162 - Experiments that changed fire science pt. 9 - Jin's experiment on visibility in smoke

Transcript

WEBVTT

00:00:00.502 --> 00:00:02.307
Hello everybody, welcome to the Fire Science Show.

00:00:02.307 --> 00:00:09.867
In this episode we're going back to the series Experiments that have changed the fire science, and I'm here just alone.

00:00:09.867 --> 00:00:31.655
I'm not going to be talking about an experiment of my own, but an experiment that I truly believe fulfills the definition in the title of the series, an experiment that really has changed the fire science and gave us one of the most powerful tenability criterion that is used all around the world to design buildings every day.

00:00:31.655 --> 00:00:33.281
That is, the visibility in smoke.

00:00:33.281 --> 00:00:40.182
And the experiments that I'm going to talk about are the experiments by Jin in Japan from 1970s.

00:00:40.182 --> 00:00:57.073
It's quite funny to think that a single experiment carried out in 1970 could have such an impact over the built environment and, in all honesty, I don't think Gene realized how powerful thing he has created.

00:00:57.073 --> 00:01:03.173
The story is a little bit longer, so I'll save that for the episode itself.

00:01:03.173 --> 00:01:09.143
Before I start, I'll tell you my relation with this research and visibility in smoke.

00:01:09.143 --> 00:01:18.778
Of course, I've not been around when Gene was doing his experiments I was not even born then but I know his work fairly well because of my professional interests.

00:01:18.778 --> 00:01:25.588
Visibility in smoke is something that I also like to research, and obviously his research is a cornerstone of whatever we're doing.

00:01:25.588 --> 00:01:34.066
So it's always been interesting to me to study his work and the access to his work is to some extent limited.

00:01:34.186 --> 00:01:36.992
We have a very good seminal paper from IFSS.

00:01:36.992 --> 00:01:43.894
Gene has received Howard Emmons award from the IFSS and gave a plenary lecture in 1997.

00:01:43.894 --> 00:01:49.668
Award from the IFSS and gave a plenary lecture in 1997.

00:01:49.668 --> 00:01:51.653
So that's where a lot of knowledge in English is stored.

00:01:51.653 --> 00:02:01.545
A lot of people know his very valuable chapter in the SFP handbook and up to the third edition of the handbook it was the original chapter written by Jin and for me that was also an interesting source of knowledge.

00:02:01.545 --> 00:02:20.870
And there's five papers actually there's six papers, but there's five numbered papers on visibility in smoke, published in Japanese, the original papers from the original experiments, and I actually had them professionally translated to English to study them in detail because they were so influential to me.

00:02:21.431 --> 00:02:26.689
In this episode we're going to talk about the contents of the papers one and two, which cover the visibility in smoke experiments.

00:02:26.689 --> 00:02:32.224
The other ones drift more into the human behavior in smoke, almost as much interesting.

00:02:32.224 --> 00:02:34.330
But let's narrow it down a little bit.

00:02:34.330 --> 00:02:50.431
I've also always felt that the model, how we're using it, has some limitations, has some issues with it, which led to a grand proposal, something that we're doing together with Lucas Arnold, and you've heard many, many, many episodes ago in the podcast.

00:02:50.431 --> 00:02:58.812
We're redoing the GINS experiments and doing some experiments of our own to hopefully build a better understanding of how visibility in smoke works.

00:02:58.812 --> 00:03:19.330
Perhaps one day, when that work is finished, I'll be able to proudly sit in here and make an episode on experiments that changed fire science on my own, but at this point I'm happy to tell you all about the genes research and how we got the model of visibility in smoke that we are using today to design buildings.

00:03:19.330 --> 00:03:22.128
So let's spin the intro and jump into the episode.

00:03:29.879 --> 00:03:30.439
Welcome to the Firesize Show.

00:03:30.439 --> 00:03:31.921
My name is Wojciech Wigrzyński and I will be your host.

00:03:31.921 --> 00:03:51.437
This podcast is brought to you in collaboration with OFR Consultants.

00:03:51.437 --> 00:03:55.403
Ofr is the UK's leading fire risk consultancy.

00:03:55.403 --> 00:04:05.213
Its globally established team has developed a reputation for preeminent fire engineering expertise, with colleagues working across the world to help protect people, property and environment.

00:04:05.213 --> 00:04:21.012
Established in the UK in 2016 as a startup business of two highly experienced fire engineering consultants, the business has grown phenomenally in just seven years, with offices across the country in seven locations, from Edinburgh to Bath, and now employing more than 100 professionals.

00:04:21.012 --> 00:04:32.685
Colleagues are on a mission to continually explore the challenges that FHIR creates for clients and society, applying the best research experience and diligence for effective, tailored FHIR safety solutions.

