216 - What do we measure and how? with David Morrisset

Transcript

Introduction to Fire Measurements Understanding Temperature Measurements Pressure Probes and Flow Measurements Heat Flux Gauges and Challenges Heat Release Rate Measurements Mass Loss Rate Measurement Data Processing and Smoothing Techniques Advanced Measurement Methods

WEBVTT 00:00:00.201 --> 00:00:01.907 <v Wojciech Wegrzynski>Hello everybody, welcome to the Fire Science Show. 00:00:01.907 --> 00:00:10.150 <v Wojciech Wegrzynski>So you're a fire safety engineer, you're working on a project, you have to solve a specific fire safety case, a problem. 00:00:10.150 --> 00:00:14.026 <v Wojciech Wegrzynski>You need to find a solution, you need to use some tools modeling. 00:00:14.026 --> 00:00:15.865 <v Wojciech Wegrzynski>What do you put in those models? 00:00:15.865 --> 00:00:29.899 <v Wojciech Wegrzynski>Usually, the golden standard would be to find an experiment that someone has done in a setting that's relevant to the case that you're studying and use the data from that experiment in your modeling, in your research, directly. 00:00:29.899 --> 00:00:33.246 <v Wojciech Wegrzynski>Because, hey, experiments are the the truth, right? 00:00:33.246 --> 00:00:34.850 <v Wojciech Wegrzynski>So, yeah, they. 00:00:34.850 --> 00:00:38.606 <v Wojciech Wegrzynski>They are the manifestation of laws of physics. 00:00:38.606 --> 00:00:42.359 <v Wojciech Wegrzynski>The laws of physics in the real world are not optional like in your simulation. 00:00:42.359 --> 00:00:45.066 <v Wojciech Wegrzynski>You don't have to turn them on for them to work. 00:00:45.066 --> 00:00:53.823 <v Wojciech Wegrzynski>But to have a really good value out of an experiment, one thing needs to be done correct, and that is the measurements. 00:00:53.823 --> 00:00:57.470 <v Wojciech Wegrzynski>And measurements boy, this is not a simple topic. 00:00:57.470 --> 00:01:06.819 <v Wojciech Wegrzynski>While it looks easy, you just stick a thermocouple into a fire, there is surprisingly, surprisingly a lot of things that have to happen. 00:01:06.819 --> 00:01:24.728 <v Wojciech Wegrzynski>Well, for that measurement to provide you with the information that you are seeking and this is the easiest one there will be more complex ones that we will be talking about in this episode because, as you can imagine, this episode is all about what do we measure in fires and how we do that. 00:01:25.250 --> 00:01:27.412 <v Wojciech Wegrzynski>I've invited a guest, dr David Morissett. 00:01:27.412 --> 00:01:29.174 <v Wojciech Wegrzynski>David has already been in the podcast. 00:01:29.174 --> 00:01:32.001 <v Wojciech Wegrzynski>We had a Far From the Moles episode together. 00:01:32.001 --> 00:01:45.233 <v Wojciech Wegrzynski>Young researcher from Queensland University, a brilliant mind with a good upcoming career in this space and also someone who's very passionate about fire experiments and loves to talk about it, which you will clearly see. 00:01:45.233 --> 00:01:59.665 <v Wojciech Wegrzynski>And this episode will give you a bit of guidance on how to interpret the data that you see in research papers, in journal papers, in test reports that you are supposed to use in your engineering. 00:01:59.665 --> 00:02:12.224 <v Wojciech Wegrzynski>We'll talk about temperature, we'll talk about flow, we'll talk about heat fluxes, we'll talk about heat release rate, mass loss rate all the useful stuff that you will find in various data sources that guide your decisions. 00:02:12.224 --> 00:02:26.883 <v Wojciech Wegrzynski>So I actually think this is a very important episode because measurements are things that not that many of us are carrying, but every single one of us is using in their engineering practice, and we need to understand how do they work. 00:02:26.943 --> 00:02:28.808 <v Wojciech Wegrzynski>That would be it for the introduction. 00:02:28.808 --> 00:02:30.070 <v Wojciech Wegrzynski>The episode is fun. 00:02:30.070 --> 00:02:30.832 <v Wojciech Wegrzynski>I promise that. 00:02:30.832 --> 00:02:32.223 <v Wojciech Wegrzynski>Stay with us. 00:02:32.223 --> 00:02:34.569 <v Wojciech Wegrzynski>Let's spin the intro and jump into the episode. 00:02:34.569 --> 00:02:40.901 <v Wojciech Wegrzynski>Welcome to the Firesize Show. 00:02:40.901 --> 00:02:44.324 <v Wojciech Wegrzynski>My name is Wojciech Wegrzyński and I will be your host. 00:02:44.324 --> 00:03:13.872 <v Wojciech Wegrzynski>The Firesize Show is into its third year of continued support from its sponsor, ofar Consultants, who are an independent, multi-award-winning fire engineering consultancy with a reputation for delivering innovative, safety-driven solutions. 00:03:13.872 --> 00:03:27.610 <v Wojciech Wegrzynski>As the UK-leading independent fire risk consultancy, ofar's globally established team have developed a reputation for preeminent fire engineering expertise, with colleagues working across the world to help protect people, property and the plant. 00:03:27.610 --> 00:03:43.706 <v Wojciech Wegrzynski>Established in the UK in 2016 as a startup business by two highly experienced fire engineering consultants, the business continues to grow at a phenomenal rate, with offices across the country in eight locations, from Edinburgh to Bath, and plans for future expansions. 00:03:43.706 --> 00:03:51.949 <v Wojciech Wegrzynski>If you're keen to find out more or join OFR Consultants during this exciting period of growth, visit their website at ofrconsultantscom. 00:03:51.949 --> 00:03:54.385 <v Wojciech Wegrzynski>And now back to the episode. 00:03:54.385 --> 00:03:59.491 <v Wojciech Wegrzynski>Hello everybody, I am joined today by Dr David Morrisset from Queenston University. 00:03:59.900 --> 00:04:03.867 <v David Morrisset>Hey, david, good to have you back in the podcast hey, wojciech, thanks for having me back, appreciate it. 00:04:04.661 --> 00:04:05.866 <v Wojciech Wegrzynski>Yeah, I appreciate it. 00:04:06.020 --> 00:04:19.769 <v Wojciech Wegrzynski>I always love to talk with a fellow FAR geek about FHIR geeky stuff, and today a topic that you have actually proposed to cover in the podcast is stuff that we measure in FAR experiments and how do we measure them? 00:04:20.279 --> 00:04:46.944 <v Wojciech Wegrzynski>Why, as a practitioner and someone who's like always misses a thermocouple location or always has trouble with some kind of data logger and converting files and messing with all this stuff, it brings me a lot of joy that there are others who consider this a significant scientific problem, are not really exposed that much to the world of measurements and the world of how we actually quantify stuff in our experiments. 00:04:46.944 --> 00:04:49.870 <v Wojciech Wegrzynski>A lot of people are blindly believing in experiments. 00:04:49.870 --> 00:04:51.543 <v Wojciech Wegrzynski>That's my experience. 00:04:51.543 --> 00:04:56.122 <v Wojciech Wegrzynski>A lot of people are taking experimental results and they just believe this is the truth. 00:04:56.122 --> 00:05:06.952 <v Wojciech Wegrzynski>You know the true truth While me, as an experimentalist, I know there's a lot of hardcore work to be done to have a good experiment and measurements are a big part of that. 00:05:06.952 --> 00:05:18.536 <v Wojciech Wegrzynski>So let's start with actually what is the act of measurement and how do you perform such an act in a scientific experiment. 00:05:18.939 --> 00:05:49.127 <v David Morrisset>I mean that's a great question and a great place to start right, Because in FHIR F, fire is a unique field of study where a lot of what we do ends up being experimental, whether that's proper novel experiments trying to understand physics or a lot of complexities in fire phenomena that we reduce to standardized test methods to just actually, instead of trying to assume something about a material or assembly, we actually light it up and see how it performs. 00:05:49.127 --> 00:05:53.942 <v David Morrisset>So all of these kinds of processes include the requirement for measurement. 00:05:53.942 --> 00:06:03.122 <v David Morrisset>So a question that I start to see as a relevant question the more I do experiments is not necessarily how do you make a measurement, but how do you make the right measurement? 00:06:03.122 --> 00:06:18.853 <v David Morrisset>The more time you spend in a fire lab, you see not only is there like a wide range of choices at your disposal to make a measurement, but there's also the more you see experiments, you get a little bit of the context behind what makes a good measurement a good measurement. 00:06:18.853 --> 00:06:26.213 <v David Morrisset>So you'll see, for example, you look at something like, let's say, like temperature is something that we try to measure all the time. 00:06:26.213 --> 00:06:30.803 <v David Morrisset>Now I can take a mercury thermometer, right, and I can measure temperature. 00:06:30.803 --> 00:06:34.329 <v David Morrisset>Right, you can measure the temperature of, say, a beaker of water? 00:06:34.329 --> 00:06:34.810 <v David Morrisset>Absolutely. 00:06:34.810 --> 00:06:42.057 <v David Morrisset>Now, does that mean that I should be sticking a mercury thermometer in the upper layer of a, you know, a smoke layer in a flashed over compartment? 00:06:42.057 --> 00:06:42.538 <v David Morrisset>Right? 00:06:42.538 --> 00:06:43.886 <v David Morrisset>Is that the right measurement to make? 00:06:43.886 --> 00:06:50.021 <v David Morrisset>How do we even interpret that, right? 00:06:50.021 --> 00:07:03.288 <v David Morrisset>But I mean in terms of, uh, yes, there are certain things that we can use to make measurements, uh, but I think it's it's actually the context behind why we use certain instruments that gives us the insight to the true value of a given measurement. 00:07:04.410 --> 00:07:09.096 <v David Morrisset>I think another element is also, measurements can be intrusive, right? 00:07:09.096 --> 00:07:15.708 <v David Morrisset>So if I put a pressure probe to measure flow, right, let's start talking about different things that we might measure. 00:07:15.708 --> 00:07:18.886 <v David Morrisset>For temperature, right, we might use a thermocouple. 00:07:18.886 --> 00:07:23.290 <v David Morrisset>If I want to look at flow fields in a fire experiment, I might use a pressure probe. 00:07:23.290 --> 00:07:28.869 <v David Morrisset>If I want to look at heat fluxes from a flame, I could use a heat flux gauge. 00:07:28.869 --> 00:07:35.264 <v David Morrisset>But all of these things are physical instruments that take up space. 00:07:35.264 --> 00:07:39.932 <v David Morrisset>Most of them are metallic in nature, right, and so they are. 00:07:40.194 --> 00:07:40.514 <v David Morrisset>Naturally. 00:07:40.514 --> 00:07:50.074 <v David Morrisset>If I stick a pressure probe into a duct that is on the same order of magnitude as the pressure probe itself, it's going to interrupt that flow right. 00:07:50.074 --> 00:07:52.242 <v David Morrisset>There's going to be an intrusive nature to that. 00:07:52.242 --> 00:08:12.867 <v David Morrisset>In the same way that if I'm taking a sample and I want to measure the heat flux acting on that sample, if the heat flux gauge that I'm using is physically a large portion of that surface, then that's going to start becoming intrusive and for some of your applications that matters and some of your applications that doesn't matter. 