220 - Test vs experiment with David Morrisset

WEBVTT

00:00:00.320 --> 00:00:02.879
<v Wojciech Wegrzynski>Hello everybody, welcome to the Fire Science Show.

00:00:02.879 --> 00:00:17.920
<v Wojciech Wegrzynski>I am very happy when people reach out to me and ask me difficult questions or drop interesting thoughts on me, especially if those people are potential guests for the podcast and they have something they would like to talk about on the air.

00:00:17.920 --> 00:00:20.879
<v Wojciech Wegrzynski>And this is the case of today's episode.

00:00:20.879 --> 00:00:33.280
<v Wojciech Wegrzynski>I am joined once again by David Morrisset from Queensland University, and David has dropped on me a very interesting thing to consider, and that is differences between testing and experiments.

00:00:33.280 --> 00:00:50.640
<v Wojciech Wegrzynski>And he did it in a very clever way by dropping on me an extremely interesting paper outside of the fire science world about dropping chickens into turbojet fans in terms of testing the turbojet fan uh ability to sustain a bird injury.

00:00:50.640 --> 00:00:57.039
<v Wojciech Wegrzynski>It's a very interesting read, it's in the show notes and we're discussing it more in the podcast, so maybe I'll not spoil anything else.

00:00:57.039 --> 00:01:04.640
<v Wojciech Wegrzynski>But uh indeed uh the differences between testing and experiments are something very important in our profession.

00:01:04.640 --> 00:01:18.159
<v Wojciech Wegrzynski>Actually, it's more important than you would uh probably think in the world of testing regimes and where fire safety is essentially largely dominated by codes and standards and standardization.

00:01:18.159 --> 00:01:23.680
<v Wojciech Wegrzynski>So tests are inherently a part of the fire safety system out there.

00:01:23.680 --> 00:01:27.359
<v Wojciech Wegrzynski>They're they're one of the most important key components out there.

00:01:27.359 --> 00:01:40.400
<v Wojciech Wegrzynski>And uh also fire safety engineering is a thing where you need data, when you need models, when you need to understand real-world fire phenomena, and you have to base that fire safety engineering on something.

00:01:40.400 --> 00:01:42.079
<v Wojciech Wegrzynski>And here comes the clash.

00:01:42.079 --> 00:01:46.959
<v Wojciech Wegrzynski>Can you base fire safety engineering on outcomes of fire tests?

00:01:46.959 --> 00:01:56.079
<v Wojciech Wegrzynski>This is probably the key difference between tests and experiments, the context, the reason they were performed, and what you can actually learn from them.

00:01:56.079 --> 00:02:00.400
<v Wojciech Wegrzynski>In this podcast episode, we will learn when test stops being a good test.

00:02:00.400 --> 00:02:14.800
<v Wojciech Wegrzynski>We will learn what it means to run an experiment, what it means to design experiments, and what caveats are there when you try to get some data or important knowledge outside of a test and put it directly into your engineering.

00:02:14.800 --> 00:02:20.319
<v Wojciech Wegrzynski>I think it's just a bunch of really interesting thoughts dropped in this podcast episode.

00:02:20.319 --> 00:02:23.599
<v Wojciech Wegrzynski>A conversation between two good friends who are both Fire Kicks.

00:02:23.599 --> 00:02:28.800
<v Wojciech Wegrzynski>So I really hope you will enjoy it as much as I did when I've recorded that.

00:02:28.800 --> 00:02:31.759
<v Wojciech Wegrzynski>So let's spin the intro and jump into the episode.

00:02:36.479 --> 00:02:38.319
<v Wojciech Wegrzynski>Welcome to the Fireside Show.

00:02:38.319 --> 00:02:42.000
<v Wojciech Wegrzynski>My name is Wojciech Wegrzynski, and I will be your host.

00:02:57.919 --> 00:03:04.400
<v Wojciech Wegrzynski>This episode is brought to you in partnership with OFR Consultants, the UK's leading independent fire engineering consultancy.

00:03:04.400 --> 00:03:15.280
<v Wojciech Wegrzynski>With a multi-award-winning team and offices across the country, OFR are experts in fire engineering committed to delivering pre-eminent expertise to protect people, property, and the planet.

00:03:15.280 --> 00:03:20.159
<v Wojciech Wegrzynski>Applications for OFR's 2026 graduate program are now open.

00:03:20.159 --> 00:03:26.719
<v Wojciech Wegrzynski>If you're ready to launch your career with a supportive forward-thinking team, visit OFRconsultants.com to apply.

00:03:26.719 --> 00:03:32.960
<v Wojciech Wegrzynski>You will join a worldless organization recognized for its supportive culture and global expertise.

00:03:32.960 --> 00:03:37.439
<v Wojciech Wegrzynski>Start your journey with OFR and help shape the future of fire engineering.

00:03:37.439 --> 00:03:42.319
<v Wojciech Wegrzynski>Hello everybody, I am here again with David Morrisset from Queensland University.

00:03:42.319 --> 00:03:42.800
<v Wojciech Wegrzynski>Hey David.

00:03:42.800 --> 00:03:44.319
<v Wojciech Wegrzynski>Hey Voice, thanks for having me back.

00:03:44.319 --> 00:03:46.000
<v Wojciech Wegrzynski>Yeah, well, welcome back, welcome back.

00:03:46.000 --> 00:03:47.919
<v Wojciech Wegrzynski>A swift comeback to the show.

00:03:47.919 --> 00:03:50.719
<v Wojciech Wegrzynski>You seem to be really into podcasting, my friend.

00:03:51.759 --> 00:03:54.080
<v David Morrisset>I mean, I just like I like having a good conversation with you.

00:03:54.080 --> 00:03:54.800
<v David Morrisset>What can I say?

00:03:55.039 --> 00:03:56.879
<v Wojciech Wegrzynski>Yeah, well, thank you, thank you.

00:03:56.879 --> 00:03:58.159
<v Wojciech Wegrzynski>And uh and vice versa.

00:03:58.159 --> 00:04:10.240
<v Wojciech Wegrzynski>I really enjoy having conversations uh with you, especially when you give me an interesting topic and then you put me into rabbit hole reading about chickens hitting jet fans for an hour.

00:04:10.240 --> 00:04:16.560
<v Wojciech Wegrzynski>And here we are uh now discussing this further and what it means for the Broader Fire community.

00:04:16.560 --> 00:04:21.920
<v Wojciech Wegrzynski>What an interesting pathline of my life that put me into this uh position right now.

00:04:21.920 --> 00:04:27.439
<v Wojciech Wegrzynski>But anyway, yeah, you brought out like a very interesting topic: test versus experiments.

00:04:27.439 --> 00:04:32.319
<v Wojciech Wegrzynski>And whenever I think this, I hear uh a very loud voice of Guillermo Rain in my head.

00:04:32.319 --> 00:04:34.639
<v Wojciech Wegrzynski>Uh Jake, this is a test, not an experiment.

00:04:34.639 --> 00:04:38.160
<v Wojciech Wegrzynski>And uh this is something very strongly embedded in my consciousness.

00:04:38.160 --> 00:04:41.519
<v Wojciech Wegrzynski>So let's let's probably start with the very hard question.

00:04:41.519 --> 00:04:45.600
<v Wojciech Wegrzynski>What is the difference between a test and experiment?

00:04:46.160 --> 00:04:48.879
<v David Morrisset>Okay, so this is this is really the crux of what I mean.

00:04:48.879 --> 00:04:51.439
<v David Morrisset>We kind of got to this point in the last time that was on the show, right?

00:04:51.439 --> 00:04:53.759
<v David Morrisset>And sort of discussing how do we make measurements, right?

00:04:53.759 --> 00:04:59.120
<v David Morrisset>Because we we make measurements in the context of let's say live fire tests or experiments.

00:04:59.120 --> 00:05:08.240
<v David Morrisset>And the really where we draw the line of what becomes a test versus an experiment comes down to basically what is the, I guess what's the the goal and the philosophy behind it, right?

00:05:08.240 --> 00:05:21.120
<v David Morrisset>Because we mentioned this last time too, but something, a place to start with this is I think most would accept this being true, but I think we've accepted that fire is intrinsically a very complicated process, right?

00:05:21.120 --> 00:05:30.319
<v David Morrisset>It combines all these various gas phase and solid phase phenomena, and these include heat transfer, you know, mass transport, chemical reactions, all sorts of complexity.

00:05:30.319 --> 00:05:39.279
<v David Morrisset>And to resolve all of this from first principles, to say I want to be able to assess blank systems uh and and put them on buildings, right?

00:05:39.279 --> 00:05:45.199
<v David Morrisset>There's a degree of complexity there where a lot of what we do is we say, let's just light it up and see how it performs.

00:05:45.199 --> 00:05:45.680
<v David Morrisset>Right.

00:05:45.680 --> 00:05:52.000
<v David Morrisset>And so uh there's a this actually there's a great quote by Howard Emmons that sort of rings my head whenever I start thinking about this, right?

00:05:52.000 --> 00:05:54.319
<v David Morrisset>That in terms of this intrinsic complexity, right?

00:05:54.319 --> 00:05:56.959
<v David Morrisset>And there's a there's a paper, I think it was called The Growth of Fire Science.

00:05:56.959 --> 00:06:00.639
<v David Morrisset>And he he said something to the effect of, now what is fire science?

00:06:00.639 --> 00:06:01.519
<v David Morrisset>You know, quotes.

00:06:01.519 --> 00:06:06.720
<v David Morrisset>It is it's certainly not something as simple as basic chemistry or physics or something to that effect, right?

00:06:06.720 --> 00:06:15.920
<v David Morrisset>And then then he goes on to talk about the interactions between all these physical and chemical phenomena, right, and and how how complex that is.

00:06:15.920 --> 00:06:17.199
<v David Morrisset>And it's really interesting, right?

00:06:17.199 --> 00:06:26.800
<v David Morrisset>And and so I think one one way that that manifests itself is to really fully resolve these complex processes, is we utilize experimentation and testing.

00:06:27.040 --> 00:06:35.120
<v Wojciech Wegrzynski>And you used a very important uh caveat in in when you were introducing this, because you said, I want to do a test or experiment to put something on my building.

00:06:35.120 --> 00:06:44.240
<v Wojciech Wegrzynski>So, besides like scientific curiosity and our need to understand the underlying physics, which I guess is a driver for many of us fire researchers.

00:06:44.240 --> 00:06:55.120
<v Wojciech Wegrzynski>We simply love it, and it's so complex, it's so intellectually engaging to study fire phenomena because of how endlessly complex they are.

00:06:55.120 --> 00:07:02.480
<v Wojciech Wegrzynski>Like, literally, there are layers of complexities, you can just add them and add them and add them beyond human comprehension.

00:07:02.480 --> 00:07:14.639
<v Wojciech Wegrzynski>This is what a lot of us find attractive in fire science, but it's also a practical field where in the end you need to place things on your buildings, you need to have people in those buildings.

00:07:14.639 --> 00:07:21.360
<v Wojciech Wegrzynski>Those buildings will encounter fires, and you want those people to not suffer from those fires.

00:07:21.360 --> 00:07:34.879
<v Wojciech Wegrzynski>So there's a whole space of responsibility and you know, this whole machine of making sure that we are sure of what we're doing.

