me some sort of visual difference of really huge change in the fire dynamics or in fire behavior when the flow path has changed ?

That justifies this type of research .

So in 2018 , 2019 , we burned some acquired homes , some single-family homes , single-story , multi-story , some garden apartments so three-story within the closed interior stairwell and then also in a strip mall .

So we burned somewhere I'll speak in English units I apologize for being in feet Units that are 60 , 70 feet wide by 70 , 80 feet deep , so very , very large structures . And what's fascinating with all of these is each time we increase our scale , the volatility or the response kind of also increases with it .

It's not something we would expect differently , I don't know , but when we talk about it it is still kind of interesting . When we did our strip mall units , we were in a , we'll say , 50 by 70 square foot structure and we had a front door open and we had eventually 64 square feet of vertical ventilation .

And when we popped that vertical ventilation to simulate in the firefighter tactic of vertical ventilation in this scenario , it was uncoordinated . So we vented just to see what would happen , independent of suppression , and we had north of 30 miles an hour and 2,000 degrees Fahrenheit coming out of that vertical vent .

I would not like to cross that boundary , and 30 is definitely way above that .

So we saw in our search and rescue study we moved away from flood windows to real glass , so we were family windows . This was built basically like a real house . We had full HVAC systems and now when you start taking windows , it's hey , we're taking this window to potentially rescue an occupant .

So we've got fire in a bedroom or we've got fire in a kitchen . Break those windows for entry to singulate Firefighters go into search for occupants . What does that do in terms of changing the fire dynamics on the structure ?

So what might have been a single room fire in a bedroom , now we've got a second low pressure vent across the hallway and if they're not isolating that room , we're driving fire across the whole of the 1600 square foot house . In first , flames at the window in 30 seconds and flash over and under 90 . You know , in that second room at night .

You shall do this . You can consider this as an option , as part of many tactics you might consider as part of your fire response .

And so when we talk about , you know , search and rescue and the study that's ongoing is , if we're going to break a window for entry to search for an octopus , what happens when we break that window in terms of the fire response ? And then what can we do as firefighters to minimize those negative outcomes ? And the first one is go and isolate that door .

So if there is a bedroom door in place , close it , because you know shutting off that , you know disconnecting that from the primary slope path right , meaning you're no longer in the exhaust portion of a flow path from that high pressure fire compartment .

That does a tremendous amount of good for either keeping conditions as they were or actually sometimes improving them , especially if you've gotten out this local exterior ventilation right , you're no longer a firefighter . In this event , take advantage of that vent , close the door and now your flow path is just that room .

You've got bidirectional flow of the window , your smoke layers lifting , fresh airs coming in , and things are improving when you don't close that door . Now we've got a low pressure exhaust and a roof full with volatile fuels . All of those hot fire gases coming out of the original fire compartment will continue to move down towards that window .

You've got the heat , you've got the fuel and now you've provided oxygen . We've shown time and time again how that time window continues to shrink , especially when that gent location is closer and closer to the fire compartment .

They just want to know hey , if I perform action A , we're action B which one's going to have a net positive impact , because there's only so much we can tell about a fire scene from the outside . We have to suppress the fire .

We have to go in and occupy the space to determine if there are occupants , their shields under a banner , behind a closed bathroom door . There's multiple instances of that occurring every month here in the States . How do we provide them the best possible information to make the safest decisions ?

We've been preaching this fire safety message close before you doze for a long time now . Of occupants are more likely to have a bedroom door closed .

Well then , we need to make sure that our tactics reflect that and give them the confidence that if you can identify that the door is closed from the outside , whether it's opening the window instead of taking it or just visually inspecting hey , I've got high visibility when I look in this window and I can see the doors closed , that becomes a safe point to enter

and also a place of refuge , because if you search beyond that room , you can re-isolate it and keep that room disconnect from the high pressure sources . That's only part of the

story . There were some of the challenges now with HVAC systems . Right , we've got this duct network of supplies and returns in the house , so how does that impact it ? And yet we see some additional flow , but not significant . Eventually the pressure balance is out and then at the same time in a state we see some HVAC systems will have supplies and returns .

Some of them will have just a supply in the room and then what we call a transfer vent , so they basically pass through and above the door and then a common return in a hallway or a living room . And so we've now started to run all of the different scenarios to reflect all of the different types .