00:08:12.867 --> 00:08:18.310 <v David Morrisset>But I think some of the intricacy of trying to set up an experiment is figuring out. 00:08:18.310 --> 00:08:25.773 <v David Morrisset>How do you balance what we can practically do with what do we do to get the most value out of the experiment? 00:08:26.334 --> 00:08:29.689 <v Wojciech Wegrzynski>I mean, you've opened so many kinds of forms in this opening. 00:08:29.689 --> 00:08:32.403 <v Wojciech Wegrzynski>Perhaps we shouldn't have done this episode, but let's do this. 00:08:32.403 --> 00:08:36.990 <v Wojciech Wegrzynski>I mean yeah uh one experiments versus tests. 00:08:37.030 --> 00:08:48.548 <v Wojciech Wegrzynski>This is very valid and at some point the quality of your measurement becomes truly the measure of the fire properties of assemblies or materials that you're using like. 00:08:48.548 --> 00:08:58.561 <v Wojciech Wegrzynski>Take a plate thermometer measurements that guarantee that you had a standard time temperature relationship in your furnace test when assessing resistance to fire. 00:08:58.561 --> 00:09:17.177 <v Wojciech Wegrzynski>This this is fundamental and actually actually the story of the plate thermometer is exactly the story of reducing the uncertainty of a measurement and making sure that every exposure in a furnace in the world is more or less the same, because they were not at some point of time. 00:09:17.177 --> 00:09:18.826 <v Wojciech Wegrzynski>I think this is beautiful. 00:09:18.826 --> 00:09:26.724 <v Wojciech Wegrzynski>Perhaps we'll go back there, and I also love the question of how they actually match. 00:09:27.285 --> 00:09:32.287 <v Wojciech Wegrzynski>How do you get a perfect experimental setup, measurement setup for your experiment? 00:09:32.287 --> 00:09:48.331 <v Wojciech Wegrzynski>And I really like the school of Guillermo Reyn, because whenever I do experiments with him, he's never crazy about the number of measurements, because he also takes into account the capacity to analyze the outcomes. 00:09:48.331 --> 00:10:01.832 <v Wojciech Wegrzynski>One could say a better experiment is the one that has more measurements, but more measurements means more intrusion the thing that you just said and also means like you're going to spend so much more time processing data. 00:10:01.832 --> 00:10:03.154 <v Wojciech Wegrzynski>So much more time processing data. 00:10:03.154 --> 00:10:13.860 <v Wojciech Wegrzynski>I remember absolutely magnificent the toll building experiment in Edinburgh where they had this three-dimensional array of thermocouples in a large open plant compartment. 00:10:13.860 --> 00:10:15.927 <v Wojciech Wegrzynski>That was like 2,000 thermocouples. 00:10:15.927 --> 00:10:21.129 <v Wojciech Wegrzynski>There were like 30 YouTube videos of them setting up this experiment for a month. 00:10:21.129 --> 00:10:23.386 <v Wojciech Wegrzynski>Absolutely beautiful work. 00:10:23.386 --> 00:10:30.751 <v Wojciech Wegrzynski>But it took them years to publish the first paper because the amount of data, the amount of stuff to process was so insane. 00:10:31.440 --> 00:10:37.942 <v David Morrisset>So yeah, so many things but one thing, that sort of just off the building off some one of those ideas there, right is. 00:10:37.942 --> 00:10:46.462 <v David Morrisset>You mentioned things like the development of plate thermometers and different kinds of methods to measure something right, whether it's temperature, whether it's whatever. 00:10:46.462 --> 00:10:56.263 <v David Morrisset>But I think also, something that I think we lose sight of sometimes is to say okay, well, I'm using this device to measure temperature, therefore the output of this is temperature. 00:10:56.263 --> 00:10:59.600 <v David Morrisset>It is the temperature of my compartment, the temperature of my gas, and so on. 00:10:59.600 --> 00:11:18.268 <v David Morrisset>But we're always making an assumption because let's say, I just take a thermocouple, I take a plate thermometer, I take take a thermocouple, I take a plate thermometer, I take a few thermocouples, I put them in a compartment and I'm measuring the temperature right, even if each of these things are exposed to the same gas temperature at any given time, because there's fluctuations in the compartment and so on. 00:11:18.940 --> 00:11:28.702 <v David Morrisset>Let's say you use a 0.25 millimeter thermocouple bead versus a 1.5 mil thermocouple bead versus, you know, a large plate, and so on. 00:11:28.702 --> 00:11:31.610 <v David Morrisset>All these things are going to measure different outputs. 00:11:31.610 --> 00:11:34.549 <v David Morrisset>That doesn't mean the gas temperature is physically different. 00:11:34.549 --> 00:11:35.615 <v David Morrisset>If they're all you know. 00:11:35.615 --> 00:11:39.995 <v David Morrisset>Let's say, in a perfect scenario all these three devices should be measuring the same gas temperature. 00:11:39.995 --> 00:11:41.461 <v David Morrisset>As a function of time. 00:11:41.461 --> 00:11:42.482 <v David Morrisset>They won't. 00:11:42.482 --> 00:11:48.073 <v David Morrisset>That doesn't mean that, yeah, that's just a consequence of using different technology, right? 00:11:48.073 --> 00:11:55.048 <v David Morrisset>So all of these things have inherent benefits and, I guess, consequences, depending on what you choose to use. 00:11:55.048 --> 00:12:06.168 <v David Morrisset>Now it's up to us, as the engineers, to then interpret that and say are we truly measuring the gas temperature or are we just, we're just measuring the output of this thermocouple? 00:12:06.168 --> 00:12:10.216 <v David Morrisset>How close is that to what we assume is, say, the gas temperature? 00:12:10.740 --> 00:12:12.226 <v Wojciech Wegrzynski>I would rephrase what you've said. 00:12:12.226 --> 00:12:17.706 <v Wojciech Wegrzynski>The thermocouple shows you what the reading is and that's my interpretation. 00:12:17.706 --> 00:12:21.951 <v Wojciech Wegrzynski>It's not necessarily the temperature of the gas wall or whatever else. 00:12:21.951 --> 00:12:28.787 <v Wojciech Wegrzynski>It's just an outcome of a thermal equilibrium at which this device is with the environment that you're measuring. 00:12:28.787 --> 00:12:46.774 <v Wojciech Wegrzynski>So if you understand the heat transfer processes between those devices and the environment that you're measuring, you have a very good chance to understanding what kind of temperature of the medium that this device is exposed to will create this thermal equilibrium. 00:12:46.774 --> 00:12:50.431 <v Wojciech Wegrzynski>On the contrary, you don't really understand those properties. 00:12:50.759 --> 00:13:05.993 <v Wojciech Wegrzynski>You end up with a very rough measurement because, yes, the thermocouple could technically be at the temperature of the gases in your compartment If it was exposed long enough, if it was like very steady temperature, which is very unlikely in fires. 00:13:05.993 --> 00:13:09.269 <v Wojciech Wegrzynski>Fires are turbulent, fires are chaotic in their nature. 00:13:09.269 --> 00:13:15.791 <v Wojciech Wegrzynski>Fires are very complex in three dimensions, so it's very unlikely you will have the same temperature everywhere. 00:13:15.791 --> 00:13:25.870 <v Wojciech Wegrzynski>But at some point those are just minuscule details and noise in a measurement and sometimes they can lead to a severe misinterpretation of the results. 00:13:25.870 --> 00:13:38.321 <v Wojciech Wegrzynski>And that perhaps was one of the causes of the plate thermometer requirements for the furnaces, because those differences if you put a tiny, tiny thermocouple versus a plate thermometer. 00:13:38.321 --> 00:13:41.008 <v Wojciech Wegrzynski>They could have been huge in the furnace. 00:13:41.789 --> 00:13:48.203 <v David Morrisset>But also absolutely, I think, an element of that that's also really important is the context behind. 00:13:48.203 --> 00:13:49.226 <v David Morrisset>What are you interested in? 00:13:49.648 --> 00:13:51.392 <v David Morrisset>If you're interested at the scale of a compartment. 00:13:51.759 --> 00:14:02.642 <v David Morrisset>That's very different than if I'm trying to use some sort of device to measure the temperatures of a flame on the scale of a Bunsen burner right, and so all of a sudden those fluctuations matter, right. 00:14:02.642 --> 00:14:09.673 <v David Morrisset>All those things I think a big part, I guess, in terms of as an engineer or as someone using data. 00:14:09.673 --> 00:14:19.091 <v David Morrisset>I guess one of the biggest things that you learn when you're in a fire lab doing these experiments is everything is highly dependent on what you choose to make that measurement right. 00:14:19.091 --> 00:14:30.847 <v David Morrisset>And I really like what you said about basically a thermocouple isn't if you put it, you stick it in a compartment, it's not actually measuring the gas phase temperature, it's just measuring the temperature of the thermocouple. 00:14:30.847 --> 00:14:41.894 <v David Morrisset>Yeah Right, whatever that might be, and we can maybe get to a good example of that later too in some of my flame spread experiments where we use thermocouples and you got to, is it really the temperature of the solid? 00:14:41.894 --> 00:14:51.769 <v David Morrisset>Probably not right, and so we're making an assumption there, uh, and just being sort of honest with yourself, I think is a important part of that. 00:14:51.789 --> 00:14:52.712 <v Wojciech Wegrzynski>But that's like any, any measurement. 00:14:52.712 --> 00:14:57.225 <v Wojciech Wegrzynski>But perhaps a nice addition to this would be to explain the how the hell does the thermocouple work? 00:14:57.225 --> 00:15:03.184 <v Wojciech Wegrzynski>I'm not sure if everyone understands, uh, that on a high level yeah, yeah, sure, I mean. 00:15:03.426 --> 00:15:11.139 <v David Morrisset>The simplest explanation, right is a thermocouple is sort of the standard issue technique we use to measure temperatures and fire. 00:15:11.139 --> 00:15:18.354 <v David Morrisset>Yeah, and it's basically a joint of two dissimilar metals, so a wire of two dissimilar metals. 00:15:18.354 --> 00:15:29.572 <v David Morrisset>Depending on the type of the metals you'll get different types of thermocouples, um, I forget exactly what metals go into k-type thermocouples, um, but k-type nickel chromium. 00:15:29.572 --> 00:15:47.927 <v David Morrisset>There you go, nickel chromium, um, they give you a very specific kind of thermocouple and basically what that does is at that junction, the through the seed back effect, basically the temperature of that joint will induce naturally a voltage and we can correlate that voltage to temperature. 00:15:47.927 --> 00:16:08.365 <v David Morrisset>So basically, by measuring a voltage we're not measuring a temperature, we are measuring a voltage, a very small voltage, and then we correlate that to the temperature at the thermocouple joint, and so that effect is basically how we measure temperatures, and so that is our workhorse in fire safety for measuring temperatures. 00:16:08.466 --> 00:16:12.784 <v Wojciech Wegrzynski>I'll ask you a question that I had to answer so many times where exactly does it measure? 00:16:12.784 --> 00:16:14.148 <v Wojciech Wegrzynski>Does it measure at the point? 00:16:14.148 --> 00:16:21.216 <v Wojciech Wegrzynski>Does it measure at the length, where I mean thermocouple can be a long metal rod, where, where the rod measures? 