111 - Experiments that will change fire science pt. 7 - CodeRed with Panos Kotsovinos

Speaker 1: 0:00

Hello everybody, welcome to the fire science show, the experiments that will change the fire science, because the one we are covering today is a fairly recent advancement. The history will judge if it was a breakthrough or not in our understanding of compartment fires, but I got a good feeling on this one. It's a really, really impressive effort. My guest today is Dr Panos Kosovinos from Company Arup, who has led a research called Code Red, which was an experimental investigation on full-scale compartment fire dynamics in a very large open-tank compartment with combustible CLT ceiling and gluolum columns, combining the experimental approach that has been used previously to research the traveling fire behavior in previous experiments carried out in Poland and I was a very proud member of the research team back then when we carried these OBORA tests in Poland and I had Professor Gilaer Morain on the podcast talking about them as well In this round Arup has tested combustible ceiling to investigate how that changes the fire behavior, how changing opening factors, how providing some encapsulation, how introducing water mist will change the fire behavior. A very impressive experimental effort carried by the large team at Arup, surrey Bath, france, and Imperial Haislep at the Imperial College London. You have to hear it directly from Panos. So yeah, let's spin the intro and jump into the episode. Welcome to the Firesize Show. My name is Vojci Winksiński and I will be your host. This episode of the Firesize Show, as all of them in this year, are brought to you in collaboration with OFR Consultants, a multi-award winning independent consultancy dedicated to addressing five safety challenges. Ofr is the UK's leading fire risk consultancy. Its globally established team has developed reputation for preeminent fire engineering expertise, with colleagues working across the world to help protect people, property and environment. Internationally, it works ranges from the Antarctic to the Atacama Desert in Chile, to a number of projects across Africa In 2023,. Ofr is growing 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 Code Red experiments. Hello everybody, welcome to Firesize Show. I'm today with Panos Kosovinos from Arup. Hello, panos. Hi Gostek, good to see you, man. It's such a long time I haven't seen you in person. Pre-covid times. This is absolutely crazy.

Speaker 2: 3:03

Yeah, it is crazy. We are in post-COVID era. There are a lot of people that will sing it's other all the time in conferences and events, and we will start from the beginning.

Speaker 1: 3:14

There's a lot of catching up and we will surely need to do that. But I brought you here not to discuss pandemic, even though that perhaps is interesting and may even get me more clicks than fire safety To discuss the super interesting project you've carried out in COVID times in Syrip in France by Arup with collaboration of Imperial College London and the Hazelab Group, and that the Code Red experiments. The experiments are huge, open facility experiments with CLT, c-lings, gllam columns Super exciting thing. Please tell me how the idea of these experiments were born in your head. Why did Arup pursue this type of massive experimental plan?

Speaker 2: 3:59

Gostek, first of all, I would like to thank you for inviting me and I would like to thank the audience for withstanding my Greek accent and being interested on this topic. The idea for Code Red really has its inception back to the Polish or Bora experiments that were carried out by Imperial College with support from myself, from Arup and Mohamed Hedari from Syrip that I know Guillermo Reyn covered before in one of your episodes. These experiments happened in Poland and you were also involved, so you know very well Happened in Poland, initially in 2017 and in 2019, being planned from back to 2016. So in these experiments, we had an open plan compartment and we started traveling fire dynamics, and the question that we had back in 2019 was, now that we start to better understand traveling fire dynamics in non-combustible compartments, what would happen to an open plan compartment that would otherwise experience a traveling fire if it was non-combustible, if it had an exposed timber ceiling?

Speaker 1: 5:10

It all ties back to the traveling fire concept. I know Arup was one of the first users of the traveling fire for structural design and, if I'm not wrong, it still is an important part of your structural design routine. So the idea was okay, we did these experiments in Poland and you write a rost experiment on traveling fires with Guillermo and we've covered, I guess, a lot of Bora, three great experiments, had great fun doing them with you guys in Poland. And now you are faced with the idea that, okay, you have exposed combustible surfaces in the compartment. They will participate in a fire in one way or another and you would like to understand if the traveling concept still holds, how is it changed? That's the idea.

Speaker 2: 5:55

Exactly, yeah, and as always in industry, this type of problem is normally said by project requirements and demands. So, going back in early 2019, we completed the second series called Dexto in Poland of the non-combustible experiments. That was early I think it was around April time that we did Dexto, and a few months after we were faced with a project in Arup where a client had a concrete office building and they wanted to add additional parts of the floor plate with a silty ceiling, and when this question was posed to us, we immediately understood that there is a lack of available design fire tools so that we can model this problem. Because if it was an combustible building we know we have traveling fires we are able to analyze the problem, but once you start adding timber, then this affects the fire dynamics and there was an uncertainty on what would that would mean for the design fires that you would assume for this building. And it happened that the person that led the first Oboro experiment in Poland, x1, which was Eglera Kauskaita, was an Arup funded PhD student at Imperial and then later joined Arup, and we both of us were working together on this project. That was concrete and they wanted to extend with a timber ceiling and also given our background on traveling fires, we identified all the limitations that the traveling farm methodology has with respect to application to timber buildings and identified the need to undertake further experiments to understand the issues.

