080 - Adaptive Fire Testing: A new foundation stone for fire safety (ERC StG Grant) with Ruben van Coile

Hello everybody. Welcome to the Fire Science Show. Having this podcast, I really like how it allows me to react on what's happening around in the fire science world. And just two weeks ago, I received a very, very good news and that was the announcement of the winners of E R C starting grants in Europe. And this year, actually for the first time in the starting grand scheme, there was a grant about fire. This is huge This is amazing that it has happened. The grand benefactor is, uh, no other than Professor Ruben van Co Multiple guests of the podcast, a friend of the podcast. I'm super happy for Ruben's achievement and yeah, you may think getting a grant, that's not a big deal. Everyone's getting a grant all the time and it's, it's just something that runs the coquilles of, of science. But no, this scheme is, is special. This is something meant to fund breakthrough research. Research that can truly change the. Discipline, their research represents and the project you can imagine. I, I've invited Ruben and we're gonna talk about it all the episode, this project. Is something that can change fire engineering and fire science, and I'm super happy about it. To get a ERC grant, you get through an insane process. There's, I think up to 12 review as per each grant application. The applications must be absolutely excellent. Good is not enough for erc. They must be excellent. They must be super well thought. They must be super well prepared. They must be extremely well laid out. So, even submitting a grant is already an achievement because you had to really think out what do you want to do, and, getting one is approved. That, that your Sound and it may change the world we live in. So I'm really excited that such a grant has happened , in the world of Fire. It used to be my dream to get an ERC founding one day, but then I started a podcast and I. prefer this, this pathway actually rather than, fighting for funding schemes. But I'm happy that there are researchers at the absolutely highest level who do fight for these schemes and they win it. And today, yeah, the episode is all about Ruben, his achievement. He's your C grant. Let's go. Hello everybody. Welcome to the Fire Science Show. I'm here today with, Ruben Van Coil again. Hey, Ruben.

Good to see you.

man. That's such a nice, uh, opportunity to meet because you just received ERC starting grant. Yes.

Amazing. Thank you very much.

Amazing. Absolutely amazing. And, for those who don't know, this is the, um, European Research Center grants that are awarded yearly. are five years projects. probably the biggest research grants. single researcher can, can receive in Europe. And, it's very rare that the fire, I mean, it's very rare for a civil engineer to get one. And for fire, I think, for fire engineering only Guillermo had had, e R C funding and we actually covered that in, in a podcast in episode two. if you want to check out the, the, the scale of a grant towards the end. Today we're gonna talk about the scale of a grant at the early day. So I'm gonna read the title and you give me, uh, your elevator pitch because you, I know you must have had, had trained that. So I, I read the title, you give me elevator pitch about what it is, and then we go into details. Adaptive fire testing and new foundation Stone for Fire Safety,

Indeed. I would say my official elevator pitch, I'll, leave that for now cause it was like, it's very, let's targeted and also for people who actually already read the proposal, right. So what we want to do is we want to develop, Tools and methods to just change how we, demonstrate fire safety and buildings. And there are different aspects to this. So first of all, if you think about how, Demonstrating safety and buildings is done. Now we have this prescriptive, uh, fire testing, which with pass fail criteria, and what we want to do is want to move towards testing, which actually teaches us or how the build the building product will behave in a real fire. And, and this, there, there's lots of uncertainties there. So this is definitely, it's not an easy question. You want to quantify how the building, the product behaves depending on the. Itself. Secondly, when you have that, we also want this to be incorporated in demonstrating safety for a building. So we don't want to just have prescriptive rules and saying, ah, it has to be this kind of product and we don't want to propose better classification methods. No, we want to have an evaluation, a real-time evaluation of the risk profile of buildings and what. allows you to do that. I mean, it allows you to do many things, but what it allows you to do is, just step away from what you think of as prescriptive guidance. So there are no rules anymore. What you now have to have is a building, which passes safety criterion. And we can talk about safety criteria later. Cause I mean it's not an easy statement there, but fundamentally you evaluate the risk of a. and if the risk is too high, you have to change something. but if you think about this, say like, okay, this, uh, this all sounds very nice, but you have to change something. And then how can I now possibly evaluate what to change? Well, yes, it's a very difficult question. You need to do medical calculation. So we're also going to do surrogate modeling. We're going to do, basically, you can call this machine learning, but we're going to do physics based machine learning. So physical constraints, physical backgrounds to the model. So getting to machine learning, which actually make, where we make sure that our results, comply with the basic physics really. So that's the core of the project. There is one additional point, I like very much. abdo would be a challenge as well, is that there's a work package on, let's say convincing stakeholders really. So we're going to do a cost benefit evaluation of compliance frameworks, the existing framework, what we propose and alternatives possibly so that then at the end, hopefully. Show, or at least we, we provide a method to show that this approach, which we are proposing actually has a benefit to society compared to just sticking what we have done forever. So, and that's the starting point of, of the proposal really saying like, look, We need to change this. We need to have a new way of testing, and this new way of testing should not be one with pass fail criteria, but one where we test to increase our knowledge, so either it's very specifically product development. We're going to increase our knowledge and reduce uncertainty in the behavior of a building product. In fire or very case specific and targeted. You have a specific building, could be even an existing building, and you want to perform a test so that you demonstrate that your building is safe. Now, this might raise lots of questions, but we can talk about that later, of course. Now, to enable all of that, you need lots of stuff you need to do. Uncertainty, quant. You need to, do what is actually a value of information assessment. What that means is that you are going to evaluate before you do a test, what you think the test is going to give you, and that way you can compare different tests. I mean, this might sound very weird when you hear first time