The supply is high versus low , the return is high versus low . All those different things are now being stacked to better understand the complexities of the system .

So , as part of our larger mission is , yes , we're still at our core doing work with the fire service , because at the end of the day , when the fire happens , we're calling them to come out .

But if we could do things on the prevention side , on the building code side , on the standards side , that's part of our larger mission is all of fire safety , and so we're running these experiments and then we'll publish our papers , like we do in either journal , open accents or self-publish , especially when it comes to the fire service , to ensure that people have

better information so that when these things are up for reviewing the building code and other standards , there's more data .

Now , when it comes to fire Because oftentimes , right , fire is kind of the secondary or tertiary point in the conversation , unless we force our way through the door to say , hey , we've got fire concerns here and we're trying to do that more and more is going to work out in all avenues to ensure that we save public and fire service at response .

It starts burning , people escape , but I have this underpinning assumption that the doors to this compartment are self-closing , what notoriously happens in Poland at least and I wouldn't visit this happening all over the world .

Someone will block these doors because they annoy them to hell , and in Poland they usually block them with fire extinguishers to make it more interesting . Anyway , this person has , in their mind , just blocked the door , but from perspective of what we're touching in this talk , they perhaps changed the complete flow path behavior in the building .

So let's say it's a compartment at some story of a larger building how big that changes the fire behavior or the fire itself in this compartmentalized setting . You can give me some examples .

So when we've done some experiments in your typical residential compartment so like 11 by 12 , 11 by shift , windows closed , door closed a lot of times that fire will self extinguish Even within each back system . You might get moldering if you've got the right ignition fuel .

But once that door's open , that fire now has access to all of that air , even at the rest of the buildings closed .

And so what we've seen then is now that fire is getting large enough that it can either self-vent the windows in the room , which now allows you to have post-flash over fire , because those windows are going to be more than enough to supply oxygen to that compartment .

But we also get a lot of flame spread into the hallway , and so now we're spreading along our carpets and our carpets are very good at spreading fuel , they're synthetic plastic materials basically and so we get this flame spread either down the hallway towards other bedrooms or into this larger common space .

And now , if you were going to make entry to the structure first as part of your fire support your suppression actions , well , now you've got a fire in its meeting , kind of where you are , because it's traveled down that carpet into the next room . All it needs now is that additional burst of oxygen . So these open doors facilitate a ton of flame spread .

They also facilitate a ton of smoke movement . And so now , when the doors closed , initially the rest of the rooms are pretty clean . I mean you have some transport through the duct network if there's an HVC system . But if that door is open , I mean you'll start to see smoke .

The smoke layer will bang down to basically the floor , almost every room in the house , and we know that that smoke is filled with a lot of toxic gases , especially as we've transitioned the synthetics .

And now if you have anything with a battery , your room might transition to post-flash over your cell phone battery , a laptop battery , who else , what else could it be Right and like now that the combination of that toxic soup is far worse .

And if you do go further and further , under ventilated , we know that there's a direct correlation with toxicity and under ventilation .

And you've previously mentioned also the acceleration To what extent this effect of acceleration is changing the fire scene . We've seen a change .

That might still be cotton and you'd have 10 , 15 , 20 minutes and before that that sofa gets full involved and the velocity of the fire coming off of the smoke production is far less . Now everything is basically polyurethane foam , polyester , batting , polyester , microfiber . Those things will transition and produce flame heights that are much different .

We've also now got a pool fire burning under the sofa as well , so we've got two factors simultaneously and that drops your time to about three minutes or so before that might flash over .

And then , what's even more difficult to comprehend you know what we think about the timeline we had is we recently wrapped up a series of like e-scooter experiments , like a small electric moped , and when those got going we went from smoke-showing to peak heat release rate in about 19 seconds , Okay , and that's about 1.2 to 1.3 megawatts and that's also with , you

know , a jet flame coming out now . So if you fail your ceiling right , you also come with a little bit of a pressure rate for these gases . You could tell us you're about 0.2 psi on an overpressure , so you can break some windows , things like that .

But if you start to fail your drywall or you're living in an environment where some of your passive fire protection is already compromised in addition to you know the rest of the things igniting because you've ejected cells , you've got a flame short , you've got a jet flake that's probably 8 to 10 feet hot , and so you think about velocity in that aspect .

Right , Like , not only are we spreading fire faster , our actual outcome is a little bit of a jet fire coming off this bike .