Speaker 1: 7:39

When I'm discussing the timber in fire or in general combustible structures in fires, it often comes back that in this case you really cannot take away the fire from the structure. You know the fire and structure are the same thing. You cannot solve them in complete separation because they just interact with each other. The growth of fire will be different and the consequences of that fire to the structure will be different than it is all tied in a very crazy feedback loop. So, designing Code Red I know that there were four experiments carried out, or you know what, maybe. First let's go back to the glorious Abora times and let's just briefly mention what the compartment was and then we can go into specificities of Code Red. So let's first introduce the battleground. The Compart Ed was used for tests Because, if I understand correctly and what I've saw in the pictures, you've built a pretty god and impressive replica of our beautiful Abora somewhere in France.

Speaker 2: 8:42

Exactly, yeah, somewhat simplified replica of the Abora building. Of course a building in real is not perfect and has a symmetry. For example, the building in Abora didn't have an exact same compartment height all over the compartment, so in the end it was a replica that was very close to the original but not exactly the same. Also, the original one had in some locations some small beams that we didn't have in the replica compartment. But of course using that facility in Poland to undertake the testing with CLT would have been impossible because of the complexity of introducing the CLT ceiling inside the building. So we had to build something from scratch.

Speaker 1: 9:29

Yeah, I was in the middle of this discussion and it's a miracle we haven't collapsed the Abora with the first three experiments, and if we introduce CLT in that, oh boy, that would be a collapse. The Abora building was 34 by approximately 10 meters wide, giving something like 350-ish square meters of compartment area. The height was approximately three meters a little less than three meters, and the original building was basically built with non-combustible brick and mortar and some sort of concrete ceiling. It had multiple openings. There were rows of windows on two longer walls of the compartment, there were large entrances on both short ends of the compartments and we've built a wood creep that spanned all along the compartment inside. So it was perhaps the biggest wood creep I've seen in my life. Yes, indeed. Well, I guess all these key characteristics of the building were copied and transferred into the French one, but you've used CLT ceiling, so tell me about the structural design of the ceiling of the building.

Speaker 2: 10:40

Yeah, more or less the building was replica. With regards to the area, the Polish experiment were 380 meters squared, while the code rate was 352. So it was slightly less. So some of the dimensions were normalized a bit based on the panel dimensions and because of the optimization in terms of budget. So they were not identical but very similar for the far dynamics phenomena that we wanted to capture. The facility design was done in collaboration with Serib and Imperial College and really we tried to replicate the design of a Bora as much as possible within practical limitations and I think in the end we were really close to that.

Speaker 1: 11:32

Seeing the pictures that were posted, I was like whoa, that's almost identical. The only thing you didn't capture was the pond with beavers next to the facility. Exactly that detail was not captured. But also building the building from the scratch allowed you to do interesting things like facade extensions and playing with openings, which we will come soon in the episode. So I'm bringing this up because it's very rare that experiments on such scale are pursued in the world of fire and because these experiences are so scarce, it's important that we learn from each other. You know and we understand what can be built, how can you utilize it, how to plan such a huge experiment. So even we didn't touch the findings of your experiments yet. I think there is already a huge value in learning how scientific experiment of that scale can be planned and conduct. But definitely there's a great explanation of both the original Bora and the French Bora in the European link in the show notes and for non Polish listeners, a Bora is a Polish word for cowhouse which somehow fitted the Spanish leaders of the project very well because of the very nice and distinct are in inside, and the best part of it was how well 1920s cowhouse represented the modern office structure. This still is surprising to me, how close the cowhouse was to an open plan office. Anyway, let's move to code red. So for experiments you have planned out, tell me what you've done. More importantly, why you've planned it like this.

Speaker 2: 13:08

Yeah, also let me comment. You said it's really important that we should not underestimate the level of planning required to undertake these experiments. To give a sense of scale, I said that the initial idea came up in 2019 with the project application. That project opportunity in the end did not progress, but the idea was still progressed and through internal ARUB funding we managed to undertake the work bit by bit, in stages. The planning of the project before any experiment was undertaken took about at least a year. So the whole 2020 during the pandemic, was involved into planning and that was several people within ARUB, serib and Imperial, including a very large project team at ARUB, a steering group with different type of people involved from different industries, so that they can guide us on what is useful for the industry and for the approval authorities or even the insurance sector. So the amount of planning was even more than we were expecting originally.

Speaker 1: 14:12

And you still had a head start because we've done the three experiments in Poland, so it already went through the process of designing the data acquisition systems, the creeps and everything beforehand twice. So one could say you had it easy, and still it's an immense effort, right? Exactly.

Speaker 2: 14:30

So we had also the previous experience from Bora that helped us and still required all this time of planning in order to go from doing an unaccompanible experiment to something from CLT, which required a lot of technical details to be sorted, including the selection, for example, of the CLT itself. That we wanted to have CLT because our focus was purely on fire dynamics and we only wanted to see how the impact of fire would affect the fire dynamics inside and outside the building. We wanted to have a CLT that wouldn't be susceptible to blue-induced failure so that there is not any uncertainty that would impact the results of the experiments. And there was a long trial and error process through a furnace testing of CLT samples in order to derive the sample that we are up with.

Speaker 1: 15:29

So this was the reason behind using melamine, urea, formaldehyde glues and the lamella thicknesses, due to not have a char falloff in your experiments.

Speaker 2: 15:39

Exactly so. This was a potential risk that we wanted to avoid because studying a char falloff was not within the primary objectives of the experiment and therefore we thought that any potential char falloff would potentially impact the value of the experiments in terms of overall fire dynamics and being able to compare between different experimental setups, as we would introduce an additional uncertainty that could impact potential comparisons.