We we're gonna jump into that in very details. So my attempt at summarizing that, are designing buildings based on. century assumption of what fire performance is, fragmented into elements, parts of it. Like not a building at all. I just had a Cardington episode in the podcast where it was 1990s, where they, for the first time, I think, looked at the building as a whole, not as a sum of its elements. Right. 1990s. given that fire testing is a hundred something years old, right. so, we need, to at it in a, in a different way. And, and, uh, current fire testing that's based on fail pass and just, single element test is not meeting that at all. So, so you would rather keep the fire testing. I appreciate that as someone who is running a fire lab thank you, It's great that you see a future for us too. but you would like us to not be the judge and say, this failed this past, but you would like us to build up any information,

Yes, we can maybe even be a bit more targeted there. So you had the, the episode of On Fire Resistance, some, say, weeks ago. And at, at some point in that episode, if I remember correct, you. You're talking about, okay, you can do testing and you can do numerical evaluations, and uh, if I remember that correctly, it's like, well, numerical evaluations are the alternative and we, we know this is not the way to go. There cannot be an alternative to testing. I would say, look, let's not split those things fully.

Hmm

Actually, you integrate them and it is your numerical evaluation, which defines which test you should. And it is your test, which then informs your numerical evaluation. I mean, it's, it's logical, but, and this is the, the core of the proposal as an, okay, it's easy to say that, but how are we actually going to do it?

Yeah, yeah, yeah. you got me at the pass file criteria, like, uh, I, I really see no point if I'm doing a fire test for a, a person and, they have a 60 minute wall and, uh, the test stands at 59 minutes and they haven't passed it. And, uh, then they have. To come again, the same wall, 58 minutes, then again the same wall. 62 minutes. Yes. Now this wall is good And, and that's, uh, that, that's ridiculous. that's a system that existed for Century. And has led to, whole industry focused on passing the tests that that's what we have today. it doesn't mean that it doesn't work, our buildings are safe. . But then again, we, we did, uh, with my friend Grzegorz um, he was also on the podcast a very unformal survey with, Building managers, uh, construction managers in Poland who were doing, uh, large investments like shopping malls, airports, like very large building skyscrapers. we were very informally asking them like, do you think, how, how much of the whole building cost that you are building, how, how much of that would be like fire safety, like, or everything fire safety related. The, the passive active fire protection, everything design. And the, the answer was very consistent between 10 to 15% of the building. So we are actually spending quite a lot on fire safety. If you think that 10 or 15% of the building industry, which is one of the biggest industries on the planet is related to fire safety. So that's a tremendous amount of, of resources, global scale that we may be using in. In a stupid way, to be honest, because, what's the difference between two and a four hour wall in a skyscraper? Like really for the, the love of God? If the fire lasts, seven or eight hours to destroy a four hour wall, the fire resistance of that wall is the least of your problems. To be honest. so, I guess your, your grant is also looking in into that because, you want to change the paradigm like completely right.

