Fire Science Show
Fire Science Show
207 - Fire Safety of Balconies with Mike Spearpoint and Konstantinos Chotzoglou
As a consequence of the Grenfell Tower disaster, some strong legislation was proposed, such as a combustible ban on building walls. This, however, affected more than just the building facades, as it excluded materials such as laminated glass used as balcony balustrades.
Today, the path forward demands evidence that could inform decisions on the future of laminated glass in this use. In this conversation with Mike Spearpoint and Konstantinos Chotzoglou from OFR Consultants, we dive deep into their groundbreaking experimental research on balcony fire safety that emerged in response to the Grenfell Tower disaster.
Through experiments involving three-story balcony setups and multiple configurations, the team quantified how different materials and designs affect external fire spread between floors.
What makes this research particularly valuable is how it transforms gut feelings into measurable facts. The researchers tested various combinations of balustrade materials, decking options, and balcony contents to create a comprehensive picture of fire behaviour. Their findings confirmed some expectations while providing surprising insights into flame dynamics around balconies. Most importantly, they established a clear ranking of safety performance: from non-combustible systems and laminated glass (which performed remarkably well) to the dangerous combination of HPL panels with timber decking (which produced fires so intense they had to terminate testing).
The implications extend beyond regulatory compliance. This research empowers architects, engineers, and manufacturers to make evidence-based decisions about balcony design while maintaining the essential outdoor spaces people value in high-rise living. It demonstrates that with appropriate material selections and protective measures like non-combustible soffits, balconies can remain both safe and functional.
You can read the balcony survey paper here: https://link.springer.com/article/10.1007/s10694-023-01467-8
A paper summarising three balcony fire incidents:
https://link.springer.com/article/10.1007/s10694-021-01154-6
As more research is published, I will try to keep this up to date.
----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.
Hello everybody, welcome to the Fire Science Show. Today we have an episode on experimental fire science. We will be talking about a very impressive research project that was carried out by the OFR on the balconies, fire safety of balconies and, as I've been visiting them for their team's event and at some talks out there, it was a great conference. I also had the chance to steal Mike's viewpoint for a minute, along with his colleague, konstantinos Czozoglou, and sit down and discuss this magnificent project. So, yeah, I'm always traveling with my microphones. I saw a chance and I took it, and the outcome is the interview that you are about to hear.
Speaker 1:In this project they were tasked with identifying or providing evidence for fire safety or unsafety of balconies in relationship to laminated glass used as the balustrades of the balconies, but also in presence of timber deckings, in presence of combustible materials used on the balconies, and also to investigate the influence of the stuff that people have on their balconies on the general fire safety of the building and fire spread pathways. So indeed, quite a big research task and it's a very large project actually, if you think about it, it's a very large project In this podcast. What's interesting is the outcomes are not very shocking, so they've confirmed a lot that was suspected, but I guess that's a good thing. A great thing is that we've moved from opinions to evidence, which in lawmaking, which in professional fire safety engineering, is a critical thing. I think there will be more and more need to prove the obvious. The magic numbers don't really make it anymore. We need evidence, and evidence-based fire science is what we should strive for, and this is definitely what is presented in here. We will also discuss about some perhaps unexpected consequences of banning combustible materials. The reason how laminated glass got excluded from balconies Very interesting case. The reason how laminated glass got excluded from balconies very interesting case we will be talking about.
Speaker 1:How do you set up such a huge research program? What are the choices, what are the dilemmas of people running those programs? I make those podcasts so you understand the context of the research that's happening all the way around. You're not going to learn everything about balconies in here, and also there will be papers. There are papers where you can read it all the way around. You're not going to learn everything about balconies in here, and also there will be papers. There are papers where you can read it. But here you can learn why they've made this research choices and I think it's very interesting. And all this in this podcast episode. Mike Spearpoint, konstantinos Chosyoglou, please join us in this conversation on the balcony fire safety. Let's spin the intro and jump into the episode.
Speaker 1:Welcome to the Firesize Show. My name is Wojciech Wigrzyński and I will be your host. The FireSense Show is into its third year of continued support from its sponsor, ofr Consultants, who are an independent, multi-award-winning fire engineering consultancy with a reputation for delivering innovative, safety-driven solutions. As the UK-leading independent fire risk consultancy, ofr's globally established team have developed a reputation for preeminent fire engineering expertise, with colleagues working across the world to help protect people, property and the planet. Established in the UK in 2016 as a start-up business by two highly experienced fire engineering consultants, the business continues to grow at a phenomenal rate, with offices across the country in eight locations, from Edinburgh to Bath, and plans for future expansions. If you're keen to find out more or join OFR Consultants during this exciting period of growth, visit their website at ofrconsultantscom. And now back to the episode a massive research project that you have completed on fire safety of balconies. But perhaps you better give me an introduction what the Balcony Project was about and can you give a two-minute overview of the ideas?
Speaker 2:Yeah, sure.
Speaker 2:So the project kind of came out from the tragic events of Grenfell Tower and in doing that, following the fire, the government in England introduced a ban on the use of combustible materials on the facade of buildings, and in doing that it also uh, looked um incorporate the fact that that would include elements fixed to the building, such such as balconies, and one of the products that was being used at the time on balconies was laminated glass, particularly with a pvb interlayer, and because of the fire performance of pvb and the way it might fall under euro class, it wasn't the non-combustible and so it was a kind of you might say a bit of an unintended consequence of a decision that was made.
Speaker 2:Who would have expected and look these things, these, these things happen because of the response to grenfell tower. And then, following uh, dame judith hackett's report, the government had a call for research projects, things, things that people felt that would need to be looked at, particularly to maybe update the guidance in approved document b, and in that call for evidence and call for research, the government identified a group of sort of a package of project which was called balconies, spandrels and laminated glass. So the project was then tasked at the outset to kind of look at the what might be the contribution on external fire spread of balconies and maybe I say maybe to look at this issue of the balustrade construction and whether the use of laminated glass on a balcony balustrade, how much it may or may not contribute to external fire spread.