Speaker 1: 16:08

Okay, that's interesting. We'll touch back on that when we join the conclusions, because this definitely has a significant impact in using the study for design or guiding design. Please give me the details. Come on, man. It's the third time I'm asking Experiments one to four. Let's go the details is so.

Speaker 2: 16:25

The first experiment was identical to the obora building but it had a fully exposed timber ceiling, 350 meters squared, and also two glulam columns. One was located around the middle of the compartment and the other one around the end of the compartment. And an additional aspect that was not included in the previous abort experiment was that we also included a protected steel column that was positioned in between the two timber columns, so between the middle and the end of the compartment. And this protected steel column was added following some similar lodging bit by bit out in previous post-lassovr experiments in the BRE in the 90s, so that one can compare the relative structural versatility between a combustible member and a non-combustible member.

Speaker 1: 17:14

So the crib layout would be the same as in X1, x2 experiments. The only difference would be the glulam columns and the COT and minor differences in structural design of the columns and then minor elements in the compartment and opening factor, the same as in X1.

Speaker 2: 17:31

Exactly, yeah, and X2 the fuel load was exactly identical to X2 part one because in X1 there were two layers of fiberboard included in the wood crib that was not included beyond the first meter of X2 part one and COT load exactly the same fuel load design as X2 part one. Then in COT load, as you say, in the ventilation was exactly the same as in the ABOR experiments. Then in Code Red 2, we reduced the ventilation by half and this was in all four sides, because all four sides had a number of openings and particularly on the shorter sides, at the front and back of the compartment there were very big openings that were similar to large doors. And in the second experiment we reduced the ventilation by half, almost approximately the same in the different side of the compartment. So the ventilation was reduced in all four sides.

Speaker 1: 18:31

But still, as I remember, x1, or O'bora as we prefer to call it. I'm not sure if Guillermo approves the Achille of O'bora, so let's call it O'bora. If I recall, in O'bora we had really really huge openings.

Speaker 2: 18:45

If we look at the opening factors and there's something that also there is a paper by Harry Mitzel published recently, a review paper by Harry Mitzel on Fire and Materials Journal that was published on the same special issues, the Code Red experiments that looked in detail the opening factors of the experiment and whether they would fit into the ventilation control or fuel control regime. So in Code Red 1 that had the identical ventilation to O'bora, we were on the fuel control regime and once the ventilation was reduced by 50% we were on ventilation control in accordance with the harmathy criteria and in both cases we're in the near the ranges of the transition between fuel control and ventilation control. In that we had a lot of ventilation in O'bora and in Code Red but even that amount of ventilation is nowhere near typical open plan offices. That would have much larger percentage of ventilation in comparison to what we started.

Speaker 1: 19:44

So in real offices you would expect that you would be more into the fuel control regime if the windows fail, that is if they do not fail, or the windows, or the facade, yeah, or the curtain walls, for example. Okay, very interesting. I'll try to link to Harry Mitzel's paper on that. Factors as well in the show. It sounds like a very good research and I think I saw his presentation as a Berlin on that. It was really nice. Okay, two was the ventilation effect. Now number three what was?

Speaker 2: 20:14

it In number three and four. They were funding internally after one and two carried out, because initially we had some funding only for two experiments and then in the progress of undertaking one and two we were able to find additional funding internal funding and progress with three and four. And the focus, after doing one and two and understanding fundamental for dynamics phenomena, goes on protection and mitigation measures and therefore a Code Red III focused on the impact of a water mist on controlling the fire with timber-fledged compartment, and the Code Red IV looked at the impact of encapsulation and involved protecting 50% of the ceiling with a fire-rated encapsulation.

Speaker 1: 20:58

Fantastic. Okay, let's move to the conclusions, because there's so much to talk on. So the idea behind the experiments was to find what would the fire dynamics be in such a case of, let's say, traveling fire, but perhaps accelerated by the fire of a ceiling. Tell me, how did it end up in Code Red? I Did it surprise you, and to what extent it was surprising.

Speaker 2: 21:22

It really did. Yeah, before the experiments, we were making beads internally about what would be the spread rate and really I think, as always, science tends to surprise even the most informed people. Of course, the background for dynamics and the fundamental phenomena we would expect what happened, but it was difficult to calculate exactly what would be the final outcome, as it really depends on so many parameters.

Speaker 1: 21:50

My guess is, it was faster, but the question is how much faster and to what extent faster, right? Yes, exactly.

Speaker 2: 21:59

So, as you say, what we observed was that once the timber ceiling was ignited, the flames spread much faster than in a non-combatible compartment, particularly because of the very rapid flame spread through the ceiling itself. So once the ceiling got ignited, the flames reached the other end of the compartment within a couple of minutes and, having the ceiling ignited, this led to, as one would expect, high heat fluxes to the wood grip below and very rapid ignition of the bed below the ceiling as well, which meant in the end that the fire development that we had was very rapid. We calculated the fire growth that was similar to an ultra-fast fire and essentially it had the characteristics more of a full-developed fire, rather than the typical traveling fire phenomena that we experience in non-combatible compartments, where one part of the floor plate is flaming and moving gradually, even at an accelerating nature, as we observed in a board up there. Gradually it was the end of the compartment. And another observation was that of external flaming that was much greater than in the non-combatible experiments and covering more windows simultaneously in comparison to the non-combatible experiments.

Speaker 1: 23:22

I have a question for that. And did you observe like independence spread on the ceiling and then the grip followed, or was it just one front moving ahead?