00:16:21.878 --> 00:16:29.533 <v David Morrisset>yeah, yeah, so sometimes they look like a rod right, but basically it's all about where those wires are connected right and most well. 00:16:29.533 --> 00:16:36.861 <v David Morrisset>I'll say with almost almost every, every single thermocouple, it'll always be the tip, because it'll be where those joint, where that joint is made. 00:16:37.283 --> 00:16:46.812 <v Wojciech Wegrzynski>But it's all about where those wires come together exactly so it's about the connection between those two metals that creates this, this electrical current of some sort responding to temperature. 00:16:46.812 --> 00:16:52.753 <v Wojciech Wegrzynski>So this current can only appear at the connection point between those metals. 00:16:53.461 --> 00:16:54.446 <v David Morrisset>Yeah, absolutely. 00:16:54.446 --> 00:17:11.944 <v David Morrisset>And again, something that you could measure, something like boiling water within the precision of 0.1 degrees C. 00:17:13.221 --> 00:17:20.026 <v Wojciech Wegrzynski>I can give you an answer if you tell me the pressure to a very high degree based on fundamental physics. 00:17:20.441 --> 00:17:23.041 <v Wojciech Wegrzynski>That's how we calibrate them actually many times. 00:17:23.041 --> 00:17:30.839 <v Wojciech Wegrzynski>But yeah, in fact you're correct the bulkiness of the device, the way how it's connected, the type of the device, I mean. 00:17:30.839 --> 00:17:38.674 <v Wojciech Wegrzynski>On one hand, it feels kind of silly to discuss what the K-type thermocouple is in the podcast episode. 00:17:38.674 --> 00:17:49.570 <v Wojciech Wegrzynski>On the other hand, I've seen so many research papers where people would be using thermocouples that are meant to temperatures up to 400 degrees, for example, to measure fires. 00:17:49.570 --> 00:17:52.540 <v Wojciech Wegrzynski>That's a massive error in your instrumentation setup. 00:17:52.540 --> 00:17:54.023 <v Wojciech Wegrzynski>That leads you nowhere. 00:17:54.023 --> 00:18:02.226 <v Wojciech Wegrzynski>That will not allow you to create a useful experimental output that could be used by fire engineers one day. 00:18:02.226 --> 00:18:06.294 <v Wojciech Wegrzynski>I've seen them connected in the worst way. 00:18:06.294 --> 00:18:13.848 <v Wojciech Wegrzynski>I've seen the thermocouple part disconnected from the plate because someone did not understood that there's a thermocouple in the plate. 00:18:13.848 --> 00:18:15.280 <v Wojciech Wegrzynski>That needs to be Hell. 00:18:15.280 --> 00:18:28.711 <v Wojciech Wegrzynski>I've seen, you know the plate thermometers hanged up with the insulative layer targeting the open space and the, the steel plate, you know, back to the wall completely opposite way and it just happens. 00:18:28.750 --> 00:18:41.824 <v David Morrisset>So I think there's a there's a huge value in discussing this, yeah and I mean even things like I mean we all I know from talking to other experimentalists at conferences, but everyone's aware of these things, but you know you don't see it written down many places. 00:18:41.824 --> 00:18:59.695 <v David Morrisset>But things like you're measuring a voltage so electromagnetic interference can mess with your measurements, so things you don't even think about, but uh, I mean everything that is normal equipment would be sensitive to, for electromagnetic interference can completely mess up your thermocouple temperatures if you're not careful. 00:18:59.695 --> 00:19:07.792 <v David Morrisset>But yeah, just things like that, like that most people don't think about, are our important context behind how we might use these measurements. 00:19:07.792 --> 00:19:08.920 <v David Morrisset>Right, what? 00:19:09.240 --> 00:19:15.311 <v Wojciech Wegrzynski>other measurement techniques you would say are common in the fire related experiments. 00:19:15.311 --> 00:19:18.726 <v Wojciech Wegrzynski>You have name dropped some, so let's perhaps try to make a list. 00:19:19.347 --> 00:19:19.750 <v David Morrisset>Yeah, sure. 00:19:19.750 --> 00:19:21.708 <v David Morrisset>So let's just, let's rattle off a few, right? 00:19:21.708 --> 00:19:26.510 <v David Morrisset>So if you're trying to measure temperature, most people would use would default to a thermocouple. 00:19:26.731 --> 00:19:27.212 <v David Morrisset>I think that's fair. 00:19:27.212 --> 00:19:36.567 <v David Morrisset>If you want to measure a flow so let's say I want to measure velocities in a gas most people would default to using a pressure probe. 00:19:36.567 --> 00:19:47.267 <v David Morrisset>Right, and you know you can go back to the original work by McCaffrey coming up with this idea of the bidirectional probe, which was more robust than like a pitot tube, but in principle it's the same idea, right? 00:19:47.267 --> 00:19:55.301 <v David Morrisset>You have a probe that measures the difference in pressure over, basically, from a stagnation point in a flow and from that pressure. 00:19:55.301 --> 00:20:02.209 <v David Morrisset>If we simultaneously have a temperature measurement, we can then correlate that value to flow velocity. 00:20:02.209 --> 00:20:09.608 <v David Morrisset>Right, and you'll see these big, chunky steel pressure probes in many fire experiments. 00:20:10.099 --> 00:20:15.022 <v Wojciech Wegrzynski>I would rank them very high on the list of measurement devices that I hate. 00:20:15.022 --> 00:20:16.930 <v Wojciech Wegrzynski>There's going to be a second list after this episode. 00:20:16.930 --> 00:20:20.228 <v Wojciech Wegrzynski>All the types of measurements that I despise. 00:20:20.248 --> 00:20:20.897 <v David Morrisset>They're the worst. 00:20:21.701 --> 00:20:23.607 <v Wojciech Wegrzynski>They're so difficult in use. 00:20:24.059 --> 00:20:28.132 <v David Morrisset>But what's funny is the probe's not the problem, right? 00:20:28.132 --> 00:20:32.586 <v David Morrisset>It's because you need to hook up the probe to a pressure transducer, and those are always the problem, right? 00:20:33.181 --> 00:20:34.446 <v Wojciech Wegrzynski>So I know what you mean. 00:20:34.446 --> 00:20:42.490 <v Wojciech Wegrzynski>Actually, for us, the transducers, they were never the issue, because we have our good ones for flow-related experiments in the lab. 00:20:42.490 --> 00:20:54.086 <v Wojciech Wegrzynski>I find it more difficult because one it's sensitive to its orientation, so you have to be perpendicular to the flow and that means that you have to understand where the flow will be. 00:20:54.086 --> 00:21:06.666 <v Wojciech Wegrzynski>If you do not understand where the flow will occur, there is no way you can put a probe in there in the correct orientation and there's very little you can do to change the orientation of the probe when you are in the middle of the far experiment. 00:21:06.666 --> 00:21:15.929 <v Wojciech Wegrzynski>That's one thing, and the other thing is that because it's a pressure signal, the pressure signal travels through some little pipes with air. 00:21:15.929 --> 00:21:19.444 <v Wojciech Wegrzynski>Basically it's like, basically like for aquarium. 00:21:20.134 --> 00:21:25.046 <v Wojciech Wegrzynski>So those are those little nasty plastic pipes and the plastic pipes. 00:21:25.046 --> 00:21:27.601 <v Wojciech Wegrzynski>They don't really work that well with fire. 00:21:27.601 --> 00:21:33.219 <v Wojciech Wegrzynski>So if you have them in sensitive locations, welcome to the world of welding copper and God. 00:21:33.219 --> 00:21:41.407 <v Wojciech Wegrzynski>It becomes really annoying to to build an array of those devices, uh, for for a large scale of our experiment. 00:21:41.407 --> 00:21:45.520 <v Wojciech Wegrzynski>So that that's why they're high on my list of why I don't like them. 00:21:45.520 --> 00:22:02.513 <v Wojciech Wegrzynski>They require ridiculously a lot of work to set up and unless you've done the experiments once before and exactly understand the flow field in your experiment, then there's a good chance you're not going to get any useful output of those devices. 00:22:02.513 --> 00:22:07.613 <v Wojciech Wegrzynski>I wish there was a simple technique for measuring velocities, really like an optic. 00:22:07.613 --> 00:22:13.086 <v Wojciech Wegrzynski>Like you know, I know piv exists but no one, no one's saying, does pav in full-scale fires right? 00:22:13.467 --> 00:22:16.242 <v David Morrisset>so I wish there was we can talk about things like piv later, but, um, I think, yeah, no one I saying, does piv in full-scale fires, right? 00:22:16.242 --> 00:22:16.288 <v David Morrisset>So I wish there was. 00:22:16.288 --> 00:22:23.460 <v David Morrisset>We can talk about things like piv later, but, um, I think, yeah, no one, I mean it's, it's, I mean it's difficult just to say that that can be readily implemented across. 00:22:23.460 --> 00:22:26.996 <v David Morrisset>Yeah, you know, especially fire testing at length scales, but I think what? 00:22:27.195 --> 00:22:39.948 <v David Morrisset>What you mentioned about pressure probes, I think is an important one for the listeners to again have context behind some of these measurements where, if you're not exactly in line with the flow, right, you can get very misleading results. 00:22:39.948 --> 00:22:46.678 <v David Morrisset>So we again, if you're putting this in a doorway where you know that it's going to, you basically know what direction the flow is going, that's fine. 00:22:46.678 --> 00:22:54.146 <v David Morrisset>But if these are just placed in a compartment, it's very difficult to say with any certainty what flow you're actually observing. 00:22:54.146 --> 00:23:05.345 <v David Morrisset>Um, so it's another important point to remember about some of the limitations and and I guess again, context is the word I keep saying, uh, behind, behind these measurements let's go further. 00:23:05.526 --> 00:23:07.819 <v Wojciech Wegrzynski>You you've said heat flux gauges before. 00:23:07.980 --> 00:23:11.053 <v David Morrisset>Let's try those uh, yeah, so I mean heat flux gauges. 00:23:11.053 --> 00:23:11.855 <v David Morrisset>They're an interesting one. 00:23:11.855 --> 00:23:15.179 <v David Morrisset>There are different techniques by which you can measure a heat flux. 00:23:15.179 --> 00:23:23.549 <v David Morrisset>I think one of the most robust techniques that people will use are water-cooled heat flux gauges Exactly the reason why we hate them. 00:23:23.549 --> 00:23:24.631 <v David Morrisset>Yeah, yeah, exactly. 00:23:24.631 --> 00:23:28.577 <v David Morrisset>So you got a water coolant, which becomes a logistical nightmare. 00:23:28.577 --> 00:23:31.866 <v David Morrisset>But if you see this in the literature, right, that's the device being used. 00:23:31.866 --> 00:23:38.597 <v David Morrisset>That's the device being used Basically. 00:23:38.617 --> 00:23:47.527 <v David Morrisset>You have a slug of material, a metallic slug that basically, as you point in the direction of, say, some sort of radiating body, you can back out the heat flux at the surface of that radiometer, right, and you can get two different. 00:23:47.527 --> 00:23:49.278 <v David Morrisset>You can get different kinds of heat flux gauges. 00:23:49.278 --> 00:23:52.125 <v David Morrisset>You can get some that give you a total heat flux. 00:23:52.125 --> 00:23:58.875 <v David Morrisset>So, basically, what is both the combined radiative and convective heat transfer at the surface of that gauge? 00:23:58.875 --> 00:24:10.840 <v David Morrisset>Again, with the water cooling, it should mitigate certain elements of heat transfer, right, but you can also, if you want to completely remove convective heat transfer from the surface, you can add things like a sapphire window. 