Fire Science Show
Fire Science Show
236 - Fitting an efficient smoke control system in a confined space
A tight, historic cellar. Arched ceilings. Long corridors. Tiny shafts. We faced a design wall: to keep routes tenable, we needed twice the extraction that the building could carry. At that point, I've failed as an engineer - I've reached my limit and could not find a solution.
Some time later, a solution appeared in my head from nowhere —what if the fan changed with the fire? Not in a crude on-off way, but by tracking temperature, exploiting density changes, and chasing constant mass flow instead of fixed volume.
We unpack the moment this clicked, the fan physics behind it, and why hotter smoke can actually make extraction easier if you use the margin correctly. You’ll hear how we oversized the fan, ran it at a lower frequency in ambient, then ramped as temperatures rose to keep kilograms per second steady. That adaptive control boosted cubic meters per second right when the layer needed support, eased plug-hole entrainment, and stabilised makeup air velocities. We walk through the thermodynamics, the electrical and pressure implications, and how these pieces form a practical control strategy for retrofits and new builds.
To ground the idea, we share two paths to proof. First, CFD with user-defined control that reads gas temperature each time step and updates fan frequency with smoothed delays to prevent oscillations—capturing the real feedback loop between fire and system. Then, full-scale container burns with live control showed the same trends from 20 to over 500 degrees: falling duct pressures, lower fan power at heat, and the headroom to increase volumetric extraction without breaking limits.
Thinking about it now, this idea is a part of many other concepts that I describe together. To show a way how we come from the simple framework—Smoke Control 1.0 (empirical, static), 2.0 (CFD-informed, still static), into a new smoke control 3.0 (adaptive, feedback-driven)—and explore how this thinking can reshape underground venues, car parks, tunnels, pressurisation, and natural ventilation.
If you care about safer evacuation, smaller shafts, lower velocities, and systems that work with physics rather than against it, this story is for you. Subscribe, share with a colleague who designs smoke control, and leave a review with your toughest question so we can tackle it next.
Reading material:
- Can smoke control become smart?
----
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. The previous week I've been interviewing Lazaros Felippidis on the fires in underground club venues, music clubs, etc. Very difficult episode to record because of the recent tragedy in Switzerland. And while during the talk, we were talking about various aspects of fire safety in the underground spaces and in such venues. And while we were discussing it, I actually remembered that I had a case study like that in the past, and it's actually a very interesting story for myself, perhaps my pinnacle as a smoke control engineer uh doing buildings, that's years 2017-18-ish, pretty much. Back then I I phased a project like that to provide uh fire safety in an underground retrofit venue, and I I failed. I could not provide a smoke control system in the venue in a way that I would be happy with it, in a way that I would say that my fire safety engineering goals are met what we've ended up. We just told the client, well, you need this much air to extract, and there's no way we can get this much of air extracted in this limited space in those circumstances. And that was kind of an engineering failure. I was not able to engineer a system for that building, but this stick with me. And uh sometime later, I've kind of made a discovery. It was quite quite funny to make a discovery, but uh I've made a discovery, and with this discovery I realized I could actually make this system work. And you know what? We've uh we've done a research paper on that and we've shown that we could have had actually made it work with clever use of fire physics, with clever use of measurement systems, with completely changing the paradigm of how smoke control works worldwide. Back then I called it the smart smoke control idea. Today it's something I would probably coin as smoke control 3-0. And perhaps if you like this, I'll I'll record a larger uh podcast episode on the whole smoke control 3-0 idea. But today I would love to share with you my personal story, how I felt as an engineer, how I succeeded as a researcher, as a scientist, and how understanding compartment fire dynamics allowed me to uh do something that was impossible, do something that we thought is not possible within the constraints given, and yet we found a way to fit more smoke through the same deck, pretty much. But that that kind of sums up the idea. I've shown this in SFP conference in Malaga 2019. Uh some time has passed, as I said, it settled down somewhere in the back of my head. Lazarus brought it back. It's again a fresh idea in my head, it was an important part of my career. I would love to share those thoughts with you today. So let's spin the intro and jump into the episode. Welcome to the Firescience Show. My name is Wojciech Wegrzynski, and I will be your host. The Fire Science Show podcast is brought to you in partnership with OFR Consultants. OFR is the UK's leading independent multi-award-winning fire engineering consultancy with a reputation for delivering innovative safety-driven solutions. We've been on this journey together for three years so far, and here it begins the fourth year of collaboration between the Fire Science Show and the OFR. So far, we've brought through more than 150 episodes which translate into nearly 150 hours of educational content available, free, accessible all over the planet without any paywalls, advertisement, or hidden agendas. This makes me very proud and I am super thankful to OFR