00:07:34.879 --> 00:07:50.399
<v Wojciech Wegrzynski>And in here, it it kind of narrows it down because you stop experimenting in a blank space of curiosity, but you have a functional goal in the end to have uh something that I would call a safety framework or fire safety framework.

00:07:50.560 --> 00:07:52.720
<v David Morrisset>And I think, wow, that's it's it's an interesting point, really.

00:07:52.720 --> 00:08:04.079
<v David Morrisset>Now that we're going to the weeds here, of really there's like a spectrum of whenever we're doing some sort of, let's say, testing or you know, experiments in a lab, there's a spectrum between practical outputs and fundamental insight.

00:08:04.079 --> 00:08:08.560
<v David Morrisset>Now, if you do them right, you can do both pretty effectively, right?

00:08:08.560 --> 00:08:17.120
<v David Morrisset>Um, because the more that you get fundamental insight, sometimes you need to go to the scale of understanding, you know, the structure of a diffusion flame.

00:08:17.120 --> 00:08:24.000
<v David Morrisset>Sometimes there's a lot of practical insight of understanding things like the actual chemical reactions that produce emissions in a fire, so on.

00:08:24.000 --> 00:08:37.279
<v David Morrisset>Some of that fundamental insight is done on a scale, or the experiments to gain that insight are done on a scale that isn't immediately apparent of what the practical outputs are, but the insight you get from that informs engineering judgment, right?

00:08:37.279 --> 00:08:44.320
<v David Morrisset>And so something that I feel a lot of a lot a lot of criticism can come around experiments not being practical enough, right?

00:08:44.320 --> 00:08:48.559
<v David Morrisset>People doing experiments on tiny little pieces of, you know, PMMA or whatever, right?

00:08:48.559 --> 00:08:49.519
<v David Morrisset>Uh what have it?

00:08:49.519 --> 00:08:53.440
<v David Morrisset>But the insights that come from that are extremely impactful, right?

00:08:53.440 --> 00:08:58.399
<v David Morrisset>Because if you can gain generalizable insight, right, that informs engineering judgment.

00:08:58.399 --> 00:09:00.559
<v David Morrisset>The engineering judgment gives you practical outputs.

00:09:00.559 --> 00:09:00.799
<v David Morrisset>Right.

00:09:00.799 --> 00:09:06.000
<v David Morrisset>That's kind of like the way that I've always sort of been taught is the progression of the development of knowledge.

00:09:06.000 --> 00:09:08.240
<v David Morrisset>Anyway, all this dance around the initial question, right?

00:09:08.240 --> 00:09:11.759
<v David Morrisset>We haven't even which I haven't even addressed yet, is what is the difference between a test and an experiment?

00:09:11.759 --> 00:09:22.720
<v David Morrisset>And I and I think it's up there the there's multiple ways to cut that definition, but I think for the sake of just today's discussion, I'll define it, then I'll let you define it, and then we'll we'll probably, you know, let's meet somewhere in between.

00:09:22.720 --> 00:09:27.919
<v David Morrisset>I would say a test particular it generally follows a sort of standard procedure.

00:09:27.919 --> 00:09:34.240
<v David Morrisset>We're looking at something that is effectively a standardized procedure by which you can achieve some sort of output, right?

00:09:34.240 --> 00:09:51.279
<v David Morrisset>Uh generally, the idea would be to look at some fire performance under a live fire condition, whether that's in a sort of simulated thermal environment or exposure to a real fire source, um, which allows the output of those tests to be benchmarked against other materials or systems, right?

00:09:51.279 --> 00:09:54.480
<v David Morrisset>And we can probably rattle off dozens of examples of that.

00:09:54.480 --> 00:09:56.480
<v David Morrisset>That's probably a good place to go after this.

00:09:56.480 --> 00:10:04.559
<v David Morrisset>But on the other hand, an experiment would be something that doesn't follow a sort of standardized, agreed upon consensus-based procedure, right?

00:10:04.559 --> 00:10:11.519
<v David Morrisset>Um, instead, you investigate some sort of phenomenon purely for the uh quantification of some element of that, right?

00:10:11.519 --> 00:10:29.679
<v David Morrisset>So you go in with a question, you say, I want to develop an ex an experiment to articulate this relevant bit of physics, whether that's just to explore it for the sake of knowledge, whether that's to validate a theory, or whether that's to compare to something like numerical simulation, all of which are very valid reasons to do an experiment.

00:10:29.679 --> 00:10:39.039
<v David Morrisset>But in terms of the process, it's it's less of a structured, I guess prescribed is the right word, there's prescribed process by which you conduct said experiment.

00:10:39.360 --> 00:11:06.000
<v Wojciech Wegrzynski>When you dropped uh this question at me, my brain immediately went back to an episode which I had not that long ago with Mike Spearpoint and Constantinus from afar about uh balconi fires, where they've done a set of experiments on balcony fires, and uh the outcome was to some extent expected, you know, more combustible material on a balcony equals to worse fire.

00:11:06.000 --> 00:11:10.080
<v Wojciech Wegrzynski>It's uh it's a conclusion that is safe to be made even without one experiment.

00:11:10.080 --> 00:11:28.240
<v Wojciech Wegrzynski>But in the end, they've done those beautiful experiments, and the outcome of those experiments was quantification of the hazard, it was a tangible proof of what the hazard is, it was a ranking of hazards present in that setting.

00:11:28.240 --> 00:11:44.399
<v Wojciech Wegrzynski>Like what they've created, I mean, they've created some new knowledge and they've seen some interesting things during those experiments, but they have created a measurable proof that something happens in one way or another.

00:11:44.399 --> 00:11:58.159
<v Wojciech Wegrzynski>And I think as our profession matures the less will be an outcome of community wisdom and opinions of people, the more it should be based on the proof.

00:11:58.159 --> 00:12:07.440
<v Wojciech Wegrzynski>And here we reach a point where you can gain proof through testing, you can gain proof of experiments, though they will be different proofs.

00:12:07.440 --> 00:12:14.559
<v Wojciech Wegrzynski>For me, if we talk testing, two words come into my mind repatibility, reproducibility.

00:12:14.559 --> 00:12:23.279
<v Wojciech Wegrzynski>Those are two things that characterize tests for me, because if test is not reproducible or repeatable, it's not a good test.

00:12:23.279 --> 00:12:34.879
<v Wojciech Wegrzynski>And in the end, the outcome is some sort of ranking or consistency within the testing framework, or placing something within a predefined framework.

00:12:34.879 --> 00:12:55.840
<v Wojciech Wegrzynski>While experiment, while it can also provide you a proof and a guidance and knowledge necessary to move your product project forward, it gives you an exploration and perhaps answers questions you have not asked and allows you to give insight, but inherently makes comparison between different experiments extremely challenging.

00:12:56.080 --> 00:13:01.759
<v David Morrisset>And I guess this might be a good stage to rattle off a few examples of standard tests that come to mind, right?

00:13:01.759 --> 00:13:04.879
<v David Morrisset>Yeah, let's get for the for the listeners just sort of uh context.

00:13:04.879 --> 00:13:07.679
<v David Morrisset>Let's drop some things on the playground to play with.

00:13:07.679 --> 00:13:08.159
<v David Morrisset>Definitely.

00:13:08.159 --> 00:13:14.720
<v David Morrisset>So I mean, I don't know if this is a a controversial sort of way to frame the problem, but let's lump them into two sort of larger categories.

00:13:14.720 --> 00:13:20.240
<v David Morrisset>I mean, again, this is people from testing agencies are probably going to be rolling their eyes at this oversimplification, of course.

00:13:20.240 --> 00:13:21.600
<v David Morrisset>Like I fully appreciate that.

00:13:21.600 --> 00:13:27.279
<v David Morrisset>But let's let's lump things into some applications of testing include furnace testing, of course, right?

00:13:27.279 --> 00:13:29.519
<v David Morrisset>So looking at standard furnace testing, right?

00:13:29.519 --> 00:13:31.039
<v David Morrisset>Kv4, let's go.

00:13:31.039 --> 00:13:33.039
<v David Morrisset>And then there's another large category.

00:13:33.039 --> 00:13:36.799
<v David Morrisset>We could look at sort of the general reaction to fire classification testing.

00:13:36.799 --> 00:13:37.120
<v David Morrisset>Right.

00:13:37.120 --> 00:13:50.399
<v David Morrisset>So whether that's anything from you know the SBI, the single burning item, uh, whether that's, you know, you can look at there is there are standards out there for things like cone calorimetry, um, whether those are applied in many regulatory environments, there's exceptions to that.

00:13:50.399 --> 00:13:52.159
<v David Morrisset>There are places where that that can be used.

00:13:52.159 --> 00:14:00.639
<v David Morrisset>Um, there's other tests looking at things like the LIFT test, the lateral ignition and flames transport or spread test, however you want to define the acronym.

00:14:00.639 --> 00:14:02.639
<v David Morrisset>You know, again, it's a standard test procedure.

00:14:02.639 --> 00:14:07.360
<v David Morrisset>Uh, and then the outcomes from that are, you know, are standardized in terms of what you should be looking for.

00:14:07.360 --> 00:14:10.080
<v David Morrisset>And we can go on and list off dozens of them, right?

00:14:10.080 --> 00:14:23.679
<v David Morrisset>But the two sort of major categories I would think of is one that has sort of a temperature boundary condition and one that looks at either a sort of the world of exposing things to heat fluxes or exposing things to sexually direct contact with something like a burner.

00:14:23.679 --> 00:14:34.159
<v David Morrisset>Um and also from uh from a cladding perspective, we can even incorporate things like the large-scale cladding tests, looking at you know, BS8414 type testing in all D European equivalents.

00:14:34.159 --> 00:14:42.799
<v David Morrisset>Um these are when we're talking about standard test methods, we're talking about these kinds of tests where we can take a material, a product, or a system and use them in this test.

00:14:42.799 --> 00:14:52.240
<v David Morrisset>And the outcomes from those tests should allow us to more or less index them against one another within the the, how do I say this, within the context of that test, though, right?

00:14:52.240 --> 00:14:55.440
<v David Morrisset>The outcomes are limited to basically the scenario that we're looking at.

00:14:55.440 --> 00:14:59.759
<v David Morrisset>Because when we're doing something like a standard test, we're accepting a certain scenario, right?

00:15:00.159 --> 00:15:10.159
<v Wojciech Wegrzynski>And in here, I I put forward the question: should that scenario be closely representative to real-world fire?

00:15:10.159 --> 00:15:22.960
<v Wojciech Wegrzynski>Because, you know, especially for perhaps I'm I'm I'm oversimplifying it, but I but I find fire resistance is especially difficult to understand for by people who are not working in in fire testing.

00:15:22.960 --> 00:15:32.000
<v Wojciech Wegrzynski>For me, you know, fire resistance, a class of minutes, REI 60, this is a very precise thing.

00:15:32.000 --> 00:15:55.279
<v Wojciech Wegrzynski>This is an extremely precise evaluation of the performance of a given assembly in a very specific testing conditions, in a very specific device, measured in an extremely highly specified way, in a very robust system, in a very reputable and uh reproducible manner by an accredited laboratory that has competencies to do so.

00:15:55.279 --> 00:15:56.960
<v Wojciech Wegrzynski>And it's only that.

00:15:56.960 --> 00:16:01.600
<v Wojciech Wegrzynski>And for layman, it's usually, oh yeah, this can resist 60 minutes of fire.