Yeah, so you, you raised an important point there. So when, when we talk about. our prescriptive, regulation, but regulation can also be guidance. That terminology there doesn't matter too much for now. we are not so clear on what we want to achieve. That's, that's one. And very importantly, if we think about how innovation is moving so fast, and this is subjective, I admit, but I'd see no way how a government body or any other institution can just keep up with it and keep on, improving or adapting their prescriptive guidance. Right. So something has to give. So either your prescriptive guidance just lags behind. Maybe that's the traditional way where you wait for failure to happen and then we'll, we'll try to sort things out. Or either you are hampering innovation, right? You, you just say, ah, this is not allowed, and why not? Well, because it doesn't meet this requirements, although it doesn't make sense. So what we also want to do within this proposal, is use this adaptive fire testing we talked about earlier where will now get a much better view on the behavior of a building product to get a risk-based evaluation of the fire performance of a building. So if you do this in a proper way, and this so conceptually, make it it simple, but it'll be very difficult to achieve. an architect. Design the building, submits the plans electronically to the building authorities, and then their computer system was probably a surrogate model. Then evaluates the risk of this building, taking into account all of the information it has on building products, previous designs, and. so, and then you get a evaluation of the risk of this building, and this is what has to be accepted. I mean, this is the only thing that really matters

You, you, you've brought something new now. there we, we talked about the, the product. So I guess it would be the, um, manufacturer of the product in interested in this, uh, approach or, or, or one of the stakeholders involved. There's an building designer and now, now we've mentioned building author, like, what is their place in your. Like today and, and within the adaptive framework. So you said that the goal would be to provide them some sort of automated tool metric, uh, system.

Yes. Although we have to, make a distinction maybe for the, for this project between, so the, the vision, right? And I can imagine many people have this kind of vision where vision, where of a instantaneous evaluation of fire risk for buildings. , which means that the authorities do not have to prescribe things anymore because you have the risk evaluation. The prescription is just some kind of substitute, right? So what you actually care about is fire performance. So this is what we want. That's the vision. Now, how do we get there? Well, therefore you need some tools. You need methodologies within the proposal. Different work packages, we are developing some of those methods, some of those tools, but you should not think them as, uh, a finished product. It's more like, okay, if we manage to do this prove of concept, then hopefully we can convince other people, institutions, researchers, government, and so on industry to invest in this and. With our, uh, proof of concept, build this new way for, uh, demonstrating fire safety.

Fantastic. Fantastic. okay. let's talk the case study because I, you, you picked an interesting case study, which is, glazing and, um, and Lord burying glazing. What, what, um, made you. Pursued this path. Like why, why, why glazing as your case study to demonstrate this new methodology?

Well, that's .that's a funny question.

I find it intriguing. It wouldn't be my first choice.

so writing the C proposal. I go through different stages, right? So almost a year in advance, uh, some preliminary ideas, but don't go into depth. And then I, I think in, in the summer before, submission really going into detail and went to different versions of it, the glazing was not there in the first version. And then you ask feedback. And feedback is because I was working with concrete, which is the typical thing which we do in our lab. and some Gibson burst, as well and said like, why concrete? Like, oh, cause it makes me feel comfortable, , and then in, right. So what, and, and say, well, Highlight that within the E R C, it's about high risk, high gain. What is challenging? What is, a breakthrough and swamp? . Okay. That's tricky what we have now. And so there are different arguments than in favor of the glazing. What we have now on the structural engineering side is that there is a new euro code coming for glazing. This will not have a fire part yet, so this fire part has to be developed. There's a lot, lot of research which is needed there. So, okay. That's, one reason why glazing is, is relevant on the fire part then. Our glazing in buildings, not load bearing glazing, but just, genital glazing, defines how our fires develop. Glass breakage. So from that perspective, although there has been a lot of research on glass breakage, there is still a lot which can be done, to just improve how we, how we do fire dynamics calculations, really. So provide input there.

Mm-hmm.

this, uh, these are two arguments in favor of the glazing. Actually, there is also a third. Just being complete here. Is that within, the department? One colleague of mine is, uh, let's say expert in glazing. So that's also really nice and convenient.

That's

Yeah.