Speaker 1:But it was going further than just the glass, it was more into that variety.
Speaker 2:Well, yeah, as the project developed and discussions with the client, then there were some interesting discussions around the timber deck or the decking material in timber, and there was quite a well-known fire prior to Grenfell Tower where a timber balcony on a building had quite a large fire.
Speaker 2:And then there was some discussion around other balustrade materials and in the during the project a fire happened. It wasn't particularly sort of landmark, but a fire occurred on a building in south london where there was a combination of a hpl panel system and a timber deck and the fire spread. It was a fairly low-rise building but spread to multiple balconies. So in the course of the project and discussions with the client a number of other parameters became of interest, other than just the question of the laminated glass balustrade. But there's a, you know, there's a sort of combination of materials that you might put in a balcony, and the other part of the project was also to look at, well, how much combustible fuel load does someone put on their balcony, because obviously that may also contribute to the project. So it had a sort of number of aspects that we tried to package together.
Speaker 3:Just to add on this a bit of the background. In the UK the balconies are considered as attachments according to the latest guidance, so they are attachments to the external wall. So regulation treats balconies as the external wall nowadays, and that's why we have this project as well, because, finally, any materials, as Mike mentioned, like whether it is timber, whether it is HPL, all these materials actually may have an impact, and not only on the balcony itself but also on the external wall, which is one of the main hazards anyway.
Speaker 1:I think you told me that people are expected to build balconies in tall buildings, that it's an expected feature of a building to have one.
Speaker 2:Yeah, so there's things like the London Plan, which is a kind of planning guidance documents that use for construction in London, and in that there is an expectation that people in a high-rise building have an outside space. I mean, you kind of think of the time of COVID and that people were wanting some kind of outside space. And of course, if you've got a high-rise building, yes, you could walk down the stairs and maybe there's a park nearby, but the balcony provides you with an out your own private, generally a private outside space. So there is a, you know, as part of the utility of buildings in terms of the general use, providing facilities for people to have an outside space is seen is seen something as part of sort of planning guidance.
Speaker 1:In those fires that you've mentioned, was the spread of the fire kind of traced to the balconies, like, were balconies determined as some sort of causal factor of the fire spread in the building?
Speaker 2:Well, definitely the fire on the timber balcony, that fire, given it was constructed of timber balustrade, timber decking, that that would clear clearly. Uh, the, the construction of the balcony was, was the major contributor. Some other fires, this one in south london, again there was a question about which I mean obviously the fire spread between balconies so clearly there was fire spread. The wall itself was, was pretty much non-combustible, so the wall had no contribution. You can see pictures afterwards. There's essentially a bit of smoke, flame and damage, but there's no wall contribute so clearly. The balcony, in that case, in terms of the items the, the balustrade, the decking and the content spread fire.
Speaker 2:There are other fires and one of my colleagues, a guy called jamie clark, has been looking at where we got fires that have been reported by fire brigades and they have reported fire spread between balconies and maybe there's a sense of what maybe the wall construction might have been and what the contents would be. Sometimes it's it's not always clear how much the balcony was a contributor over and above any other element in the building. So sometimes it's obvious, sometimes it it's not so clear cut but at least it's a suspected pathway.
Speaker 1:Okay, and but for me, a balcony like I'm a user of a balcony, very happy user of a balcony one thing is is the structure, how you build balcony, what are the components of the balcony, and the other thing is what people keep on their balconies, and from my observations it can be quite a scatter.
Speaker 3:Yeah, indeed, and this is what we had many discussions on the project for this purpose, like, what do we include on the investigation at the end of the day, because we need to be somehow realistic In many countries, for example, they use a complete concrete structure for the balcony, so it's part of the overall structure of the building. However, you can see, like in the UK at least, there are many balconies which are fit after a post, let's say construction, as actually I must say attachments I use the word of attachments or you can have like different structure, like a metallic structure as well, maybe stacked balconies in which the investigation in one of the fires was actually shown that the balconies were stuck on top of the other using a metallic structure. So we had a lot of discussion on that. What we include finally and Mike mentioned as well one of the contributing factors is indeed the contents. How do we measure that and how do we quantify this? And there is work done previously. There is a survey.
Speaker 2:Mike, I think you can tell more on the survey Part of that project was to get an understanding of how much combustible content people might put on their balcony. And if you look in the literature there's very little data. I mean there's lots of data. People have done all sorts of surveys of combustible contents inside buildings, but very, very little. We did find some work that had been done in Japan where it was part of a sort of bigger project.
Speaker 2:So particularly my colleague, sam Bryan from the Edinburgh office, did a survey looking at 1,000 balconies it was actually during COVID, so even the idea of accessing balconies, but we could use Google Street View and she looked at over 1,000 balconies, assessed each balcony in terms of the content potentially Did it have plant pots on?
Speaker 2:And she assumed they were all plastic. Did it have some kind of furniture? Did it have clothes, people drying clothes and whatever, and from there she could do a statistical analysis of a 90th percentile fuel load. She also assessed how much of it might have been cellulosic, how much it might have been plastic and what the dimensions of the balcony might be, and so we could use that to then separate the question of the of the materials to construct the balcony versus the things that me people might put on the balcony. So now, whenever I'm in a city, I'm always looking up at balconies, going, oh look, they've got, they've got, they've got that on their balcony. And look at another one, go, there's all that. One's quite empty. So now I can't stop looking at balconies and assessing how much fire load there might be watch out if dr spearpoint is in your town because your balcony is being observed and assessed.
Speaker 1:I would like to move to the experimental part, but I still have to tie one more thing because, uh, again the contents of the balcony. That truly makes sense to me, but that laminated glass is such an odd issue. Can you tell me why exactly, like you said that it doesn't have the glass? But what is the known behavior of laminated glass in fire, like before the project? Not perhaps what you found intermediately in the project?