It peaks and then , while you might be able to handle that , that first wave from the kind of energy dissipation , the bike itself is filled with a whole bunch of plastics and foams and so the bike can get a second peak of another 1.2 to 1.5 megawatts .

So these things store anywhere between 1 to 3 megawatts of energy in this really compact swarm factor and our timeline is shrunk again . So we've got the velocity of . We need to either isolate ourselves or evacuate the structure combined with just the straight fire behavior and fire linearity of the exhaust bank up there . So it's a two-fold problem .

Love to see to what extent this flow path is really the defining element in the potential fire .

The more energy we're transferring , we're failing things faster . Right now , imagine if you're a firefighter or an occupant and even if we kept delta T the same and we put them in the exhaust portion of the flow path , right , their inability to survive that is at skyrockets Because we're transferring energy so much more efficiently .

We've got these high velocity flows . So all of these things are . We're talking ambient conditions . Right Now we add wind to the equation , as you kind of alluded to earlier , and your fire can generate between yours , in a single story , single family residential environment , probably about five miles per hour of velocity just from fire growth .

Now we add another five to 10 miles per hour from wind engine's equation and you've got a much more hazardous situation way quicker than you probably anticipated .

. Another thing that actually struck me when I was watching the streams that they sent me was that you had a fire in some sort of a bedroom and it's a multi-compartment house , a corridor that connected to multiple rooms . You had a lot of cameras and you've set a fire in a bedroom . It was like a nicely growing fire .

I don't even remember was it the armchair or was it the mattress .

Now , eventually this room transitioned into a flash over and from that point in time that was ridiculous like in , like I don't know 20 , 30 seconds , the corridor was full of smoke , not a smoke layer anymore , it was literally full of smoke and you could see this layer , dense black layer through which you cannot see within seconds going down .

So this transition like of course everyone knows a flash over is a brutal transition in fire dynamics . Of course every fire engineer knows that the fire after flash over and before flash over will be two different animals . That's why we distinguish them by the fact if they went through flash over or not .

But damn , this was intense , like this transition was insane .

And now , if I like , take this knowledge , this observation , into my project and let's say I'm designing an office space anywhere and my assumption is there's a one megawatt fire in there which perhaps would be fiddling at the flash over limit of that compartment and if flash over happening in that compartment , all my other assumptions of this design fire , they're

pretty much worthless because the fire is so much different . At that point it is something that I , as an engineer , I mean . I understood it happens , but , boy , you have to witness it to really appreciate this transition , absolutely so .

We've got some feedback , all these things In our real-world experiments . Those windows start to crack and fail Kind of an unknown time . It depends on whether we're screens in place to start or whether there are no screens . We single-paying , dual-paying windows . We've got a concurred study looking at that from the wild end .

Exposure from the outside in All of that stuff plays a role . What we think could be a well-contained fire because hey , we've only got a partial window up or something like that it's limiting our fire growth .

As soon as we've got that change in behavior in the structure that's , the window breaking or say the occupant escaping and leaving the door open , the transition of flash over goes from yeah , we've got some fire on the bed or the mattress or this chair , it's contained there for two , three , four minutes From window break to flash over .

Then , right , we see , probably within the next 15 to 20 seconds . That takeoff it's hard to describe when you try to model that or to pull design curves . Yeah , our heat release rate is going to go from maybe a couple hundred kilowatts to a couple of megawatts in 15 seconds . You're like no way , that's not possible . That doesn't seem right . Why would we ?

We have that historic fire growth curve right . It's this nice clean parabolic shape . We've simplified it down to that . We see that in all the textbooks , especially when we first teach where an underventilated fire is in carbon fire dynamics . That's all with this fixed ventilation kind of profile .

When we go underventilated prior to flash over and now we've got this change in ventilation this ventilation-induced flash over is much more volatile . It happens . We see it in the bedroom really quickly . We flash over somewhere in five to five and a half minutes . But as that last 30 seconds that room takes off . I think you also saw our kitchen experiment .

Now we've gone from an 11 by 14 room to something that's probably 20 , 40 by 30 . It's an open floor , clean kitchen , living room . We ignite with a small pan of canola oil , basically to simulate an undetended cooking fire . We let that reach its auto ignition temperatures . Now we've got this kind of oil burning on a pan .