Speaker 2: 23:31

That's a very good question. Yeah, thanks to the very intelligent camera setup that was done by our team at Imperial College the PhD students that were the ones that actually traveled to undertake the experiments during the pandemic. So we owe some huge thanks for that and the knowledge that we gained from undertaking the board experiments and how not to burn all of our cameras and completely get as much data as possible for the leading and railing edges of the fire. We were able to visually observe separately the leading and railing edges for the CLT and the fuel bed and being able to understand how the flaming of the CLT affected rapidly the flaming of the rib.

Speaker 1: 24:19

I'm really intrigued now but it's difficult to do this interview because I was there in Obora and I have many things that we call from memory, so I need to be careful to explain to the listeners what I'm thinking about. So in Obora the idea was we have 30 meter long compartment, we have 30 meter long crib on the ground of the compartment, like literally 30 meters of wood crib, continuous wood crib. We set the fire at one edge of this crib, it forms, let's say, a line fire and this line starts moving towards the end of the compartment and it starts with a very small fire, then gradually increases in size, increases in size. Eventually we reach a point where it just eats out meters and meters in matters of seconds. Because the end of Obora X1, the acceleration of the fire was really intense. Maybe one minute to cover the last 10 meters of the crib, yeah, but you could really distinguish this linear, this line that marked, you know, the beginning of the fire. Now the question in your co-treat experiment, did you still observe the line moving? Or perhaps at some point you've observed like a flash over behavior? Like a flash over, as we would classically define would be when the fire changes from surface to volumetric combustion in the compartment, meaning that all materials ignite at the same time. The theory, of course, was devised for small compartments where you can ignite the whole compartment altogether. Now the question is with the ceiling being on fire, did the floor like at some point ignite altogether, at the whole length or till the end? You've observed this moving front, at whatever velocity.

Speaker 2: 26:07

Yes, that's a very interesting question, rostek, and I think this is where the understanding from fire dynamics in open plan compartments and with timber ceiling starts to affect our knowledge of traditional fire dynamics from smaller compartments and where we start having divergence between different conditions that make it much more complex. So in an open plan compartment, what we show is that we don't have conditions that can easily be categorized as in, let's say, the traveling fires domain or the typical post-flash overfire domain. We have characteristics that resemble both types of scenarios and naturally, depending also on the different times during the fire. With regards to your specific question, on the line of the fire, when we had the exact same ventilation as in a Bora and the fire was more fuel controlled, we were able to observe a more typical line fire, as you would expect, that just was spreading very fast and the surface of the woodcrip was flaming and then was spreading very fast throughout the compartment. In Code Red 2 that we reduced the ventilation and we were more in the ventilation control domain, what we saw was that the flames spread more rapidly at the end of the compartment where there was more available ventilation, and then back to the middle of the compartment and the shape was not as much as of a line fire as it was when more ventilation was present. So really we had a mix of characteristics.

Speaker 1: 27:51

We are simple scientists and we know like to put complex physics into certain shelves on our library, like these are the flash-overed fires, these are pre-flash-overed fires, these are post-flash-over fires. Unfortunately, physics does not respect our categorization of fires. So it's very difficult to shelf those fires, especially for the first time you observe behavior that does not fit either of them and, let's face it, like ventilation controlled fires. This is something that perhaps needs 50 or 100 more years of research to be fully understood. You may recall there was an experiment in Cardington with like 24 meter long compartment which was also some quality, even a traveling fire experiment that was not considered traveling back then where they had an opening on one end of the compartment and what they ended up with was what the fire went through hold of the room, but then the flames moved more towards the entrance because that's where the oxygen was, which is logical, and they kept burning at the entrance. So the back of the room was generating fuel and the fire was at the entrance. In both a borough and your code read experiments your ventilation was very nicely located around the perimeter of the building. Did you have experience, and in non-uniformity, of fire behavior that would be related to how air enters the building or you cannot tell from the data it was uniform inside?

Speaker 2: 29:24

With regards to flame spread, what we experienced was, as discussed earlier, that the encoded, the flames, moved quickly to the end of the compartment and then, once the more of the end of the compartment had been burned, they moved more towards the middle of the compartment where there was initially less ventilation. So what we observed from that is that the presence of ventilation when you have an open plan compartment, where some areas will be closer to ventilation and some areas more remote ventilation by a big distance in comparison to a smaller compartment where you can say that you have an opening factor for the whole area, that this affects the local conditions in that location. And this something that we also understood from the charting depth measurements that were taken that in many cases there is a big variation on the actual charting depth measurement in different locations and mostly having to do also with the presence of ventilation near the location where the charting depth was taken, the measurement.

Speaker 1: 30:33

Do you think this impact of ventilation is perhaps important for the whole of traveling fire methodology? Because the traveling fire methodology very directly connects the peak temperatures with the locations of the flame front. And now, if this flame front may be, let's say, adjusted by the availability of oxygen near your openings, that perhaps changes the exposure you would have. However, as a disclaimer, traveling fire is and I don't think it was ever meant to be, a replica of exact physics of the fire. It was rather a very useful approximation that can be used to simulate complicated scenarios In a rather simple manner for structural design and give the designers another Tap of understanding that behavior of their structure, in different than flash over to pre, flash over the fires sure, because we know what did you hit a certain member for a long time and then you really push fire on that member behaves different than when you experience stun fire. Correct me if I'm wrong. I hope Girmann doesn't kill me for that. So do you think this? So ventilation relate behavior Can be important for the whole of traveling fire methodology and in particular for the exposed combustible traveling fires? How important is that?