00:16:01.600 --> 00:16:08.080
<v Wojciech Wegrzynski>Well, no, it it nowhere says that in resistance uh to fire of 60 minutes that it's resist fire for 60 minutes.

00:16:08.080 --> 00:16:10.159
<v Wojciech Wegrzynski>But but it it's it's kind of the thing, right?

00:16:10.399 --> 00:16:17.120
<v David Morrisset>I think this is this is a really important one to discuss, at least in one aspect of furnace testing, right, is the idea of the of the minutes, right?

00:16:17.120 --> 00:16:20.399
<v David Morrisset>Of the output of this, because every test will have an output, right?

00:16:20.399 --> 00:16:23.039
<v David Morrisset>And the way that you frame that output is really important, right?

00:16:23.039 --> 00:16:32.320
<v David Morrisset>Where you draw the line between something passing or failing, where the where you draw the line of the difference between an A-rated material and a B-rated material, these all have huge significance, right?

00:16:32.320 --> 00:16:36.240
<v David Morrisset>And and the output of a furnace test is basically the exposure time, right?

00:16:36.240 --> 00:16:45.200
<v David Morrisset>And if you go back to the original work, you know, that you can read Angus Law and Luke Bisbee's paper from the University of Edinburgh looking at the rise and rise of fire resistance, it's an excellent paper.

00:16:45.200 --> 00:16:50.320
<v David Morrisset>But looking at sort of some a bit of a historical perspective of the development of furnace testing, right?

00:16:50.320 --> 00:17:02.159
<v David Morrisset>And you see the need that arised at the time, you know, over well over 100 years ago, to develop basically a testing regime for structural members and structural elements, building elements.

00:17:02.159 --> 00:17:05.359
<v David Morrisset>And actually our boundary conditions from that haven't really changed.

00:17:05.359 --> 00:17:06.480
<v David Morrisset>The curve is the same.

00:17:07.039 --> 00:17:09.119
<v Wojciech Wegrzynski>This goes back to what I've asked before.

00:17:09.119 --> 00:17:11.839
<v Wojciech Wegrzynski>Shall it be representative of a real world case?

00:17:11.839 --> 00:17:15.359
<v Wojciech Wegrzynski>And here you you're touching on a very, very important thing.

00:17:15.359 --> 00:17:22.480
<v Wojciech Wegrzynski>Because today we are perfectly aware that it is not representative of a real world fire, the exposure, the boundary condition.

00:17:22.480 --> 00:17:24.640
<v Wojciech Wegrzynski>It is representative of some fires.

00:17:24.640 --> 00:17:30.480
<v Wojciech Wegrzynski>Yes, fires exist that grow and and decay like well, the standard firecraft doesn't decay.

00:17:30.480 --> 00:17:33.119
<v Wojciech Wegrzynski>That's another podcast episode coming to you soon.

00:17:33.119 --> 00:17:36.160
<v Wojciech Wegrzynski>But they could grow like a standard fire.

00:17:36.160 --> 00:17:41.759
<v Wojciech Wegrzynski>There are class of fires that grow like that, but it's one of a million, like that, there's a lot of different fires out there.

00:17:41.759 --> 00:17:48.559
<v Wojciech Wegrzynski>So so it's very hard to state that this test uses a real-world fire exposure as a boundary condition.

00:17:48.559 --> 00:17:49.279
<v Wojciech Wegrzynski>It does not.

00:17:49.279 --> 00:17:55.920
<v Wojciech Wegrzynski>But does it mean it's worthless because we have a hundred years of using that?

00:17:55.920 --> 00:18:04.880
<v Wojciech Wegrzynski>Because we've tested countless amounts of materials with that, because uh you could not put an argument that it did not create safety.

00:18:04.880 --> 00:18:13.839
<v Wojciech Wegrzynski>It has created, it has resulted in safety, in safe applications and globalization of fire safety engineering in a way.

00:18:13.839 --> 00:18:16.960
<v Wojciech Wegrzynski>So is it the best standard that it doesn't replicate real life?

00:18:16.960 --> 00:18:17.599
<v Wojciech Wegrzynski>I don't know.

00:18:17.599 --> 00:18:21.440
<v Wojciech Wegrzynski>If we consider it an experiment, it's a horrendous experiment.

00:18:21.440 --> 00:18:26.160
<v Wojciech Wegrzynski>If we consider it a test, it perhaps is not as bad as I I would usually claim.

00:18:26.400 --> 00:18:29.839
<v David Morrisset>But even within within that context, I think it's important.

00:18:29.839 --> 00:18:37.039
<v David Morrisset>The key thing I'm getting at with the outputs of a fire resistance test is the structure of the of the output being in minutes, right?

00:18:37.039 --> 00:18:42.480
<v David Morrisset>Because if you go back to the original definition of it, they you know they created this fire condition within a furnace.

00:18:42.480 --> 00:18:45.759
<v David Morrisset>They tested an assembly for a certain amount of time exposed to that.

00:18:45.759 --> 00:18:51.920
<v David Morrisset>The the hard part is that is translated to the everyday vernacular and staying in minutes, right?

00:18:51.920 --> 00:19:00.799
<v David Morrisset>So, like you said, this is a very specific, this is a very specific condition that is maybe representative of a certain class of fires.

00:19:00.799 --> 00:19:04.400
<v David Morrisset>Not every fire, sure, but you know, you you need to benchmark against something.

00:19:04.400 --> 00:19:05.279
<v David Morrisset>I appreciate that.

00:19:05.279 --> 00:19:08.079
<v David Morrisset>But the scary part is is the output being in minutes.

00:19:08.079 --> 00:19:13.920
<v David Morrisset>To this day, there are many engineers who to prescribe that as time and then a quote unquote real fire.

00:19:13.920 --> 00:19:20.000
<v David Morrisset>Yes, so people are using this as a benchmark to compare against things like egress times and things that are based on real time.

00:19:20.000 --> 00:19:30.720
<v David Morrisset>So the time it takes to to get people out of a building, or the time it takes to for the fire service to arrive, or the time it takes to for a fire to spread from a compartment to compartment, those are real time.

00:19:30.720 --> 00:19:33.519
<v David Morrisset>Those are based on uh that is in real time.

00:19:33.519 --> 00:19:41.759
<v David Morrisset>And then those cannot be benchmarked against the outputs of something like a standard fire test, because the time in that furnace is a specific condition.

00:19:41.759 --> 00:19:45.039
<v David Morrisset>The the ranking should almost just be called points or something.

00:19:45.039 --> 00:19:45.359
<v David Morrisset>Right?

00:19:45.359 --> 00:19:55.200
<v David Morrisset>You know, like that's uh that's something I remember Luke Bisbee used to always say was if we just called it points, that would that would reduce a lot of the the miscommunication to various parties who use these.

00:19:55.200 --> 00:20:01.200
<v David Morrisset>Because of course you and I are buried in these this world of of testing and experimentation every day.

00:20:01.200 --> 00:20:09.839
<v David Morrisset>But it's it's actually kind of not immediately apparent as you're as you're working in the whether you're working in the engineering space or or you know learning about these tests from from the onset.

00:20:09.839 --> 00:20:15.440
<v David Morrisset>If it says it in minutes, it's a very easy thing to misinterpret in terms of you know what does that actually mean.

00:20:15.920 --> 00:20:18.799
<v Wojciech Wegrzynski>Which brings us to like practicality of those.

00:20:18.799 --> 00:20:26.559
<v Wojciech Wegrzynski>Like one practical aspect having some sort of ranking of materials, I see benefits of that.

00:20:26.559 --> 00:20:28.480
<v Wojciech Wegrzynski>It could be considered useful.

00:20:28.480 --> 00:20:31.279
<v Wojciech Wegrzynski>You need sometimes ranking of those to do that.

00:20:31.279 --> 00:20:44.480
<v Wojciech Wegrzynski>But if a fire engineer actually is tasked with uh designing a load-bearing structure that can survive a fire, uh fire resistance has proven to be not the worst proxy in the world uh of that.

00:20:44.480 --> 00:20:53.119
<v Wojciech Wegrzynski>But of course, for many aspects of structural design, it it's challenging and inaccurate and not really you can do better simply.

00:20:53.119 --> 00:21:05.680
<v Wojciech Wegrzynski>However, if a fire engineer is burdened with providing a proof or or uh you know uh yeah, basically a proof that that the load brain will be maintained, they may be looking into some experimental data.

00:21:05.680 --> 00:21:15.039
<v Wojciech Wegrzynski>And then when you start designing an experiment that answers your question in real-world minutes, like what's the real world minute time of collapse of my structure?

00:21:15.039 --> 00:21:35.359
<v Wojciech Wegrzynski>We this is a very relevant question today when we are building those giant uh warehouses with stacking units, uh, with with the steel structure inside that can span over multiple levels, and you put 2,000 people into that warehouse and you have to evacuate them in five minutes, and there's a good chance it's gonna collapse after 10.

00:21:35.359 --> 00:21:38.799
<v Wojciech Wegrzynski>So this is a very practical question when it will collapse.

00:21:38.799 --> 00:21:41.200
<v Wojciech Wegrzynski>How we design an experiment to answer you that question.

00:21:41.440 --> 00:21:42.240
<v David Morrisset>That's an interesting one.

00:21:42.240 --> 00:21:50.240
<v David Morrisset>Before we move on to discussing the transition to experiments, though, in the context of uh the standard furnace too, I think there's one more element to discuss.

00:21:50.240 --> 00:21:53.119
<v David Morrisset>And this is something that I know your lab is looking to too, right?

00:21:53.119 --> 00:21:57.279
<v David Morrisset>But and and I agree that there is a there's a utility to benchmarking.

00:21:57.279 --> 00:22:14.799
<v David Morrisset>I mean, I mean that's like I I am an absolutely huge believer in the ability to using you know standard tests to benchmark performance, uh so long as it's understood within the context of those tests, uh and those tests are done in a way that the outputs are actually applicable into to the context again of the test.

00:22:14.799 --> 00:22:22.640
<v David Morrisset>But the thing about furnace testing that I think we've discovered recently too is if you put, you know, say a mass timber element uh in the furnace, right?

00:22:22.640 --> 00:22:24.960
<v David Morrisset>What is I mean, you you can speak from your experience too.

00:22:24.960 --> 00:22:31.599
<v David Morrisset>I know there's been many studies that have looked at this, but we know that conditions within that furnace are following a temperature time curve.

00:22:31.599 --> 00:22:37.759
<v David Morrisset>And we know that the combustible elements of the wall assembly are contributing to that, those conditions within the furnace.

00:22:37.759 --> 00:22:42.559
<v David Morrisset>So you just inherently end up with a different amount of fuel being injected into that furnace.

00:22:42.559 --> 00:22:46.400
<v David Morrisset>So you're the the the boundary conditions from a temperature perspective aren't changing.

00:22:46.400 --> 00:22:51.440
<v David Morrisset>But from the perspective of the actual fuel being injected into that furnace, things are changing.

00:22:51.440 --> 00:23:00.400
<v David Morrisset>In the way that, you know, if you have a timber compartment on fire, if there's a couch on fire next to it, the couch doesn't know to regulate itself to burn less to match temperature time curve, right?

00:23:00.400 --> 00:23:02.640
<v David Morrisset>The couch is just gonna do what the couch is gonna do.

00:23:02.640 --> 00:23:14.400
<v David Morrisset>And so we we're introducing an interesting complexity here that just highlights the some of the this wasn't, you know, the intent of the original fire resistance framework was to look at structures that were inherently non-combustible.