for this long-lasting partnership. I'm extremely happy that we've just started the year 4, and I hope there will be many years after that to come. So, big thanks OFR for your support to the Fire Science Show and the support to the fire safety community at large that we can deliver together. And for you, the listener, if you would like to learn more or perhaps even become a part of OFR, they always have opportunities awaiting. Check their website at OFRConsultants.com. And now let's head back to the episode. Okay, retrofitting efficient smoke control in historical venues. Uh back then I coined this as a smart smoke control solution. One of the older professors asked me what's so smart about it, and that's a hell of a difficult question to be answered. But in fact, when you discover something new, when you propose something new, there has to be a name for it. And back then I thought uh smart smoke control would be very fitting. So before I tell you how I discovered something new and how I've succeeded as a scientist, let me tell you how I felt as a fire safety engineer. Because uh not being able to deliver fire safety for a project, I consider this a failure of myself. Well, uh up to you to judge if I am too harsh or it's justified. But uh let's let's talk about the the problem at hand. So we were approached by a client who was uh trying to connect two underground cellars in uh in a historical buildings, you know, large cellars that are underneath uh the buildings with quite steep entrances through steep staircases uh inside. Uh they consist of multiple smaller rooms uh connected together through some sort of corridors, and some of those rooms that were adjacent to the external facade were actually much deeper than the rest of the buildings. They had the ceiling at approximately the same height all over the place, but some of the rooms were significantly taller than other rooms. Uh this is because the role of those uh of those spaces. Back in the days, you would drop coal from the street level down to those uh rooms. Therefore they had some uh little windows that connected to the uh exterior, nothing you could use to smoke control really, but we we could have used them as sources of makeup air. We had some limited shafts in the room because there were like literally two small shafts vertical going up all over the building that we could have used. Uh we had two staircases, there was a possibility to build a third one, but the investor for them that was a massive, massive problem to uh retrofit the staircase into this layout. And in general, it was quite large. It was like multiple, I think 20-ish compartments, maybe less, but but quite a lot. It was supposed to be a restaurant and a music club. Uh, long story short, uh, the law requires smoke control in corridors in Poland. The law requires you to provide smoke control when you exceed some evacuation distances, which was the case in here. Um, it obviously looked dangerous. Like you could take a look at the picture of the layout, the layout is in the paper, and you can tell that this is a very challenging space to uh fire safety engineer around. Uh also ah, one more one more interesting caveat was that the compartments they were not rectangular, they the roofs were arched. So it was not only just you know limited space, but also the smoke reservoirs were even more uh reduced because of the arch shape of the ceilings. In this confined space, it's already difficult to provide smoke control because you don't have smoke reservoir, you need a smoke reservoir to be able to extract the smoke that is produced in the fire. If you don't have sufficient smoke reservoir capacity, the layer will descend and it will drop below, let's say, 1.8 meter, which I would consider the minimum value for safe evacuation. If it drops, if it drops down below that, there is really little you can provide in terms of smoke control. And if the smoke reservoir is not large enough, you you have to compensate with larger smoke extraction capacity. But this brings another challenge because if you don't have a big enough layer, you start to have a plug-hole in situation, which means that the smoke control system is extracting not the smoke from the layer, but the clean air from underneath the layer. The smoke extraction can be so powerful that it basically sucks the smoke around and starts sucking the cold air from the bottom layer, and that obviously is a systematic failure of the smoke control. So you end up with a situation. You have a bunch of smaller compartments, they fill with smoke quite quickly, they spill the smoke to neighboring compartments, those neighboring compartments fill up very quickly, they spill the smoke into common corridor, the corridor fills up quickly. You end up in a situation that within three to five minutes for a reasonable design fire, your evacuation routes are untenable. This is how bad it is. You could perhaps do this if you increase the smoke extraction capacity. So I think we've started at 10 cubic meters per second, um, as what you could perhaps maximally fit in this space. We've ended up uh seeing that at 20 something ish cubic meters per second, you would actually get a performance that that would work with this smoke reservoir. But uh this number, me as a fire safety engineer, I come up with a number how much air I want. I give this number to the MEP engineer, and then the MEP engineer has a heart attack because there is absolutely no way they can provide me with that. Why? Because they are limited with space. They are limited with space on new projects, because the space is money. Every square meter of uh extraction shaft is one less meter you can sell of a building and that stucks up very quickly. In historical buildings, they simply don't have space. They don't have duct space to to put more ducts in. It's it's impossible to fit more ducts in. Another challenge is with uh the