Speaker 2: 31:53

As we know, ventilation plays a big role on fire dynamics, so it will naturally play a role also on traveling fire dynamics and, as you say, it's not an input for the so called traveling fire methodology that we use in design, because that method is not aimed to capture the actual physics, as I think probably also an open, blank apartment they are very tricky to capture the first place, but it tries to put down a framework that can be used for design and this is where the balance needs to be as well, also for timber compartments on. From one side we have this complexity that happens in an experiment and we try to take knowledge out of it. But in real life it will be very difficult to be able to exactly calculate this type of variation in a large complatment of different types of depth, depending on the location and ventilation. There are too many difficult, complex parameters to analyze and from a design point of view, we need to take this knowledge and come up with a conservative but appropriate methodology that can be used in design.

Speaker 1: 33:06

Okay, but you have accelerated flame spread. You have very quick fires, like 20 ish minute fires that led to burn out of the whole fuel load which in a way, was representative of an office fuel loads. So let's say you would have the fuel that you normally have in office. It burned down in 30 minutes. That's a very rapid fire. I often in my research and still to struggle like what actually is the worst, what's what? Which fire is worse for? Still to structure like burning it very fast, very viciously, but getting done with it sooner, having Still a large fire but going there for two hours is that was scenario. So, from your perspective, how bad was this for your building and what were your feelings when you were looking at?

Speaker 2: 33:55

Yeah, what happens is that I saw ways making something better in one condition can go to anything another. So, as you say, by allowing a much more rapid flame spread throughout the compartment, provided that flaming extinguishment happens, and this is a key driver. That was an assumption, god read, because we only had one exposed, the in person in service and and the glue that was not allowing for jar for love. So that was a given in our case. Provided that you have a rapid flame spread, you will have in the end a less intense duration of fire. That's something that we know, also from a traveling fires in non combustible structures, that larger fires tend to result in a lower structural far severity from Medium or small fires because of the phenomenon of far filled heating that pre heat the structure before the flames reach that location. But equally, by allowing allowing a much more involved part of the floor plate to be flaming, we have a greater heat release rate and that means a much greater external flaming in comparison to a non combustible building. So the internal and external conditions the one influence the other.

Speaker 1: 35:16

That was my next question. Like you have a massive production of fuel inside, you add now 350 square meters of ceiling that emits flammable gases. It's kind of you said, it's ventilation regime. So it's obvious that not everything burns inside must burn outside then. So to what extent the external fire was greater and, having this experience, how it affected your design of mass timber buildings, to what extent the designers need to be careful about the external flames? But because you know we live in a world of paradigm, that you have this amount of meters of separation, you would be safe, and I guess this new introduction of combustible materials could actually shatter that paradigm a little bit. So what's your experience with the verticals exposure?

Speaker 2: 36:07

It's a very interesting and complex problem in that what we experienced from the code red experiments was that we had more external flaming, in that we had more windows involved in the fire and therefore having external flaming from them in comparison to the non combustible building, but also the flame heights are, it's where, a greater than the non combustible buildings. The challenge with external flaming is that, unlike farce vertigo in the compartments and structural design, where we can calculate capacities and the performance criteria are simpler to define, with external flaming we cannot clear performance criteria on what is and is not acceptable. And what makes the situation even more complicated is that even in all the possible buildings that follow prescriptive requirements, we have many times seen external first spread happening from floor to floor, despite all the prescriptive requirements have been met. So the analogy back to a non combustible compartment does not always mean that you will not have external fast spread in a building, but it may mean that it is less likely to have it or it will happen at a later time, which exactly makes the situation even more complicated as the performance criteria are not well defined or understood.

Speaker 1: 37:30

Before in the talk you've mentioned that you've experienced a flame out or flames going down on the on the ceiling, so maybe let's touch a bit on that in what conditions you've observed the ceiling to stop blaming, and how far into the fire was. That was before the crib burnout after the burnout.

Speaker 2: 37:49

In code red one, flaming of the CLT stopped approximately five minutes before the flaming of the crib and that is why it was still at a relatively high heat flux in the bar to what we would expect normally if a flaming extinguishment. And that was an interesting outcome that fascinated us as well and that's something that we didn't expect. In code red two and four, flaming of the CLT about the same time with flaming of the wood crib, probably black about something like 30 seconds before the flaming of the wood crib In both cases. But yeah, I know, in all three cases there was no flaming of the CLT after the flaming of the wood crib has finished. And by flaming of the wood cribs we defined once you start you don't have any, so much of a visible flames because of the flaming is also something that is difficult to define from from photographic images. But once the there were not much of visible flaming then we considered it as a send off. Flaming of the wood cribs.

Speaker 1: 38:52

As you mentioned at the beginning of the episode, the idea behind the design was to prevent char fall off, so I assume you've achieved that, or did you occur, child?

Speaker 2: 39:03

Yeah, we achieved that the selected CLT that we used had me, la men, and he's even was steady, tested to a two hour furnace fire before the experiments, and we have done many different type of tests To be able to see, with different level of thicknesses, that there was no jar for love in furnace conditions.

Speaker 1: 39:23

What was the lamela?