I'm not surprised that glazing is interesting. I, I had a podcast episode with Yu Wang and that, uh, was, fascinating, like to learn the complexities of glazing and, also have been testing multiple types of, of, of glazing. So glazing is interesting as hell. just hard it's really tough. But I imagine for e r c, you're, uh, not, uh, like concrete as a safe path.

Yeah, it depends. So indeed that was then the motivation. Although I do admit that within the set of reviewers, cause there are many reviewers for the E R C that there is at least one reviewer who said, well maybe you should have taken concrete, but.

Yeah. And you replied. Yeah, that's what I told them. Right.

No, no, there is no reply, which you actually have. You don't have this opportunity. Uh, you, I only got the comments after everything was closed, so after even getting the award.

Okay. so talk about, um, the pathways that, that you've, drawn for your, for your project, with the structure of work packages, how you want to. End up in this new paradigm for a fire's design acceptance in the 21st century. and you start with, with work package that, that's related to, to the information value of fire tests. So let's go back to the value of information theory and what, what do you really mean by, by value of information? How, how do you when an information is valuable and, what, what is.

Okay, so I'll, I'll try a sh uh, shorter answer first. We can go into more depth necessary. Yeah. Uh, that's one also, but, and I'm already expanding now. Uh, one of the things I've learned, or I'm trying to learn following the e c is, which I should, answer more, concisely . uh, so value of information. the point being when you do a. and in our field a fire test. You would want to know beforehand, which test gives you the most value? Are you going to do as now with an example, a 60 minute fire test? standard fire? Are you going to do a fire with a cooling face? What fire do you want? Because yeah, you are not going to have infinite number of fire tests. You can do test costs, money tests, have environmental impact and so on. Which one should choose? And I would argue that for the moment, at least in our field, this is not obvious. So within value of information, we will actually quantify.

Hmm.

and have an an answer to that question about which test you should do.

And here you use, examples of, risk metrics that are used related, into the value of information like FN, curves, all arp. How, how does display, is this like, done in the further steps or already like the test? The purpose of test is only to give the, the information and inform the fir the rest of the process.

that's next level, right? So, I'll, I'll park the, the risk parts for a second, let's say. Okay, so if, I hope this is make sense to, to the listener. So if you say, okay, I want to, know which test, is the most valuable to do. Okay, great. Then of course you need to have a, a metric. Which you're trying to, let's say optimize. And if you are a manufacturer within the current paradigm with standard fire resistance testing, and I'm just using structural, uh, examples here, but I mean, can change to anything else, then this could be, an evaluation of the uncertainty on your, uh, just justify resistance time. Right? And you want. Make sure that you are on the safe sides, that you have your fire resistance time is definitely more than 60 minutes. The average is 65 standard deviation and so on. So you could target something like that within the current paradigm. within product development. So if you're not even testing for compliance within product development, you could just test your specimen to reduce the uncertainty on its perform. because you're in a, you're a developer and you just want to know if, how the product is actually behaving. Now, next level is, okay, we are now within a approach of evaluating the fire performance of buildings, really quantifying that. That means that we are evaluating, for example, the risk, high risk of a building. Because of the uncertainties we have, our building does not first comply with our target.

Mm-hmm.

Which test should we do so that we think that we will switch or move our risk assessment so that now it'll comply. So you have an existing building. It has maybe an existing masonry wall or an existing timber column. Um, it does not comply with, our risk target. Which tests are we now going to do, which gives us a better assessment of the risk and hopefully if the test outcome is positive, but okay. Shows that uh, our building is safe.

And, and that is, through understanding, performance of a completely new untested, previously material composition setting of materials, building element, or through reducing us uncertainties of something that we already knew maybe right.

Exactly.

So here's the feedback loop that, that you first want to understand the full building performance as proposed and find out, okay, what I'm missing here to, to get a better information of this risk or get the risk profile that is acceptable because I received unacceptable risk profile, right?

Yeah. So there, there is always uncertainty. So uh, you have a small building, low rise, nothing special, not nothing fancy on it. And then there's quite some uncertainty on the fire resistance of some slap. But you, within the risk evaluation, it is quite possible that you don't need more information on the fire performance of the slap. That within the mean, there is uncertainty on it, it's taking into account, but actually having a better assessment of the SLAP fire performance doesn't teach you anything regarding the building. While maybe. Again, structurally, but maybe there is a column which actually this thing is the thing you should be testing,

Mm-hmm.