00:24:10.840 --> 00:24:28.487 <v David Morrisset>So then what you're getting is as close as we can get to a pure radiative boundary condition, and that allows you to just by choosing things like different kinds of heat flux gauges, you can sort of change what kind of heat flux conditions you want to focus on. 00:24:28.487 --> 00:24:37.063 <v David Morrisset>But that becomes of course difficult too, because then you need different gauges and they all need to be water cooled and that becomes a difficult process, of course. 00:24:37.083 --> 00:25:00.904 <v Wojciech Wegrzynski>course, but that's the sort of the tried and true default thing that people would go to for heat flux would be a water cooled gauge for for me that would be probably the one of the most useful measurements I could do in any large-scale fire experiment and at the same time one of the most difficult to actually get done, again, due to the logistical nightmare that it creates. 00:25:00.944 --> 00:25:03.318 <v Wojciech Wegrzynski>In the laboratory you need to get the water cooling. 00:25:03.318 --> 00:25:23.054 <v Wojciech Wegrzynski>So again, the plastic pipes or copper welding to get your water to the heat flux gauge you have like, basically the back side of it should be in non-fire exposed compartment which, uh, in some experiments your fire might switch from a compartment to compartment, for example. 00:25:23.054 --> 00:25:37.401 <v Wojciech Wegrzynski>So you're not really, unless you're willing to sacrifice them, which is not a great idea because they're extremely expensive from my perspective, and also if you just want to use them on some sort of structure to put it near the fire. 00:25:37.401 --> 00:25:57.817 <v Wojciech Wegrzynski>If you, for example, want to measure a heat flux one meter away from a solid structure, wrap it up in a lot of mineral wool, protect it. 00:25:57.817 --> 00:25:59.101 <v Wojciech Wegrzynski>Basically it has to. 00:25:59.101 --> 00:26:06.303 <v Wojciech Wegrzynski>It becomes, it suddenly becomes a really massive device that really can influence the flow field around of your structure. 00:26:06.303 --> 00:26:12.118 <v Wojciech Wegrzynski>So it's it's not easy to put a lot of them around and I think another approach you can. 00:26:12.259 --> 00:26:15.143 <v David Morrisset>You can again, like you said, these heat flux gauges are expensive. 00:26:15.143 --> 00:26:22.211 <v David Morrisset>So there have been, you know, other instruments developed like a thin skin calorimeter, right. 00:26:22.211 --> 00:26:23.780 <v David Morrisset>So tscs are used. 00:26:23.780 --> 00:26:40.047 <v David Morrisset>If you calibrate those and you calibrate them in the right conditions, they're basically a a small disc with a thermocouple on it and, in principle, if you can, if you can set it up in a configuration where you can calibrate them, they can give you a pretty good approximation of heat fluxes. 00:26:40.047 --> 00:26:40.509 <v David Morrisset>Right. 00:26:40.509 --> 00:26:44.787 <v David Morrisset>Again, there's quite a few more assumptions there than a water-cooled heat flux gauge. 00:26:44.787 --> 00:26:56.356 <v David Morrisset>But if you're looking at a large-scale test and you want different, high spatial resolution of heat flux measurements, a lot of people will lean towards those, just so they can get more information. 00:26:56.356 --> 00:27:02.298 <v David Morrisset>But still accepting the reality that we can't just put a thousand heat flux gauges, yeah, compartment test. 00:27:02.840 --> 00:27:11.589 <v Wojciech Wegrzynski>Arguably, if you accept the decrease in quality of your measurement and then some approximation to it, you could also live with plate thermometers. 00:27:11.589 --> 00:27:12.654 <v Wojciech Wegrzynski>That's my perspective. 00:27:12.654 --> 00:27:28.025 <v Wojciech Wegrzynski>You can get a lot from plate thermometers, especially if we're talking about very high temperatures, because if you're like, of course, if you're trying to get a minuscule change on a very small sample and plate thermometer, which is quite large device, it's not going to cut it. 00:27:28.025 --> 00:27:35.045 <v Wojciech Wegrzynski>But for compartment fires, large flames, this is not high for me, to be honest. 00:27:35.366 --> 00:27:37.775 <v David Morrisset>Sure, and I guess it all comes down to accepting. 00:27:37.775 --> 00:27:42.941 <v David Morrisset>What is the goal of that measurement, absolutely and understanding. 00:27:42.941 --> 00:27:44.282 <v David Morrisset>Am I choosing the right tool for that? 00:27:44.282 --> 00:27:45.746 <v David Morrisset>What are the impacts of that? 00:27:46.386 --> 00:27:51.884 <v Wojciech Wegrzynski>That's the struggle what I'm trying to measure and what I want to do with this measurement in the end. 00:27:51.884 --> 00:28:00.290 <v Wojciech Wegrzynski>If do with this measurement in the end, if I wanted to give to my fellow foreign engineers as some sort of reference, do I want to use this to calibrate my modeling? 00:28:00.290 --> 00:28:02.720 <v Wojciech Wegrzynski>Do I want to understand fundamentals of physics? 00:28:02.720 --> 00:28:05.288 <v Wojciech Wegrzynski>Do I want to understand the novel material? 00:28:05.288 --> 00:28:09.480 <v Wojciech Wegrzynski>Each will lead to different measurement setups, of course, yeah. 00:28:09.480 --> 00:28:11.826 <v Wojciech Wegrzynski>So, david, what's next on your list of measurements? 00:28:11.826 --> 00:28:13.578 <v Wojciech Wegrzynski>And I'll rank how annoying they are. 00:28:14.782 --> 00:28:15.846 <v David Morrisset>So up next we have. 00:28:15.846 --> 00:28:21.983 <v David Morrisset>Let's look, I mean, I guess something that we want to characterize quite frequently right is the burning rate of a solid. 00:28:22.526 --> 00:28:22.906 <v Wojciech Wegrzynski>Absolutely. 00:28:23.898 --> 00:28:33.597 <v David Morrisset>Solid, whether it's a piece of timber or whether it's a couch, right, yeah, and there's two sides of that coin, one being what is the heat release rate, what is the energy being released? 00:28:33.597 --> 00:28:47.729 <v David Morrisset>And we measure that typically through oxygen consumption, calorimetry, right, where, by measuring the depletion of oxygen in your effluent stream, we can then back out basically what is the amount of heat being released, right? 00:28:47.729 --> 00:28:52.503 <v David Morrisset>And, of course, there are corrections for CO generation, co2 generation, moisture, so on. 00:28:52.503 --> 00:28:55.623 <v David Morrisset>But that's the principle, right, and we've talked about heat release rates. 00:28:55.955 --> 00:29:04.555 <v Wojciech Wegrzynski>You know a lot on this on this podcast right, we should go deeper on that, because it's uh, it's easy to say you just measure oxygen concentration, but that's not everything. 00:29:04.555 --> 00:29:10.045 <v Wojciech Wegrzynski>Like if I want to burn a vehicle in my hood. 00:29:10.045 --> 00:29:12.588 <v Wojciech Wegrzynski>Now I'm thinking as an experimentalist. 00:29:12.588 --> 00:29:22.775 <v Wojciech Wegrzynski>You're my client, you come in and you say I have a vehicle to burn down and you have to give me the heat release rate, which I can't provide it because I do not have oxygen chlorometry in my large foot. 00:29:22.775 --> 00:29:26.221 <v Wojciech Wegrzynski>But if I had, there would be multiple challenges to that. 00:29:26.301 --> 00:29:37.428 <v Wojciech Wegrzynski>First of all, I really need to capture all of the smoke, which is not certain, like it's not 100% certain, that you will capture all of the smoke in the measurement. 00:29:37.428 --> 00:29:38.897 <v Wojciech Wegrzynski>That's number one. 00:29:38.897 --> 00:29:55.483 <v Wojciech Wegrzynski>If you look through a lot of videos from different calorimetry experiments performed, especially in vehicles, some of them were performed with ad hoc devices which lost a lot of smoke to the environment, which means you have not captured all of that, so you've not measured all of this precious oxygen depletion. 00:29:55.483 --> 00:30:12.883 <v Wojciech Wegrzynski>Second, I need to transport that in a reliable manner that allows me to establish how much time it takes from the fire to the measurement, which is not obvious, like it can be 20 seconds, it can be seven, depends on the flow field depends on temperature, densities and everything. 00:30:13.556 --> 00:30:40.378 <v Wojciech Wegrzynski>I have to understand the mass that is flowing through my pipe where I'm measuring the oxygen, which is not a straightforward measurement and it either includes a complicated ventilation ducting with some sort of pressure-based measurement devices or velocity probes and temperature measurements and then you can measure oxygen and compensate for your carbon monoxide, which is. 00:30:40.378 --> 00:30:50.884 <v Wojciech Wegrzynski>There are equations and this is arguably the easy part, but for me, from the laboratory perspective, the things that need to happen correctly, in the correct order to get to that point. 00:30:50.884 --> 00:31:00.079 <v Wojciech Wegrzynski>Even this is a madness, even though I find this measurement, that point, even this is a madness, even though I I find this measurement. 00:31:00.079 --> 00:31:15.724 <v Wojciech Wegrzynski>I would put ability to measure fires through oxygen calorimetry as perhaps top three things that ever happened for fire science, like seriously, like this drysdale's book and uh and fds top three for me absolutely well, let's, let's take one big step back to when we're talking about heat release rate. 00:31:15.765 --> 00:31:29.323 <v David Morrisset>right, because, like you said, it is one of the singular, most important measurements that we have been able to make as a fire scientist, and the most critical for fire engineers because they need it for the design fires. 00:31:29.423 --> 00:31:37.163 <v David Morrisset>Because everything, basically our framework of fire dynamics, has been set up in a way where the heat release rate becomes the input. 00:31:37.163 --> 00:31:43.969 <v David Morrisset>It becomes the heat pump, so to speak, in our system, where it's the thing generating the smoke. 00:31:43.969 --> 00:31:50.760 <v David Morrisset>It's the thing generating, you know, it's the input to your two zone model, it's the input to your ceiling jet correlations, it's the input to your FDS model. 00:31:50.760 --> 00:31:53.824 <v David Morrisset>Right, it becomes an essential part of the system. 00:31:53.824 --> 00:31:56.483 <v David Morrisset>But the measurement is inherently complex. 00:31:56.483 --> 00:31:57.858 <v David Morrisset>We see it all the time. 00:31:57.858 --> 00:31:59.724 <v David Morrisset>Every you know, every lab can do it. 00:31:59.724 --> 00:32:14.480 <v David Morrisset>But, like you were saying, something that isn't immediately apparent when you're reading these papers that show you the heat release rate is that you have to, literally in your mind's eye, imagine the smoke being released, you know, from this fire. 00:32:14.480 --> 00:32:17.766 <v David Morrisset>It has to travel into a duct, it needs to go through that duct. 00:32:17.766 --> 00:32:18.734 <v David Morrisset>It's mixing. 00:32:18.734 --> 00:32:26.661 <v David Morrisset>It goes quite a distance away from the fire through a duct, and then a tiny little pump will extract some of that little bit of smoke. 