00:04:32.685 --> 00:04:48.267
In 2024, ofr will grow its team once more and is always keen to hear from industry professionals who would like to collaborate on FHIR safety futures this year, get in touch at oafrconsultantscom and we're back.

00:04:48.387 --> 00:04:50.091
Let's go and talk about visibility in this work.

00:04:50.091 --> 00:04:56.562
I think a nice thing to start with is a timeline of fire safety engineering and where the genes research places in that timeline.

00:04:56.562 --> 00:05:07.680
I think it gives a very, very interesting context to the work he has carried, context to the work he has carried.

00:05:07.680 --> 00:05:10.185
So, if you think about it, the jeans research was performed in 1970.

00:05:10.185 --> 00:05:14.494
And it was preceded by experiments carried out in buildings around mid 1960s.

00:05:14.494 --> 00:05:20.612
This means that the research that he was doing was not truly related to modeling fires.

00:05:20.612 --> 00:05:24.310
I mean at that point the modeling of fires has not existed yet.

00:05:24.310 --> 00:05:35.343
The models that existed were models that connected the area of opening or the amount of air entering a compartment to the fully grown fire size.

00:05:35.343 --> 00:05:39.331
So that's definitely Kalgoorlie and Thomas in Europe.

00:05:39.331 --> 00:05:43.846
These are the very first developments in modelling fires.

00:05:43.846 --> 00:05:54.553
The zone model has not existed yet at that point, so there definitely were no engineering tools to actually benefit from using visibility in smoke.

00:05:54.553 --> 00:06:05.095
So his research was really focused on finding the visibility as a thing that defines safety for humans, not modeling at all.

00:06:05.095 --> 00:06:06.766
It all came later on.

00:06:06.766 --> 00:06:13.213
It all came when we developed zone models to allow us to model fires and use them in engineering.

00:06:13.680 --> 00:06:18.692
Then in early 90s, the first instances of using CFD in modeling fires.

00:06:18.692 --> 00:06:24.870
This is where his model was applied and where it was used and started its brilliant career.

00:06:24.870 --> 00:06:39.331
Because for CFD modeling the visibility became the number one credibility criterion, basically the first one that is passed, the first one to reach its critical value and because of that is the one that dictates design you design for visibility in buildings right now.

00:06:39.331 --> 00:06:45.930
To be able to use his model we also needed some advancements in our understanding of smoke.

00:06:45.930 --> 00:06:51.235
So here comes Mulholland and his specific extinction coefficient for smoke.

00:06:51.235 --> 00:06:57.889
That allowed us to quite easily calculate the extinction coefficient we need in Jin's model actually.

00:06:57.889 --> 00:06:59.613
So that's it.

00:07:00.040 --> 00:07:02.865
And later on there also came more advancements.

00:07:02.865 --> 00:07:21.392
There were multiple attempts in defining visibility in different terms than JIN, but despite them being very scientifically interesting, they have not succeeded in replacing the JIN's model, which is still used every day by every fire safety engineer worldwide, pretty much.

00:07:21.392 --> 00:07:40.194
If you think about technology, j Gene also did not have access to a lot of modern technology that we would have today the densitometers today, based on lasers that has not existed in 1970, the display technologies that we have today they did not exist in 1970.

00:07:40.194 --> 00:07:48.557
He's also referring to reflecting and light emitting signs, which obviously were based on completely different technologies in the 1970s.

00:07:48.557 --> 00:08:00.552
So a lot of things have changed, a lot of things have went much far ahead, but still the best experiment we have on visibility in smoke is actually this one from the 1970s.

00:08:00.552 --> 00:08:07.362
I like to tell about this timeline because it makes people think it's not that someone was sitting down.

00:08:07.362 --> 00:08:12.432
Oh, how do we model the visibility distance in numerical modeling?

00:08:12.432 --> 00:08:13.274
Let's solve this.

00:08:13.274 --> 00:08:33.553
But it was an honest attempt on actually defining the critical distance at which you can see a shape and connect that with the brightness and the smoke extinction to better understand how people perish in fires, how fires are dangerous to people, and perhaps improve that.

00:08:33.553 --> 00:08:39.072
It's interesting that it turned into a model that now shapes the built environment.

00:08:39.072 --> 00:08:49.047
So now that we have the timeline clear, let's talk about rationale and I'm not going to interpret what Jin was thinking about doing in 1970s.

00:08:49.047 --> 00:08:54.527
I'm going to read from his seminal paper from the IFSS, because he's giving that in the introduction.

00:08:54.527 --> 00:08:56.552
So let me read it out for you.

00:08:56.572 --> 00:09:03.152
In Japan since 1960s there has been an increasing number of people killed by smoke in fire resistant buildings.