So, and this is, I would say, not taken into account really in the, in the current approach and within the risk-based approach. You could do that.

Fantastic. I always thought that, that there's a space for performance to base engineering in fire lab. That, that it could, give, than just index and, you are actually showing, a realistic pathway to, to achieve that. That that's, I mean, everyone could say, okay, we could test the things in different way and then receive other things. The, the framework is necessary here to make it useful for the industry because if there's no framework, then we, we cannot, now to enable the adaptive fire testing. You also have work package too, which is surrogate modeling. I guess this is what you said in the beginning of the interview, that we shouldn't, treat them sub barely. They, their simulations and, and testing could be used all the way together. Is this circuit modeling also a part of, of this thinking?

yes. So model modeling definitely is a part of it. the point there being that if you have advanced modeling, then just the, the computational expense quickly becomes very big, right? And if you do the value of information approach, What you need to do is evaluate before doing the test, what a specific test results will teach you. So you need to actually update your model in many different ways to then find out which test you actually want to perform. And you could think of this as confirming, um, which model to choose. Yeah.

Almost like a sensitivity study before you do the test. Like which test would matter the most?

yes. This is, you could put it that way. Definitely. but it means in that way that you would have to evaluate your model lots of times. So it becomes computation. Very expensive, very, very quickly. So the obvious approach, would be to use circle modeling and called machine learning. either way. So a, a model which is computationally very efficient and which approximates very well, a more advanced model. Okay, that's great. Now I have subjectively some concerns with black box. surrogate models, and just as an example, When you ask somebody to do a, to use a machine learning methods on a fire engineering problem, then very often you will find non-physical results, as in they've trained their model and the the results are excellent and whatever, whatever you, where all the metrics. And then when you go into detail and you, you push the model a little bit, you find out that there are combinations where, for example, there is a fire and everything cools down. Just because the surrogate model has interpolated and, has been trained in a way that it gives non-physical results work package two gray surrogate modeling is the same approach. So it's, still these methods of surrogate modeling machine learning, but with physical co. So physical constraints in both within the model itself, within the cost function. So when you do machine learning, you have a cost function, which you're trying to, to minimize. and also as putting those constraints. So we're going to explore that. Can we develop again, proof of concept, ways of making machine learning tools, which actually comply with what we know about how fire behaves, how structures behave.

Man, this is, uh, it's like a dream. It's so funny. You're gonna get to do this now, another question. you were kind enough to send me the short proposal for e r c. You're sending the short one and the, the long one. So maybe this is covered in, in the long version. but I couldn't find it anywhere in the short one. I, however, however, I spotted in on a figure, so, on the pathway between the designer and architect towards the authorities. You said that they submit um, digital blueprints and, on the framework it's, it's b i m model. So I wondered like, how critical is b i m technology for your project and, how do you intend to work with that?

The easy answer is it's not included,

It's not included. Oh, no, Okay. There's a space for the second E R C grant, you take it as this.

not even that. Not even that. So, this is a bit of a, a different perspective maybe on, on the, proposal or on the, the E r C. So I, I was struggl. in the beginning to have a, let's say, a compelling story, then showing that this is research worth supporting. So I have, wrote down this, let's say, Vision where you have inin instantaneous evaluation of the risk profile of buildings. and how would that be possible? Feasible in a right would be, for example, with, as far as I understand, BIM models, being evaluated by a surrogate model. Right? So that's the storyline.

Hmmhmm.

I say, well, if we want this, we need to develop these kinds of tools or proof of concept tools. And then those are, different work packages. so the actual, how are we going to extract data from the bim and which specific, Things do we need and so on. This is not part of this proposal and actually, um, now jumping forward to some, maybe some other work package, actually, I do recognize that within the proposal in sense that I highlight a lot that, there, if this will move forward, there will be large follow up investments needed and which is why there's an entire work package really targeting, if stake call.