makeup air. We are talking about underground venue. So again, I said that there were little windows uh connecting to outside. We thought about you know combining those little windows with um uh some ducts to bring the air down to the ground and uh allow for some makeup air situation with them. However, the amount of those windows is also limited, and uh as you increase the extraction velocities uh in your ducks, you also increase the inlet velocities and your inlet points. And too high inlet velocity is also a big problem in any smoke control design because it will cost mixing of uh smoke, air. Basically, you end up with uh a little colder mixture of smoke that is still untenable, still prevents any safe evacuation. Nothing you can do about it. Also, I probably should have mentioned that uh this venue obviously had no sprinklers in it, and the investor had no intention to bring sprinklers in the venue. That was even worse than making additional staircase from the perspective. So we end up with those bunch of problems with being unable to fit uh horizontal vertical ducks, being unable to fit a reasonable makeup air strategy for the space. In general, a situation in which we know that if we want this venue to be safe, we need twice the amount of air extracted from the venue, then we can realistically maximally provide with the circumstances given. That was the limiting factor. And also, one more thing which is important later. Um, we ended up with ventilators uh for class F600 because the smoke temperature was also extremely, extremely um hot. This is how I failed. We were not able to engineer a system for this space. So that was it for this project. We delivered it. We're not sure what the client did with it. I have no idea what's the follow-up on the project. Did they build a venue or not? Did they get more air or not? No clue. I have not done a follow-up on that. But from our perspective, this was it. This was the end of the project. Uh simply unfeasible. Uh I've continued my life as an engineer for months or years, and eventually, uh one day, in a very random uh moment, a solution uh came to my head. So that day, my colleague from the office, uh Pshamek, he was working on a smooth control project. I think it was for a car park. I'm pretty sure it was for a car park. And Pshamek was uh having the same struggles as I did back then uh in the in the underground venue. He needed to extract more air than he could uh extract through the ducts. Basically, he found a number at which he was pleased with the smoke control system in the car park, uh, but he only could fit maybe 60% of that number into the ductworks of the building. And the investor was very rigid, there was absolutely no way they would allow to increase the size of the vertical shafts in the building. Like simply impossible. And and and Shemak was really struggling because he would not sign off a project with lower value of smoke extraction because uh this obviously has not met the goals of the analysis. Uh and he knew that there is a solution, it just requires more air. And he I was his supervisor back then. He will I and I was I remember I was sitting at the desk, I was super busy with another project, and he's there nagging me, I cannot fit the air, I need a way to fit more air. I uh w what I'm supposed to do with the project. And I was in a flow mode uh on another project and I just simply replied to him, Jesus Psemek, just put a larger fan on that duct, and uh when the temperature rises, perhaps you could increase the extraction on it because the losses will will drop. So technically, you in the end you could extract your 100% through the duct, which only allows you 60% in ambient. And the moment I said that, I don't even know why I said that. It it's kind of this Eureka moment, I guess, when people refer to Eureka, uh, you know. I I think that that's kind of the feeling. I mean when I said that, everything clicked in my head, and I was like, whoa, wait. And uh the first thing I did is to call my another colleague, uh Xegosh, who is an expert on smoke control. We work together, and I called Xegosh, if we put a larger fan on an extraction duct, and when the smoke increases in temperature, we increase the volumetric capacity of that fan, that's gonna work, right? And he's like, Yeah, I think it's gonna work. And I was like, wow, this is like a confirmation. I went from not having an idea in my head into into a really clever solution within like m two, three minutes. That was like really interesting feeling, uh, you know, when when things click in your head. If you don't get why I uh say this is uh a wow moment, uh I'll explain to you in the details because there is a bunch of uh thermodynamics that go into that and and a bunch of history of uh how smoke extraction systems are being designed and then tested that go into that. But uh that moment I knew that the world of smoke control has uh has changed of me like with a click of a button. So to explain why this works, uh why this changes everything, I have to take you through um kind of a history of a fire in a compartment uh from the perspective of thermodynamics and and uh fluids and heat transfer. Uh, what happens in a compartment where you have mechanical ventilation in it to various characteristics of a fire? In my paper from 2017, I uh present this as a nice uh matrix of nine plots of various properties uh of compartment physics that change with with time as the fire grows. I'll try to explain this in uh spoken content, but wow, this is this is a challenge. Um anyway. Imagine you have a fairly small compartment, let's say it's 10 by 5 meters, 3 meters tall. Maybe not the smallest compartment, but it's it's a small compartment in which you can have a mechanical extraction from with a duct and a set of extraction grills. Uh those ducts obviously connect to some extraction fan somewhere further in the building to that space. You