00:09:03.152 --> 00:09:12.731
This trend has been recognized by many people concerned with the fire safety science and technology, as well as fire safety authorities, and has emphasized the importance of the research in fire smoke.

00:09:12.731 --> 00:09:18.832
Many full-scale experiments were conducted in Tokyo metropolitan area in the years around 1965.

00:09:18.832 --> 00:09:23.971
I and my staff took part in those experiments, especially those measuring smoke density.

00:09:23.971 --> 00:09:29.011
They were the beginning of my research on the interaction between the human behavior and fire smoke.

00:09:29.011 --> 00:09:33.851
So this is the rationale as Jin presents it in his seminal paper.

00:09:33.851 --> 00:09:37.770
He also writes at first visibility in fire smoke was our major concern.

00:09:37.770 --> 00:09:42.030
However, in due course I became interested in human behavior in fire smoke.

00:09:42.460 --> 00:09:48.943
The research in human behavior was mainly conducted by interviews, questionnaires sent to evacuees from hotel and department store fires.

00:09:48.943 --> 00:10:10.101
So this is also very important to me, because the very first research that June did was based on the visibility itself, the perception of the sign, on the importance of the environmental conditions at which the sign is observed, on the fact if you can or you cannot read the sign.

00:10:10.101 --> 00:10:17.225
The sign Now.

00:10:17.225 --> 00:10:19.129
This was just a small part of his research, pretty much a very initial part of his research.

00:10:19.129 --> 00:10:27.461
Later on this research turned into research on visibility in smoke in terms of how does it influence the human behavior in fires or the human evacuation process.

00:10:27.461 --> 00:10:40.577
In the further research he's looked into things like decrease in movement speeds in smoke or the impact of irritating or non-irritating smoke on the cognitive capacities.

00:10:40.577 --> 00:10:50.197
They've been doing experiments in which they've assessed at which smoke concentration people have difficulty solving mathematical problems, for example.

00:10:50.197 --> 00:10:51.750
Very interesting research.

00:10:51.750 --> 00:10:56.268
But all this followed and all of this happened after 1970.

00:10:56.268 --> 00:11:02.198
It was 1970 that he did the first research on visibility in smoke and that was it.

00:11:02.198 --> 00:11:07.065
That was the foundation of the visibility in smoke model that we are using today.

00:11:07.625 --> 00:11:21.998
So I'm going to walk you through the experiments, but perhaps let's first try to distill what's the discovery in here, what really is the important thing the gene has realized.

00:11:21.998 --> 00:11:32.937
So, in simplest terms, gene's experiments consisted of a box compartment, a small compartment with the dimensions of 5.5 by 1.2 by 1.2 meter.

00:11:32.937 --> 00:11:46.770
At one end of that compartment there was a target which was a circle illuminated by light and on the other end of the compartment there was a person observing that circle, that shape, from behind the window.

00:11:46.770 --> 00:11:57.070
The person had control over light that was illuminating the shape and there obviously was smoke in the chamber between the observer and the target.

00:11:57.070 --> 00:12:02.265
There was a densitometer in the chamber that allowed to measure the smoke extinction.

00:12:02.265 --> 00:12:14.721
At the time the measurements were taken and the observer was reducing the illumination of the target, which was a circle, up to a point where they stopped actually seeing that.

00:12:14.923 --> 00:12:23.350
So, basically, the observer has reached critical contrast between the sign and the background at which you're not longer able to see the sign.

00:12:23.350 --> 00:12:35.078
This is why the visibility in a Jeans equation is named visibility of signs at the obscuration threshold, because it really was that that was the threshold at which the sign could be seen.

00:12:35.078 --> 00:13:19.493
Now, having those measurements, so having the distance, which was the distance between the observer and the target, having the measurement of smoke density and also knowing the illumination of the sign, he started plotting the relationships between extinction coefficient and that visibility distance at which the observer could not see the sign anymore, and he realized that there's a linear relation, or he approximated, there was a linear relation between those two that actually held quite well at different ranges of extinctions, that also held quite well for different illumination conditions.

00:13:19.493 --> 00:13:23.408
And that's when he realized you could actually correlate those two properties.

00:13:23.408 --> 00:13:45.951
So that product of extinction coefficient and the distance at which you had reached the critical visibility of the shape later became something that we know today as the k-factor, the number that defines the conditions we use, value of three for smoke reflecting signs, number eight for smoke illuminating signs.

00:13:45.951 --> 00:13:50.990
But that's pretty much the product of the extinction coefficient and the critical visibility distance.

00:13:50.990 --> 00:13:57.110
Now it's interesting because if you go into the SFP handbook you'll see that it's not three or eight.

00:13:57.110 --> 00:14:00.457
It ranges from two to four, from five to 10.