We've got some adjacent bags of chips and paper towels . We've got a full kitchen , so cabinets and everything , but we're on the lean side when it comes to fuel in that kitchen . We've got some cups in the cabinets and stuff , but typical kitchen might have a lot more fuel in the cabinets and things like that . What's interesting there is the timeline .

It follows a similar-ish timeline . We've got this slow growth for a long time , especially getting those wood fuels going versus some of the polyurethanes in our bedroom You've got to contain to the kitchen probably 10 minutes after oil ignition . 12 minutes after oil ignition We've got some flames that begin to roll the ceiling , but it's still a kitchen fire .

We've set up the smoke launch probably within 30 seconds of that oil going to ignition and then we're still a kitchen fire quite some time . And then it goes from kitchen to living room to post-clash over . This final minute again is hey , we've either filled the kitchen window or the top pane of one of the living room windows cracked and started to fail .

Now we've got a little more ventilation and then we're off to the races . That part is predictable once we see that glass fail . But the time to glass failures is not necessarily that repeatable . So that's the challenge . When you're looking at your design curve is okay . Well , when is this going ?

When is this next piece , this next domino going to fall on this scenario , because that's what we have to plan for is the unintended ventilation ? I don't know , because that's going to happen at some point down the line is something else is going to fail in the system ?

Because if the size of the fire depends on the fact that the glass breaks or not and if it happens in 17th or 19th minute on which we have no power over , and I also think in design space , perhaps underestimate the threats of very quickly developing fires .

Because if you fire can develop in two , three minutes in a confined space and reach a flash overstate at which it feels a corridor with smoke within literally 30 seconds , if this happens in large floor of complicated building and I often would go to two office as an example if you put that on the timeline of a building , you know how long does it take to

detect . How long does it take for fire alarm to pass , because it's not just sensor detecting , it has to go to a control panel . There is a person that perhaps can be obligated to verify the fire or not . It is automated but not immediately . It's not like the second you detect . The second people know it takes time , sometimes minutes .

If your timeline shrinks to minutes and you have to add the uncertainty of the grass process on top of that you know decision making , the movement itself you suddenly enter quite dangerous timeline because in your design fire scenario , when it is medium growth and you can keep it like that for a few minutes and then it slowly spreads the smoke , the fire doesn't

spread . You perhaps are okay , but if in a scenario where your armchair is put against a cabinet with paper and you have some loose paper on your desk and perhaps the window is slightly opened , fire dynamics of that compartment will be completely different than the ones that you could assume and this really scares the hell out of them .

So we've got on the design side in recognizing all these challenges and at the same time , that's why we do so much work with the fire service is so that they can recognize , if they're going to create a ventilation opening as part of their suppression or search and rescue operations , what are the consequences of set action .

And so if we've got two points to open , one near the fire , one far from the fire , recognizing that distance is buying them time for a negative , a potential negative outcome , right , it's if we can vent further that allows us to get into that room to isolate it first , you know , because we might not be able to search past that living room kitchen that's on

fire , right , if the rescue operations are getting there before the suppression . Well , how do we get to those rooms ? Well , we've got to get from the outside in , so the front door might no longer be the natural point of entry .

And so if we choose an alternative point of entry , recognizing what the consequences of that is , and if we break those windows and don't isolate , we're going to bring the fire down the hole towards that ventilation . There's more than enough fuel at any residential structure these days to facilitate that flame spread .

And so we have consciously been working with the fire service for the last decade to kind of put all these pieces together whether it's the suppression actions , time-lapse ventilations , ventilation , time-lapse search , search , time-lapse suppression so that all of these things can be independently recognized but then also recognizes how the system operates together .

So now let's say we've done everything correct . Any of your building design or decisions are correct , but there's this external factor . How does that play in the grand equation of the fire in the compartment ?

Wind will completely reroute those fire gases within the compartment , so what would be a natural bi-directional flow of potential unidirectional exhaust heading up where what other vents might be open ? The wind can turn that to unidirectional in and that becomes a huge concern for us because it's a different scenario .

Put that into the real world now and firefighters might be breaking that window and they put firefighters in now what they thought would be the intake portion of the flow path and it's in the exhaust portion of the flow path . And we talked earlier about velocity .

Right Like now , there's too much higher convective heat flux , which is really really bad for them and the environment .