Very good. that's, that's a smart move. And, uh, I think there's a lot of, uh, movement in BI and fire, uh, mainly in, um, Code compliance. this is a path for BIM for sure where, prescriptive code compliance could be assessed through the BIM model almost automatically. I think that would be a very compelling, uh, thing to see, such an assessment done in parallel with, estimation of the risk profile of the building. You know, the, the fundamental question that that lies, uh, is. According to Low Safe and how safe it is, uh, which we don't have a good answer for that in, in your framework. I, I don't think that, uh, that's, that important is safe enough, uh, with pro because you, like, you are not interested in the prescriptive. You want to break this, uh, You want to shift it completely

Well, I still call it prescriptive, what I call it, uh, let's say the 21st century prescriptive. Right.

okay. Okay.

So,

you replace just a set of rules with, uh, risk profiles and inform the metrics and, surrogate

exactly. Yeah. And then if, Somebody who hears this is things, well, this is definitely not possible for every sit single situation and so on. Yes, fully acknowledged. So within this called vision again, 21st century prescriptive risk evaluation. then you still have your performance based design approach. But now with better data and better input. Also the adaptive fire testing approach, where if something is not working out all within your performance based design that you can know or at least assess which test you should do to get a better assessment. So definitely I am not saying that there is no space there for fire engineers. On the contrary, I think within. Framework, I concept fire engineering is much more engineering than maybe sometimes it is today.

I had a discussion with Xinyan Huang at some point, and, We were talking about machine learning in, in general, and, he, said something that really. Resonated with me that, you know, we can describe nature with, an equation or a set of equations. We can describe nature with, with a three-dimensional CFD model or something. And, And, there's a neural network that describes the, the, the reality or, you know, some phenomenon in a way we don't completely understand. But it's a different language of, describing. If you can, build that. It works. it's not the same thing. And in here again, you can describe your building with. you know, requirement that walls shall be R 60 and the windows shall be R 30 or EI30 30, and the doors shall be this. And you can have 40 meters of your path and, and stuff like that. And you can describe the building with a completely different, uh, language that you would like the risk profile to look not worse than this. And, when assessed in, in this particular way. And, also giving a, a tool how. Adjusted risk profile. So if you want to adjust this risk profile, you don't, change the fire resistance of your wall from 60 to 90 minutes as you could do today. Like today, if I want to, uh, I don't know, reduce the, the fire resistance of my walls, I have to put sprinklers in it. And that's my cost benefit analysis. The world cost this, the sprinklers cost. This I benefit by, by making the building cheaper, not, not. Probably safer, but I, I don't know that in, in reality, in your approach, if you succeed with that, you, you get, a full package of tools, to know the risk and then based on the, on the risk That's, that's beautiful.

Maybe one thing to, to highlight is that, well, I think that the, the paradigm, how we. Demonstrative building a safe that should move. I am not saying at any point, let's throw away all of our existing products and, and the, the results we already have. because, so it's as if you gave an example earlier about high rise structure and fire resistance on, it's as if you, if you, you have this Okay. BIM model of this high rise structure, the risk profile is evaluated. It comes out very good. Maybe even you can reduce the classification of your, of some structural elements automatically or not, and find out that this does not influence your risk profile. . Or on the contrary, if the risk is too high because there is this low probability, high consequence event, you might find out that okay, but in using our existing classifications, the assessment of the model is now that, okay, you have now adequately reduced, this event.

Hmm.

So if a, if it's too much of a revolution, as in throw everything away and we start from scratch, this is just not, not reasonable, as in it's not going to happen.

And. how do you feel like about blind spots of this approach? Like, if you are measuring based on, on a set of, assumptions, even, even if you try to do it holistically, and, okay. The current paradigm is not very innovation friendly as well, but something might escape the, the set of rules. So you, you hope this will be continuously evolving? Uh, we will be building these, measures to, to incorporate new, new challenges because if, if you base, the safety and the risk profile, you really narrow down the safety margins. You know, in, in prescriptive fire engineering, the safety margins are, are enormous in most of the things. So, There are chances that if something like comes, beyond the radar uh, safety mar margins would cover that and, and we are still protected. Maybe that's the trick why the current paradigm worked for a hundred years because we, we were not aware, uh, about the safety margins in your approach. By data mining. The, let's say true fire safety of your building. you are very devoid of this safety margins. This, this one, the ones that we don't know about because you want to understand the performance completely. So there's less, let's say, uh, for a mistake. How, how do you feel about that?