provide some makeup air through mechanical or natural means. So there is, let's say, a fan that pushes some air back to the room, there is an open door to that room through which you can suck air from outside. Now imagine a fire starts in that compartment. So you you have a fire growth uh which we are used to. We would usually define it with some quadratic uh function, you can have linear functions of fire growth, uh whatever you choose to define your heat release rate as a function of time. Usually the presumption is that the fire will develop, it will grow to some certain stage, limited by ventilation, by fuel, by other things, uh, and eventually it's gonna start decaying. As the fire is growing, what's happening in your smoke layer is that the hot air and smoke produced by the fire is entraining a cold air from surroundings, basically creating a large amount of smoke, which is basically your air with some fire products in it. And this smoke starts to feel the uh reservoir underneath the ceiling. Uh, the amount of smoke produced will follow the growth of the fire. So, of course, the higher the heat release rate, the more smoke you're gonna produce. It's not a linear relationship, there's a power low. We call those power lows the axisymmetric smoke plume models, and I had a whole podcast episode about them, uh, I think. Uh, as your fire grows, as you have more smoke produced in the plume, also the temperature of the smoke in the room changes. So this changes also in a power law, it goes a little different than increase of mass of smoke produced, but basically you get a very rapid early growth of the temperatures of the smoke, and then the temperature stabilizes at some point. If you reach a steady state where you have a certain amount of megawatts produced in your room, those uh properties will technically stabilize. So for a certain amount of megawatts at a certain layer height, you will produce the same amount of smoke, and the temperature will eventually reach some sort of state of equilibrium in the compartment. So basically, as fire grows, you get more smoke and it becomes hotter. That's very easy. Now, as I mentioned, you have an extraction in that room. You have your traditional mechanical smoke extraction system fit in that room, which means that the fire is detected. The building knows that it has to start the extraction, so dumpers open promote the flow of smoke. From that particular compartment that we try to extract the smoke from, a fan starts to operate and it starts sucking the hot gases from the compartment. Now, the fan is a machine which basically rotates the impeller of its own following the changes of alternating current that flows through coils which are in the engine. Depending on how many coils you have and how you set them up, you will reach a different kind of rotational velocity RPMs based on the frequency of the alternative current that you supply with it. In Poland we would use 50 Hz alternating current, which translates into rotational speeds of 960, 1500 or 3000 RPMs. I think those are the default values for 50 Hz that we usually get in here. Regardless, you supply power to the fan, so the fan turns around. It spins at the same velocity all the time because the velocity of the spinning of the fan is dictated not by the contents of what's transferring through the fan, but it depends on the frequency of the alternating current that you supply to the fan. So it spins all the time the same thing. The characteristic of the blades of the fan, they do not change. Technically, there are fans which can change them, but normally they do not change as the fire progresses. So basically, from the moment you start your fan till the end of the fire, this the fan is supposed to spin the same way with the same velocity each time the blade sweeps the same amount of air and uh as it works, it extracts the same volume of air with every spin. This is important, this is the most important thing in it in this design, that every time uh an impeller rotates, the same amount of air is transferred through the fan. Therefore, the fan operates at constant volumetric capacity. The volumetric capacity of the fan is not changing with the changes of temperature of the smoke or any other uh measures that happen beyond the fan. The fan is dictated only by the power supply and it always provides the same. And I know that I I knew that that's the thing why this Eureka moment happened to me in particular, because I am also someone responsible for testing those fans. I run fire tests, I'm personally responsible in my laboratory. I am the person who tests the fans. This is my responsibility. And I I've tested like I don't know 40 of them in high temperatures on that on the furnace, and you even have a feature in EN12101 Part 3 standard that tells you you can measure this volumetric capacity and you should prove that the volumetric capacity has not fallen less than 10%. So it this fact that the van is a volumetric machine, always has a volumetric capacity which is steady. You you even have to there's even a point in the standard that says that that because you have to check it. So that's how I know that they work like this, and that's why I was so sure that they work like this. Now, this is the volumetric capacity. But what about the mass flow? The the kilograms per second, not cubic meters per second, but kilograms per second. Because the temperature is increasing in your compartment, the air density is dropping. So, you know, the air at uh ambient will have something around 1.2 kilogram per cubic meter density, but when you hit it up to let's say 300 degrees, this density falls by by a half. If I'm not wrong, it's 0.6 kilograms per cubic meter of air at 300 degrees. So the density of the air sharply decreases. You are extracting the same volumetric capacity, therefore, what's happening is that the mass flow through that fan decreases. You