Speaker 2:Yeah, so there's been some work on laminated glass. I mean, as part of the work we did a little history of laminated glass from its original conception around about the early turn of the century. It was a French guy and I think it was sort of discovered by accident and it started off particularly in things like the automotive industry. So it was the time when people were putting glass in vehicle windows. And then you find, particularly during the second world war, the fighter and bomber planes and that, and then it became a product that we used in construction and there has been some work done on the performance of glass. Pretty certain one of your podcasts you've had you want yes, that's right.
Speaker 2:So so you know, it's been a product, it's been around and there's been other researchers I think there's a paper with Michael Klassen on that and they looked at some different laminated glass products and there's quite a bit of work on glass, single-pane float glass.
Speaker 2:There was stuff that was done at Holster many years ago, there was stuff that was done by Skelly and there's some others, but it was was. You know, like a lot of these topics, there was some work and we'd done a little bit ourselves, sort of even before we got this project. We actually um, got some laminated glass and we and we went up to the lab at the university of edinburgh and did a few initial sort of investigations, trying some stuff. And and particularly my colleague ewan who had finished his phd, they had set up the h tris and we could use that and we and we've done some stuff there. So there was some work around in the literature and so there was an understanding somewhat about the performance of there's different laminates within. You can put within the glass pvb, eva, um, sgp. So there was some work on that. It was, you know, like many things in fire science there's not a thousand papers on it.
Speaker 1:But if I understand correctly, it's just a really thin layer of foil that keeps the pieces of glass. It's not like an intermediate sandwich of acrylic into pieces of glass. How big is that foil?
Speaker 3:The thing is that, as you mentioned earlier, at the end of the day, for regulatory purposes, it cannot achieve the class needed for the regulation. So even though, yes, indeed, it is like very thin layering between the two panes, you can have like a description of how it behaves and everything, but at the end you cannot use it, at least after this ban, although it has been for many years in construction in the UK as well, like everywhere in the world. It has been used a lot, also on balconies, you see, if you see like balconies use a lot of balustrades made of glass. So, yes, indeed, there is some research, but at the end, our target, our investigation, was how we incorporate this laminated glass into balconies, how is it incorporated in reality, and whether this actually plays a role on the fire spread, on the, on the dynamics during a fire but to answer your question in terms of the thickness, I mean typical laminated glasses you might see on balustrades the glass, the two.
Speaker 2:I mean we're looking at two layers of glass sort of in the eight, ten, twelve millimeter thick and then the, the interlayer, normally sort of two millimeters actually, maybe a bit less than that. So you know, as a proportion, I mean I cut what two, one, one and a half mil of interlayer. Some interlayers might be a bit thicker, the cast in place type products, as I remember, the interlayer might be thicker just because of the manufacturing process. But in terms of the proportion, I mean obviously you've got, so if you've got, two layers of 10 mil glass plus the two of that, and it's what's that sort of 10%? Yeah, all right, 10%, yeah, that sort of number.
Speaker 1:I mean obviously, obviously it doesn't sound like a big fire issue. Uh, I mean, okay, if it was a thicker layer, like like in the sandwich panels, and then it's a different story. But because that's a lot of fuel on something that peels off, but as a part of the glass, we also know that it, when shut, as it stays in the same place, it doesn't, uh, crush into a million pieces that fall off, if I understand yeah, I mean that's the purpose of the laminated glass.
Speaker 2:So we've obviously got a potential for the glass to break just in normal use. Someone might crack, knock something into it and the glass can break. And I and I found a paper from the night around about 1970s, 1980s, published by someone who worked in a hospital and was showing talking about the fact at that point it was monolithic glass which would break, and there's some not very jolly photographs to look at of people's arms sliced open and that. And this author was saying this is you know, this stuff, we can't have glass and allowing it to fall and that. But there's laminated glass. So why, you know, we should be using laminated glass? Because it prevents this, these terrible lacerations that people gain.
Speaker 3:So so it performs a, you know, has a safety benefit, obviously that you know the fire is part of it, but it has other reasons why we might use it and again, it's a safety benefit because of that, day-to-day things just get knocked into it and and that in itself presents a hazard and it was one of our thoughts, let's say, during the project, that when it breaks, what happens when it breaks during a fire, and whether actually another thing was on the intermediate layer, the PVB or whatever it is, if it melts, how it melts internally, and whether this can create some other hazard as well. So, yes, indeed there is a lot of work done, but in a bigger, let's say, picture how all these little details, let's say, can play a role in a balcony fire.
Speaker 1:Okay, so let's move to the experimental onset. It's always difficult to present those in a podcast material, so you give me the outline and I'll try to recap and summarize. But how did the research plan look like? So, first of all, how did you intend to test it Like? What was the experimental plan in terms of the facility and in terms of the experimental setup? And secondly, what were the variables that you were looking into? Perhaps let's cover the experiment first.
Speaker 2:Okay, I mean the intention of the experiment ultimately was to do some full scale balcony fire spread. But in before we wanted to do that, we wanted to get a bit more of a particularly an understanding of laminated glass and the different combinations of glass thickness, laminate type, exposure conditions or different heat fluxes. So the first part of the project is we did a whole series of experiments. The guys at the effectus lab in northern ireland set up a radiant panel system and we looked at a sort of you might say a parametric study. So many of these variables. So we had different glass thicknesses, different laminate types, different exposure conditions because we wanted to try and choose a reasonably conservative representative of the glass we're going to put on the full-scale balconies. We didn't have the resources to build full-scale balconies and look at a whole range of different combinations of glass, and so that was the first part and then from that we used that to help us develop the full-scale experiments. Use that to help us develop the four-scale experiments.
Speaker 3:And then we had also some input, of course, from industry, because it would be, let's say, a waste of money, a waste of resources to use something which will never be on a typical, realistic, let's say, balcony design. So we had this input as well. We had the results from the small-scale experiments that we did and then we ended up on the large-scale, which took like months and months to design. But anyway, that was the hard part of this project. I think we ended up with trying to see, okay, which scenario we are starting with.