00:14:00.457 --> 00:14:07.868
If you go through his original research you see that there's a scatter of points that give those correlations.

00:14:07.868 --> 00:14:25.229
But like with many magical numbers, we've lost track of that and I think from convenience we settled down on the discrete values of three and eight as our defaults because it's just convenient to have one number instead of range.

00:14:25.229 --> 00:14:27.735
So if you think about it.

00:14:27.735 --> 00:14:36.948
So he put the line through a scatter of points, he said that it's a relation that holds and then it led to building a model that is this naive?

00:14:36.948 --> 00:14:40.015
I wouldn't say Gene was naive, I would say we are naive.

00:14:40.557 --> 00:14:53.417
Gene has identified and researched multiple things that influence this relation, that influence this visibility, this perception of a sign, that it's critical visibility threshold.

00:14:53.417 --> 00:15:01.058
So first he investigated the impact of the brightness of the background versus the luminance of the sign.

00:15:01.058 --> 00:15:08.784
So this is important because it's not just the sign that matters but the environment in which the sign is placed.

00:15:08.784 --> 00:15:15.541
The lighting conditions within the room at large will be impactful as well.

00:15:15.541 --> 00:15:20.572
So he definitely realized that and he also studied that in detail.

00:15:20.572 --> 00:15:24.927
Gene was also very well aware that we have different types of smoke.

00:15:24.927 --> 00:15:33.875
That's also something that he has researched and he understood that some types of combustion will yield a different type of smoke.

00:15:34.716 --> 00:15:41.133
Actually, let me go back to his seminal paper from 1997 and read you another part which highlights that.

00:15:41.133 --> 00:15:51.504
So Jin writes In the range of visibility from 5 to 15 meters, the product of visibility at the obscuration threshold and the smoke density is almost constant.

00:15:51.504 --> 00:15:56.913
The visibility in black smoke is somewhat better than in the white smoke of the same density.

00:15:56.913 --> 00:16:03.456
This remarkable difference in visibility is not recognized among smokes from various materials For reflecting signs.

00:16:03.456 --> 00:16:07.493
The product of visibility and smoke density is also almost constant too.

00:16:07.493 --> 00:16:12.032
The product depends mainly on the reflectance of the sign and the brightness of illuminating sign.

00:16:12.032 --> 00:16:31.770
The visibility at the obscuration threshold is found to be and here we get the relation that you take the k factor, which in Jin's model the original ones was a range subdivided by smoke density or extinction coefficient of smoke, and you receive the visibility.

00:16:32.774 --> 00:16:33.839
But he has not stopped there.

00:16:33.839 --> 00:16:41.879
He has not just defined that the type of a smoke or the type of combustion actually will influence the visibility in smoke.

00:16:41.879 --> 00:16:48.297
He studied that in depth In his 1971 paper, that is the paper named Part 2,.

00:16:48.297 --> 00:16:54.817
He writes the intensity of illumination light in smoke depends on the amount of light scattered by the smoke.

00:16:54.817 --> 00:17:02.914
Therefore the amount of scattered light was investigated for various building materials and smoke was produced under various conditions.

00:17:02.914 --> 00:17:13.137
A little bit later in the paper he continues in the case of smoldering smoke, the amount of light scattered is almost the same for wood and various plastic materials.

00:17:13.137 --> 00:17:24.897
On the other hand, the amount of light scattered in the case of flame-burning smoke is not so much affected by heating temperature but is significantly affected by the type of building material and the amount of air supplied.

00:17:24.897 --> 00:17:34.386
So he realized that the burning conditions will heavily change the smoke properties and this is also reflected by the contents of the paper.

00:17:34.867 --> 00:17:43.157
This difference between the scattering smoke the white smoke and light-absorbing smoke the black smoke, was so impactful.

00:17:43.157 --> 00:17:53.676
In his equations you actually have a factor it's named small k, not to be confused with the large k, which is a product of the visibility distance and the smoke extinction coefficient.

00:17:53.676 --> 00:18:03.297
The small k is a ratio between the total scattering coefficient and the extinction coefficient for the smoke, and this is something that actually was used to derive the first correlation.

00:18:03.297 --> 00:18:15.337
So he definitely realized that different types of smoke will produce different visibility and in different conditions you will have a different product of that relation.

00:18:15.337 --> 00:18:22.798
What I find kind of interesting is that we've never actually used that way of thinking.

00:18:22.798 --> 00:18:26.664
We've never actually took that further from the genes perspective.

00:18:26.664 --> 00:18:36.035
The way how we incorporated the differences in smoke in our modeling studies, in our engineering, goes from a different way.

00:18:36.035 --> 00:18:39.253
It goes from the way how modeling is performed itself.