are extracting the same amount of cubic meters per second, but you're extracting less and less kilograms per second through a system because of the changes in density of the hot smoke. Now let's look at the um properties inside the ductwork and what happens to the fan. Because you are extracting less dense air, the air is lighter than in ambient because it's hot, it has less density, the pressures in your shafts fall down. You observe that in the in the fire testing of fans as well. You very quickly observe a sharp drop in pressure in all the ductworks in the test as soon as you start the furnace and increase the temperature of the smoke. Uh, that's a consequence of the change in density of air. And again, the standard tells you how to adjust that to the ambient density and how to measure that during an experiment. I've observed this countless times that if you if you increase the temperatures, the the pressures in the shafts decline. And uh what happens to the to the fan? As I said, the the rotational speed of the fan is dictated by the frequency of your electricity, alternating current, not by the uh circumstances in which the fan operates, but those circumstances make it easier for the fan to work. The fan has to push less kilograms per second through its body, which means it has to do less work with uh extracting that, which means that the power supply required by the fan also sharply drops following the increase of temperature and decrease in density. And again, it's something we see in fire testing of fans. It's very, very normal that you observe this when you test fans in high temperatures. Now, where's the discovery? What does it mean? As you see, the volumetric flow of your extraction fan, this remains constant over the operation of the system. The mass flow falls down, the pressures fall down, the power required by the fan falls down. To me, those drops in pressure and that drop in electrical power needed by the fan is my margin of operation that I could actually put to work. Because you see, my ductwork was, let's say, designed to handle 500 Pascal uh under pressure. I could provide 30 kilowatts to the fan, and suddenly my pressures is like 300 Pascal in high temperature, and my fan only takes 15 kilowatts. This means I could actually deliver more to the fan. And that was the basis of my discovery. That basically, if I have started not with my fan that that operates to the point in ambient temperature, so let's say I need 10 cubic meters per second, I put therefore a 10 cubic per meters per second uh fan in in my system. If instead of that, instead of that 10 cubic meter per second fan, I've placed a 20 cubic meter per second fan on the same DACTWork, interesting things happen. Obviously, I cannot start a 20 cubic meter per second fan on the DACTWork that's uh meant to carry 10 cubic meters per second because I'm gonna create two huge underpressures, I'm I'm probably gonna destroy some things, the electrical um needs for that fan will be tremendous, and I'm not gonna be able to supply this fan with electricity. But in normal conditions, I could operate that fan at lower frequency because the frequency of alternating current is what defines the capacity of the fan. You know, the frequency defines how fast the blades spin and that defines how much air I extract. So I can put a frequency inverter that changes the frequency, and instead of operating the fan at 50 Hz, I can operate it at 30 Hz, for example, in my normal conditions. So my fan now is a fan that theoretically could have uh the capacity of 20 cubic meters per second, but works at 10, and that is ambient mode of operation. When the fire happens and the temperatures increase, I can use some sort of device like a thermocouple, or there are other devices that you could use to track that, but you can use a device to track the change in of the temperature of the smoke that you extract, and you see that the temperature is increasing, therefore the density is decreasing, therefore I can safely increase the frequency of the fan to extract more volume. And what you end up, what is the dream state of operation, is that instead of extracting the constant amount of meters per second, instead of constant volumetric capacity, you start to work at constant mass flow rates on the fan. You try to increase or decrease the momentary performance of the fan in such a way that every single second you remove the same amount of kilograms of hot gases, regardless of their density, which in the end means you start to extract different amounts of uh volume of smoke from your compartment, and this is the profound change. Because suddenly, instead of extracting 10 cubic meters per second in the compartment of the fire where I am worried with the smoke filling up my room, I start to extract 20. And this 20 means that suddenly my inefficient smoke reservoir, my not working solution for an underground cellar in a retrofit building, suddenly I have the amount of extraction air that I wanted. And the best part, if you think about the makeup error, the makeup error is applied in an ambient condition. So basically, you're always uh submitting the same amount of kilograms per second and the same amount of cubic meters per second into the room. However, the balance within the compartment is based on mass flow rates. So if my fan, let's go back to the traditional system, the fire has increased its heat release rate, the temperatures have increased, and the densities have dropped in the room. I start extracting the same amount of cubic meters per second, but less and less kilograms per second from the room because of the change of density. On the inlet end, I supply the matching amount of kilograms per second to what I could extract because the balance is on the mass. And therefore, I bring up less and less air into the compartment because it's simply not needed. If I had a mechanical inlet and uh and