Speaker 3:So for us it was developed fire internally, the flames come out from an opening or so and then we have the flames ejecting on the balcony area. So we started from that point and therefore we had to choose what type of fire design we do, what type of fire we have. So we ended up with the BS8414 crib source, which was it's something which is used a lot, it's something which is used on the testing for external wall systems. So we thought like, okay, we know what to expect, we have measured the heat, we know what to expect from the crib itself. So it represents probably a scenario that we wish to have on these large-scale experiments. And then we said okay, now it's time to work on the balcony itself, on the structure of the balcony.
Speaker 2:Yeah, so we ended up with a. In terms of the structure, we ended up with a stacked balcony arrangement and this reproduced.
Speaker 1:How many?
Speaker 2:So essentially three balconies, so the floor of the lab kind of represented Balcony number one, balcony number one or balcony zero. Then we had the 8414 combustion chamber which we raised up compared with the normal, the normal test setup so it was.
Speaker 1:It was just underneath the next balcony, so balcony one, and then we had, we had, balcony two but by chamber you mean it's like this one meter deep, like two and a half wide, something tall yeah, it's a standard just a small chamber in which you put this massive, massive wood crib and then you expect all the flame to go out yes, exactly I assume in your experiment it projects on onto the balcony directly yeah, and on the underside of the balcony.
Speaker 2:So so we've got, we've got three balconies, so zero one. Two we were. You know, there was an element, we because the if the guys that affect us in the lab, we've got a calorimeter. We could put the whole setup under the calorimeter measure, heat release, which is, you know, fairly unique to be able to do that. But obviously, like anything we've got lab, there's only so much height in the lab.
Speaker 2:We wanted to represent, obviously, a spacing between each floor of the building, the balcony to balcony. So there was obviously discussion about what that should be. We wanted to represent a kind of residential building but we wanted to use a, you know sort of a lower height or a shorter distance, because we wanted to challenge obviously, if we put the balconies further and further apart, it makes it less likely for fire spread. So we chose. But we wanted to be realistic. Right, you know, there's no point putting making the balconies one meter high because no one to do that. So there's always that, there's always those discussions I say compromises, but but somewhat is a compromise trying to work obviously around the lab, what the lab can do, what is realistic, what is reasonably conservative, and all those discussions at the end, when you are limited with from resources, you need to make decisions and all these decisions need to have, like, some thought behind and some compromise.
Speaker 3:obviously, as Mike said, the main thing was what's the distance between the opening and the balcony right above, so the underside of the balcony and that was one of the main things I discussed. We ended up with some specific distance that we thought it was realistic and it also followed some of the literature review we did. So we wanted to be as much as possible, let's say, in line with some literature we found on that. And, of course, as Mike said, we wanted to look on the balconies themselves. So we decided to have a non-combustible facade. So the projection of the flames would affect, let's say, only the balconies themselves, not the facade. So we had boards calcium silicate boards on the facade, extending both the combustion chamber of the BS8414, let's say, the standard combustion chamber.
Speaker 1:And you mentioned some specific structural challenges. Mike, can you tell me more about where the balconies? Did you use some attachment systems? How did you attach them to the wall?
Speaker 2:In the end, we used a stacked balcony system rather than using them as an attachment. So what does it mean? Well, I mean there was firstly some practical questions. So we were a little bit concerned. If we used an attached balcony, we didn't know how severe the fire was going to be. We might have got failure of that and have to replace it each test.
Speaker 1:That would be fun.
Speaker 2:We might have got failure of that and have to replace it each test and, as you know, running a lab. All that takes a lot of time. It costs a lot of money.
Speaker 3:There are health and safety issues as well. We wanted to make sure that there is nothing internally in the lab.
Speaker 2:And, of course, and also different suppliers of boundaries, of different attachment systems. So in the end, if we were going to do it that way, we'd have to choose one, and this was not an assessment of an attachment systems. So in the end, if we were going to do it that way, we'd have to choose one, and this was not an assessment of an attachment system from a particular supplier. So we end up with a stacked balcony arrangement, which pretty much represented the fire that happened in South London. But there was practical reasons that meant we could have pretty good confidence that we could do the series of tests without having failure of the balconies and a lot more time and expense doing it. So again, it was a compromise that says it does represent a balcony system that we see on real buildings. No, it doesn't exactly test the performance of an attached balcony.
Speaker 1:yeah, you know I mean, we probably spend more time than necessary discussing those things in the podcast and I'm not so sure if it's super exciting to the listeners. But I find this element of discussion critical so that people who are not laboratory researchers they understand the way how the job is done. You know, because if you look from a side on any research project, you can pick up a hundred things that someone could have done differently. And even in our research we were getting oh this is a stupid, why have you done that? You should have done something else. And in the end it's all choices and ultimately you have to have one. You have to choose something.
Speaker 1:You're not really usually in a position where you can increase the parametric space of your tests. I I mean, of course you probably would love to test also an attached balconies and perhaps different types of attached balconies, but if you're already running a project that costs a lot of money and tripling it or quadrupling it just to add a variable is difficult. I sympathize with this. And also when you mentioned balcony falling in the lab, I would not like that in my lab.
Speaker 2:Yeah, there's expense, but there's also the time right. So the longer the more, the bigger you make the test matrix, the longer it's going to take, and people are wanting to be able to update guidance to things. So, you know, there's an element of timeliness and even this project is taken from it, it's an initial funding agreement to to where we are now. It's taken, you know, several years and and in industry, and that, of course, there's a desire to be able to resolve these things. You know and I appreciate that somewhat, that people are keen to hear what the results are and and what might be the response to it. But unfortunately, you know, I would love to wade the magic wand and say we can do it all in a day and we can do it for tuppence-half a knee.
Speaker 3:Do a lot more, check a lot more factors, a lot more parameters. It would be ideal, but at the end we ended up with eight tests and I think it was like a lot. Considering all these factors that we had to look at.
Speaker 1:It was a lot of Maybe let's go to the experimental matrix actually. So what kind of variables you were looking into between those sets of experiments, like what was maybe? Let's start with the baseline what was your baseline experiment? And then what has changed in the subsequent experiments?