00:18:39.253 --> 00:18:55.154
So in modeling today, in CFD modeling today, the way how you would calculate the visibility in smoke is by measuring the mass density of smoke in a given volume in your model, in your CFD model.

00:18:55.154 --> 00:19:01.953
So you know how many grams you have of smoke in the cell that you're investigating.

00:19:02.615 --> 00:19:15.757
Then in early 2000s, mulholland from NIST has done a brilliant attempt on integrating the knowledge in the extinction coefficients of smoke, specific extinction coefficients of smoke.

00:19:15.757 --> 00:19:22.346
So he's found the values that correlate the extinction coefficient to a mass of smoke pretty much.

00:19:22.346 --> 00:19:32.592
And from that we have two values one which is commonly used for the smoldering smoke, one that is commonly used for the product of flaming combustion.

00:19:32.592 --> 00:19:41.194
If I'm not wrong, the value for the flaming smoke was 8.7 square meters per gram and the smoldering was something like 4.3 square meters per gram.

00:19:41.194 --> 00:19:43.271
I hope I didn't make a mistake here.

00:19:43.271 --> 00:19:46.595
Please double check and correct me if I'm wrong.

00:19:46.595 --> 00:19:53.576
And with this you can correlate Like I have one gram of smoke, how much of extinction coefficient is that?

00:19:54.285 --> 00:20:01.769
Now, another driver that we use to control that in our CFDs is by controlling how much smoke you have in those cells.

00:20:01.769 --> 00:20:06.977
And you do that by defining the size of your fire, the heat release rate of your fire.

00:20:06.977 --> 00:20:12.472
You define the heat of combustion, so you know how many kilograms of mass have just turned into energy you just released.

00:20:12.472 --> 00:20:19.076
And then you define the suitability of your fuel, so how much mass of the fuel goes into the smoke, how much goes into your other products.

00:20:19.076 --> 00:20:37.528
By defining how much grams of smoke you release, you basically drive how much smoke there will be in your elements of your model and by using a specific extinction coefficient for smoldering or flaming combustion, you eventually receive the extinction coefficient that, multiplied by visibility, is the K factor.

00:20:37.528 --> 00:20:39.653
So you're back to the genes equation.

00:20:39.653 --> 00:20:46.958
That's how we combine those technologies together, those models together, to get an engineering tool that we use every day.

00:20:46.958 --> 00:20:55.089
However, as I said, the original findings of GIN, how different smokes would have different reflecting and absorbing properties.

00:20:55.089 --> 00:20:58.857
This is not something very well recognized in our everyday modeling.

00:20:59.444 --> 00:21:15.251
Jin has also realized the importance of the contrast between the observed sign, the observed shape and the background, so basically the conditions at which the sign is observed, the evacuation sign is observed.

00:21:15.251 --> 00:21:23.398
He has plotted multiple plots for different external light conditions, from 20-20 lux to 180 lux.

00:21:23.398 --> 00:21:46.275
He was able to define the conditions in his chamber because it was filled with 24 10-watt fluorescent lamps that were installed inside as an equipment so he could change the amount of lamps that are shining, meaning that he changed the conditions in the compartment at which the observations were performed and, of course, the person who was observing.

00:21:46.275 --> 00:21:54.516
They had the ability to change the light intensity of the evacuation sign up to the point where they could not see the sign anymore.

00:21:54.516 --> 00:21:57.907
Evacuation sign up to the point where they could not see the sign anymore.

00:21:57.907 --> 00:22:15.517
So you could actually measure the ratio between the background and the sign itself in a way and that's actually what Jin realized that you get different visibilities or different extinction coefficients for the sign at which the sign is visible, and the darker it is, the easier it was to see the sign.

00:22:15.517 --> 00:22:36.413
But if you actually collapsed those relations by dividing the background intensity by the sign intensity, you would get one line and it would hold very well, and he basically later used that relation to limit the impact of the conditions at which the sign was observed.

00:22:36.913 --> 00:22:57.148
I honestly believe that it's an interesting observation that we've never, ever used in fire safety engineering that the conditions at which the sign is observed will heavily, heavily influence the visibility in smoke, if you think about it, the visibility is carried in an emergency situation usually at the power shutdown in the building.

00:22:57.148 --> 00:23:04.032
I'm not sure how it works in your countries, but in Poland the minimum light intensity by evacuation lighting is just one lux.

00:23:04.032 --> 00:23:06.858
The lowest Jin had was 22 lux.

00:23:06.858 --> 00:23:22.458
So your space is very, very dark at which the evacuation takes place, and the darker it is, the easier it is to see the illuminated signs because you obviously have bigger contrast between the sign and the exterior.

00:23:22.458 --> 00:23:34.510
However, this is absolutely not reflected in our engineering because that would require you to use a different K factor that would correlate the K with the background properties.