We were doing an experiment a few months ago now and we had a living room fire with a open floor plan , kitchen living room and we had two windows on the kitchen side and two windows on the living room side , so they're kind of connected but open floor plan and right inside the front door which connected to the living room , we had a e-scrooter so that that

moped e-bike and we lit the so fun fire . The fire started to spread and we got flashed over in our living room , eventually our kitchen , because of the vents . But we had about 10 to 12 mile per hour wind coming from the kitchen side . So the living room side now becomes the low pressure exhaust and the kitchen's a high pressure supply .

So we basically had unidirectional in on our kitchen windows and when we eventually ignited this e-bike , we didn't even know we did it visually . Our instrumentation told us , hey , this thing probably failed because the cell temperature got up to a critical value .

We had so much fire coming out of our vents at such a magnitude that we didn't know we set an e-bike on fire and so that's what wind can do . We had quite the wind driven fire and I mean suppression became okay , well then , we'll tack from that side and put water in that window and knock the whole thing down .

In that regard we took advantage of the wind but at the same time , like that definitely changed our entire experiment , and fortunately for us it's an experiment .

But now imagine that being someone's house or someone's building and you take that wind side vent and it's gonna be much harder fire to control , especially if you can't get suppression where it needs to be , and so it plays a huge role , and then at times you can play much less noticeable aspects .

So we were doing some in our search and rescue houses and we had a slight wind , probably a couple of miles per hour , and so you couldn't necessarily see it visually in the windows .

So , like I mean , you're still not pressure in the structure that we're having some exhausts come out early in fire group before that room kind of reached flash over , the wind played a role where we had much more fire in our hallway than we expected .

Okay , so while it didn't necessarily meet the true conditions of a wind driven fire , in the sense that we had this well defined change in gas flows within our flow path , it was subtle enough that those gases that would normally have flown out the window because that was the nearest vent to our ignition , you know a lot of that it kept enough pressure at that

window so that we had a lot more fire in a hallway . So when the firefighters came down to suppress the fire , you only turn the corner and the entire hallway is flashed over and they're like what happened ? Why isn't this different than every other fire we had ?

And so , while as experiments , we obviously instrument the whole structure , but we also have wind sensors and weather stations up so we could actually see the wind change to the point where we had several line-hour gusts , and the gusts were enough .

Right , we didn't notice it in the time , because it's that because of our microscopic impact on the macroscopic problem and it was in no sense that we started to break through our subfloor and things in the hallway we were damaging some of the other doors because the door directly across from the fire problem was closed , but we actually started to burn it through

because we had so much more of a thermal insult in that hallway . And so when we watched the videos back , we had this good one that we need to release at some point , side by side . From the exterior , you have no idea the difference . And then we have the helic cams in the interior , cams for the suppression crew . They turned .

They're like these are drastically different fires inside , and so not only is it a problem for designers right and the fire protection community , but also for the responding firefighters is . They might not know that that size up from the exterior tells them a lot , but it doesn't tell them everything . And that was one of the more fascinating ones . Was that subtle ?

So we lined up . You know here's our wind and then here's our heat flux in the floor in the hallway and you can start to see . Oh yeah , this is why these two things are different . And then , once it got going , the wind died down a little bit , but also there was enough fire in this interior hallway that the wind was no longer going to impact .

It was just spreading down the hall towards the other bedrooms and also towards our kitchen and living room . So we had a much different buyer for one of the unknown exterior factors .

If you want to have very high EAM or lead certificate for your building , you may actually have some sort of natural openings in your building which can be randomly opened by the people inside your office . And what you've said here you probably won't have significant flow path .

Suddenly a jet stream of air in your you might , if the wind is strong , you know , but this little nudging of the fire towards like one direction or another can cause a very interesting outcome . If you also mentioned something that captured my attention , you said that you had a scooter . It was .

You were not actually setting the scooter on fire , it was just there . And , as I understood from my interview with Adam Barovie , you held us a lot of this test where the lithium ion batteries or electronical devices would just be present in the fire to just see what happens to the fire scene by their presence .

So not necessarily it's serving as the trigger for the fire or the first item ignited , but just being there . So what are your experiences so far with this type so devices ? To what extent it changes the outcomes ?

I've only done a few , so not you know , relative to the everything else that we've said on fire at least me personally . I've only done a handful where the battery's been kind of the victim and not initiator . And for one , you know , we had our wind-driven incident .