00:23:14.400 --> 00:23:23.920
<v David Morrisset>That was one of the original intents, and and to and that the fire resistance framework was designed so that the compartment couldn't withstand the burnout of the contents of that compartment.

00:23:23.920 --> 00:23:25.680
<v David Morrisset>Those were like inherent assumptions.

00:23:25.680 --> 00:23:30.400
<v David Morrisset>And I just think it's interesting that we're starting to sort of challenge those assumptions in the context of what we're doing now.

00:23:30.720 --> 00:23:39.359
<v Wojciech Wegrzynski>Uh yeah, so so I've already multiple times I I've uh said that's a really horrible way to assess the properties of timber through fire resistance testing.

00:23:39.359 --> 00:23:41.279
<v Wojciech Wegrzynski>And uh I I have papers on that.

00:23:41.279 --> 00:23:44.799
<v Wojciech Wegrzynski>I truly uh despise this way of thinking about it.

00:23:44.799 --> 00:23:58.240
<v Wojciech Wegrzynski>But if you for a second forget about the utility of it as in those elements reaching the building, if you consider it just within the testing framework, it kind of puts the timber in a rank with other elements.

00:23:58.240 --> 00:24:16.319
<v Wojciech Wegrzynski>Like you you apply the same repetitive thing, you apply the same uh way of testing, you provide pretty much the same uh consistent metrics of performance in terms of minutes, in terms of load bearing, in terms of integrity, in terms of insulation.

00:24:16.319 --> 00:24:23.519
<v Wojciech Wegrzynski>So you kind of perform the same thing to this to the slightly different element, but in a very same way.

00:24:23.519 --> 00:24:24.960
<v Wojciech Wegrzynski>You just perform a test.

00:24:24.960 --> 00:24:29.920
<v Wojciech Wegrzynski>From this perspective, the test has not failed yet because it has been done in the same way.

00:24:29.920 --> 00:24:33.839
<v Wojciech Wegrzynski>And in a way, it allows you to have verified the timber.

00:24:33.839 --> 00:24:48.559
<v Wojciech Wegrzynski>The problems with it is one, the timber kind of games on the boundary condition of the test, so it actually does interfere with the boundary condition in a way that no other tested element does.

00:24:48.559 --> 00:24:55.759
<v Wojciech Wegrzynski>It's like you know, you play blackjack, but you're allowed to look at other people's hands or or the or the casino hand.

00:24:55.759 --> 00:24:57.759
<v Wojciech Wegrzynski>So you're gaming the system in a way.

00:24:57.759 --> 00:24:58.559
<v Wojciech Wegrzynski>That's one problem.

00:24:58.559 --> 00:25:13.519
<v Wojciech Wegrzynski>And the second, the end world utility of that test is completely different than the end utility of different materials where you could perhaps use it as a proxy of your structural fire safety.

00:25:13.519 --> 00:25:16.880
<v Wojciech Wegrzynski>In here, I would say, um no, not truly.

00:25:16.880 --> 00:25:25.680
<v Wojciech Wegrzynski>There are too many other things to be considered, which again have been so like if you just consider this as a testing framework, it it kind of works.

00:25:25.680 --> 00:25:32.640
<v Wojciech Wegrzynski>It's just, you know, the goal of the test suddenly is misaligned and though the boundary condition is broken.

00:25:32.960 --> 00:25:48.160
<v David Morrisset>I mean, like you said earlier, uh, we have a track record that shows, for particularly if we're looking at, you know, let's say the tried and true non-combustible materials within the fire resistance framework, we have an idea of whether we're just getting lucky or whether we've actually truly provided implicit safety, right?

00:25:48.160 --> 00:25:50.960
<v David Morrisset>There is an element to which there is at least a track record there.

00:25:50.960 --> 00:26:00.400
<v David Morrisset>Uh and and I think you mentioned a really interesting point that with the introduction of timber, it's fair to say that we're still matching the conditions of the test, right?

00:26:00.400 --> 00:26:03.200
<v David Morrisset>We are still achieving the standard temperature time curve.

00:26:03.200 --> 00:26:10.400
<v David Morrisset>We're s we're achieving some the standard conditions, but like you also said, we know that the conditions in the furnace are just different, right?

00:26:10.400 --> 00:26:15.599
<v David Morrisset>So it it raises an interesting question of it's not that there's anything intrinsically wrong with the tests.

00:26:15.599 --> 00:26:17.519
<v David Morrisset>Like you said, the test is what the test is.

00:26:17.519 --> 00:26:23.920
<v David Morrisset>It comes down to our interpretation of is this an adequate condition for what we are then using those outputs for.

00:26:23.920 --> 00:26:24.319
<v David Morrisset>Right.

00:26:24.319 --> 00:26:27.680
<v David Morrisset>And I think that's a really uh that's a it's an interesting question.

00:26:27.680 --> 00:26:36.000
<v David Morrisset>And I'm not and I'm not convinced that, like you said, that in all structural applications we have enough information from fire resistance alone for something like timber elements.

00:26:36.559 --> 00:26:45.920
<v Wojciech Wegrzynski>I I just find this single information to be perhaps interesting, but not a complete proxy uh in which you can encapsulate the problem.

00:26:45.920 --> 00:26:58.720
<v Wojciech Wegrzynski>Again, also in this podcast, I had multiple guests with with whom we've agreed that you need to test the products in their end-use condition, like they are supposed to be used in the building.

00:26:58.720 --> 00:27:10.240
<v Wojciech Wegrzynski>And with an assumption that your boundary condition, the thermal boundary condition, the exposure condition, whatever it is, that it kind of represents the hazard.

00:27:10.240 --> 00:27:17.599
<v Wojciech Wegrzynski>So like SBI and uh and external walls, I mean, it's a test method that's very well established.

00:27:17.599 --> 00:27:25.839
<v Wojciech Wegrzynski>It's just uh characteristic of the test itself, are nowhere close to what external walls are exposed to when uh they face venting fires.

00:27:25.839 --> 00:27:32.319
<v Wojciech Wegrzynski>So uh a test is a test, it works, it's just doesn't really answer the questions that you're asking.

00:27:32.319 --> 00:27:47.920
<v Wojciech Wegrzynski>No, the question is like as we build up our knowledge and we try to modify those tests, we also kind of lose you know the backtrace the history of outcomes that those tests created.

00:27:47.920 --> 00:28:12.799
<v Wojciech Wegrzynski>So by changing and narrowing and fine-tuning those tests, moving closer to individual experiments, not repetitive, you know, framework of testing, you start to be closer to that reality, whatever reality is and whatever you define reality is, but you lose more and more of that comparativeness, uh, this compatibility of those tests altogether, you know.

00:28:12.799 --> 00:28:19.519
<v Wojciech Wegrzynski>And I think it's it's a very challenging thing to balance out, like when it's worth to lose that.

00:28:19.519 --> 00:28:27.440
<v Wojciech Wegrzynski>And uh and I'll I'll give you an answer me that and I'll give you an example where I found it extremely interesting uh case.

00:28:27.759 --> 00:28:33.200
<v David Morrisset>Yeah, I think understanding the the history of of a testing regime, I think, is is really important, right?

00:28:33.200 --> 00:28:55.279
<v David Morrisset>You you mentioned the SBI being applied to cladding systems, which we know that you know, if you if for anyone who's been followed followed the Grenfell Tower inquiry, right, that led to you know a myriad of issues that led to combustible clouding being effectively gaming that those testing regimes to show that certain products were able to pass those tests, but you know, in a real world application, those those performed terribly, right?

00:28:55.279 --> 00:29:04.960
<v David Morrisset>And this comes down to again, this there's a lot more context actually provided in um there was a paper written by Angus Law and colleagues at the University of Edinburgh, who was my supervisor for my PhD, right?

00:29:04.960 --> 00:29:07.839
<v David Morrisset>So that's why I'm so familiar with all of the bits that he's done.

00:29:07.839 --> 00:29:12.160
<v David Morrisset>But uh that was about the sort of the black box of the Euroclass system.

00:29:12.160 --> 00:29:14.559
<v David Morrisset>And that was it's an interesting paper, it's a fascinating paper.

00:29:14.559 --> 00:29:20.559
<v David Morrisset>But one element to sort of to think about with the with the SBI in particular is really if you go far enough back, right?

00:29:20.559 --> 00:29:26.559
<v David Morrisset>It's it's basically it's a it's a corner test within a you know, a it's a single burning item within a a wall corner.

00:29:26.559 --> 00:29:31.759
<v David Morrisset>Uh it's akin in a lot of similarities to things like an almost like an ISO room corner test, right?

00:29:31.759 --> 00:29:34.160
<v David Morrisset>I mean again, obviously they're they're separate tests.

00:29:34.160 --> 00:29:44.559
<v David Morrisset>It wasn't a direct translation necessarily, but um, and if you look at the context of those tests, that is very different than the scenarios in which you see in an exterior wall system, right?

00:29:44.559 --> 00:29:51.680
<v David Morrisset>Those things are designed to look at a single burning item within a corner, to look at basically interior finishes on a wall within a within a compartment.

00:29:51.680 --> 00:29:53.759
<v David Morrisset>It seems to be the more applicable space there, right?

00:29:53.759 --> 00:29:59.839
<v David Morrisset>Um I wasn't there as some there are people who've been on your podcast who were there for these discussions in the year in the development of the year.

00:29:59.839 --> 00:30:02.000
<v David Morrisset>Class system, but I wasn't there for those.

00:30:02.000 --> 00:30:06.559
<v David Morrisset>But I'm imagining that was sort of the you know the original intent for some of these sort of tests.

00:30:06.559 --> 00:30:14.799
<v David Morrisset>And so to apply those to clouding systems, uh you've already broken the sort of the context that was established for these, right?

00:30:14.799 --> 00:30:21.119
<v David Morrisset>And so so it's just that is an intrinsic assumption in using the test is that we're using this in a way that you know they should be applied.

00:30:21.839 --> 00:30:23.599
<v Wojciech Wegrzynski>I would say again, it's nothing wrong with the test.

00:30:23.599 --> 00:30:27.279
<v Wojciech Wegrzynski>Like if you just consider tests being a test, it's it's just a test.

00:30:27.279 --> 00:30:29.359
<v Wojciech Wegrzynski>It's the way uh what do you use it?

00:30:29.359 --> 00:30:39.440
<v Wojciech Wegrzynski>I I promised you an example from my end, how this chase of reality was abandoned, and that was in uh computer simulations of car park fires.

00:30:39.440 --> 00:30:45.200
<v Wojciech Wegrzynski>So I have a long history of uh fighting with everyone else about uh car park fires and design curves.

00:30:45.200 --> 00:30:58.640
<v Wojciech Wegrzynski>And uh actually, what what we are doing in the end is that we apply still always the same curve that comes from one single experiment done by TNO in the Netherlands in 1999, if I believe.

00:30:58.640 --> 00:31:18.240
<v Wojciech Wegrzynski>And we still use this curve, and we're facing a load of backlash with the community around us because they're this curve is so wrong, it doesn't represent a car, it doesn't represent an electric vehicle, there's like a ton of different studies done in different ways, and now I've analyzed all of them.

00:31:18.240 --> 00:31:19.839
<v Wojciech Wegrzynski>Like I had a podcast episode about it.