00:23:34.510 --> 00:23:40.762
So it's kind of funny that when doing this research, a different k factor that would correlate the k with the background properties.

00:23:40.762 --> 00:23:44.648
So it's kind of funny that when doing this research, it seemed very clear to him that it is an important variable.

00:23:44.648 --> 00:23:49.077
However, us today, using his model, we're not really recognizing this at all.

00:23:49.746 --> 00:23:58.173
Now let's try and touch on the experiment itself and how some choices for the experiment has driven the development of the model later on.

00:23:58.173 --> 00:24:17.737
So one thing I really find interesting and I think it was lost somewhere either in the translation or in passing forward this knowledge Jin made his observers observe a shape that was a circle with diameter from 5 to 15 centimeters.

00:24:17.737 --> 00:24:22.345
The observers, as I mentioned previously, had the ability to change the illumination.

00:24:22.345 --> 00:24:35.634
So the shape was placed on a smoked window and there was an illumination device behind the window which you could alter how strongly it illuminates the sign.

00:24:35.634 --> 00:24:42.726
This is how the observers were looking for the critical threshold at which the sign is no longer seen.

00:24:43.487 --> 00:24:52.730
The interesting thing is that observations were carried from distances of 5, 5.5, 10.5, and 15.5 meters.

00:24:52.730 --> 00:24:58.488
So if you see the gene's plots, the visibility distances are discrete rather than continuous.

00:24:58.488 --> 00:25:08.058
In his research that's because he had people at very specific distances from the target and the distance was made by using mirrors.

00:25:08.058 --> 00:25:09.845
So he had mirrors in his chamber.

00:25:09.845 --> 00:25:22.708
You could view the target directly and that would be 5.5 meter distance from the target, or you could view the target through mirrors and this would increase the distance at which the target would be seen.

00:25:23.308 --> 00:25:32.320
An interesting thing that that was lost or perhaps not realized by people is that the viewing angle was constant in his experiments.

00:25:32.320 --> 00:25:40.596
That's actually quite interesting because obviously if you move back from the sign, the further you're away from a sign, the smaller the sign will be.

00:25:40.596 --> 00:25:55.109
In case of Jin's research, the target the circular target was 5 centimeters for the 5-meter distance, 10 centimeters for 10.5-meter distance and 15 centimeters for 15.5-meter distance meter distance and 15 centimeters for 15 and a half meter distance.

00:25:55.109 --> 00:25:59.694
So he kept the viewing angle constant to limit that as a variable in the study.

00:25:59.694 --> 00:26:15.234
It kind of makes sense because his study was very, very much oriented on the impact of external conditions on the differences between the background and the illumination of the sign, so this was chosen as a constant.

00:26:15.234 --> 00:26:24.268
However, it has significant consequences to modern engineering Because if you think about it, if there's a larger sign it's much better visible than a smaller sign.

00:26:24.268 --> 00:26:30.807
I felt that this is a lackluster in the visibility in smoke model that we don't include for that in our engineering design.

00:26:30.807 --> 00:26:41.262
I cannot today use the visibility in smoke model in my CFD to justify use of a different size of an evacuation science to make my evacuation better.

00:26:41.262 --> 00:26:46.401
I think it's an important thing that we're losing from the model because that's an obvious consequence.

00:26:46.401 --> 00:27:04.826
If you have not enough visibility in smoke, perhaps instead of putting another FON and a 1 million euro installation to remove more smoke from your compartment, perhaps you can install better signage that will simply be visible and you get back again above the visibility threshold.

00:27:04.826 --> 00:27:11.045
So I think this is something that we did not realize and we've kind of lost.

00:27:12.106 --> 00:27:20.923
Another detail from Gene's experiment is related to the types of smoke, so he was using a filter paper.

00:27:20.923 --> 00:27:24.230
It's named Toyo furnace paper, number two.

00:27:24.230 --> 00:27:26.722
I have no access to Toyo paper.

00:27:26.722 --> 00:27:31.431
If you can find me some I'll be very happy, because I'm trying to repeat Jin's experiments.

00:27:31.431 --> 00:27:41.257
Anyway, he was burning five to 10 grams of completely dry paper that was bent to a length of approximately 10 centimeters.

00:27:41.257 --> 00:27:58.269
He mentions that he's using paper because it was somehow more reputable than wood and he was placing it in a furnace where he was heating the paper up to 400 degrees at specific oxygen concentrations and with that he was releasing the smoke.

00:27:58.269 --> 00:28:08.234
In some paper he mentions that by driving the oxygen concentration he could change the properties of smoke from smoldering smoke to flaming smoke, black smoke.

00:28:08.234 --> 00:28:14.383
He mentioned that at low oxygen concentrations the smoke was more white, at high it was more black.