Speaker 2:Yeah. So the baseline experiment was to take the 8414 crib, have a non-combustible balcony floors, have no balustrade, and what we want to do there is get a characterization. Now, as constantina says, we've got a really good handle on. If you put an eight, four, one, four crib, have a facade, non-combustible facade, so so what happens if you put the balcony projection there? What so? We had heat flux gauges, we had thermocouples and obviously we had heat release. We also had mass loss, although unfortunately the the load cell fires were quite severe and it got damaged. But because we're measuring heat release anyway, that wasn't the end of the world. So we wanted to get a baseline characterization of what does the crib in the combustion chamber, with no contribution other than flames being projected as it impinges under the balcony.
Speaker 3:And the BSA 414 crib actually provides enough, let's say, flames to extend to our balcony. And this is what we wanted from the start to have enough to extend.
Speaker 2:Yes, so the flames extended underneath the balcony and kind of a meter and a half from the balcony edge. So you know it's a fair. When the crib is burning at its maximum, you've got a fire, that's, you know, fairly severe.
Speaker 1:You could say it's attacking the balustrade actually.
Speaker 2:Definitely definitely attacking the balustrade.
Speaker 1:Okay, so that's the reference. And now, how did you build up the experimental setup to account for more variables? What were you looking?
Speaker 3:into One of the main tests that we did after it was adding the contents, or at least a representation of the contents, in a balcony. So the next one was to add some fuel load on each balcony, and we did it by using a crib incorporated both wood and some plastic that we had. So we ended up with adding this kind of crib providing heat, based on the survey that Mike mentioned earlier, on the work that Mike mentioned earlier. So we added this on every balcony to see what's the impact of adding extra fuel load on each balcony, and this was, let's say, the next step of our experimental work.
Speaker 2:And in doing that we also characterized the crib by itself. So that was burnt separately under the calorimeter. So we knew the heat release, we knew the heat flux you'd get from that crib and we could see that the flames from that crib were tall enough that that fire in itself would have the potential to spread fire from one boundary to the other. The crib was obviously a crib. Some would say, well, it doesn't really represent real fuel load. But as we know, the crib was, I mean, obviously a crib, some would say, but it doesn't really represent, you know, real fuel load. But as we know there's, you know the practicalities. We could have gone and bought some chairs or whatever, but they're more difficult to characterize.
Speaker 2:But we know cribs are repeatable, we understand them. But what we made sure was that what we did was place the crib at the front edge of the balcony, so rather than on the back wall. So we wanted to try and encourage the potential for fire spread between the cribs and we've concentrated the energy into a crib. And you look at real balconies, there's bits spread around. There'll be a palm plant over here and there. But as Constancino says, the next step was to take our non-combustible floor, non-combustible, with no balustrade, and what just this representative fuel load would that? How much does that contribute to the fire dynamics?
Speaker 1:We'll get to the results in a second. Let's just quickly wrap up the rest of the experiment. So, movable load, that's one variable that you're looking into. What else have?
Speaker 3:you been looking into. So the next one was timber decking. We added the timber decking in order to represent a combustible deck on each balcony floor, which was one of the main, let's say, objectives as well on this research project. And just like summarizing, let's say that there was also different balustrades. So we had two different types of balustrades, except the open one which Mike mentioned on for the base scenario. The open one, which Mike mentioned for the base scenario, we had laminated glass as balustrade and HPL panels for using, let's say, two different types of balustrade. And then we had the last parameter, which was a soffit, a non-combustible soffit underneath the balconies, let's say protecting the deck.
Speaker 2:It was a timber deck. So in the approved document there is a solution provided for if you've got an exposed, underneath exposed timber deck. One solution is to use a non-combustible soffit. And so the client was interested to kind of get a bit more understanding how that might work in combination. And now it worked. We did one where we had the soffit and a timber deck and then we had another one with a soffit, the timber deck and the HPL balustrade. So we've got eight tests and we've got several variables there fuel load content, two types of balustrade, two types of decking, the presence or not of a non-combustible soffit, and so trying to decide which combination of experiments to get the most value out of each one, knowing full well you know, as a scientist would say, you only keep change one variable at a time. But we know you know, financially, time-wise, that was impossible. So again, there's a lot of discussion, decisions that made to make sure we we could leverage the most value from each experiment and get the most learning.
Speaker 3:And we were able, let's say, to move from one to the other and knowing exactly, let's say, the contributing factors from the previous test. So we moved from the base scenario, we started building on that, so we knew exactly what was the impact from the additive load on its balcony and then what was the impact of HPL panels, Even though that we had, let's say, limited number of tests, let's say done. Finally, we were able to somehow quantify, also using the calorimetry, which was quite helpful to quantify what's the contribution from each factor.
Speaker 2:And we're also fortunate because although we had a general vision of what the experimental program would be, there was room to, as we learned from one to another, sometimes slight decisions were made. That says okay, at one point we were thinking about maybe we would or wouldn point putting the fuel load on that that. So that got removed. Oh, so we had. I think we had at least enough glass to probably do another laminated glass balustrade because we weren't quite sure how things would go. So there was some flexibility in the matrix, there was a general view of what we wanted.
Speaker 2:But within that, as we learn from each test, there was a bit of a logic. We should do this first. We understand that, then we can. From each test there was a bit of a logic. We should do this first. We understand that, then we can make a bit more of a decision about how we do the next one. So when you read the research paper and you think, or maybe they had that plan already on and it was like, no, it evolved as part of the learning process and that's again, that's the way sometimes these things have to be done.
Speaker 1:And it's probably the most efficient way. And, to be honest, let's talk findings and let's talk conclusions, because I don't want to run out of time before we tell our listeners the conclusions of this study. So I see you have been gradually increasing, let's say that, the fire load, the fuel load on the balconies, by introducing movable load, and the timber decking, hpl, the source of combustible material. So you've ended up with, let's say, different types of balconies with different relative combustibility, however one would define that term. Did the final ladder of performance reflect your initial expectations of which is going to?