00:32:26.661 --> 00:32:41.307 <v David Morrisset>Take that through another duct into a gas analyzer, and that's where it'll actually and it'll cool it down, it'll pass it through things like desiccant, whatever, and through all this process then it'll give you an oxygen and CO and CO2 reading. 00:32:41.307 --> 00:32:53.457 <v David Morrisset>Now all that time, even if we're saying, okay, cool, now this tiny little sample, which I'm assuming is mixed and is representative, let's say that gives me the oxygen concentration In order to actually get oxygen depletion. 00:32:53.457 --> 00:32:57.200 <v David Morrisset>Another big assumption is I can measure the flow through that duct. 00:32:57.200 --> 00:33:10.710 <v David Morrisset>So again we run into all the issues that we just talked about, where I need to know the temperature and the pressure in that duct, and that comes up to all the issues and uncertainties that we have with thermocouples and pressure probes, because we need to be able to understand the flow. 00:33:10.710 --> 00:33:30.986 <v David Morrisset>So it turns out, if you really sort of look at the propagating uncertainties in heat release rate, one of the biggest, if not the biggest, is the uncertainty in knowing our flow rate in that duct Because actually the uncertainty of the oxygen measurement we have that down pretty well, but it's actually the flow rate that becomes a bit difficult. 00:33:30.986 --> 00:33:46.685 <v David Morrisset>There was a paper recently published in FSJ by the researchers at NIST looking at different ways to quantify the actual flow rates through large scale calorimeters and that becomes really interesting in trying to reduce those uncertainties. 00:33:46.685 --> 00:33:49.063 <v David Morrisset>But that becomes a big part of this equation. 00:33:49.063 --> 00:33:55.498 <v David Morrisset>But now let's say we have that data, let's say that we're confident in our flow rate and say we're confident in our oxygen reading, right. 00:33:55.878 --> 00:33:58.266 <v David Morrisset>Then comes another assumption, right? 00:33:58.266 --> 00:34:12.911 <v David Morrisset>The beauty of oxygen consumption calorimetry is that for most things that we burn, we can assume an energy constant of about 13.1 megajoules per kilogram per kilogram of oxygen consumed, precisely. 00:34:12.911 --> 00:34:20.286 <v David Morrisset>So that's this beautiful number where, if we know the kilograms of oxygen consumed, we can get energy release right. 00:34:20.286 --> 00:34:23.117 <v David Morrisset>But, like all things, right, that's still an assumption. 00:34:23.117 --> 00:34:25.465 <v David Morrisset>Now, for most things that works really nicely. 00:34:26.135 --> 00:34:38.438 <v David Morrisset>But there are lots of fuels that actually, if you look in the back of the SFB handbook, there's an entire table tabulated for energy constants for different materials, and most of them hover around 13.1. 00:34:38.438 --> 00:34:44.818 <v David Morrisset>But they're even for individual materials, like things that are phenolics, even some species of timber, right. 00:34:44.818 --> 00:34:51.516 <v David Morrisset>You start to deviate significantly 10, 20, 30% from those energy constants, right. 00:34:51.516 --> 00:35:03.842 <v David Morrisset>And so the principle of measuring the heat release rate means I can take any material, I can throw a car, I can throw an office space, I can throw a house underneath my hood and I can measure the heat release rate. 00:35:03.842 --> 00:35:19.523 <v David Morrisset>But I'm assuming that all the gases produced, all the pyrolysis process products being produced, are going to behave somewhere around that 13.1, right, if you're not specifically adjusting for it, which is the you know again, but it's something to think about. 00:35:19.523 --> 00:35:22.284 <v David Morrisset>It becomes another assumption in that list of assumptions. 00:35:23.215 --> 00:35:26.219 <v Wojciech Wegrzynski>Politically difficult question Would you apply that to a battery? 00:35:26.740 --> 00:35:27.563 <v David Morrisset>Oh, that's tricky. 00:35:27.563 --> 00:35:30.081 <v David Morrisset>I think we have a lot more. 00:35:30.081 --> 00:35:36.606 <v David Morrisset>I mean short answer is I think there's a lot more work to continue doing on that. 00:35:36.606 --> 00:35:55.601 <v David Morrisset>So we're doing battery testing at UQ and one thing that I think is continuously challenging right is to understand can we quantify enough the composition for any given particular manufacturer of battery or whatever, how it changes the state of charge, all these different variables. 00:35:55.601 --> 00:36:01.438 <v David Morrisset>Do we understand the effluent stream coming out of this well enough to say yes, without a shadow of a doubt? 00:36:01.699 --> 00:36:15.170 <v David Morrisset>We have the energy constants that we're assuming, and you know I'm not necessarily convinced that that's across the board yet, because the second you have say like, let's go to a simpler solid Instead of of a battery. 00:36:15.170 --> 00:36:20.724 <v David Morrisset>Let's look at polyoxymethylene, right, pom, which has oxygen and you know, embedded into its molecular structure. 00:36:20.724 --> 00:36:30.375 <v David Morrisset>We know that the energy constants for pom don't work out to be 13.1 right, because it in itself has oxygen embedded in it. 00:36:30.375 --> 00:36:40.681 <v David Morrisset>So any kind of material that would release oxygen in any capacity through any sort of chemical reaction, that's something you got to pay attention to, it's something you got to look at right. 00:36:40.681 --> 00:36:42.239 <v David Morrisset>And can we quantify that? 00:36:42.239 --> 00:36:50.688 <v David Morrisset>Can we account for that and not saying that we, you know that, whether it's batteries or any other technology, not saying that we can't do that. 00:36:50.688 --> 00:36:52.262 <v David Morrisset>We have the framework to do it. 00:36:52.262 --> 00:36:54.945 <v David Morrisset>I'm just not totally confident that. 00:36:55.548 --> 00:37:01.206 <v Wojciech Wegrzynski>Uh well, we're completely across every case, I would just say it's not your easiest straightforward measurement. 00:37:01.206 --> 00:37:08.929 <v Wojciech Wegrzynski>Just drop a battery in a cone and and have the number that it shows and then praise it as the true value, right? 00:37:08.929 --> 00:37:11.179 <v Wojciech Wegrzynski>That that's where I was leading, it's not that? 00:37:11.179 --> 00:37:12.161 <v Wojciech Wegrzynski>It's not that easy. 00:37:12.161 --> 00:37:14.487 <v Wojciech Wegrzynski>Um, let's talk another. 00:37:14.487 --> 00:37:35.217 <v Wojciech Wegrzynski>I mean, if I rank things about how annoying they are, I I mean, I don't know how annoying oxygen calorimetry is, because I don't have one and I would love one and I probably would be willing to sacrifice a lot of my comforts to have oxygen calorimetry, whereas I cannot say the same about the bidirectional probes, especially for obvious configurations. 00:37:35.217 --> 00:37:48.728 <v Wojciech Wegrzynski>But yeah, I think it's fundamental, especially if we are talking about research, if we're talking about science, if we are trying to use those experiments to build engineering on top of them. 00:37:48.728 --> 00:37:50.842 <v Wojciech Wegrzynski>Of course, not always possible. 00:37:50.842 --> 00:37:53.483 <v Wojciech Wegrzynski>There are other ways which we'll talk right now. 00:37:53.483 --> 00:37:55.458 <v Wojciech Wegrzynski>What are the other ways to measure fires, david? 00:37:55.458 --> 00:37:56.461 <v Wojciech Wegrzynski>The other ways which we'll talk right now. 00:37:56.481 --> 00:37:57.987 <v David Morrisset>What are the other ways to measure fires, david, the other ways? 00:37:57.987 --> 00:38:01.981 <v David Morrisset>And, and I think before I get into that, I want one more sort of comment on the battery in particular. 00:38:01.981 --> 00:38:07.164 <v David Morrisset>Yeah, this actually ties really well to the philosophy of measurement right where heat release rate is amazing. 00:38:07.164 --> 00:38:16.876 <v David Morrisset>Yet the default is let's try to measure the heat release rate of this thing right where I think there's for anyone who's seen videos of battery fires, right. 00:38:16.876 --> 00:38:29.936 <v David Morrisset>One thing that we know is when you see a battery fire, characteristically what you're seeing is fundamentally different than a hydrocarbon fuel, like a couch fire, a pallet fire. 00:38:29.936 --> 00:38:32.543 <v David Morrisset>You're seeing, you know, jet flames. 00:38:32.543 --> 00:38:36.722 <v David Morrisset>You're seeing really energetic, really rapid growth rates of fire growth. 00:38:36.722 --> 00:38:39.322 <v David Morrisset>So I find it something to think about. 00:38:39.322 --> 00:38:44.987 <v David Morrisset>Take a step back, and the first question we asked on the podcast is is this the right measurement? 00:38:44.987 --> 00:38:47.563 <v David Morrisset>Are we even comparing apples to apples here? 00:38:47.563 --> 00:38:56.764 <v David Morrisset>So I think understanding the consequences of a battery fire are different, and sure, we might be able to manifest that through the heat release rate. 00:38:56.764 --> 00:39:12.143 <v David Morrisset>But I think there's also more information there, because if everyone's saying that the most important reason we need the heat release rate is to use it in our engineering models, but are our engineering models, even validated for battery fires, because these flows are different, because these flames are different. 00:39:12.143 --> 00:39:17.865 <v David Morrisset>It's just something to think about, right, and I think there's a lot of exciting work to be done there. 00:39:17.865 --> 00:39:23.217 <v David Morrisset>But before we assume that this is the right measurement to make for every case right Is there? 00:39:23.217 --> 00:39:25.445 <v David Morrisset>What is the reason we're doing it right? 00:39:25.445 --> 00:39:29.146 <v David Morrisset>Does it actually capture the differences in the consequences produced? 00:39:29.146 --> 00:39:30.016 <v David Morrisset>Right, anyway. 00:39:30.416 --> 00:39:37.550 <v David Morrisset>But to move away from oxygen consumption, calorimetry, heat release rate, is one way to look at the burning rate. 00:39:37.550 --> 00:39:44.041 <v David Morrisset>The other I kind of say like the other side of that coin, is the rate at which you're losing mass, so the mass burning rate. 00:39:44.041 --> 00:39:49.327 <v David Morrisset>So they're kind of two sides of the same coin if you look at most common fuel packages. 00:39:49.327 --> 00:39:55.188 <v David Morrisset>So, basically, the rate at which you're losing mass, the rate at which you're let's assume that most of that is pyrolysis. 00:39:55.188 --> 00:39:56.840 <v David Morrisset>That's also an assumption, right? 00:39:56.840 --> 00:40:02.427 <v David Morrisset>But let's say that most of that's pyrolysis, then that is sort of the input to your flame, right? 00:40:02.427 --> 00:40:26.322 <v David Morrisset>Your pyrolysis gases feed that flame and there's a link, basically, you know, between your heat release rate should effectively be your mass loss rate times some heat of combustion, whether it's a true heat of combustion or an effective heat of combustion, some obviously that's based on combustion efficiency, but some sort of heat of combustion, right, and that kind of links the two ideas together. 00:40:26.322 --> 00:40:29.103 <v David Morrisset>But already by saying that I've made a few assumptions, right. 