00:28:14.383 --> 00:28:17.277
It's also quite profound Later on.

00:28:17.277 --> 00:28:30.102
This is something we know today as equivalency factor, and there's a whole area of fire science and toxicology of smoke that uses similar relations between the oxygen concentration and the properties of smoke.

00:28:30.102 --> 00:28:34.419
Jin has just casually observed that and noticed that he's using that.

00:28:34.920 --> 00:28:40.519
Another thing that I want to tell you about his measurements was what he was actually able to measure.

00:28:40.519 --> 00:28:42.442
I think that's very interesting.

00:28:42.442 --> 00:28:44.586
Again, we're talking 1970s.

00:28:44.586 --> 00:28:57.223
So it's not that you had access to all the fancy measurement tools that we have in today, but still he was able to capture a lot of details in those experiments.

00:28:57.223 --> 00:29:01.977
So obviously he was able to measure smoke extinction coefficient.

00:29:01.977 --> 00:29:09.811
Smoke extinction coefficient is actually a very convenient measurement you can carry and we do that in laboratories.

00:29:09.811 --> 00:29:21.702
We do that pretty much on a daily basis Because it's a product of the light intensity light intensity where there is no smoke and light intensity where there is smoke present.

00:29:21.702 --> 00:29:29.138
So if you measure that, the ratio between those, if you put that into equations, gives you the extinction coefficient.

00:29:29.138 --> 00:29:35.717
Basically you can interpret it as how much light did the smoke take?

00:29:35.717 --> 00:29:51.412
Now the problem with that, or a small problem with that, is that it captures the merged effects of absorption and scattering and in some cases it makes it challenging to interpret the results of your experiments if you have too much scattering.

00:29:51.412 --> 00:29:53.623
So Jin was also very well aware of that.

00:29:53.623 --> 00:30:04.449
Actually, in later experiments, in 1971, he was measuring the scattering intensity using a Shimidazu PG-2 type light scattering photometer.

00:30:04.449 --> 00:30:08.025
So he was very, very devoted to measure this properly.

00:30:08.474 --> 00:30:15.044
Jin also, of course, measured the luminance of the sign, or the one that the observer had the control over.

00:30:15.044 --> 00:30:19.903
And for that I'll read you out from his paper because I find it simply fascinating.

00:30:19.903 --> 00:30:27.544
The luminance meter is a modified single-lens reflex camera with a 400mm telephoto lens.

00:30:27.544 --> 00:30:36.369
A circular slit with frosted glass was attached to the film surface so that the field of view became 3cm diameter at 5m away.

00:30:36.369 --> 00:30:48.861
A sensitive photomultiplier tube was installed behind the slit and its output was connected to an automatic balance recorder slit and its output was connected to an automatic balance recorder.

00:30:48.861 --> 00:30:51.308
This is fascinating, how much effort you have to do to make a simple illuminance measurement.

00:30:51.308 --> 00:30:56.402
We're so blessed with the technology today that you just purchase a device and it does it for you.

00:30:56.402 --> 00:31:11.785
So those were basically things that had the ability to measure First the background and sign intensity, and the sign intensity was recorded at the moment the observer reached the visibility threshold.

00:31:11.785 --> 00:31:27.338
He was of course knowing the distance at which the sign was observed, because it was a discrete value coming from the design of the experiment, and he knew the smoke extinction coefficient and in the end he of course combined all three to formulate his observations and conclusions.

00:31:27.980 --> 00:31:39.752
In his later work, in 1971, he actually went a little bit further with the impact of smoke on that visibility and he started capturing smoke.

00:31:39.752 --> 00:31:57.816
He started capturing smoke on the very fancy traps and then was observing and picturing the smoke under microscopes so he was able to get some sort of size distribution of smoke particles for different smokes.

00:31:57.816 --> 00:31:58.719
That's one thing.

00:31:58.719 --> 00:32:05.340
He was able to derive very good ratios between scattering and extinction for different types of smoke.

00:32:05.340 --> 00:32:06.276
That's the second thing.

00:32:06.276 --> 00:32:20.009
And then again he repeated his experiments with multiple types of smoke to get new relations, basically to confirm the relations from his first paper.

00:32:20.009 --> 00:32:27.348
And in fact you can see that the relation, the main relation, still kind of holds but it's not as sharp as we use it.

00:32:27.348 --> 00:32:39.747
As I mentioned, he proposed 2 to 4 range for K-factor for light reflecting science, 5 to 10 for light emitting science, and this is actually the outcome of those observations with different types of spoke.

00:32:40.055 --> 00:32:45.248
So now let's move into the potential sources of uncertainty in the model.