Speaker 3:be worse. We definitely knew from the start, I think, or at least during the experiments, which one will be the worst Because, as Max said, we had the HPL and Tiber decking combined and we didn't even have to add the fuel load. We thought like it's going to contribute a lot, so it's going to increase a lot the heat and all these measurements. And we were right on this one at least, because we ended up with stopping. That was the only test actually. We stopped during the test due to safety and also due to really high heat measured on the calorimetry. We were afraid of damaging the equipment so we just had to stop. I think it reached about 14, 13 or 14 megawatt fire, which is like massive, and it was fully involved. All elements were fully involved in fire on that one. So I would say definitely that one was the worst case that we tested.
Speaker 1:And the basic ones like movable load and laminated glass. First, like I would not expect, anything happened with that technology. So how did the test look like?
Speaker 2:Yeah. So when we had the movable fire load, by itself the fire spread to what we call balcony one, but didn't spread any further. When you look at various metrics heat release rate, other heat fluxes and that the contribution of the laminated glass well, in terms of energy release, hardly anything. Fire didn't spread upwards. Eventually the panels did fall away. We also had to make some decisions about the way that the balustrade was constructed, with different connectors. So again, there was some choices made, sort of 30 minutes into the fire. Sometimes one of the panels would fall away. The laminate might be still burning and there was a potential that if it dropped onto a lower balcony, that may have, you know, if you had a timber deck or something there, you know you might have got some fire spread, as Constantino says. You know, as we added, you might say unprotected timber HPL balustrade, then the more burny stuff you put on, the bigger the fire. So there was kind of, you know, no surprise.
Speaker 3:At least for the laminated glass test.
Speaker 3:We were interested also how the edge of the laminated glass was protected, whether it was protected, how it was protected and whether there was actually droplets from the PVP layer in between and whether it was, let's say, sufficient. We didn't have any measure, let's say any metric, actually taken for the droplets. We had some small pull fires definitely. We observed some pull fires downwards, but the location of them was not consistent during the test. So we had, like some, let's say, closer to the, to the balcony below or slightly far from the edge of the balcony below. So we tried the main thing was we tried to have the protection of the edge of the laminated glass as realistic as possible, but also we tried to have the worst case scenario. So we had the edge unprotected, completely Uncapped, uncapped yeah, because many people in industry they use a kind of metallic cap on the edge, on the bottom edge of the laminated glass, and we didn't want to introduce this during the experiments just to see whether droplets actually and the other thing was the way we mounted the glass In some balconies.
Speaker 2:The glass would be in a channel, so the bottom edge would be inside a channel. We used a fixing system where the glass was. I can't remember the gap, but there was a gap from the floor to the underside the glass that was uncapped. And so we, again, we wanted to be conservative in terms of the potential for fire spread. Talking to industry, they would say a lot of times actual practical projects, it's likely the glass would be in a channel, therefore that bottom edge would be not, wouldn't be exposed.
Speaker 1:So so yeah, decisions again, decisions are made to try and and did you have the scenario in which you do not have a timber deck but you have an HPL ball straight? No, no okay.
Speaker 2:No.
Speaker 1:Okay, so from the experiments done with the HPL ball straight, would you say it contributed a lot. Comparing them with each other's?
Speaker 2:Yeah, I mean. So we did the one with the timber deck by itself, so we did the one with the timber deck by itself and we did the one with unprotected timber deck and then we did the one unprotected timber deck and hpl balustrade. If you look at the heat release rate, the initial curve, it would look to be mainly the timber deck because the curves pretty much line each other. But of course then, as constantino says, there's a lot more energy release from the with the HPL as well, and that just kept that, just kept going up, whereas the timber deck peaks and starts to decay, so that the HPL contributed to the overall energy release. In terms of the initial fire spread, it would seem to be having an unprotected timber deck is the main mechanism of fire spread in those early stages.
Speaker 3:Yeah, because at the end it was observed that using the timber deck, it was one of the main contributing factors to fire spread. Once you have a timber deck completely unprotected, then it gives a lot of energy, a lot of, let's say, a rapid fire spread to upper layers. And we had this soffit. When we had this soffit protecting it was actually a factor delaying fire spread, not completely avoiding the fire spread, but at least delaying for some period of time.
Speaker 1:And the performance objectives, like how you were assessing those. When would you say the solution is safe? Let's say what were the criteria?
Speaker 2:Yeah, so we used a couple of different metrics. And we've got heat flux, we've got time of fire spread between either balustrades or the content or the deck, so we used one. Well, and we've got heat release. So one metric would be to apply the FIGRA type approach that we use for standard testing. We used a time of fire spread between elements of the construction and there we used a 15 minute.
Speaker 2:So when you look at some of the older work that's been done, there was some work done in the 60s and 70s and it was considered that if you gave the people in the building, the fire service, 15 minutes window between story to fire spread, that would be sufficient for people to move away and the fire brigade to be there and start tackling it.
Speaker 2:We're not saying it will be 15 minutes, but that was so.
Speaker 2:We used that 15 minute because what we wanted to say is if we can show that the contribution to the balcony is no worse than we might expect if you had a building with a whole bunch of open windows.
Speaker 2:That we're again, we know we can get fire spread. And then we also used a heat flux and we used 12.6 kilowatts per meter squared, which again is a is a number that you'll find in sort of design guidance as a means to look at fire spread, building to building fire spread, and so from that you can incorporate all those different metrics and from that you can rank, and so you end up with, I mean, not surprising if you just have a non-combustible floor with nothing else, that no, there's no fire spread. If you just have fuel load, as our cribs, it's slow enough or doesn't spread any further. That's fine. And then, of course, the other end of spectrum if you have timber deck and hpl, it every. All the metrics basically show it's quicker than 15 minutes, the heat fluxes exceed the heat release and the figure you would say this doesn't seem to perform as we would like.