Speaker 2: 39:24

thickness in this case. So the lamela thickness was 40 millimeters. We had the five plies CLT that had the total thickness of 150 millimeters. Okay, when we did the jar for love test in the furnace, we also looked at lower lamela thicknesses just to ensure that jar fall off wouldn't happen for this particular arrangement also when we were planning for the experiments back in 2020. There were not the type of products and suppliers available that we have today, so we had to go through a lot of internal testing with the CLT manufacturer in order to derive with a sample that we're happy with, and essentially the chart that Changdeb didn't reach the group lines it didn't follow, so you didn't have this secondary effect.

Speaker 1: 40:16

Yes, exactly, but in the commercial projects I'm not sure everyone would go to the same length. Maybe some people would be comfortable losing a layer of CLT and having the chair follow. How would you? Okay, we are now not talking about your experimental findings, but we're brainstorming what would happen in the experiment, the reason why we wanted to avoid it was that it could potentially impact the comparison.

Speaker 2: 40:45

It would be an additional uncertainty that would impact the comparison between different case studies considered. So, as we see, sometimes there are experiments that when a chart follow happens with everything else being identical, because it's a complex phenomenon that we don't understand fully, we have different results and we knew that from a previous experience being involved previously with some experimental testing of CLT of compartments that with everything else identical, we still had somewhat different results because of chart follow in smaller compartment fires and as our focus goes on fire dynamics and being able to understand exactly what parameter influences another phenomenon, we wanted to reduce that uncertainty as much as possible, which in a large compartment there is already uncertainty because of the weather conditions, wind conditions and all other things that are going into a large scale experiment. So we wanted to avoid another uncertainty of this kind and that is also the reason why. One other reason is why something I forgot to mention before that our structure was unloaded so that we don't have any potential structural failure that would again impact the fire dynamics, and also because we wanted to capture for the first time smouldering happening in an open plan compartment in a large scale experiment of this kind with exposed timber ceiling.

Speaker 1: 42:16

You're really good at leading the industry. You should be a podcaster. So I was about to ask about the smouldering, because it is truly a unique observation. So, to put things first, you did not put water on the structure.

Speaker 2: 42:30

No, we didn't In this experiment. So what happened? What happened was that once flaming finished, we actively monitored what will happen to the structure without an intervention for at least 48 hours. It was 48 hours from the first experiment, 60 hours from the second experiment and several days from the last experiment, code red 4. What we noticed is that in certain locations we had the appearance of hotspots after flaming finished, and these were normally in locations where there was a junction of the panel with an exterior wall or with a beam, normally where you have high insulation and the presence of some sort of cavity, air cavity. Some of these hotspots extinguished on their own and some of them continued to smolder the timber and lead to the formation of large holes after several hours and days. Any particular observations for the connection between the glulam column and the ceiling when we had the glulam column and the ceiling, we didn't see in our experiments any particular hotspots. With regards to smouldering, we saw increased the heat fluxes because we had plate thermometers and we were measuring the radiation at these locations in comparison to other locations where we didn't have glulam columns. What we saw was that there was flaming extinction despite the presence of the timber columns. In our experiments there wasn't any smouldering in these particular experiments at the top of the column and the connection with the ceiling, Despite there was a 50mm area where it was insulation. So 50mm connection with insulation between the top of the glulam column and the timber ceiling. But at the bottom of the glulam columns we saw increased charring because of the presence of the wood crib, even after flaming was really done. Because of the presence of the smouldering wood crib itself. There was increased charring at the bottom of the columns and also in one of the experiments there was a smouldering hotspot at the bottom of the column that continued to smoulder the base of the column until it fell off to the ground.

Speaker 1: 44:54

Very interesting observation that perhaps we should do more often when we're doing CLT experiments. I wondered what extent you would still have the hotspots if you actually actively pursued extinguishing the surface, because, from what you say, they're not even material related. I would expect them to be mostly at the glulam CLT interface because there's the most fuel, but it really may be related to having the most hidden space in your ceiling with enough ventilation In.

Speaker 2: 45:24

Code Red 4, I think it's interesting that you mentioned about the hidden locations In Code Red 4 that we had the encapsulated ceiling. There was a hotspot near the junction of the very first panel in the compartment, near the ignition of the fire and the exterior wall. That hotspot progressed below the encapsulation and finally led to a formation of a very big hole through both the thickness of the CLT and the encapsulation below and that hotspot would be very difficult to track due to the presence of the encapsulation. So the real risk with small daring is particularly when it happens in hidden locations, for example in risers or behind the encapsulation.

Speaker 1: 46:10

Let's talk about the mitigation strategy, because you're Code Red. 3 and 4, where you focus on finding how can we reduce the hazard of CLT, let's say, exposed combustible material in a structure and it's on fire dynamics. So experiment 3, you've used what type of water mist was it?

Speaker 2: 46:29

We used a low pressure water mist system that was compliant to the FM Global Standard, a typical design for an office building in the UK. The idea behind that was to see because water mist normally takes some time to control the fire gradually without throwing a lot of water to the fire there was the potential risk that while this control phase happens, the CLT ceiling is ignited and then the fire becomes out of control. What we saw was that in our case, there was not an ignition of the ceiling.

Speaker 1: 47:07

Oh, so the water mist activated before even the ceiling become involved, so you did not have this spread on the ceiling that could escape the control zone Exactly.

Speaker 2: 47:18

If it had, we would have a Code Red 4.

Speaker 1: 47:23

And you installed the water mist on the whole area or just in the corner I was installed on quite an extensive area of the compartment, so it was a corner test.