a natural inlet, there could even be a point where by mechanical means I supply more air in terms of kilograms per second than I extract, which could perhaps lead to reversal of flows on some openings. This is very dangerous. Uh, however, if you have a smart smoke control system, if you have a smoke control 3-0 in your building and you adopt the performance of your extraction fan to the temperatures that you have, to the conditions that are within the compartment, and you increase the volumetric capacity and you you maintain the same mass flow through the fan, on your inlet side, you always have the same mass flow, you always have the same velocity because this is constrained. Therefore, you always have the same performance of a makeup air, and you don't create those hazardous changes of flows in your compartment. So we kind of used the fire physics against the fire. The fire increases the temperature of my smoke, makes it easier to extract that smoke because it's lighter, and I can just use more power. And what's funny, the the hotter the fire, the bigger the effects of the fire are, the higher the temperatures of that fire are, the easier it becomes to extract more and more. Because as the density drops, as the temperature increases, I can get the more flow through the same fan by constraining the same mass flow rate. This is the whole idea of smart smoke control, to use the thermodynamic effects of the fire against the fire itself. And you may remember I mentioned that uh back then in that building we ended up with fans for F600 class. Actually, when you apply the concept of smart smoke control, we're able to show that you could do the same with F400 fans because uh actually, because we are so efficient in extracting the smoke from the room, the maximum temperature we end up in the compartments is actually lower. And why F400 is such a profound difference? Because fans for F600 require external cooling uh for their performance when F400 fans do not, so it's a tremendous change in technology of the fan. So, yeah, that that's how the thermodynamics allowed us to solve the issue of smoke control in that particular cellar in uh Psemek's uh car park, and uh it became a concept that uh I thought I could really change the world of smoke control with. So at that point, we obviously were not involved in the cellar project anymore, and uh it was not about delivering that project, but I was quite sure we are on something bigger. And uh, while I was convinced that the logic that I've just described makes sense, I needed some sort of proof that it actually works like that. At that point, the easier way to obtain such a proof was obviously through computational fluid dynamics, because it allows you to test things like this very quickly. It was not an easy CFD project, it was not your everyday CFD engineering, but I came up with some clever ways of doing CFD that allowed me to capture those effects of uh smart smoke control system. So uh when you do your normal CFD and you define the boundary conditions for your extraction uh devices, you usually would define either some sort of surface with constant volumetric uh extraction rate on it, or you would define some fan properties and just basically model HVC network and model a fan, which is much more difficult. But then you would provide some properties of the fan in that system. In my case, the fan would not operate in a steady state, but it would more react to the changes of thermodynamics in the compartment from which we extract the air. So I am more interested in the response of the fan rather than having a constant boundary condition that I would normally use. Uh, therefore, because I'm an ANSYS user, I resorted to something in ANSYS we call the user-defined functions. User-defined functions, UDF for short, is basically a custom code written in uh C language that allows you to do some interesting things uh within your models. In ANSYS, I have a capability to trigger an action which reads a state of specific variable point statistic in my model, and I have an action that allows me to refine a value of a boundary condition, and those actions happen in between every time step of my analysis. So every time let's say a second passes in my numerical model in my CFD modeling, the CFD software would read up the temperature of the smoke gases near my fan, then I would have a mathematical algorithm that would tell the model how much I should increase or decrease the capacity of the fan based on the current temperature, and then this new value of the new capacity of the fan would go into the model back again, fitting the new boundary condition of the fan. If you do it too rigidly, you can end up with oscillations. So I had to soften down this algorithm, include some delays, running averages of temperatures, etc. You know, to make sure that the response is not instantaneous, but rather happens over the course of 10-20 seconds. So the fan is slightly lagging behind the changes of temperatures in the rooms, but it also creates less abrupt effects in terms of uh hydraulic bumps in the in the network. It basically makes the operation much more smoother. The main point is that when I defined this in my CFD, when I provided those functions to dictate the performance of the fan, I was able to capture those effects. I was able to observe everything I've just described to you in a compartment fire, but within a CFD model, within a simulation that exactly show me how those uh properties of uh air change as the fire grows, and also how the fan is capable of responding to the momentary needs in the system. It was very, very interesting. I have two case studies published like that actually, one from the original 2017 paper where I've introduced the mechanics of the transient response to change of temperatures. It was published in International Journal of Hidden Mass Transfer. It's I think my only solo paper I've