Speaker 3:So and then you get a spectrum in between adding that we wanted to see whether the, the upper, the top, let's say balcony would be also involved in fire, because, finally, we, we know, we expect that probably the, the one right above, will be involved quickly enough, let's say. But the issue is whether and when exactly the, the top balcony will be involved in fire as well, and and that's one of the observations we need to make during that, considering the 15 minute period of time that Mike mentioned. So all these kinds of observations are taken into account in order to analyze the performance, not exactly stating which one is safe or not, but at least comparing in between those two and, as Mike said, to come up with some sort of rank, not ranking, but at least a sort of understanding of which factors contribute the most. On the upper water fire spread.
Speaker 1:So maybe let's reiterate them. So how would you rank those factors in terms of how much they contribute, from the least to the most impactful?
Speaker 2:So, as I said, obviously if you've got nothing combustible, that's the best, but if you've got laminated glass as a balustrade, that doesn't make much difference. That still seems to be acceptable.
Speaker 1:So laminated glass not that different from a completely non-combustible system.
Speaker 2:And then, and then the timber deck with a soffit that performed. I mean, as Constansino said, it slowed. It didn't necessarily prevent fire spread, but it slowed it to a point where one would think if there would be time for fire and rescue services and people to move away, evacuate. Obviously there's other parts of the fire strategy that need to work as well. We're not just relying on that and that would seem to be still in the realms of At least during experiments, it slowed down, at least, if I remember correctly, for 20, 25 minutes.
Speaker 3:Maybe during the test the fire spread in comparison with the timber, unprotected timber only, which was obviously the next and the worst case scenario.
Speaker 2:So that's kind of the ladder, and I say that in ways not surprising One might. If you'd ask someone like yourself just to rank them, you would have probably looked at it and went, oh, there's a lot of stuff there combustible that would surely be worse than something that doesn't have that much.
Speaker 1:Okay, very good, very good. One thing that's very tricky for me in this consideration, now that I'm looking at the whole thing One thing that's very tricky for me in this consideration, now that I'm looking at the whole thing, is that you may have power over the structure of the balconies, but you really have no power on what people put on them. So you found the timber deck is an important factor, but I can imagine if someone keeps two electric scooters on their balcony or just has a large cushioned piece of furniture item, that's also perhaps something that could strongly contribute to that right.
Speaker 3:But this one is also a matter of management as well of the building. We see, especially now, that there is limited space internally within the apartment, especially in block of flats or high rise buildings. Internally within the apartment, especially in block of flats or high-rise buildings many people tend to store a lot of materials in the balcony, which is what people do now, and I think it's going to be a bit difficult to stop this unless there are strict measures, strict restrictions from the management of the building. Obviously, if you stack e-scooters or so, you have different hazards to look at. But at the end, for this specific study, we want to have some, let's say, base scenario for contents. We wanted to look at maybe some furniture that people have on the balcony or this kind of fencing. Sometimes they have fencing on the balcony, bamboo type fencing. So these kind of things we wanted to look at. Obviously, then you can build on that, but I think it could be like some future work.
Speaker 2:Yeah, I mean I'm not sure when you said there's a lot of stuff for people. When you look at Sam's survey and when you look at balconies, I mean a lot of balconies don't have that much. So yes, going back to it, I quite agree it is a management. I mean if you've got rental flats you might have some rules about what you can and can't put on your balcony. You obviously have to police that. I mean, obviously the balcony provides a utility.
Speaker 2:I mean one could say let's just not allow anyone to put anything on their balcony, but then you're taking away the utility of the balcony. Want anyone to put anything on their balcony, but then you're taking away the utility of the balcony. So you've got to balance the, the fact that it's there for people to use and they might want some plants, they might want a picnic table and that. But if someone decides to put a three couches and a wardrobe on it and a motorbike, you might kind of go. Maybe that's too much. Obviously there's advice or requirements in terms of the activities. So so many high-rise residential buildings might have clauses about whether you can put a little barbecue on that and that might be not. So there's ways maybe to police it. And clearly the other way is kind of communication education.
Speaker 2:So if the residents or the users of the building understand why certain things might not be a good idea to do on their balcony, hopefully that engagement with the occupiers they'll be able to understand why do we set certain rules and that and how it will protect their neighbours from particular consequences so so there's, you know, there there's, there's rules, regulation and that, but there's also education and understanding of the building and I think those and an appreciation that people want to and should be able to use their, their balcony, within reason to do, to live, live their life and do and have activities on their balcony, because I mean, in the end it's a large project and the the big findings are pretty much in line with what you could have expected, kind of, but you've managed to quantify them.
Speaker 1:Like Konstantinos, you said laminated glass. You know that it probably is not the worst type of material, but it doesn't have the class.
Speaker 3:I'm actually quite happy with it. And obviously, okay, if you ask me now, probably I would have done some things differently, or I would have asked, let's say, as a consortium, to do some things differently. I don't know. However, I believe you know, looking at the big picture, we were able to provide and I think that's the most important thing we were able to provide, as you mentioned, evidence on how things work in such situations. Evidence on how things work in such situations, maybe limited for now, but I think it was especially large-scale. Tests are not easy to perform, they are quite expensive, they are difficult to organize, but I think with this sort of experimental work, we were able to provide some evidence which was not there to use until now.
Speaker 1:Find that in fire science there is a lot of stuff that has been placed in the guidance documents, in the requirements, in the laws that were obvious but never evidenced, and today we call them magic numbers, you know, and they are kind of a source of a problem. So I find immense value in providing experimental evidence for even the well-formulated opinions on the far behaviour of different things.
Speaker 2:I agree with that and as an example is, following Grenfell Tower, there's been the introduction of an assessment process for facades of buildings and there's a document called PASS 9980. Within that there's some guidance on how different factors might contribute to fire spread. Now it would seem to be that those are based on the wisdom of the authors and that and they seem sensible and obvious. What we've done is hopefully with this experimental program is again produce evidence. It doesn't contradict what's in that guidance, it supports it. But we've now got some more evidence from some research that allows people to have more confidence that the use of that document and the guidance there is supported by some, some actual evidence so it's not really about discovering new things.
Speaker 1:Well, you could have discovered something along the way this is how things happen in fire science but it's not. It was not about discovery.