Speaker 2: 47:34

The specific Code Red 3 experiment had greater fire load density in comparison to what we had in Code Red 1, 2 and 4. So it tried to mimic a little bit the FM Global Test. We had a 570 mZ per m2 fuel density, which is corresponding to 8% for offices in accordance with the UK National Annex to Eurocode 1. And we had the wood grid located in the corner and the fire ignited in the corner by two methanol pannes located in the corner arrangement. So it was not exactly the same ignition and position of fuel load as in the other experiments in O'Bora, but same compartment.

Speaker 1: 48:17

And Code Red 4, where you've used encapsulation of the ceiling. How did it change the outcomes of the spread and the fire itself?

Speaker 2: 48:24

Code Red 4 encapsulated 50% of the ceiling. Then encapsulation and encapsulated part was at the middle portion of the building which was also exactly above where the wood grid fuel load was located, Because the fuel load was only covering also approximately 50% of the compartment, as it did in the O'Bora experiment In which it was driven by the location of cow feeders.

Speaker 1: 48:47

actually, exactly, yeah, that's my interest, yeah.

Speaker 2: 48:50

But we had to follow that in order to be able to make a more meaningful comparison and really, in an ideal world, you need many more experiments to study all these different parameters and how they affect the final results. In this case, as the experiment happened in December and it was winter time, the moisture content of the wood grid was despite. It was initially in a controlled environment, one's positioned in the facility and when it was the time to do the experiment because they were happening during night for practical reasons the moisture was greater than Code Red 1 and 2 and O'Bora, which I think about 4% greater, which resulted in 20% to 30% additional energy required in the vaporization of the moisture in comparison to the previous experiments. That meant that the initial fire growth in Code Red 4 was less intense in comparison to Code Red 1 and 2 and it took 15-20 minutes until the fire was growing gradually, similar to X2 Part 2, the second part of the second experiment in Poland, where the fuel load density was 30% less and we didn't have intense flaming that would touch the ceiling. So we had similar conditions in Code Red 4 until at about 20 minutes the timber ceiling got ignited While the flame height was still quite low in the first ignition. The first ignition resulted to extinction quickly. It was from one side of the encapsulation, the exposed part. Then the other side of the exposed part of the encapsulation ignited and then extinguished again on its own and then, where it was the first time, it ignited with the third try and that sustained flaming that very rapidly spread to the end of the compartment in a similar fashion to Code Red 1 and 2. So with very rapid flame spread across the ceiling.

Speaker 1: 50:56

So you had a very, let's say, gentle growth of the fire on the crib itself. It was warming the ceiling. Then you had a puff of flame on the ceiling. It went out. You have a second puff and then the third one finally ignited the ceiling surface and then it went rapidly.

Speaker 2: 51:13

Yeah, and then we had very similar phenomena to what we had in Code Red 1 and 2. The flame spread on the CLT was similar to Code Red 1 and 2. The flame spread on the wood grips was as one would expect 50% of the ceiling was encapsulated was lower than Code Red 1 and 2. So the fire dynamics were something between the non-cababable experiment and the fully exposed timber ceiling, but more torwards the fully exposed. But one thing that also Code Red 4 has shown us was that a fire that would not otherwise become fully developed If it was a timber ceiling, it became fully developed. So it was an additional, although we didn't plan initially for a higher, a greater moisture and therefore more time for the fire to develop and more energy getting into the fire. In the end we had a meaningful observation that was also the impact of the timber ceiling making a fire that would otherwise not be fully developed to make it a fully developed fire.

Speaker 1: 52:18

So you're a fire engineer. You're working for a large fire engineering company. Who funded this research? What's the most important takeaway for an engineer and for Arup from pursuing this experiment? What was the number one takeaway for you guys?

Speaker 2: 52:33

So, for us, key knowledge that we gained from the experiment is with regards to design fires for exposed mastimber buildings and how this differ from non-combustible buildings and therefore we have to account this phenomena when designing the buildings. The impact of rapid fire spread on evacuation and firefighting needs to be understood, particularly on the escape of people, because we see quite rapid fire spreads, but also, since we have a large part of the floor plate involved in fire at any given time, and also the flaming of the ceiling, that would also affect firefighting activities and the amount of water that they have to throw into the fire. External flaming is an important consideration and how a vertical compartmentation, which is an important strategy for tall buildings, needs to be adequately addressed and potentially through encapsulation or not having as much timber exposed or other measures, smoldering as well of timber is an important phenomenon that needs to be understood, both in design but also for operation of buildings, and I think this again another area where, as fire engineers, we need to talk to firefighters and understand how this type of building needs to be considered from a firefighting perspective. What would happen with hidden locations, for example, that we said before, and what they have, for example, to remove the encapsulation in order to be able to track hotspots, things like that.

Speaker 1: 54:05

And as a researcher who is doing large experiments, I very well know that whenever you do an experiment to answer a question, you get two new questions after an experiment. So by solving this gap in knowledge pursuing code red, which new gaps in knowledge that are urgent you found?