ever written. This is 100% me and my chain of thoughts, my programming before chatbots. It's my scientific achievement, one could say. And this and then this in this paper contains a case study of a car park. And there's a second one which contains the case study of the underground venue on which I record this podcast episode as well. Uh, both are available online, and the links are in the show notes if you would like to uh see the CFD. What happened later? I I guess it's to an extent disappointing because uh I really thought that I've discovered a huge thing and I could change the world, and obviously the world has not changed, and there's a great chance it's the first time you hear about this concept, even though 10 years have passed almost. Um, reasons for that are multiple. One, uh, this was definitely something that one could patent and one could perhaps build a business out. I mean, I found a way to fit much higher extraction capacities at a smaller shaft size, which means I also discovered a way to significantly decrease the sizes of shafts in new buildings, which has significant monetary advantages to that. And I could technically monetize it, but because I'm a public servant, I work in the in a public uh research institute. Um, I'm not really allowed to design systems. And we work in I also work in a laboratory. I'm I'm uh as I said, I'm responsible for testing those fans. So I cannot design a fan system for manufacturers that monetize that and be a uh a judge at the same time. I had to pick one and I've picked my existing career. Unsure if that's the greatest choice of uh of a person has ever made, but uh it kind of worked out and I'm here where I am based on that decision that day, so I'm happy with that. Uh basically, I could not monetize on that, I could not patent it really. Uh, the choice I made back then is to release this to the wild, to release this to the public through scientific papers. And I written, I think, three or four papers on this, and they're all available online. If anyone wants to do anything with this, be my guest, do it. It's it's it's I think it's a great concept, and and it's such a shame it has not, to the best of my knowledge, it has not been used really in practical engineering yet. I also Had not the greatest, you know, chances to promote this worldwide. Yes, I went to the SFP conference in Malaga with the talk about this system. I was super excited by that conference. I thought that this is the conference where I will show this to the world and I will make a massive impact afterwards. And uh, two things made it unfortunate. One, I was giving a talk directly after Professor Viegas, who was talking about wildfires in Portugal, and that was such a powerful talk. Like, really, people were crying in the room when he was mentioning the stuff that happened back then in Portugal. Portugal had one of the biggest waves of wildfires, and he was giving an exact description of what happened, how it happened, how people died, how responders had a hard time answering the threat, etc. It was such a moving presentation, very emotional, very like really powerful. And I was next, you know, a young guy from Poland with a funny attitude. It was not a good time for jokes. Um, I delivered my presentation, but I think a lot of people uh attention, their their heads were elsewhere, they were not really following the complicated compartment fire physics that I wanted to give on the conference. And the second thing uh it's it's kind of funny, but uh the room next doors, there was parallel sessions. There was uh a talk uh by Christian Malouk. I think he was talking about CLT. Christian Maluc uh at that time he was uh shortly after winning IFSS Award for Best Thesis in his prime researcher days and uh big name, and uh I just saw you know uh exodus of people from the room that were going to the other room, so I lost a big chunk of the attendance to the other talk, so yeah, not that many people attended mine. Uh but uh yeah, I thought I'm gonna change the world at that conference, and I obviously did not. At that point, I only had numerical proof that it works, I only had simulations to show that it works. Uh, in some years later, I've actually managed a collaborative project with a company, Flaktwoods, uh, from the UK, uh, who also have a Polish office, and back then I was working with the Polish people from that company. And together with them, we've actually set up an experiment. So we've purchased a C container, I fitted it with burners, they've uh made a control panel that could do this smoke smart smoke control action on it, and we've actually tested in full scale the concept, the whole of the concept. Like can the fan adjust its performance to the um momentary temperature changes? Can uh the fan react? Can we extract more air? What happens to the air velocity in the duct? What happens to the air velocity at the makeup air? We went all the way from 20 degrees ambient to I think 500 something degrees high temperature flows, uh, whole spectrum, two different fans, a small one, a huge one, able to capture the the whole spectrum of effects of this smart smoke control. And actually the experiments actually have confirmed this operation. Unfortunately, we this was something that we have never really finished. Something that we have that that those experiments happened shortly before COVID. I think it was maybe one of the last experiments I was capable to do before COVID 2019, end of the year. Uh, then we were buried with the COVID problem, and I've never really had the chance to sit down on the Riesels and process them to a paper. Uh then the war in Ukraine started, it also caused a massive disruption in Poland. FLACT at the meantime tried to build it into a commercial project that is uh possible to patent for them. So they uh I mean I had nothing against that. What they wanted to patent was the control algorithms that they