00:40:29.403 --> 00:40:40.956 <v David Morrisset>Yep, because if you actually put a solid on a load cell and you measure its mass loss over the duration of a fire, you're accounting for. 00:40:40.956 --> 00:40:43.001 <v David Morrisset>The bulk of that is pyrolysis, let's say, or the release of flammable gases in some way shape or form. 00:40:43.001 --> 00:40:46.978 <v David Morrisset>But it includes other things too, right, like even things like moisture loss. 00:40:46.978 --> 00:40:51.170 <v David Morrisset>You will be losing moisture in that effluent stream, right? 00:40:51.170 --> 00:41:00.782 <v David Morrisset>You'll be losing things like if you burn timber, eventually bits of ash will fall off right, bits of bits of timber might fall off right. 00:41:00.782 --> 00:41:06.498 <v David Morrisset>So all these little finite losses of discrete mass will be accounted for in that load cell. 00:41:06.498 --> 00:41:14.760 <v David Morrisset>But for most applications it's a pretty good surrogate for the rate of pyrolysis, right, which gives us a nice way of quantifying the burning rate. 00:41:14.760 --> 00:41:19.927 <v David Morrisset>We just need to sort of keep in mind if there are other large sources of mass loss. 00:41:19.927 --> 00:41:23.536 <v David Morrisset>Those will also be part of that measurement. 00:41:23.918 --> 00:41:28.130 <v Wojciech Wegrzynski>For me as a laboratory, this is a measurement that I default to. 00:41:28.130 --> 00:41:28.550 <v Wojciech Wegrzynski>Usually. 00:41:28.550 --> 00:41:32.320 <v Wojciech Wegrzynski>I really like doing mass loss rate measurements. 00:41:32.320 --> 00:41:35.887 <v Wojciech Wegrzynski>From my perspective, they're manageable to set up. 00:41:35.887 --> 00:41:42.469 <v Wojciech Wegrzynski>We are very commonly designing custom load cells for our clients. 00:41:42.469 --> 00:41:49.943 <v Wojciech Wegrzynski>We've designed like a scale on a scale so you can measure the crib inside of a building and the building structure independently. 00:41:49.943 --> 00:41:52.623 <v Wojciech Wegrzynski>We've measured hanging CLT slabs. 00:41:54.016 --> 00:41:57.226 <v Wojciech Wegrzynski>There's a lot of funny measurements you can do with master straight. 00:41:57.226 --> 00:42:29.088 <v Wojciech Wegrzynski>However, one thing that you really need to be aware of, and you kind of pointed to that if I want to measure a slab and know that over 20 minutes of a steady state fire where steady state is, of course, a very big word for a fire, but a fire that plateaued on some sort of size and did not really change that much over the course of that time, I had approximately five megawatts of heat release rate. 00:42:29.088 --> 00:42:32.902 <v Wojciech Wegrzynski>That's the type of information I can get from my mass loss rate system. 00:42:32.902 --> 00:42:46.643 <v Wojciech Wegrzynski>If you want a detailed analysis, second by second, of what was the heat release rate every second, that's a hell of a measurement to make and I know you've done research on that. 00:42:46.643 --> 00:42:48.623 <v Wojciech Wegrzynski>You've investigated that case. 00:42:48.623 --> 00:42:51.735 <v Wojciech Wegrzynski>How do we, what do we get out of that measurement? 00:42:51.735 --> 00:42:54.684 <v Wojciech Wegrzynski>Because at some stage it gets it becomes chaos. 00:42:55.315 --> 00:42:57.519 <v David Morrisset>So I think I mean I for the same reasons, get out of that measurement because at some stage it gets, it becomes chaos. 00:42:57.519 --> 00:42:58.375 <v David Morrisset>So I think I mean I, for the same reasons. 00:42:58.375 --> 00:43:18.878 <v David Morrisset>I like mass measurements in fire because, like you said, I've done mass measurements on the scale of a entire compartment, you know, on load cells, down to you know, a tga crucible full of, you know, milligrams of powder, right, and you can see some really interesting results across scales. 00:43:18.878 --> 00:43:19.400 <v David Morrisset>Right. 00:43:19.400 --> 00:43:23.067 <v David Morrisset>But like you said, it depends on what you're trying to look at. 00:43:23.467 --> 00:43:31.659 <v David Morrisset>So let's go back to the drawing board of what I defined the mass burning rate as like the opposite side of the same coin as heat release rate. 00:43:31.659 --> 00:43:39.782 <v David Morrisset>So, yes, you, you can use a mass loss measurement to then get to a heat release rate if that's the output you want. 00:43:39.782 --> 00:43:45.181 <v David Morrisset>But I also think it's really valuable using the mass loss rate for what it is right. 00:43:45.181 --> 00:43:56.246 <v David Morrisset>So in some of our work we will just keep it as the mass loss rate because, let's say, I'm trying to look at the change in the rate of pyrolysis, then the heat release rate is, you know, again, it's a surrogate. 00:43:56.246 --> 00:44:07.916 <v David Morrisset>But by measuring the heat release rate you're making again, you're making an assumption, that's the right surrogate, whereas if you measure the mass loss, you're actually making one less assumption and that's actually maybe the better measurement to look at. 00:44:08.436 --> 00:44:13.327 <v Wojciech Wegrzynski>And perhaps making a mental shortcut in here, because of the role of this podcast. 00:44:13.327 --> 00:44:23.039 <v Wojciech Wegrzynski>It connects to engineers and I see very little or even no applications of direct massless rate measurement in engineering daily. 00:44:23.039 --> 00:44:32.181 <v Wojciech Wegrzynski>I think it would be extremely difficult to apply, maybe in some really really specific projects, whereas heat release rate every single project. 00:44:33.978 --> 00:45:02.992 <v David Morrisset>That's why this is a connection in my brain that's very strongly enhanced over the years of practice I mean it's fair enough, right, and and you're right, but I also think it's worth the practicing engineers I know, like when I was, you know, when I worked before my phd I was working at an engineering firm too, right, and so, yep, having experience working in an engineering firm, I wish I I had known to look, if I looked at a paper and looking at the difference between this is actually looking at the rate of pyrolysis, right? 00:45:02.992 --> 00:45:10.465 <v David Morrisset>The mass loss rate has so many significant things because it has links to things like ignition theory, it has links to extinction theory. 00:45:10.465 --> 00:45:13.706 <v David Morrisset>So, looking at like, self-extinction of a timber compartment will not be determined by the release rate. 00:45:13.706 --> 00:45:15.594 <v David Morrisset>Fundamentally, they not be determined by the heat release rate. 00:45:15.594 --> 00:45:20.463 <v David Morrisset>Fundamentally, they'll be determined by the mass loss rate and things like that where, yes, those are subtleties. 00:45:20.463 --> 00:45:24.525 <v David Morrisset>Many engineers will not be applying that in the day-to-day work. 00:45:25.275 --> 00:45:44.742 <v David Morrisset>But I hope we can build tools like that in the future that require the level of understanding that there is a difference, right, and I think that's what we should be aspiring to as researchers to develop those tools that are based on the right physics and in a problem like an ignition problem or an extinction problem or those are just two I can think of off the top of my head. 00:45:44.742 --> 00:45:46.942 <v David Morrisset>Those are led by pyrolysis. 00:45:46.942 --> 00:45:50.740 <v David Morrisset>Another one actually come to think of it is charring, right? 00:45:50.740 --> 00:45:57.822 <v David Morrisset>So charring of, say, a timber slab, that's just a pyrolysis process, that's the progression of a pyrolysis fund. 00:45:57.822 --> 00:46:02.318 <v David Morrisset>So if you can measure the mass loss rate, you can get a surrogate for charring. 00:46:02.318 --> 00:46:16.235 <v David Morrisset>So, actually, you know, being able to decouple these ideas I think are really important, and if we want to develop engineering tools, they need to be based on the right physics, right. 00:46:16.235 --> 00:46:20.010 <v David Morrisset>So not to say that we have those tools widespread at the moment, but I do think it's worth having that, that distinction there absolutely, absolutely. 00:46:20.030 --> 00:46:23.641 <v Wojciech Wegrzynski>Let's talk about, perhaps, the averaging, because I know that you had papers on that. 00:46:23.641 --> 00:46:27.262 <v Wojciech Wegrzynski>I found them very interesting and they highly relate to the mass loss rate. 00:46:27.262 --> 00:46:34.141 <v Wojciech Wegrzynski>So, uh, maybe you can explain the problem and the solution and you have like five minutes yeah, perfect, all right. 00:46:34.202 --> 00:46:35.244 <v David Morrisset>Right, I got this Easy. 00:46:35.244 --> 00:46:49.878 <v David Morrisset>I mean, I'll ask anyone who's ever done a mass loss experiment or used mass loss data in fire science what is the one issue you always have with mass loss data, wojciech, if I want to get mass loss rates, what's your issue? 00:46:50.237 --> 00:46:57.788 <v Wojciech Wegrzynski>20 grams zero grams, 17 grams, five grams, 25 grams, 20 grams zero grams, 17 grams, five grams, 25 grams. 00:46:57.807 --> 00:46:58.750 <v David Morrisset>That's second by second Noise, right Noise. 00:46:58.750 --> 00:47:00.311 <v David Morrisset>So basically it looks like chaos. 00:47:00.652 --> 00:47:01.512 <v Wojciech Wegrzynski>Like the first. 00:47:01.512 --> 00:47:08.367 <v Wojciech Wegrzynski>It looks like chaos, yeah, the first time you open it and you plot the differences between the intervals, it's chaos. 00:47:09.135 --> 00:47:10.179 <v David Morrisset>Yes, totally. 00:47:10.179 --> 00:47:17.179 <v David Morrisset>So what I'll say is basically, for anyone who can't imagine this at the output of a, let's say, when you're measuring mass, you're measuring. 00:47:17.179 --> 00:47:23.739 <v David Morrisset>Let's say, I start with something that's 100 kilos and then by the end of the experiment it's 40 kilos, I don't know, but you're looking at a more of a continuous drop. 00:47:23.739 --> 00:47:25.164 <v David Morrisset>It's a very smooth line. 00:47:25.164 --> 00:47:30.775 <v David Morrisset>If you just look at the mass, generally speaking it's relatively smooth, depending on your load cells and so on, Exactly yeah. 00:47:31.860 --> 00:47:35.898 <v David Morrisset>But the second it becomes an absolute nightmare. 00:47:35.898 --> 00:47:38.197 <v David Morrisset>It's a scatterplot of just noise. 00:47:38.197 --> 00:47:38.998 <v David Morrisset>Basically. 00:47:38.998 --> 00:47:42.559 <v David Morrisset>Right, because most of these load cells operate. 00:47:42.559 --> 00:47:46.956 <v David Morrisset>Basically it's about the operation of the load cell, right, and some of it is our fault, let's be real. 00:47:46.956 --> 00:47:59.869 <v David Morrisset>But it's basically the load cells that we have are not optimized to measure a reasonable resolution between the timescales that we want them to is basically what it comes down to. 00:47:59.869 --> 00:48:10.818 <v David Morrisset>So in between these time scales you'll go from instantaneously losing five grams to zero grams to 10 grams to zero grams again to one gram to. 00:48:10.818 --> 00:48:12.661 <v David Morrisset>You know, it just becomes noisy. 00:48:13.222 --> 00:48:18.300 <v David Morrisset>So that's something that I sort of came across when I was doing my master's degree research. 