Speaker 2: 54:26

That's a very good question, gostek. I saw I'm a person that really believes in research and large scale experimental research, in that it helps us to frame the problems that we then try to solve in more detail in, for example, smaller scale or more fundamental type of work. So I think that there is a very important place for this type of experimental work and much more experiments should happen in the future. This is something that we hoped for when we did code red. We were hoping that more and more people would then follow the same route and, by the way, the facility in Serb is available. Whoever wants to go, they can go and use it, so it's free for everyone to use. So, with regards to additional experiments, there are many more parameters that need to be studied that also came up as part of our discussion, for example, for flame spread through the CLT. Our CLT was untreated. The impact of treatment on flame spread would need to be understood. The impact of different arrangements of ventilation. We had pre-existing openings in our experiments and large ventilation openings at the end of the compartment, which naturally would affect flame spread from the ignition location to the end of the compartment. So it would be interesting to understand how different arrangements of ventilation or even more realistic facade systems that are used in offices would perform in such a far condition. As we know, the fire and the ventilation is a chicken and egg situation, even for non-combustible compartments. It's an area that has really been studied how realistic facades would impact the fire dynamics observations that we see, as the ventilation is always static. Then we also had quite relatively low, far low density in our experiments, which originated from the police or more experiments and the suggestion by the local fire service there not to damage the building and be able to control the fire. Therefore, we would need more experiments with more realistic fire load densities, as the ones that you would expect in an office.

Speaker 1: 56:38

I always wondered. The non-continuous fire load could be very interesting, because you would have parts which are much higher like local fire load and spaces with essentially none. That could be perhaps very interesting. The world of fire research is vast and there is a lot of space for explorers, so everyone is welcome to jump on this bandwagon. The facility in Serip is free for work. I doubt it's completely free, but it exists there. Arab has spent a lot of resources, time and effort on pursuing this. Was it worth it? Would you do it again?

Speaker 2: 57:19

For sure, despite what we did, in the end it was much greater than what we thought originally. It started smaller and then gradually increased in size as a commitment. There was a large team of us within Arup, but also the students at Imperial and people at Serip, that devoted a lot of time to this project. They didn't initially think it would last as long and would need as much commitment as it did. As an example, probably for two, three years 50% of my time went solely on this project, which for a consultant is a significant amount of time to devote to one project, and this thanks to the support from Arup and other colleagues. But definitely despite all the difficulties and the challenges with the pandemic and communications and also having a big team different partners I would love to do it again. If anyone has fun out there, I'm happy to do more experiments in the future.

Speaker 1: 58:27

Come on, there had been hundreds of structural fire experiments on concrete and steel, and everyone goes back to Cardington. There's a huge value in pursuing the most ambitious and biggest fire research because sometimes not even sometimes often the scale matters. You just have to build it this big and this great to really find out what to research in the future. Many of the things you've mentioned couldn't be done now in a smaller scale with lower resources spent. But we wouldn't know that if you haven't done a contract. So thank you very much for that and thanks to Arup for funding this great piece of research and choosing to publish it, not hide it in the company archives as an internal report which we'll never see in the daylight. Very thankful for publishing and sharing it with the scientific community.

Speaker 2: 59:20

That was a very important driver from the beginning for us. So we wanted to contribute to the community. So it was not for a specific project, but that was later, didn't progress that project. The whole objective was to be able to contribute to the body of knowledge and actually scratch the surface, initially to understand what's happening, how the fire dynamics could be if you have an open-plan compartment with timber, really in the hope that other people would then continue the work and take these findings and study them in more detail. Arup's idea was not to try to solve the problem on its own or like it's a company in the end it's not a research institute. So the principles from the beginning was we want to try to do something to contribute to our community and then others are willing to study this in more detail and help us back.

Speaker 1: 1:00:18

Fantastic Panas and thank you for now sharing these experiments with the community of Fire Science Show. I appreciate your efforts and looking forward to next experiments and catching up with you.

Speaker 2: 1:00:32

Definitely. I want to thank you, Gostek, for the invitation and also the audience, for the rain that's in our work. I'm very approachable. I'm happy to answer any questions that anyone has through my email or LinkedIn, media and similar.

Speaker 1: 1:00:52

And that's it. Thank you very much, panas. This is an excellent team effort of so many people involved in your experiments and I really appreciate you highlighting the role of all those involved, from Arup, from Surrey, from Imperial College London, a group of very dedicated researchers trying to understand phenomena that we've never observed in full-scale research or in Fire Science. And there is a huge value in doing research at this scale because it really shows us where do we need to look into details of this phenomena and after you've done that, we know much more on what interesting phenomena are occurring in the fires. Some findings perhaps one could expect that the fire will be quicker, but it's not just observing that, it's about how much quicker and what changes and especially observations like one where you've said that the fire would not have grown to the fully developed fire without the combustible ceiling, but due to the participation of the ceiling it actually grew to that size in Experiment 4. If I recall correctly, that's a very interesting observation and a very strong point that everyone needs to understand when they try to design large compartments with exposed silty. Also, we're finding regarding the speed at which the fire traveled through the room and how quickly the growth, the acute, combined with what we know about the vocation process and how many people can be in these offices To some extent interesting, perhaps even disturbing definitely something that needs to be taken into account. A lot of golden nuggets like that. I guess you really need to go into the papers to find all the interesting things that were found during Code Red experiments. I appreciate these papers being published. They are open access. Anyone can look into them. So I would highly recommend that the links are in the show notes. And I guess that's it about the Code Red experiments. Tell me if you think this experiment will change the fire science. I think it has a good chance to be not forgotten by history but referred for many years, if not decades, as one of the most important experimental attempts on exposed timber in office compartments. So, yeah, that's it for today. Thank you for listening. See you here next Wednesday for another, hopefully great episode of Fire Science Show. Cheers, bye.