00:31:19.839 --> 00:31:26.480
<v Wojciech Wegrzynski>We've analyzed all of them, like literally every single available data source that we could find in the existence.

00:31:26.480 --> 00:31:30.000
<v Wojciech Wegrzynski>There's no unbiased car park design fire curve.

00:31:30.000 --> 00:31:31.920
<v Wojciech Wegrzynski>Simply does not exist.

00:31:31.920 --> 00:31:34.400
<v Wojciech Wegrzynski>Every of them comes with caveats.

00:31:34.400 --> 00:31:55.519
<v Wojciech Wegrzynski>So if all of them are wrong, you could say, then let's just like keep one that we're using and be consistent with it, because at least I have ability to compare the outcomes of this simulation with 200 other simulations that I've done in my life, and it gives me a reasonable point of evaluating whether this smoke control system is good or bad.

00:31:55.519 --> 00:32:05.839
<v Wojciech Wegrzynski>And it's not because of the realism of my design curve, it's because of the back catalogue of times it has been applied.

00:32:05.839 --> 00:32:18.799
<v Wojciech Wegrzynski>Uh and and for me, this was an example of why seeking the truth, you know, the realism, the real fire is simply failing you because define what the real fire is.

00:32:18.799 --> 00:32:19.440
<v Wojciech Wegrzynski>Not possible.

00:32:19.440 --> 00:32:19.920
<v Wojciech Wegrzynski>Yeah.

00:32:20.079 --> 00:32:22.880
<v David Morrisset>And I think there's, I mean, we've said this before, right?

00:32:22.880 --> 00:32:28.000
<v David Morrisset>But there isn't there is immense utility in being able to have a benchmark for things, right?

00:32:28.000 --> 00:32:28.640
<v David Morrisset>Absolutely.

00:32:28.640 --> 00:32:33.599
<v David Morrisset>And so I think that on one level, it's important to have things that you can benchmark against, right?

00:32:33.599 --> 00:32:48.559
<v David Morrisset>But I think it's also really important to like you just said, if you choose something as a comparative purpose, for a comparative purpose, I should say, then there is utility in that, but not to convince yourself that this is every fire you'll ever experience within this within this compartment, right?

00:32:48.559 --> 00:32:52.079
<v David Morrisset>Or within this you know context, whether it's a car park or or or a building.

00:32:52.480 --> 00:33:01.119
<v Wojciech Wegrzynski>In a car park, the vehicle can burn in so many different ways that are like completely different causes of fire, completely different consequences.

00:33:01.119 --> 00:33:05.279
<v Wojciech Wegrzynski>It can be like one vehicle in the corner, five vehicles in the middle of it, you know.

00:33:05.279 --> 00:33:08.240
<v Wojciech Wegrzynski>It can be a very tall car park, very low car park.

00:33:08.240 --> 00:33:09.519
<v Wojciech Wegrzynski>It can have smoke control.

00:33:09.519 --> 00:33:12.880
<v Wojciech Wegrzynski>Like there are thousands of ways a vehicle can burn.

00:33:12.880 --> 00:33:22.000
<v Wojciech Wegrzynski>So there is no way you can say one fire is represent just like the standard time temperature relationship is not representative of every single fire out there, right?

00:33:22.240 --> 00:33:25.200
<v David Morrisset>But I mean, just go back to that Howard Emmons quote that we started with, right?

00:33:25.200 --> 00:33:30.640
<v David Morrisset>Of like, you know, it's you know, fire science is not a simple subject, you know, it's this complex mix of physics and chemistry.

00:33:30.640 --> 00:33:38.079
<v David Morrisset>And and the reason that we have to we have to resolve the physics down to things like testing and experiments is because it's so complex, right?

00:33:38.079 --> 00:33:43.599
<v David Morrisset>In the same way that if you have that's why we need benchmarks in a lot of uh a lot of cases, yeah, right.

00:33:43.599 --> 00:33:49.200
<v David Morrisset>So it is one of the intrinsic assumptions that we um we sort of started talking about from the out from the outset.

00:33:49.200 --> 00:33:53.359
<v David Morrisset>But I think let's let's maybe even move away from fire testing for a second.

00:33:53.440 --> 00:33:53.599
<v unknown>Yeah.

00:33:53.759 --> 00:33:59.039
<v David Morrisset>And let's talk about, just for a moment, let's talk about shooting chickens into jet fans.

00:33:59.039 --> 00:34:01.279
<v David Morrisset>Because I think why not?

00:34:01.279 --> 00:34:01.599
<v David Morrisset>Why not?

00:34:01.680 --> 00:34:03.039
<v Wojciech Wegrzynski>Yeah, that's a great example.

00:34:03.039 --> 00:34:11.280
<v Wojciech Wegrzynski>It's it's in the show notes, and they're like it's uh I would say it's not yet Angus Law level of writing, but still good enough.

00:34:11.280 --> 00:34:13.599
<v Wojciech Wegrzynski>I highly recommend uh to read it through.

00:34:13.599 --> 00:34:14.800
<v Wojciech Wegrzynski>It's in the show notes.

00:34:14.800 --> 00:34:16.719
<v Wojciech Wegrzynski>Uh check check the paper on your own.

00:34:16.719 --> 00:34:18.800
<v Wojciech Wegrzynski>But let's let please show me the lesson.

00:34:18.800 --> 00:34:20.079
<v Wojciech Wegrzynski>Show me the lesson, David.

00:34:20.320 --> 00:34:25.119
<v David Morrisset>Well, I mean, I'm uh you you haven't heard this until you've heard uh presentation from someone like Angus on this one.

00:34:25.119 --> 00:34:30.320
<v David Morrisset>But um but the the principle of this paper is this paper's called When the Chick Hits the Fan, right?

00:34:30.320 --> 00:34:36.800
<v David Morrisset>And it's a it's an interesting study that's looking at basically the philosophy of standard testing, right?

00:34:36.800 --> 00:34:48.320
<v David Morrisset>And the idea behind this is really if you look at the if you look at the industry of developing basically the turbofan jets for for jet airliners, like for airplanes, basically, right?

00:34:48.320 --> 00:35:13.280
<v David Morrisset>One of the final tests that they have to do to to sort of determine the readiness for bird strikes, so whether or not uh an engine can actually strike a bird and continue on, is they literally will take a compressed air, basically, cannon, right, and they'll spin up one of these turbofans, and they will put effectively what is a frozen chicken in this in this giant air cannon and launch it into the into the engine.

00:35:13.280 --> 00:35:25.039
<v David Morrisset>So, you know, this process by which you're developing these multi-million dollar or pound pieces of equipment, these amazing works of feats of engineering, these turbojet fans, right?

00:35:25.039 --> 00:35:36.239
<v David Morrisset>The final test, really just look at look at readiness, is to literally just chuck a chicken at it, basically, and see if they can like chew it up and continue on without you know completely uh failing the engine.

00:35:36.239 --> 00:35:40.320
<v David Morrisset>And what's really interesting is the paper is marvelously written.

00:35:40.320 --> 00:35:51.679
<v David Morrisset>Where one of my favorite quotes from it is basically, all technological tests, right, um, unavoidably contain irreducible ambiguities that require judgments to bridge, right?

00:35:51.679 --> 00:35:55.119
<v David Morrisset>And so to show these judgments can have real consequences, right?

00:35:55.119 --> 00:36:04.559
<v David Morrisset>Um, and so basically the point of that being that the re the reality is is these engines must be tested to understand if if they can withstand bird strikes, right?

00:36:04.559 --> 00:36:21.440
<v David Morrisset>But if you really take a step back and you look at the in the the absurdity of taking a a you know a chicken and launching it into a jet engine, it's just really sort of it's an interesting sort of thing to comment on because yes, you need a test standard to look at these, right?

00:36:21.440 --> 00:36:28.559
<v David Morrisset>But by assigning that test standard, by by codifying that knowledge to say this is the threshold it must meet, there are implications to that.

00:36:28.559 --> 00:36:32.079
<v David Morrisset>There are consequences where if that that is now what it's designed for, you know.

00:36:32.079 --> 00:36:38.480
<v David Morrisset>So if if if that's your threshold, what happens if the bird strike that it actually experiences is slightly more severe?

00:36:38.480 --> 00:36:39.840
<v David Morrisset>How did we choose that chicken?

00:36:39.840 --> 00:36:40.239
<v David Morrisset>Right.

00:36:40.239 --> 00:36:42.320
<v David Morrisset>What about like, what about other surrogates?

00:36:42.320 --> 00:36:44.960
<v David Morrisset>What if you hit a flock of birds, right?

00:36:44.960 --> 00:36:56.400
<v David Morrisset>So they actually come up with additional tests to look at multiple smaller items or you know, of various combinations of different, you know, sizes and weight distributions of these frozen birds.

00:36:56.400 --> 00:37:06.559
<v David Morrisset>But all of these are really interesting because there's a direct sort of application within fire safety where there is unavoidable ambiguity in developing a test standard.

00:37:06.559 --> 00:37:14.639
<v David Morrisset>But you got to draw the line somewhere and you have to say, you know, what is what is my threshold and what are the consequences of assuming that this is my threshold?

00:37:14.880 --> 00:37:30.000
<v Wojciech Wegrzynski>I I think uh nice parallel in here you said that uh you want this test of a bird hitting the fan as a proxy of what will happen if a real-world jet engine hits a bird while the plane is flying, hopefully not crushing and killing everyone on the board.

00:37:30.000 --> 00:37:31.760
<v Wojciech Wegrzynski>That makes a lot of sense, right?

00:37:31.760 --> 00:37:37.280
<v Wojciech Wegrzynski>Uh but does this prove you that the real bird will not destroy the jet fan engine?

00:37:37.280 --> 00:37:39.840
<v Wojciech Wegrzynski>In no way does the does it prove you, it's just a proxy.

00:37:39.840 --> 00:37:56.800
<v Wojciech Wegrzynski>In the same way, a lot of our colleagues, fire safety engineers, would be burdened with having to solve some technical problems, some you know, technical challenge within their building, how to apply a specific piece of equipment in a specific context of a building.

00:37:56.800 --> 00:38:04.000
<v Wojciech Wegrzynski>And they only have this body of knowledge that comes from tests because this is what the manufacturer gives them.

00:38:04.000 --> 00:38:15.760
<v Wojciech Wegrzynski>We will go into how they can design an experiment to solve their problem in a second, but let's for now assume they're not able to design an experiment and they only have the body of knowledge that comes from tests, you know.

00:38:15.760 --> 00:38:24.239
<v Wojciech Wegrzynski>What caveas those engineers should avoid to use that testing data to find an answer?

00:38:24.239 --> 00:38:33.440
<v Wojciech Wegrzynski>Because if the question is, will this uh wall survive 60 minutes and I only have a 60-minute fire resistance test?

00:38:33.440 --> 00:38:42.000
<v Wojciech Wegrzynski>On one hand, it seems very simple, yes, because it's 60 minutes versus 60 minutes, but as you cleverly said before, those are very different minutes.

00:38:42.000 --> 00:38:45.360
<v Wojciech Wegrzynski>So the pathway to achieve the answer is actually quite complicated.

00:38:45.360 --> 00:38:50.880
<v Wojciech Wegrzynski>So so how how not not always you will use tests to rank things up.

00:38:50.880 --> 00:38:54.960
<v Wojciech Wegrzynski>Sometimes you need to extract information and knowledge from them.