00:48:18.300 --> 00:48:21.291 <v David Morrisset>Actually I had the chance to go over. 00:48:21.291 --> 00:48:33.681 <v David Morrisset>That's how I met the guys over at the University of Edinburgh where I ended up doing my PhD, and I was doing experiments on timber in the FPA and I was trying to process the mass loss data, because actually what we wanted was information on the pyrolysis rate. 00:48:33.681 --> 00:48:36.215 <v David Morrisset>We didn't want heat release rate, we wanted pyrolysis rate. 00:48:36.215 --> 00:48:42.902 <v David Morrisset>So we did mass right, and so we were looking at the mass loss rates and what I noticed was it was, it was a bit messy. 00:48:43.264 --> 00:48:46.655 <v David Morrisset>So I did what anybody would probably do, and I put a smoothing filter on it. 00:48:46.655 --> 00:48:51.173 <v David Morrisset>I put like a moving average to sort of, you know, smooth it out reasonable. 00:48:51.173 --> 00:48:53.219 <v David Morrisset>That's what I think the average person would do. 00:48:53.219 --> 00:49:09.967 <v David Morrisset>But something that I realized in processing that data was, if I use something like, let's say, for my data set at one Hertz, if I used a uh, a five point moving average versus a 25 point moving average, the 25 point moving average looked really nice, it looked smooth, it was, it was clean, um. 00:49:09.967 --> 00:49:15.911 <v David Morrisset>But if you put those two side by side and remove the raw data, they looked like fundamentally different curves. 00:49:15.911 --> 00:49:18.438 <v David Morrisset>So they came from the same data set, right? 00:49:19.019 --> 00:49:23.900 <v David Morrisset>But, they looked really different, and so I'm sure you've seen this too, right. 00:49:23.900 --> 00:49:37.130 <v David Morrisset>But if you just have these standard procedures where you take data, you extract the raw data, you blindly smooth it and you plot that figure without the context of what the raw data looked like, how do we know that that was the right decision, right? 00:49:37.130 --> 00:49:50.458 <v David Morrisset>So if I know that I'm going from my five point moving average to my 25 point moving average I forgot what it was, but I think it was like a 25% decrease in the peak mass loss rate Like that's a substantial change, right? 00:49:50.458 --> 00:49:56.898 <v David Morrisset>And not only are you things like for timber, you're going to have this nice, distinct peak mass loss rate. 00:49:56.898 --> 00:50:01.784 <v David Morrisset>So not only am I truncating that by smoothing it, I'm also moving it physically in time. 00:50:01.784 --> 00:50:13.657 <v David Morrisset>So now, all of a sudden, if I'm trying to use that as an input, whether it's as a pyrolysis rate or as a heat release rate I'm now I'm also decreasing my peak value, but I'm now moving it back in time. 00:50:13.657 --> 00:50:15.682 <v David Morrisset>So now it no longer represents reality. 00:50:17.005 --> 00:50:20.293 <v David Morrisset>And so we were sort of sitting there like, well, what's the right one to choose, right? 00:50:20.293 --> 00:50:41.157 <v David Morrisset>And so it became a bit of a you smooth it as you try to get a smooth curve out of it, right, because you can't just use scatter, you need to do something, right, we need to take this, you know, scatter, plot of points and turn it into a curve, but how do you do that reasonably and responsibly without shifting the reality, and so this? 00:50:41.157 --> 00:50:53.056 <v David Morrisset>We really sort of went in circles on this, and something I started thinking about was well, what's convenient, though, is, even though the mass loss rate is messy, the mass data is really continuous, right, so it's nice and clean it's. 00:50:53.056 --> 00:50:53.818 <v David Morrisset>It's pretty smooth. 00:50:53.818 --> 00:50:55.061 <v David Morrisset>It is in most cases. 00:50:55.061 --> 00:50:59.639 <v David Morrisset>So from the perspective of conservation of mass, right. 00:50:59.639 --> 00:51:00.762 <v David Morrisset>Right, there should be a link. 00:51:01.269 --> 00:51:05.041 <v David Morrisset>So let's say, I take a curve that I've smoothed with some filter. 00:51:05.041 --> 00:51:08.260 <v David Morrisset>Let's say I take a and for anyone following along, take a look at the paper. 00:51:08.260 --> 00:51:09.985 <v David Morrisset>I'm sure we can link it to the show notes, right? 00:51:09.985 --> 00:51:10.469 <v David Morrisset>Yeah, absolutely. 00:51:10.469 --> 00:51:31.483 <v David Morrisset>This walks through this in a little more detail, right, but this paper that we wrote shows, if you take the mass loss rate curve, let's assume some sort of smoothing, so I have a curve I can then at between any two points in time, let's say from time zero to time, you know T if I know how much mass I've lost on my load cell, right? 00:51:31.603 --> 00:51:32.371 <v David Morrisset>So let's say I've lost. 00:51:32.371 --> 00:51:33.855 <v David Morrisset>Between that time I've lost 10 grams. 00:51:33.855 --> 00:51:47.320 <v David Morrisset>I should be able to go to the mass loss rate curve and integrate the area underneath that curve and between those same two points in time it should give me the same value, because the integral of that should be the total mass that has been lost, right? 00:51:47.320 --> 00:52:00.565 <v David Morrisset>And what that allows you to do is I can then quickly crank out 10 different smooth curves that all look a bit different and then I can compare it. 00:52:00.565 --> 00:52:03.682 <v David Morrisset>I can just integrate the area under those curves, compare it to my mass data and then I can choose which curve actually represents the mass that was lost. 00:52:04.710 --> 00:52:06.222 <v Wojciech Wegrzynski>Integrate those smooth curves. 00:52:06.222 --> 00:52:08.077 <v Wojciech Wegrzynski>Integrate the smooth curve. 00:52:08.289 --> 00:52:08.650 <v David Morrisset>Exactly. 00:52:08.952 --> 00:52:10.931 <v Wojciech Wegrzynski>Wow, that's good, that's clever, that's clever. 00:52:11.454 --> 00:52:16.811 <v David Morrisset>So by integrating, so I can crank out 10 different smooth curves, right? 00:52:16.811 --> 00:52:24.581 <v David Morrisset>And then I can say, based upon the integration technique, I can tell you which one of those is the closest to reality, based upon the mass data. 00:52:24.581 --> 00:52:29.199 <v David Morrisset>And what's really nice about that is that kind of it's a nice closed loop solution, right? 00:52:29.199 --> 00:52:37.233 <v David Morrisset>So if we're smoothing heat release rate data or we're smoothing temperature data, there's really no, there's no way to sort of figure out what was the truth. 00:52:37.233 --> 00:52:40.822 <v David Morrisset>Right is the hard part, but with mass data you have it. 00:52:40.822 --> 00:52:53.637 <v David Morrisset>The mass measurement that you made is sort of the golden standard, so it allows us to circle back and be like can I actually check my work, can I actually determine, you know, is my curve realistic? 00:52:53.637 --> 00:53:01.958 <v David Morrisset>And so anyways, yeah, so for people listening, feel free to read the paper for more detail, but it allows us to close that loop. 00:53:02.079 --> 00:53:04.835 <v Wojciech Wegrzynski>Yeah, that's pretty good for a master student. 00:53:04.835 --> 00:53:06.380 <v Wojciech Wegrzynski>Have you considered the PhD? 00:53:07.885 --> 00:53:08.166 <v David Morrisset>I don't know. 00:53:08.166 --> 00:53:08.768 <v David Morrisset>I'm still thinking about it. 00:53:08.768 --> 00:53:09.190 <v David Morrisset>I've already checked. 00:53:09.931 --> 00:53:19.798 <v Wojciech Wegrzynski>Yeah, maybe one day let's actually talk about it, because I also know you had some funny and interesting ways of measuring stuff in your PhD which are very interesting. 00:53:19.798 --> 00:53:27.335 <v Wojciech Wegrzynski>I mean, on the list, we had a list of five points to get and we pretty much made through the first one and a half. 00:53:27.690 --> 00:53:29.829 <v David Morrisset>So there's going to be a follow-up. 00:53:29.829 --> 00:53:32.351 <v David Morrisset>Sometime, I think there'll have to be a part two for EJEC. 00:53:32.351 --> 00:53:33.074 <v David Morrisset>I think that's it. 00:53:33.878 --> 00:53:40.514 <v Wojciech Wegrzynski>Absolutely Rank the future measurement devices on how much we want them and how realistic they are. 00:53:40.514 --> 00:53:56.217 <v Wojciech Wegrzynski>That would be fun, but let's maybe finish it up with some nice stories from measuring the PMMAs and doing those fine-tuned measurements in blocks of materials in which you can't really measure that well because you're altering too much. 00:53:56.217 --> 00:54:03.163 <v Wojciech Wegrzynski>So I know there was a lot of interesting problems and a lot of interesting solutions, so so perhaps give me some of those cool. 00:54:03.302 --> 00:54:09.260 <v David Morrisset>So I'll rattle off a few just for you know to, for the sake of this time for the listeners. 00:54:09.260 --> 00:54:17.693 <v David Morrisset>But, um, let's go back to three different phenomena that we were interested in, right, like temperature, flow and heat flux. 00:54:17.693 --> 00:54:18.998 <v David Morrisset>So we'll start with flow. 00:54:18.998 --> 00:54:29.163 <v David Morrisset>We couldn't put pressure probes in my PMMA samples when we were doing flame spread, because I was looking at flame spread over things on the order of like 200 mil by 50 mil, right. 00:54:29.163 --> 00:54:32.920 <v David Morrisset>So you can't instrument that with pressure probes because they become intrusive. 00:54:33.429 --> 00:54:40.494 <v David Morrisset>But there was a series of experiments that we did, looking at flame spread over spherical slobs of PMMA, so like spheres of PMMA. 00:54:40.494 --> 00:54:48.119 <v David Morrisset>There were some really funky experiments, but there were some really cool flow physics going on, and so we were looking at downward flame spread over spheres. 00:54:48.119 --> 00:54:58.574 <v David Morrisset>But what was really cool was we also simultaneously took high-speed images and what we noticed with PMMA was PMMA was giving off these discrete little. 00:54:58.574 --> 00:55:08.800 <v David Morrisset>We called it like packets, like fuel packets, fuel parcels, little bits of fuel that were ejected from the surface and then they would react through the flow around the sphere. 00:55:08.800 --> 00:55:20.376 <v David Morrisset>So what we did is we took high speed images and we tracked these little eddies all the way around the sphere and we were able to get basically 2D resolved flow measurements right by tracking these eddies in time. 00:55:20.469 --> 00:55:27.436 <v David Morrisset>We basically did very cheap you know PIV around the sphere and we were able to make flow measurements. 00:55:27.436 --> 00:55:34.117 <v David Morrisset>So you had to spatially correct for them, you had to de-warp the image, you had to do all sorts of stuff Again for these optical measurements. 00:55:34.117 --> 00:55:36.539 <v David Morrisset>Maybe we have to have another podcast episode on. 00:55:36.539 --> 00:55:39.179 <v David Morrisset>I wonder if this is scalable to a large scale. 00:55:40.552 --> 00:55:45.317 <v Wojciech Wegrzynski>I always wondered if you could actually, because what's a flame? 