developed on their own based on my general idea. So that was fair play for for them to do that. And I'm not sure if they were successful with patenting that. Eventually, um, yeah, some years have passed, some people moved companies, uh, a lot of energy went down. I still have the results of those uh of those experiments. Perhaps it's a good moment to revisit them and maybe publish them uh so you see that experimental proof that this works. But uh since then it was really not ever used in practice. So yeah, that's that's the story. Uh I'm not sure if it has a happy end. I mean, the happy end for me was the joy of the discovery. It was really a very interesting thing when uh stuff connects passively in your head when you're not thinking about it, and you suddenly find a solution for something you felt is impossible. I still think this has a future, I still think this could be used in practice. Perhaps I should focus a bit more about how to implement it in practice more now that I am in a different position, now that I have different audiences, I have you, I am able to act at different levels, I don't need a conference to spread my thoughts uh worldwide, and there is no parallel session next to me, and therefore I having full attention of the listeners, maybe I could perhaps change the world. Now, as I think about it, uh and I was thinking about it the entire week, the whole uh idea of how smoke control could evolve. And I've been in multiple conversations with multiple people in the recent days about the state of how we design smoke control in buildings. And if you think about it, not that much has changed since the times of Howard Morgan, Margaret Law. Like the same relationships are still being used to design the smoke control. Um, I tried to frame it, uh I tried to build a framework. So I would call it smoke control 1.0, where you you have your hand calculations to approximate the amount of air that you have to extract from a room or the temperature of that smoke that you're gonna extract, or you know, whatever relationships that give you a ballpark number of how much air you need. That that's what I would call smoke control 1.0. It's based on empirical evidence from countless experiments, it's based on first principles of physics, it's based on basic laws of physics, thermodynamics. It kind of works, but we all know that it delivers not the best systems in terms of their efficiency, in terms of them being best fit for the particular compartment, particular space. Then in Poland we have something that I would call smoke control 2-0, where instead of those fundamental laws of physics, those basic empirical relationships, we employ advanced modeling techniques to discover what kind of parameters of systems we need to our uh locations. This connects the concept to the architecture of the building. This makes the architecture the part of the solution because you account for your irregularities in shapes, for your smoke uh reservoirs, etc. It uh allows you to study the local effects, it allows you to study things like jet fans, things like effects of beams, spiel plumes to a better degree than simple uh spiel plume models, Harrison spear point ones. Um, there's a lot of things that you can suddenly see and optimize with numerical modeling, but the limitation is you do that for a specific design fire scenario, and that design fire scenario has no interaction, no way of interacting with the smoke control system. It it like happens in the parallel universes, the fire growth and the smoke control. Those are technically in the same space, but they don't really talk with each other. There's no feedback loops between them. And then there's something I would call a smoke control 3-0, where the smoke control system adapts to what's happening in the building, it reacts to the fire, it changes the fire, and fire changes the smoke control. It uses the thermophysical uh effects of the fire against the fire to make the extraction more efficient, to make it work better, to extract more, to extract with lower velocities, for example, to provide air with lower velocities, to reduce the outcomes of the fires. And it is kind of sensitive to what's happening around. And it's not just, you know, the smart smoke control system that I've talked in today's episode. It's also about how you can design smarter systems in tunnels, in in car parks, you can design smarter pressurization techniques, you can design smarter corridor smoke control systems. I have thoughts similar to the thought process I've uh explained in here for basically every single type of smoke control, natural ventilation, aerodynamics, effects over the roofs. A lot of those thoughts, uh my entire career was on building concepts like this, and I think it's it's time that I start to put it all together for some sort of philosophy of uh smoke control 3-0. Maybe I will recall the whole podcast episode about how this concept could look like, and I'll give it to you and you tell me if it's right or wrong, and uh maybe I'll eventually write a book or something. That would be nice target to write a book on smoke control 3-0. Ah, maybe one day. For today, I think that would be it. Uh it's uh a complicated thermo thermophysical, thermodynamical uh episode which uh shows you uh how thinking beyond your normal basic assumptions of fire safety engineering can uh give you new new limits, could give you new boundaries, could give you new performance, new capabilities that you have not even thought you could have. I didn't thought we could fit that much air through a duct, and yet uh we've shown that it is absolutely possible. If you're interested more about the concepts that I've talked about today, uh three papers are in the show notes. Please go and read them. Let me know what you think, and be sure to come back here next Wednesday for even more fire science coming your way. That's it for today. Cheers, bye.