00:38:54.960 --> 00:38:57.119
<v Wojciech Wegrzynski>How how how can an engineer can do that?

00:38:57.519 --> 00:39:00.239
<v David Morrisset>But I mean that is actually the trick, right?

00:39:00.239 --> 00:39:06.239
<v David Morrisset>Because effectively all a test provides is a test does not provide safety, it provides compliance, right?

00:39:06.239 --> 00:39:07.360
<v David Morrisset>There's a big difference.

00:39:07.360 --> 00:39:11.519
<v David Morrisset>You say that you can tick the boxes that this has been deemed compliant, right?

00:39:11.519 --> 00:39:16.400
<v David Morrisset>And that is an implicit level of safety where you can start extracting information from that.

00:39:16.400 --> 00:39:22.079
<v David Morrisset>And the person to really speak about this would be like Reuben Van Coyle and his team, right, looking at adaptive fire testing, right?

00:39:22.079 --> 00:39:40.960
<v David Morrisset>Because there are ways where you can take the frameworks that exist for fire testing, and by you know adding more measurements or or figuring out, you know, the different ways to sort of frame that test, you can still achieve sort of a standard test and extract more information that might actually be usable from an engineering context, right?

00:39:40.960 --> 00:39:47.440
<v David Morrisset>What you can't currently do, right, is basically take things like, let me take a random example that uh comes to mind, right?

00:39:47.440 --> 00:39:50.639
<v David Morrisset>There is a there is an ASDM standard for the Lyft, right?

00:39:50.639 --> 00:39:55.280
<v David Morrisset>The flame, the lateral flame spread test developed by uh Quintiri and his colleagues, right?

00:39:55.280 --> 00:39:56.480
<v David Morrisset>And it's an elegant test.

00:39:56.480 --> 00:40:00.000
<v David Morrisset>It's it's you know, as far as standardized tests go, I could talk all day about that.

00:40:00.000 --> 00:40:07.440
<v David Morrisset>There's there's some really interest, really interesting theory backed up by really clever design of a test system, right?

00:40:07.440 --> 00:40:14.800
<v David Morrisset>And it's it's very excellent what you can in terms of the the the way you can sort of drop out some very valuable information very quickly from those tests.

00:40:14.800 --> 00:40:22.320
<v David Morrisset>Now, the caveat being that, you know, one of the parameters that drops out of this is this flame spread parameter called fee, right?

00:40:22.320 --> 00:40:29.280
<v David Morrisset>And you can you can use that to index basically the flame spread behavior across different materials under that test.

00:40:29.280 --> 00:40:40.239
<v David Morrisset>One thing that I would that I I've I've seen engineers and I've just talked about engineers or talked with engineers about using is can we use those fee values to then approximate flame spread in a design fire scenario?

00:40:40.239 --> 00:40:44.320
<v David Morrisset>And I think uh the the sort of default answer to that will always be no.

00:40:44.320 --> 00:40:49.519
<v David Morrisset>There's always a huge caveat there of these are limited to the context of that test.

00:40:49.519 --> 00:40:52.639
<v David Morrisset>So unless your design fire is a lift apparatus, right?

00:40:52.639 --> 00:40:57.360
<v David Morrisset>Like I wouldn't use those values for predictive design conditions, right?

00:40:57.360 --> 00:41:02.639
<v David Morrisset>Again, it's a very, very useful, extremely useful test for benchmarking materials, right?

00:41:03.039 --> 00:41:11.840
<v Wojciech Wegrzynski>And I would say the same challenge would go with using heat release rate per unit area from your cone into your design fire in the building.

00:41:11.840 --> 00:41:25.440
<v Wojciech Wegrzynski>The same would go for using your oxygen bomb heat of combustion and using it for uh fuel burning outdoors in a in a real uh in a real space and and uh and and th and things like this uh are endless.

00:41:25.440 --> 00:41:39.440
<v Wojciech Wegrzynski>Sometimes stuff looks like looks very familiar and looks very directly applicable, yet the challenges within the context and and and how what it responded to in that test.

00:41:39.440 --> 00:41:41.920
<v Wojciech Wegrzynski>And I think this is the thing you need to understand.

00:41:41.920 --> 00:41:54.719
<v Wojciech Wegrzynski>You cannot like unfortunately the foreign engineer, when they try to apply outcomes of a test in some form of engineering beyond just ranking, beyond just applying a certificate, code speak, etc.

00:41:54.719 --> 00:42:06.079
<v Wojciech Wegrzynski>If they want to use this in engineering, unfortunately they must be very familiar with the testing regime, standard assumptions, and the practice of how the test is performed.

00:42:06.079 --> 00:42:06.800
<v Wojciech Wegrzynski>That's my opinion.

00:42:06.800 --> 00:42:07.519
<v Wojciech Wegrzynski>Absolutely.

00:42:07.679 --> 00:42:13.840
<v David Morrisset>And and not to say that we we can't get to a stage where we can do engineering analysis with various tests, right?

00:42:13.840 --> 00:42:16.639
<v David Morrisset>But just as as a that shouldn't be the baseline assumption.

00:42:16.639 --> 00:42:23.440
<v David Morrisset>The baseline assumption shouldn't be I can use this test to do additional engineering design, right?

00:42:23.440 --> 00:42:24.880
<v David Morrisset>It should be the other way around.

00:42:24.880 --> 00:42:28.000
<v David Morrisset>The outcomes are limited to the test itself.

00:42:28.000 --> 00:42:34.639
<v David Morrisset>And if there is applicability, if I can if I can fully get around the context, then perhaps there's additional insight I can gain from that.

00:42:34.639 --> 00:42:35.440
<v David Morrisset>Perfect.

00:42:35.679 --> 00:42:48.159
<v Wojciech Wegrzynski>And uh what if like we've took the ability to experiment from the hands of our engineer, let's give it back into the hands of uh the fire engineer, the uh ability to experiment.

00:42:48.159 --> 00:42:53.360
<v Wojciech Wegrzynski>How can the foreign engineer design an experiment to guide them?

00:42:53.360 --> 00:42:58.159
<v Wojciech Wegrzynski>Because you're an experimenter, you know it is very difficult to design those.

00:42:58.159 --> 00:43:08.400
<v Wojciech Wegrzynski>So what what would be the pitfalls of designing an experiment and and and what one has to think when they really need a specific answer to a very specific question?

00:43:08.400 --> 00:43:11.679
<v Wojciech Wegrzynski>How do they get to that ex to that answer through experiments?

00:43:11.920 --> 00:43:12.159
<v David Morrisset>Right.

00:43:12.159 --> 00:43:29.360
<v David Morrisset>So in this transition now to discussing experiments, I think it's uh let's revisit that definition of basically an experiment is we'll say a it's to explore fire phenomena, uh various fire phenomena or a specific fire phenomenon that is not followed by a standard procedure, right?

00:43:29.360 --> 00:43:42.960
<v David Morrisset>So you have the the freedom to sort of explore whatever you want and or to explore whatever you want and design the experiment in the way you want to, but the pitfall of that is you don't have a track record of a standardized uh procedure by which people spend.

00:43:43.519 --> 00:43:43.920
<v Wojciech Wegrzynski>Why not?

00:43:43.920 --> 00:43:49.760
<v Wojciech Wegrzynski>Why cannot I do like if you do you can do cone experiments and they're like Yeah, of course.

00:43:49.840 --> 00:43:50.800
<v David Morrisset>No, no, you're right, you're right.

00:43:50.800 --> 00:43:51.679
<v David Morrisset>And that's an important point.

00:43:51.679 --> 00:43:55.599
<v David Morrisset>But what I'm saying, well, all I'm saying is is for example, that's a great example actually.

00:43:55.599 --> 00:43:58.880
<v David Morrisset>You can take the cone calorimeter, a very standardized piece of equipment.

00:43:58.880 --> 00:44:01.119
<v David Morrisset>And I have used it for experiments all the time.

00:44:01.119 --> 00:44:05.280
<v David Morrisset>Because the second you deviate from one thing from the standard, you can, it's no longer a standard test, right?

00:44:05.280 --> 00:44:08.079
<v David Morrisset>You're using this apparatus to explore something, right?

00:44:08.079 --> 00:44:14.639
<v David Morrisset>And I don't, I mean, I agree, I think that's actually standard testing kit is sometimes the best place to start for an experiment, right?

00:44:14.639 --> 00:44:15.360
<v David Morrisset>Absolutely.

00:44:15.360 --> 00:44:20.320
<v David Morrisset>All I'm saying is now you have to basically once you change one of those things, right?

00:44:20.320 --> 00:44:23.840
<v David Morrisset>Those is you need to sort of understand what are the implications of that, right?

00:44:23.840 --> 00:44:29.760
<v David Morrisset>If I'm changing the size of my sample so it's no longer the standard, but what are the heat transfer effects within that sample?

00:44:29.760 --> 00:44:33.360
<v David Morrisset>Am I truly getting things like in the context of the cone calorimeter?

00:44:33.360 --> 00:44:41.920
<v David Morrisset>The reason the size is the sample size is specified to what it is is to achieve a uniform heat flux to create as close to one-dimensional heating conditions as possible.

00:44:41.920 --> 00:44:46.239
<v David Morrisset>I've done experiments where I've put a sample way bigger than that under the cone.

00:44:46.239 --> 00:44:53.039
<v David Morrisset>And that's okay, but you have to acknowledge the fact that you're probably not going to get nice one-dimensional heat transfer anymore.

00:44:53.039 --> 00:44:54.559
<v David Morrisset>There's going to be other issues at play.

00:44:54.559 --> 00:44:56.880
<v David Morrisset>Which maybe that's a problem, maybe that's not, right?

00:44:56.880 --> 00:45:01.119
<v David Morrisset>I think uh this is one that we actually we get kind of, it's funny, right?

00:45:01.119 --> 00:45:11.679
<v David Morrisset>When we're when we're working on these kinds of things where we're doing experiments on a standard piece of kit, like the cone, we've gotten comments from from reviewers and so on that, you know, that we're not following the standard, right?

00:45:11.679 --> 00:45:12.639
<v David Morrisset>It's an interesting one.

00:45:12.639 --> 00:45:13.760
<v David Morrisset>And I'm sure you've seen some of these.

00:45:13.760 --> 00:45:16.400
<v David Morrisset>Some similar comments to this too, right?

00:45:16.400 --> 00:45:21.119
<v David Morrisset>And it's and it's funny because like nowhere uh we're not we're not necessarily trying to follow the standard.

00:45:21.119 --> 00:45:28.880
<v David Morrisset>We appreciate the the amount of effort that went into the to the the structure of of the standard procedure and the equipment that came from it.

00:45:28.880 --> 00:45:38.880
<v David Morrisset>But I don't I don't see a downside then at all to using that kit to run a different experiment, to, to change things like the pilot location, to remove the pilot and look at auto ignition, right?

00:45:38.880 --> 00:45:42.559
<v David Morrisset>To to you know, to change the size of your sample, to change the orientation.

00:45:42.559 --> 00:45:49.199
<v David Morrisset>My during my master's thesis, I ran a bunch of experiments angling the cone at angles in between horizontal and vertical, right?

00:45:49.199 --> 00:45:53.360
<v David Morrisset>So I would go 45 degrees and 30 degrees and 60 degrees with the cone.