That we need to be, careful in what we do. And this might, this, it's a bit of a, a kinda liberal statement maybe, but the less, I would almost say, government intervention, the better. And, and what I, what I mean with that if you prescri , Wojciech Wegrzynski: that. if, if you prescribe specific target, You are a hundred percent, you're, you are right. Right. You prescribe a specific value. and yeah, then it, what is the cheapest way to just pass again? But what I hope and which we will is automatically enabled if we. Do what we are trying to do. What I hope will be, uh, will happen is that you do a la evaluation and, or you can use any other words, whether you want, but you look at the costs and benefits of fire investments, right? in, in the best case. Right. And which I actually also believe it actually fulfills the promise of performance-based design much more. Right? So you will have, you could have cheaper buildings because you will remove things which are actually not helping, which have better performance because you're going to add things which actually do improve safety. So, and, and safety, both for people. It could be for the environment, it could be for property protection, whatever you want. So, if you set a target, okay, this is always tricky. Which, what should the target be? actually, I like those targets for thinking purposes. Cause it helps reasoning, but what you should do is always think about, okay, but in normal situations you always end up into alarm. and you invest to the level, which is, the risk is as low as reasonably practic. So, yeah, you, you should not be close to that target anyway.

Fantastic.

I'm also just being a bit cautious because it's, it's one thing to, to put a vision out there, which I can imagine many people have, could have similar visions and then say, okay, these are the things which we need to develop. am I sure that we are going to develop some nice things? Yes. There, I mean, there's risk mitigation the project like in every project. So yeah, we are some things. Happen. do I know now where we're going to end in five years? No, of course not.

Now this is the, this is the ERC projects that high risk, high gain, by saying multiple people might have Similar ideas of course, because it's evident that our approach is not perfect. I guess Kunio Kawagoe I had similar ideas. I, I guess a lot of people on the road had similar ideas. There's a big difference between having an idea, to having, Well thought out pathway, how to go from idea to a solution, which is where you are now. And I hope in five years you are at the, at the solution. Demonstrate for one case study at least place, and maybe, maybe somewhere where you are, um, where you have something very general that that can just work with the whole industry. And I think it's a, it's a very good shot. Okay. I guess, we've covered project, quite well. I mean, it's, it's a little chaotic because the projects in e r c, they are very not linear. to the listeners, it's not that Ruben is proposing I will do this test, which will give me this coefficient, which then I will use in this model and I'll receive it. But yeah, I hope you guys got a really good Where this is heading, or more importantly, why and how Now, I, I would love to ask you, because it's an opportunity for the, for the fire, for you as a, as CRC grand recipient. This is obviously a, i, I guess a career changing, moment. But you also need, need people to work with, so, let, let's talk about pragmatic stuff. There's a lot of, future potential postdocs, uh, listening to the show. So, so let's talk about the resources you're gonna need to, to fulfill that. how, how do you plan out this part of the project managing all, all of this.

Well, I had to, um, write down a proposal, as part of the RC application. It's currently foreseen that there are five people working on the project. Those are, three PhDs. Four years each, and then two posts, three years each now. But there is flexibility there. So if in the end it turns into three posts, two PhDs, I mean, there is nothing stopping us from doing that.

Hmm.

Uh, one thing, which I mean, if we are talking. Pragmatical realistic here. One thing which is quite specific for, for Belgium maybe and some other countries as well obviously is we have, so the u c is 1.5 million euros, so that's a lot of money. part of it goes into just general overhead, but then we have the rest of the money is basically all of it is for people because we, um, we already. test equipment and, and we'll manage on, on other aspects. But in Belgium, it's, Belgium is probably one of the best countries in the world to do your PhD as in the, it's, the wages, are very high, notably for, for PhDs. And, think this is why , we have a place for, for five people within the project and, and.

Okay. If you did in Poland, you could afford like 10 sure. Uh, that's an opportunity. It can open and oversee, uh, department of of the project if you

So if the

just kidding.

as long as we stay in Europe, there are some things, uh, no joking there, but um,

No, we're not leaving. No worries.