Speaker 3:It was more about narrowing down the uncertainty, I guess yes, yeah, yeah, I think that's yeah, I think that was one of the main outcomes of these resets and I think what provided at the end, what we delivered at the end, it was some sort of assurance that some, let's say, indications that we had before they are actually right or, let's say, they are actually accurate enough until now.
Speaker 1:But come on, it's fire science. There must have been something surprising. Anything surprising happened.
Speaker 2:Well there was a little bit. Some of the fire dynamics was quite surprising when we had the combustible fuel load the crib and when it ignited the one on the ground floor, how it changed the flame dynamic around the balcony, which was when you think about that little crib versus the 8414 crib one I looked at and quote. Oh well, it would, it will add a little bit of heat release, but it actually changed the way that the flames spread around here we go.
Speaker 2:Fire science, here we go and that was, that was a surprise. I genuinely looked at and went, oh, that I wasn't expecting.
Speaker 3:Yeah that was surprising during the test. And then just to note on that, the, the small crib, the added load, it was about a 900 kilowatt peak and obviously the crib of the BS8414, if you compare it it was about four megawatt or so or similar. So you're seeing like this deflection of the flames, instead of coming out projected out in front of the opening, deflecting on the sides. It was quite interesting.
Speaker 1:I knew there has to be. It would not be fire science if it all went as planned, so I'm relieved it gave a human dimension to the perfect project. Okay, guys, any final words of wisdom.
Speaker 2:We appreciate there's limitations. I mean quite rightly, rightly, someone would say well, what about wind? That we in all our cases had no wind involved, and we know that wind can either potentially enhance fire spread or it actually actually can reduce the likelihood of fire spread. It all depends on a whole bunch of factors. We had only one balcony size, only one distance between the balconies, only three balconies, so there's always those constraints. But, as Constantine says, I feel as though we've done some. The work has contributed to an understanding it would have seen that it's stuff that not, you know, hasn't been done before be caveated with an understanding that, like everything, it can't, it can't cover everything else, but it, I'd like to think it's added a little bit to our pile of knowledge that people can use and makes. It is there as a tool to help make, make other engineers and regulators and to make, hopefully, some sensible um and informed, you know maybe the but it's informed right.
Speaker 2:We use that to make informed decisions. We know it doesn't cover everything, but it allows us to not have to quote guess, but we use that to inform a decision. But it's part of our toolbox and there's other things that one might need to think about that not just use this in isolation.
Speaker 1:Science-based regulations is definitely something, and evidence-based regulation is something that I can definitely sign up for.
Speaker 3:Just to add a last thing. I think, expect that the regulation and how you can use all this for technical guidance. It's also, I think, important for manufacturers of different materials. They can also base their production, their manufacturing processes, their development of different materials on all this kind of experimental work, so they can understand even better how the use of their materials is done in reality and what's the impact, finally, on fire safety.
Speaker 1:And I can just say I'm super jealous 14 megawatt fire calorimeter, three-story tall experimental setup in the lab. I really would love to have heat release rate measurement capabilities in my lab, maybe one day. Okay, guys, thank you very much for coming to the Fire Science Show and sharing their research outputs. If the papers come up, I will keep updating the show show notes. I'll link people to the ones that are already published. Yes, and hopefully the pipeline will deliver more.
Speaker 2:That's right so there's a, there's a paper by sam bryan on the uh and myself and ewan on the fuel load. Uh, there is um we haven't talked much about. We did do some work on looking at the structural capacity of the balconies. Danny Hopkins did analysis using equivalent fire severity and we're going to publish some stuff at Interflam and there was a little bit of again. I haven't mentioned all the work.
Speaker 2:There was some work again with a bit more collaboration between us and the University of Edinburgh, between us and the University of Edinburgh. Also, we had a collaboration with University College London who provided kind of an internal external reviewer to challenge some of our thinking and that sort of thing. And there might be, hopefully, a paper in Fire Safety Journal on the small-scale laminated glass. It's still going through the review process. We like to think there might be a couple more papers that people might find useful in the future on the small-scale laminated glass. It's still going through the review process. We like to think there might be a couple more papers that people might find useful in the future.
Speaker 1:It looks like that and I'll keep you informed. Thank you very much for coming.
Speaker 3:Thank you so much for having us.
Speaker 2:Yes, enjoyed it. Thank you very much. Thank you.
Speaker 1:And that's it. I hope you've enjoyed this balcony conversation. One thing I have to reflect on, something I've said in the episode when we were discussing the laminated glass, I said it's obvious that it's not a problem, and I found a problem with my statement because perhaps someone used the same wording about the material that ended up on Grenfell Tower, perhaps. So even if my gut feeling tells me it's safe, I guess it's not enough for modern fire safety engineering. I guess that doesn't cut the deal and all of us can be wrong about the material. Therefore, we really need those sound, well-performed, well-justified, well-created experimental programs to have a true confirmation that behavior of something is as we expect. And this is what these guys delivered to us in this balcony project. We had expectations that more combustible material you will have on the balcony, much worse it will be. They shown that. We had expectations that if you use laminate glass, it's not going to add that much to the balcony fire load. It did not. What they did more is they quantified some of those. They've measured the spread time, they've measured the heat fluxes. They've measured how fast the fire will get through the balcony and how big the fire will be in terms of heat release rates vigorous and so on. And then you can use those metrics to have a relative comparison between the solutions, rank them in the ladder and from that you can work with some really good fire safety engineering. That's why I really appreciate this project. I think money well spent and the research we did not have we have.
Speaker 1:Now. Many answers are given to us in form of experimental evidence, experimental proof. It's not a hypothesis anymore, I think. Today it is knowledge. So very happy to have Mike and Konstantinos in the podcast. We had a lot of good discussions during the OFR events and this one was one of the better ones. It was a grand finale for me of the event and I'm very happy that I captured mike with the microphone and got this interview recorded for this week. That would be it in the fire science show and if you are looking for more fire science, you will find it here next wednesday. Thanks for being here with me. Cheers, bye. Thank you.