00:45:53.360 --> 00:45:57.039
<v David Morrisset>And of course, that's a non-standard thing, but you have the ability to do that, right?

00:45:57.039 --> 00:46:04.880
<v David Morrisset>And I think there's a lot of benefit in using that kind of equipment to do non-standard, um, truly experimental procedures.

00:46:04.880 --> 00:46:09.280
<v David Morrisset>The question becomes if you change something from that standard, why is that?

00:46:09.280 --> 00:46:11.920
<v David Morrisset>Why, why do why are you making the measurement that you're making?

00:46:11.920 --> 00:46:14.000
<v David Morrisset>Why are you changing the conditions that you're looking at?

00:46:14.000 --> 00:46:19.760
<v David Morrisset>Because when you're establishing an experiment, the first question should always be, what is the what is on, what is it that I'm trying to articulate?

00:46:19.760 --> 00:46:21.280
<v David Morrisset>What is it that I'm trying to explore?

00:46:21.280 --> 00:46:21.760
<v David Morrisset>Right?

00:46:21.760 --> 00:46:23.760
<v David Morrisset>And then the question becomes, how do I achieve that?

00:46:23.760 --> 00:46:30.400
<v David Morrisset>And one of the big questions to look at is, depending on my question, what are the boundary conditions I'm gonna use, right?

00:46:30.400 --> 00:46:37.920
<v David Morrisset>Is it sufficient to do something on a really small scale and understand what a heat flux exposure at the surface would tell you?

00:46:37.920 --> 00:46:38.239
<v David Morrisset>Right?

00:46:38.239 --> 00:46:39.760
<v David Morrisset>Can you gain insight from that?

00:46:39.760 --> 00:46:41.360
<v David Morrisset>Do you need to go to a large scale?

00:46:41.360 --> 00:46:46.480
<v David Morrisset>Do you need to actually look at, you know, fire behavior of uh at the compartment scale, right?

00:46:46.480 --> 00:46:49.440
<v David Morrisset>And if you're gonna do that, how do you change those boundary conditions?

00:46:49.440 --> 00:46:51.440
<v David Morrisset>What kind of initial fire do you use?

00:46:51.440 --> 00:46:57.360
<v David Morrisset>Do you use things like you know wood cribs within the compartment to sustain a fully developed fire at some stage?

00:46:57.360 --> 00:47:05.840
<v David Morrisset>But all these you you actually assign these boundary conditions in an experiment, there's nothing stopping you from completely changing any of those things, right?

00:47:05.840 --> 00:47:10.320
<v David Morrisset>And sometimes it can be a bit of a decision paralysis kind of situation, right?

00:47:10.320 --> 00:47:21.599
<v David Morrisset>Where you want to, you know, there's so many options to choose from, but you want to choose to choose the combination of boundary conditions and measurements, they're gonna give you the outputs that you're looking for.

00:47:22.000 --> 00:47:35.599
<v Wojciech Wegrzynski>I resonate with you and your challenges with the cone because we went through a hell when we proposed the heat release rate uh control regime for a furnace, and uh the first response was you're not allowed to do that, it's a time-temperature relation.

00:47:35.599 --> 00:47:36.880
<v Wojciech Wegrzynski>And by the way, it's standard.

00:47:36.880 --> 00:47:38.880
<v Wojciech Wegrzynski>So that that was a fun one.

00:47:38.880 --> 00:47:45.440
<v Wojciech Wegrzynski>Uh I think uh another thing in those experiments, like I also cherish the variable control.

00:47:45.440 --> 00:47:54.239
<v Wojciech Wegrzynski>I also cherish uh ability to like change one thing in the system and uh find out the outcomes.

00:47:54.239 --> 00:48:00.639
<v Wojciech Wegrzynski>Unfortunately, in the realm of fire, it's not always possible to fiddle with a single variable at a time.

00:48:00.639 --> 00:48:04.880
<v Wojciech Wegrzynski>Uh, the costs and scales of experiments are often way, way too big.

00:48:04.880 --> 00:48:12.880
<v Wojciech Wegrzynski>And also eventually you also get into space of something that I like the most, which are exploratory tests, experiments.

00:48:12.880 --> 00:48:14.559
<v Wojciech Wegrzynski>I'm not sure they're experiments.

00:48:14.559 --> 00:48:16.400
<v Wojciech Wegrzynski>I think they're experiments, not tests.

00:48:16.400 --> 00:48:25.360
<v Wojciech Wegrzynski>Exploratory in a way that you try to do something that has never really been done, and you're absolutely not aware of what will be the outcomes.

00:48:25.360 --> 00:48:29.360
<v Wojciech Wegrzynski>Like literally, you have no expectations towards the experiment.

00:48:29.360 --> 00:48:43.519
<v Wojciech Wegrzynski>You are doing it for the sake of it to be done to provide you this first reference point from which you can move onwards and start your play with the variables, start your play with with stuff that for me those are the most rewarding experiments, really.

00:48:43.519 --> 00:48:49.119
<v Wojciech Wegrzynski>And uh surprisingly, there's a lot of things we don't have exploratory experiments in fire science yet.

00:48:49.360 --> 00:48:51.440
<v David Morrisset>Yeah, no, it's that's really interesting, right?

00:48:51.440 --> 00:49:08.559
<v David Morrisset>It is the idea of if there's a if there's a question, if if I'm an engineer and I'm like, I want to know how a fire would behave in this really random, complex condition uh that's in my building, I think it's really satisfying to try to develop a an exploratory experiment and just see how it goes, right?

00:49:08.559 --> 00:49:10.559
<v David Morrisset>And you can gain so much insight from that.

00:49:10.559 --> 00:49:13.519
<v David Morrisset>Again, let's say, let's say let's go with that example.

00:49:13.519 --> 00:49:30.239
<v David Morrisset>And let's say we have uh an atrium space, I don't know, in a building that's a bit complicated and and you you have a good idea of what the fuel load might look like, and you just want to know how would, you know, what would the fire behavior of this X fuel load actually look like, you can try it out.

00:49:30.239 --> 00:49:35.840
<v David Morrisset>Just do an exploratory test, bring it into a lab, light it up, and see what kind of fire behaviors you observe.

00:49:35.840 --> 00:49:42.960
<v David Morrisset>The hard part becomes how you do that, right, is gonna be very uh the outcomes from that are gonna depend on how you do it, right?

00:49:42.960 --> 00:49:48.960
<v David Morrisset>So you're not gonna develop a magical design condition, but you'll see how this particular trial went, right?

00:49:48.960 --> 00:49:52.559
<v David Morrisset>You depending on how you ignite it, depending on how you you know arrange your fuel.

00:49:52.559 --> 00:49:55.360
<v David Morrisset>Say you're looking at a Christmas tree, right, in an atrium.

00:49:55.360 --> 00:49:55.599
<v David Morrisset>Yeah.

00:49:55.599 --> 00:49:59.360
<v David Morrisset>The way that you light that Christmas tree is gonna change the heat release rate behavior, right?

00:49:59.360 --> 00:50:01.679
<v David Morrisset>Because you're gonna get different rates of fire growth and so on.

00:50:01.679 --> 00:50:11.360
<v David Morrisset>And all those things create a single fire uh condition, which gives you insight for sure, but it doesn't necessarily automatically become a gold standard benchmark.

00:50:12.000 --> 00:50:25.039
<v Wojciech Wegrzynski>And in almost no circumstance you will be able to claim that this is how a fire will develop, and the consequences of the fire will be exactly like that in the future.

00:50:25.039 --> 00:50:28.320
<v Wojciech Wegrzynski>It will just give you a reference data point.

00:50:28.320 --> 00:50:32.159
<v Wojciech Wegrzynski>The more you do, the better your cloud of points is.

00:50:32.159 --> 00:50:37.360
<v Wojciech Wegrzynski>You can start playing with statistics, uncertainty, you can start controlling for that.

00:50:37.360 --> 00:50:48.639
<v Wojciech Wegrzynski>But again, you will not, even if you set a fire in an office uh build in the same way, like uh you would have it in a real building.

00:50:48.639 --> 00:50:49.840
<v Wojciech Wegrzynski>Look, look, Dalmerlock.

00:50:49.840 --> 00:50:53.199
<v Wojciech Wegrzynski>I had Wolfram and Guillermo on the podcast here again.

00:50:53.199 --> 00:50:58.320
<v Wojciech Wegrzynski>A simple rug changed an outcome of an experiment uh in Dalmerlock time.

00:50:58.320 --> 00:51:05.599
<v Wojciech Wegrzynski>So that's something that I would love the engineers to be absolutely fully like aware.

00:51:05.599 --> 00:51:11.760
<v Wojciech Wegrzynski>Even if you do it and it looks real, it doesn't mean that the fire will grow like this.

00:51:11.760 --> 00:51:20.239
<v Wojciech Wegrzynski>And this is a struggle I have all the time with car park fires, you know, because PRE has burned a real car, which means the cars burned like that, right?

00:51:20.239 --> 00:51:23.360
<v Wojciech Wegrzynski>Well, no, it means in that experiment it burned like that.

00:51:23.360 --> 00:51:27.599
<v Wojciech Wegrzynski>It doesn't mean the vehicles at large will always behave like that.

00:51:27.599 --> 00:51:49.519
<v Wojciech Wegrzynski>This is a very big pitfall, I think, where sometimes people would pursue experiments because they need to have a proof of how the truth will look like, whereas in real outcome they just get a sample, you know, they just get the data point, and the interpretation of the data point is very difficult.

00:51:50.079 --> 00:51:50.400
<v David Morrisset>Definitely.

00:51:50.400 --> 00:51:58.239
<v David Morrisset>I mean, I think so, yeah, there's a lot of there's a lot of benefit in having something like uh an exploratory experiment, right?

00:51:58.239 --> 00:52:00.639
<v David Morrisset>Let's imagine go back to your car park example, right?

00:52:00.639 --> 00:52:04.960
<v David Morrisset>And let's say that there's a we want to understand what's the dynamics of a fire in this car park.

00:52:04.960 --> 00:52:12.559
<v David Morrisset>And one of the say one of the core elements of this engineering analysis is understanding the heat flux exposure to nearby objects, right?

00:52:12.559 --> 00:52:18.480
<v David Morrisset>You can actually get a really good idea of what kind of heat fluxes you can expect from an actual laboratory experiment, right?

00:52:18.480 --> 00:52:22.559
<v David Morrisset>We go back go back to our previous episode of how do we measure things like heat fluxes.

00:52:22.559 --> 00:52:24.159
<v David Morrisset>You can you can make these measurements, right?

00:52:24.159 --> 00:52:32.480
<v David Morrisset>And and again, that that doesn't mean that this is going to be your this isn't the fire that's gonna happen every time for your design condition, but you can still gain a lot of insight from that.

00:52:32.480 --> 00:52:40.559
<v David Morrisset>But the but then you start entering this really slippery slope where if if let's say you want to measure the radiant heat flux from this object, right?

00:52:40.559 --> 00:52:48.159
<v David Morrisset>And that's a really integral part of your your analysis, but then you're in a position where what if uh how how do you interpret those results?

00:52:48.159 --> 00:53:01.280
<v David Morrisset>What if one of your heat fluxes is, you know, wildly high because actually you had an EV fire and there was a jet flame coming out of one side of it and it was, you know, the jet flame was basically impinging on your heat flux gauge.