It struck me so odd because came in we looked like we would normally for any kind of fire patterns on the wall or the floor or anything that would indicate that , hey , this ignited differently than everything else where there's a pre-flash-over pattern , things like that , and there was nothing Granted .

The wind-driven case , you know , obviously magnifies all of that , and so we're hopeful the next time we get to completely destroy a building that we can do it again , because that was one of those ones where we learned a lot from the wind-driven case .

But we really wish it wasn't a wind-driven experiment , but we were locked in a timeline , so you kind of have to go forward . Sometimes when we've gone to some of the smaller batteries , it always depends A lot of our experiments .

We tend to go to a post-flash-over state and so you don't necessarily see a lot of the impact because that fire's just overwhelming some of these things . But other colleagues have done more . So it's kind of out of my scope a little bit in terms of what I've usually said in a fire , or at least the fire sizes I've been dealing with recently .

So we removed a BMS from the equation and we're going to send it , you know , some current and voltage to allow this battery pack to go to thermal , only One . The timing is so wide . As an experimentalist it's very frustrating when you're like three and a half hours into overcharging this thing and you're still not getting ignition . But the failure mode is .

What was kind of fascinating to me is that sometimes we would get this kind of full . The definition of an explosion is kind of always one of those which definition we're gonna choose . But this over pressure event where we actually , you know , broke some windows and doors where we had this kind of more uniform type you know fire behavior .

And then there's times where we get this jet flame , okay , and it just shoots a plane 8 , 10 , 12 feet straight up or straight out the side , and when it straight up it's not as hazardous , at least in my opinion , because you're going up to drywall , ceilings and and stuff and it pulls the ceiling but the fuel isn't there .

When we failed it and shot it sideways , you know , we shot it straight into a couch , okay . And so that was like , okay , well , this is bad . And then we get some overheating experiments as well .

So we just took like a pipe heater or bucket heater and slapped it right on the side of the cells and overheated it that way , and Sometimes we got this explosive type behavior and sometimes we got this jet fling . So that was kind of interesting . When we , you know , started to look at that in terms of , well , how predictable really is this ?

Our time windows were all over the place , as short as you know , 30 minutes or even less , and in a couple of overheating experiments to four and a half hours . And so when we talk about design aspects , right is , how do you control for that such large variability ? Where is the bike damaged ? Is it brand new ? What's the charging when it ? What all happened ?

That's a kind of a little bit of a black box at the moment , you know , in terms of the fire protection community . And so the type of hazard and the failure mechanism was was all over the place in terms of then what it did to the rest of the Structure .

I mean , we broke some walls , we broke some windows , we broke drawers and then we did run some experiments with a sprinkler system . They had some very encouraging , very positive results and so we're currently writing that up for both the fire spread side and the toxicity side .

So hopefully you should have that paper done by the end of the year to share with everyone in terms of what that looks like with a residential , you know , 13d system and some of these electric vehicles inside .

So we're electric scooters where the e-mobility type stuff that's exciting Not yet ready to share because there's still some , some stuff we need to analyze from from the downside .

I mean , when we did our , our strip mall experiments , those still all strike me as some of the most exciting Experiments we've ever done , at least me personally .

I mean we , when we had a double-wide unit that was like about 70 by 80 or maybe using 80 by a little bit larger , we were using like group , a plastic commodity boxes , so 125 polystyrene cups and cardboard boxes .

We had them stacked to think for high and Hmm , and I treated them almost like aisles , like you would see , in a typical like on a strip mall , tight occupancy , and we knitted in the far corner from the open front door and we started to get fire spread across these boxes and so we went from , you know , kind of one group to the next group .

So we had a little bit of a honey definition . I might , you know , some of the audience might not agree with a traveling fire , so to speak , but we had this flame front moving and we had temperatures , you know , in these local localized areas that were , you know , floor to ceiling , 1100 degrees Celsius .

So whether you call that , you know flash over or not , and I had a good conversation with Guillermo recently

about that . But we had this flame front move across and what was so wild to me is that we put we basically had a unit directional exhaust .

We're at this front door this entire time , so it was was shooting out like a jet 75 feet from the front door while the door was unit directional exhaust as this fires building , because with the open floor plane Strip mall unit there's more than enough oxygen to support this fire before it actually started to come truly under ventilated .

So the flame front moves to the front door and as soon as they get to the front door the exhaust stops . We get nice bi-directional flow and it looks like anything you'd ever see before .