00:55:45.317 --> 00:55:50.278 <v Wojciech Wegrzynski>When you look into a very large flame, it's a bunch of vortices that grow. 00:55:50.278 --> 00:56:03.775 <v Wojciech Wegrzynski>If you have a slow motion video, you can pick up a vortice that happens at the bottom of the flame and observe how it travels across the flame and grows and eventually turns into smoke and flows. 00:56:03.775 --> 00:56:25.639 <v Wojciech Wegrzynski>Further away, but perhaps using some convolutional neural networks, you could maybe put some sort of origin point or central point of the vortice and use that as some sort of tracking uh, you know marker, because unless you have particles to trace, there's not much you can. 00:56:25.639 --> 00:56:26.902 <v Wojciech Wegrzynski>You can click to that. 00:56:26.922 --> 00:56:33.228 <v David Morrisset>That's my dream of future yeah, I mean I was lucky that my flames were pretty small in that context, right. 00:56:33.228 --> 00:56:37.730 <v David Morrisset>So scaling that up, there's complexity, but I'm sure if we, I'm sure we can figure out some way to do that right. 00:56:37.730 --> 00:56:46.818 <v David Morrisset>But then in terms of other measurements, right, like in terms of heat fluxes, this was one where we talked about this quite a bit, actually in the last podcast that I was on, for flame spread. 00:56:46.818 --> 00:57:05.277 <v David Morrisset>But we I was inspired in my phd by a paper by uh ito and kashawagi from the 80s where they used this technique called holographic interferometry to basically get it's a fancy way of getting 2D temperature measurements through the depth of the solid and with that information you can basically solve the heat flow through the solid. 00:57:05.277 --> 00:57:18.222 <v David Morrisset>So I did really high resolution temperature measurements with thermocouples through the solid and because what I couldn't do is I couldn't put a heat flux gauge in my solid, because again it will be like the width of my sample basically. 00:57:19.289 --> 00:57:23.260 <v David Morrisset>So I couldn't use that to measure the heat flux, but I wanted to know how much heat flux is coming from the flame. 00:57:23.260 --> 00:57:39.161 <v David Morrisset>So by measuring the temperature through the solid, you can then basically resolve the temperature field through that solid to which we can back out what is the heat flux at the surface, how much heat flow is coming from from the gas to the solid, how much heat is flowing through the solid itself. 00:57:39.161 --> 00:57:59.757 <v David Morrisset>So actually mapping out those heat fluxes through the solid, but all it took was temperature measurements, I mean a lot of temperature measurements and and a lot of uh calculation, uh, but we were able to actually look at the heat flux using something besides a heat flux gauge and a well well-defined fuel package and a great understanding of basic physics. 00:57:59.858 --> 00:58:01.652 <v Wojciech Wegrzynski>But, uh, I saw, I saw a concept. 00:58:01.652 --> 00:58:20.740 <v Wojciech Wegrzynski>I think I'm not I'm not sure if it originated from edinburgh or from queenston, but they named it a instrumented brick and it was something like you've described, like a piece of material equipped with thermocouples that you just used to, you know, figure out the heat flux from the temperature profile inside the material. 00:58:20.740 --> 00:58:22.835 <v Wojciech Wegrzynski>I really like this idea, yeah. 00:58:23.378 --> 00:58:34.559 <v David Morrisset>Yeah, totally right, Because it allows you to elegantly sort of solve all the heat flow and that really allows you to sort of look at this from first principles and say you know what is my actual heat flux. 00:58:35.692 --> 00:58:45.956 <v David Morrisset>And then the last measurement I'll tease and I think this will be a good transition into, hopefully, I guess, a part two sometime is for temperature. 00:58:45.956 --> 00:58:46.818 <v David Morrisset>Temperature is in the depth of the solid. 00:58:46.818 --> 00:58:47.822 <v David Morrisset>I was using thermocouples. 00:58:47.822 --> 00:58:50.211 <v David Morrisset>That was fine, right, but I also really needed the surface temperature. 00:58:50.211 --> 00:59:03.239 <v David Morrisset>I really needed that as part of my model, and so the best I could do at the beginning was basically taking thermocouples and trying to fuse it to the surface of the solid, and the thing is, that was okay. 00:59:03.298 --> 00:59:07.193 <v David Morrisset>But again we get into the questions that we started this, this podcast, with. 00:59:07.193 --> 00:59:09.159 <v David Morrisset>Is is what am I actually measuring? 00:59:09.159 --> 00:59:10.682 <v David Morrisset>Am I measuring the gas? 00:59:10.682 --> 00:59:12.152 <v David Morrisset>Am I measuring the solid? 00:59:12.152 --> 00:59:13.458 <v David Morrisset>Am I, am I actually? 00:59:13.458 --> 00:59:14.661 <v David Morrisset>Is this a reasonable measurement? 00:59:14.661 --> 00:59:37.422 <v David Morrisset>And so I I really started running into some problems there, and so one of the techniques that I sort of stumbled across, that I started using, was a laser diagnostics based technique called phosphor thermometry, which allowed us to get really highly accurate, highly resolved spatial temperature measurements of the surface using optical techniques. 00:59:38.065 --> 00:59:46.617 <v David Morrisset>And so, again, I don't think we have time to really go into that, but sort of just to say across those three phenomena, right, whether it's heat fluxes, flow or temperature. 00:59:46.617 --> 00:59:48.510 <v David Morrisset>There are other ways to do it right. 00:59:48.510 --> 00:59:57.456 <v David Morrisset>These are not the sort of approaches that I took in my PhD, were kind of non-traditional, but just sort of as a closing thought, right. 00:59:57.456 --> 01:00:03.998 <v David Morrisset>The ability to measure these phenomena are not limited to a thermocouple, a pressure probe and a heat flux gauge. 01:00:03.998 --> 01:00:08.436 <v David Morrisset>Right, it's about making the right measurement that gives us the physics that we want. 01:00:08.436 --> 01:00:20.833 <v David Morrisset>It's about choosing the capabilities that are most realistic and the thing that gives us what we actually want to explore versus what is our sort of default approach to, uh, to measurements absolutely. 01:00:21.215 --> 01:00:21.635 <v Wojciech Wegrzynski>I love this. 01:00:21.635 --> 01:00:25.067 <v Wojciech Wegrzynski>This will be a follow-up and for now, let's stop. 01:00:25.067 --> 01:00:33.717 <v Wojciech Wegrzynski>We also didn't talk about visual measurements, which are also very interesting, and a very, very growing field of of measurements in in laboratories. 01:00:33.717 --> 01:00:36.106 <v Wojciech Wegrzynski>There's a whole story to be told as well. 01:00:36.106 --> 01:00:44.420 <v Wojciech Wegrzynski>I mean, uh, yeah, it's surprisingly how uh, how much fun you can have talking about the thermocouple measurements in FHIR. 01:00:44.420 --> 01:00:50.099 <v Wojciech Wegrzynski>I've enjoyed this story and I hope you did as well and I hope the listeners did as well, absolutely. 01:00:50.815 --> 01:00:53.875 <v David Morrisset>No, that was a great conversation, so I'm looking forward to part two. 01:00:54.416 --> 01:00:54.797 <v Wojciech Wegrzynski>And that's it. 01:00:54.797 --> 01:00:55.559 <v Wojciech Wegrzynski>Thank you for listening. 01:00:55.559 --> 01:01:01.021 <v Wojciech Wegrzynski>Surprisingly, a lot of things go into measuring surprisingly simple things right. 01:01:01.021 --> 01:01:22.164 <v Wojciech Wegrzynski>My trust towards measurements is, let's say, questionable when I'm doing them myself, understanding all the uncertainties Not that the thermocouple is going to show me a wrong value, but is it the right thermocouple in the right location, in the right time averaging scheme, in the right, correct data logger? 01:01:22.164 --> 01:01:26.960 <v Wojciech Wegrzynski>That's all the uncertainties that scare me when I'm doing my measurements. 01:01:26.960 --> 01:01:34.862 <v Wojciech Wegrzynski>And also those are the uncertainties that scare me when I have to use somebody else's measurements in my fire safety engineering. 01:01:35.422 --> 01:01:45.802 <v Wojciech Wegrzynski>It's a trap for fire safety engineers that sometimes you are requested to incorporate directly experimental data in your fire safety engineering project. 01:01:45.802 --> 01:02:04.373 <v Wojciech Wegrzynski>I mean, usually that's the best way to do, but there are challenges in that and to some extent you need to be able to really understand what the researchers have done to truly, truly use, to get the most out of the data that you receive from them. 01:02:04.373 --> 01:02:18.820 <v Wojciech Wegrzynski>And that's the point of this episode to show you all the complications that go into good measurements and perhaps guide you towards stuff that you should be looking for when you are reading those scientific papers and research reports. 01:02:18.820 --> 01:02:29.184 <v Wojciech Wegrzynski>The more you get into standardized fire testing, the better the measurements become in terms of how well defined they are, how exactly well placed they are. 01:02:29.184 --> 01:02:36.514 <v Wojciech Wegrzynski>At this point you stop having the issues of whether I place the thermocouple correctly, because the standard defines it. 01:02:36.514 --> 01:02:47.563 <v Wojciech Wegrzynski>However, a new challenge arises, like with the bs814 facade stand, for example, where you can just build stuff up around the thermocouples. 01:02:47.563 --> 01:02:48.391 <v Wojciech Wegrzynski>You know where they are. 01:02:48.391 --> 01:02:55.994 <v Wojciech Wegrzynski>You can put stuff in there and perhaps game the system a little bit to for the thermocouples to show lower value. 01:02:55.994 --> 01:03:00.864 <v Wojciech Wegrzynski>This problems as well if you understand the word of measurement too well. 01:03:01.471 --> 01:03:08.275 <v Wojciech Wegrzynski>Anyway, we still have like three points to go through our list of talking points with David, so there will be a follow-up. 01:03:08.275 --> 01:03:32.427 <v Wojciech Wegrzynski>I am really hyped about talking about the future measurements, because what we've talked today are things that have been done in forest science since 1970s, 80s, and there are things that are coming up that are the measurement technologies of the future, more remote, more clean, more simple, perhaps covering a bigger area of your sample, not just the point, giving you a better insight into the fire physics. 01:03:32.427 --> 01:03:34.532 <v Wojciech Wegrzynski>Anyway, that would be it for today's episode. 01:03:34.532 --> 01:03:40.568 <v Wojciech Wegrzynski>Thank you for being here with me next week, slovenia, european Symposium on 5 safety science. 01:03:40.568 --> 01:03:46.829 <v Wojciech Wegrzynski>I hope to meet some of you there it's gonna be fun and, yeah, see you here in the podcast next Wednesday. 01:03:46.829 --> 01:04:00.519 <v Wojciech Wegrzynski>Cheers, bye, bye.

216 - What do we measure and how? with David Morrisset

Available Results

Quick Summary

Deep Knowledge Extract

Book Recommendations

Knowledge Blocks

Quotes and Claims

Resources Mentioned

Quick Summary

Deep Summary

Knowledge Blocks

Topic Map

Quick Summary

Deep Summary

Knowledge Blocks

Topic Map

Health Tips

Business Ideas

Book Recommendations

Methods & Frameworks

Risks and Caveats

Action Items

Quotes and Claims

Resources Mentioned

Deep Knowledge Extract

Extract Knowledge

Transcript

Chapters