00:53:01.280 --> 00:53:10.719
<v David Morrisset>Does that become now, if that is a really high heat flux, can you buy it from from across the board, say as a design standard, can we use that as the maximum heat flux?

00:53:10.719 --> 00:53:12.480
<v David Morrisset>And that's a that's a tricky question, right?

00:53:12.480 --> 00:53:17.679
<v David Morrisset>Because you need to then understand what's the, again, what's the context of why was that measurement so high, right?

00:53:17.679 --> 00:53:24.159
<v David Morrisset>And then can you overlay that context to understand the utility of the outcomes from that kind of test or experiment in that case?

00:53:24.480 --> 00:53:38.880
<v Wojciech Wegrzynski>Oh man, I I think with uh flood of people with uh a lot of things and uh in the end uh the the things are more dull than uh they have been at the beginning of episodes because now uh the difference between the tests and experiments is even more unclear.

00:53:38.880 --> 00:53:40.880
<v Wojciech Wegrzynski>But uh it kind of is the point, you know.

00:53:40.880 --> 00:53:53.119
<v Wojciech Wegrzynski>I I I really want to embed this image in in people's heads that testing regime is a testing regime that just lows your comparison between different test outcomes in whatever framework the test has been conceived for.

00:53:53.119 --> 00:53:55.039
<v Wojciech Wegrzynski>And that that's the premise of a test.

00:53:55.039 --> 00:54:07.199
<v Wojciech Wegrzynski>Whatever mole you do with that, you do it at your own risk, and you do it at your own uh you know conscious choice and your own conscious interpretation of the new context, the new application, the new idea.

00:54:07.199 --> 00:54:20.320
<v Wojciech Wegrzynski>And if you are conscious to that, we've achieved success with this with this uh with this podcast episode because perhaps uh there will be less extrapolation of things that are not extrapolatable if that even is a word.

00:54:20.320 --> 00:54:27.760
<v Wojciech Wegrzynski>And also experiments are not like your one single direct proof of what reality is gonna look like.

00:54:27.760 --> 00:54:34.320
<v Wojciech Wegrzynski>They're merely a data point crafted within the realm of variables that you have chosen to scout with.

00:54:34.320 --> 00:54:36.960
<v Wojciech Wegrzynski>That's a message I would like to believe the listener with.

00:54:37.280 --> 00:54:43.039
<v David Morrisset>And I think it's important to highlight there's there's I mean, tests are what they are and experiments kind of are what they are, right?

00:54:43.039 --> 00:54:44.480
<v David Morrisset>There's no the there's no pros.

00:54:44.480 --> 00:54:49.440
<v David Morrisset>There are pros and cons to both, right, in terms of of of what they offer in terms of utility.

00:54:49.440 --> 00:55:07.039
<v David Morrisset>I think also in terms of particularly in experiments, there's um sometimes there's an over-reliance on this being the the the magic the magic bullet of this is exactly the fire that I'm going to see in my you know my design condition, which is not true, but you can still gain a lot of insight into some of the behavior you might expect.

00:55:07.039 --> 00:55:11.280
<v David Morrisset>Uh in the same way that some experiments are, you know, really fundamental, right?

00:55:11.280 --> 00:55:18.159
<v David Morrisset>You know, I I do quite a lot in the world of you know ignition and flame spread and and I, you know, things and sort of in combustion diagnostics, too.

00:55:18.159 --> 00:55:23.920
<v David Morrisset>And it's hard to see the immediate uh engineering application of some of these measurements, right?

00:55:23.920 --> 00:55:34.559
<v David Morrisset>But I think it's really important to remember the really the the goal of that scale of experiment, right, is to understand to build engineering judgment by which we can then apply to other problems.

00:55:34.559 --> 00:55:36.320
<v David Morrisset>So it all kind of builds up.

00:55:36.320 --> 00:55:49.440
<v David Morrisset>Again, that's not a uh flame spread on a slab PMMA is not something that actually most designers are gonna have to necessarily ever encounter, but the insight you can gain from that can then help uh develop engineering judgment for other scenarios.

00:55:49.440 --> 00:56:01.199
<v David Morrisset>So basically trying to understand where on the spectrum of testing to experiments the study that you're looking at exists, and understanding, you know, what is the what how do I use this information, right?

00:56:01.199 --> 00:56:07.920
<v David Morrisset>Um, and making sure that all of it is within the context of what those tests or experiments were intended for.

00:56:08.320 --> 00:56:08.960
<v Wojciech Wegrzynski>Fantastic.

00:56:08.960 --> 00:56:11.679
<v Wojciech Wegrzynski>Uh well, thank thank you so much for this conversation.

00:56:11.679 --> 00:56:14.400
<v Wojciech Wegrzynski>As you said at the beginning, it's always nice to talk.

00:56:14.400 --> 00:56:28.239
<v Wojciech Wegrzynski>And I hope we triggered a lot of people and you have very strong opinions about how wrong we are, and we are very keen to hear about uh that from you because it's uh it's an onset for another interesting uh conversation to be had.

00:56:28.239 --> 00:56:33.840
<v Wojciech Wegrzynski>And uh as promised, laser diagnostics uh will eventually come.

00:56:33.840 --> 00:56:42.800
<v Wojciech Wegrzynski>They have not been covered in this episode, but I have a feeling David will come back to a fire science show once again for another interesting conversation.

00:56:42.800 --> 00:56:44.559
<v Wojciech Wegrzynski>Thanks, David, for coming here today.

00:56:44.559 --> 00:56:45.920
<v Wojciech Wegrzynski>Twist my arm.

00:56:45.920 --> 00:56:46.719
<v Wojciech Wegrzynski>I guess I'll come back.

00:56:46.719 --> 00:56:47.440
<v Wojciech Wegrzynski>And that's it.

00:56:47.440 --> 00:56:48.079
<v Wojciech Wegrzynski>Thank you for listening.

00:56:48.079 --> 00:56:50.159
<v Wojciech Wegrzynski>Oh David, I don't need to twist your arm.

00:56:50.159 --> 00:56:51.519
<v Wojciech Wegrzynski>I know you will come back here.

00:56:51.519 --> 00:56:54.960
<v Wojciech Wegrzynski>And I already have like three different ideas for podcast episodes.

00:56:54.960 --> 00:56:55.679
<v Wojciech Wegrzynski>It's just so nice.

00:56:55.679 --> 00:57:03.280
<v Wojciech Wegrzynski>Uh chatting with another fire nerd that loves talking about fire and actually shares very useful information uh alongside.

00:57:03.280 --> 00:57:20.239
<v Wojciech Wegrzynski>I hope, my dear listeners, we have not made uh your idea of what is a test and what is experiment even worse than it was before the podcast episode, but it actually comes down to a very simple thing, the context and how do you want to use it?

00:57:20.239 --> 00:57:28.400
<v Wojciech Wegrzynski>If you want repetability, if you want scoring, if you want classification you're dealing with a test.

00:57:28.400 --> 00:57:37.440
<v Wojciech Wegrzynski>If you want knowledge, if you want a variable control, if you want insight, well you're probably playing with an experiment.

00:57:37.440 --> 00:57:40.000
<v Wojciech Wegrzynski>Can a test be an experiment?

00:57:40.000 --> 00:57:48.000
<v Wojciech Wegrzynski>If you know how to use the data from a test in a new way, we've done that multiple times with Imperial College.

00:57:48.000 --> 00:57:50.320
<v Wojciech Wegrzynski>I would say they are experiments.

00:57:50.320 --> 00:57:52.960
<v Wojciech Wegrzynski>And can experiments be tests?

00:57:52.960 --> 00:57:54.480
<v Wojciech Wegrzynski>Yes, sometimes.

00:57:54.480 --> 00:58:06.400
<v Wojciech Wegrzynski>Uh they're just a repetition of each other to change if a different boundary condition, like a different material, different fuel, different opening factor or something changes the outcome at large.

00:58:06.400 --> 00:58:08.880
<v Wojciech Wegrzynski>Yes, at some point they start to look like tests.

00:58:08.880 --> 00:58:13.920
<v Wojciech Wegrzynski>It's all it all depends, I guess, whether they are used for any sort of classification or not.

00:58:13.920 --> 00:58:17.440
<v Wojciech Wegrzynski>That's where I would put the boundary line between each other.

00:58:17.440 --> 00:58:21.280
<v Wojciech Wegrzynski>But anyway, I I hope you got something from this episode.

00:58:21.280 --> 00:58:39.440
<v Wojciech Wegrzynski>One important lesson that I really, really, really wanted to translate in the middle of the episode was that if you are forced to do fire engineering and only thing that you have is access to outcomes of fire tests from standardized methods, from fire testing laboratories.

00:58:39.440 --> 00:58:41.920
<v Wojciech Wegrzynski>It's difficult, it's not straightforward.

00:58:41.920 --> 00:58:50.320
<v Wojciech Wegrzynski>Minutes are not minutes, hit fluxes are not hit fluxes, flame spread velocities are not flamespread velocities in terms of the real world problems.

00:58:50.320 --> 00:59:04.639
<v Wojciech Wegrzynski>You have to have a very, very good, robust understanding of the testing regime, the assumptions, the limitations, the way how the sample is built, handled, mounted, seasoned, etc.

00:59:04.639 --> 00:59:11.840
<v Wojciech Wegrzynski>to be able to actually apply anything that comes from a laboratory into the reality.

00:59:11.840 --> 00:59:19.679
<v Wojciech Wegrzynski>Whereas when you're designing an experiment, you have to be extremely good in defining those variables to match your real world conditions.

00:59:19.679 --> 00:59:23.679
<v Wojciech Wegrzynski>That is the challenge and that is the practical lesson behind this experiment.

00:59:23.679 --> 00:59:28.079
<v Wojciech Wegrzynski>I hope you found some more beyond this in this in this nice talk with David.

00:59:28.079 --> 00:59:31.840
<v Wojciech Wegrzynski>Thank you very much for being here with us in the fire science show.

00:59:31.840 --> 00:59:35.679
<v Wojciech Wegrzynski>Uh it's the first day of the academic year in here in Poland.

00:59:35.679 --> 00:59:38.559
<v Wojciech Wegrzynski>So I wish all the students all the best.

00:59:38.559 --> 00:59:44.880
<v Wojciech Wegrzynski>If you know a student who is starting fire safety engineering this year, make sure they know about the fire science show.

00:59:44.880 --> 00:59:47.440
<v Wojciech Wegrzynski>It's an important resource for them.

00:59:47.440 --> 00:59:50.559
<v Wojciech Wegrzynski>So make sure they are aware of the podcast.

00:59:50.559 --> 00:59:55.840
<v Wojciech Wegrzynski>Let's get them hooked on fire science and let's help the best for their growth development.

00:59:55.840 --> 00:59:59.679
<v Wojciech Wegrzynski>And in some years we're gonna have more fire engineers around because we are way.

00:59:59.679 --> 01:00:01.199
<v Wojciech Wegrzynski>Way way too few.

01:00:01.199 --> 01:00:07.280
<v Wojciech Wegrzynski>Anyway, thank you for being here with me this Wednesday, and I am looking forward to meet you again next Wednesday.

01:00:07.280 --> 01:00:08.559
<v Wojciech Wegrzynski>Same place, same time.

01:00:08.559 --> 01:00:09.039
<v Wojciech Wegrzynski>See you.

01:00:09.039 --> 01:00:09.760
<v Wojciech Wegrzynski>Bye.