So this you know , if your , if your fire department's arriving three , four , five minutes afterwards , they can arrive as you've got this unit directional smoke coming out , thinking that we've got a smoke explosion going on or some other Effect , when it's really just fired on a vision these large volume could combine with a very specific fuel that you had , because

that's a very , let's say , dangerous type of commodity that you had exactly there .

There's a really fascinating case in New York and I think it's in Troy , new York , the alpha lane bowling alley , if you ever get a chance to look up that video a ton of smoke coming out of the bowling alley and then all of a sudden they night's off as a unit directional exhaust , basically all self-driven .

And we've seen these kind of things happen more and more recently with these larger volume structures and these highly volatile fuels and they're fire dynamics driven , not necessarily where you know , we're taking windows , changing the ventilation pattern , things like that . It's surely , hey , we've got this large volume that could support this large fire growth .

So we're gonna be doing some townhouse fires in the next few years . Looking at now , what does it look like when you have three-story stack with an open stairwell ? Can you replicate some of this behavior in the residential fire ground ?

We saw it somewhat in some of our fire investigation work where we had a kind of a two-story great room when we could get this unit directional exhaust out the front door , and so that's still one of those fire dynamics things .

It's fascinating to me that that's still kind of you know , I'm very interested to continue to see and then as we go to this building , to building fire spread stuff , you know , looking at the different glass types , weather We've , you know the ordering of the panes and how that works and how the different gases between the panes all play a role .

But we got the best class in the world , but if we don't have the proper frames we don't have the proper side . You know it's this fire protection system . It is still so fascinating to me and how all of these pieces and how we can identify some of these things by Looking at the full scale . You know we've done a lot of small-scale stuff .

We have a materials database now with over 70 materials for their admission properties and you know we're gonna be adding their kinetics . Now We've got thermal conductivity and density and all that stuff and it's wonderful for a fire modeling perspective . Right , how can we start to improve our baseline capabilities ?

But sometimes you don't necessarily see all of the challenges until you burn it to full scale and see , oh , wow , here's this big flaw in the system that we didn't consider , because all these systems are developed by people of different expertise and once one put together here it's , it's kind of like oh , we have think about that one .

So for me that that part never gets old and it's still fun setting big things on fire and I , I I hope , with your research , we will have less and less moments of arm .

It's sometimes very difficult to imagine that and Understand , but there are videos , there's materials , there are online courses produced by FSRI that you can go and see in details , slow-paced , what actually happened and what's learning from that particular fire and what . Where should people go to find those ?

But the course is really cool because we've incorporated a lot , a lot of our video , a lot of the data , kind of in a very Interactive way that we haven't done before , and we also have a lot of our tech panel members as part of the course .

So you can hear from the firefighters that are responding to this , to these scenarios , what they Feel , how they interpret our results . So training , that FSRI , that work .

And then , on the other scale of things , plug for our materials database , materials FSRI org has a wealth of information about how all of these materials burn at the small scale to medium scale , so everything from an MCC to a colon , and then soon We'll be adding a lot of our full-scale data as well .

So that's going to be a home for for lots of material properties and we think about why , why this cabinet earns different than that , than that chair , that sofa . We can look at those material properties now and see first-class what that looks like , all in open access . Yes , everything's free , everything's absolutely free and open source .

So the database is cool because we have the front end that you guys can interact with with interactive graphs and tables . But then all of the data is available on github . So you can go to the UL address right github and see a handful of experimental data sets that are open , including the materials database .

But as I re-listen to this , I hope we've done a good job providing you voice illustration of what changing the flow paths and flows in buildings caused to do the fires , and the impact is immense , especially when Craig mentioned this some would say Non-important impact of slow winds .

That was literally just nudging the fire towards one opening versus another , but that was enough to change the the complete fire physics in the compartment . So well , this , this impacts may be considerable .

If you would like some visual cues , craig just listed a lot of interesting resources provided by FSRI , and this includes a lot of videos from their Experiments as well . So you probably would like to check that out , because that database is literally gold mine for any fire safety engineer .

So I absolutely recommend you checking out these resources and and for today's talk . Yeah , that would be it . Thank you very much for being here with me this Wednesday and the next Wednesday I see you here once again . Thank you , bye . This was the fire science show . Thank you for listening and see you soon .