The good points. So with within the projects place foreseen now is that in, let's say September two PhD students would start, but as I said, if, uh, if that would be a postdoc and a PhD would be happy as well. It's of course crucial to have, good people for the job at hand. So, and actually this is maybe very specific, but it's indeed it's an opportunity for me as well. The research project I'm seeing here, if I have my timeline nicely. so the first PhD would be working on this value of information aspect,

Mm-hmm.

actually working towards it throughout a PhD. making a model. Glazing, collecting data, so input, uncertainty, quantification input there, actually doing the tests and then quantifying upfront what the test output is going to teach us. So this is the value of information. So this a PhD, which is just doing that basically bringing us to a first evaluation value of information for a fire test. the second PhD is on the gray model. But gray modeling with a bit of, focus on risk as well, I must say.

Mm-hmm.

this is, those profiles actually very different if you think about it, right? So the, the first PhD has some, would need to know something about structures, or at least be able to, to handle, let's say, infinite element modeling as well there, because you need to do tests on a tructure. Element glazing and have the modeling there with uncertainty, quantification, second profile, thinking much more about fundamentals of risk, what are the constraints and if we do physical modeling, how are we going to implement those constraints in our machine learning approach? So very different. those two. The nice thing of the project, is that the researchers, both the PhDs and the postdoc, they're actually very nicely. integrated in the sense that, I'm in a way quite proud of, of that is as an. One PhD builds up something, a post-OC comes in, they work together to, on one hand side, finish the PhD and the post takes over and moves it forward on the next stage. So there's

uh,

yes,

of, of knowledge goods.

at least that's the what is intended. Of course that can, there will be for sure some practical issues. Um, people who start a bit later. Things happen, right? So it'll not be exactly as I said it would be

Life always comes in the way, but it's, there's a reason why this is a five, uh, grant, not a two or three funding scheme. So it, it's meant to accommodate really grant ideas and support you with funding too. to help you achieve these grand ideas. it, it looks like, uh, I mean, not even day one. It's it's day minus, 180. But, it's already, it is already looking good at Ruben. I'm really happy that, this happening and, you have a really great chance to, to change, how we work. Hopefully, you'll you'll achieve that.

Thank you. Yeah, it's, uh, a challenge, but I mean it in a positive way. It's, um, the project is what, for me personally, what I wanted to work And there's, I think where I don't see, at least, I don't see any other way to get consistent funding for different work packages, which are actually nicely linked to each other, but still very distinct.

absolutely.

Um, so now this is, this is it. Now I have to, um, make it.

Yeah, that loud, the easy part. Congratulations, Ruben. Um, once again, and if you, if you do something, tell me on on air. I don't want you to call me. I want you to come to the show and tell us what happened and, uh, to anyone listening, uh, there's, as you heard, September two nice opportunities open. Keep your. tracking this space. This is a good space to, to follow and, uh, really nice to work, in a project that may change, uh, fire science and fire engineering. Ruben, thank you very much for, for coming to the show and I, I'm crossing fingers for, your project

Thanks for the opportunity to talk Wojciech it was great. Thank you.

and that's it. It's even hard to summarize the episode. It's, it's so good. The grant proposal is so good, and he really is changing the fire science. It's not the first time someone has proposed that the whole paradigm must shift. It's not the first time someone has, proposed innovative ways of testing, but it, it is the first time where someone has put it into one. Major, well laid out proposal with steps on how to get there, non-linear steps. I'm sorry if the parts of the episode were confusing. It's really these grants are really complicated and you have to jump from one part to another. I hope we've done a good job in explaining the, the base concepts and. Yeah, I, I hope Ruben will really, really achieve, uh, what he planned in, in this grant. If you are someone at the early stages of their career in fire science, uh, watch out, uh, announcements for job opportunities at the grant, it's an exciting opportunity to work at the project that may change the world. And as Ruben said, Belgium sounds like a great place to, to work for decent amount of money. So definitely. Opportunity I would be looking for. that would be it, uh, for this episode. Um, a final announcement, a final call. The survey of listener experience is still up, in the podcast Webpage will be up for one more week. Uh, so if you would like to share your feedback with me, please do so. I, I've received amazing feedback, great ideas. In five minutes, I'm recording an episode that's directly related to the request from the survey. This is how I like to react, to, to your needs. So if there is something you would like to hear on the podcast, if there is a way how I can improve the podcast, please tell me and I'll do that. And that's for today. I appreciate you being here with me. And yeah, see you here next Wednesday. Cheers.