00:32:45.248 --> 00:32:55.407
I think this is very important for everyone who's using the visibility in smoke modeling and let's correlate those not to my thoughts about how do we do visibility in smoke modeling.

00:32:55.407 --> 00:32:58.505
That would be an entire podcast episode which actually exists.

00:32:58.505 --> 00:33:02.619
There's an episode with Lucas Arnold, I think it's something like episode 32.

00:33:02.619 --> 00:33:06.288
Around around that episode when we talk about visibility predicting framework.

00:33:06.288 --> 00:33:09.284
That's our research that we're carrying.

00:33:09.284 --> 00:33:12.203
Let's keep this in relation to Jin's experiments.

00:33:12.595 --> 00:33:18.769
So one thing that he realized there's multiple sources of uncertainty that he has realized already.

00:33:18.769 --> 00:33:22.645
One is the contrast threshold.

00:33:22.645 --> 00:33:46.365
So he understood that the contrast threshold the point at which the observer stops seeing the sign, stops seeing where the observer reaches the critical visibility distance is not a single number and he already mentions that this value can be from 0.01 to 0.05.

00:33:46.365 --> 00:33:50.070
That's 400% difference between those two values.

00:33:50.070 --> 00:33:54.803
In the first research he used the value of 0.02 as a default one.

00:33:54.803 --> 00:33:59.903
He tracks the source of this difference to the properties of eyes.

00:33:59.903 --> 00:34:04.538
Everyone has their own eyesight and eyesight of different people will differentiate.

00:34:04.538 --> 00:34:07.123
So it's not a simple thing.

00:34:07.123 --> 00:34:21.570
Actually, the critical contrast threshold, it's a very individual thing and later on in his research he was actually very careful because he started doing eyesight measurements before people were entering the experiments to actually quantify that property of eye.

00:34:21.570 --> 00:34:29.001
But in the first relations, in the first model, the contrast threshold is there and we're not really using that.

00:34:29.001 --> 00:34:31.827
It's kind of funny that we're not using that.

00:34:32.195 --> 00:34:41.913
The second thing, which I've already discussed at length, is the brightness the brightness of the room, the room that you're in and the luminance of the sign that you're observing.

00:34:41.913 --> 00:34:50.766
The contrast ratio between those two will also influence the visibility distance at a specific value of the extinction coefficient you have.

00:34:50.766 --> 00:34:51.887
That defines your smoke.

00:34:51.887 --> 00:34:58.914
And the third thing that is a source of uncertainty is the property of smoke, the scattering to extinction.

00:34:58.914 --> 00:35:07.047
So the little k factor which he defined as being in the range between 0.4 to 1.0.

00:35:07.047 --> 00:35:10.998
0.4 means the smoke is dark and 60% is absorbed.

00:35:10.998 --> 00:35:14.867
1.0 means it's mostly reflecting, so a perfect white.

00:35:14.867 --> 00:35:20.681
So those three values already bring a ton of uncertainty to that.

00:35:20.681 --> 00:35:21.902
A ton of uncertainty.

00:35:21.902 --> 00:35:30.521
Just think about the contrast threshold having 400% range, the small k-factor having 150% scatter.

00:35:30.521 --> 00:35:37.766
Here you can have, for the same extinction coefficient of smear visibility of 2 meters and 20 meters.

00:35:37.766 --> 00:35:48.155
That's completely possible within the framework of this theory, just changing the variables which we never touch in our engineering, framework of this theory, just changing the variables which we never touch in our engineering.

00:35:48.155 --> 00:35:52.079
So this is a warning, literally a warning light.

00:35:52.079 --> 00:35:54.628
It's not that Gene has not realized he realized it very well the importance of those variables.

00:35:54.628 --> 00:35:55.371
That's why he put them in equations.

00:35:55.371 --> 00:35:59.262
It's us who are ignorant to that knowledge and who are not using that.

00:35:59.262 --> 00:36:00.284
Perhaps we should.

00:36:01.025 --> 00:36:03.778
Another source of uncertainty comes from the extinction coefficient.

00:36:03.778 --> 00:36:08.449
So Gene, in his experiments, has measured the extinction coefficient.

00:36:08.449 --> 00:36:20.023
He had the densitometer in place, he measured the light intensity at the densitometer and he has calculated the extinction coefficient from the measurement that he has carried.

00:36:20.023 --> 00:36:30.971
In our modeling we're not measuring the light intensity, we're approximating it using Lambert's beer law for an absorbing homogeneous aerosol.

00:36:30.971 --> 00:36:33.322
We apply it to quite a heterogeneous mixture.

00:36:33.322 --> 00:36:48.380
That's a problem for me as well, but we approximate it, knowing the Mulholland's value of specific smoke extinction coefficient and the density of smoke that is a product of our design fire, the heat release rate and the soothills that we put into the place.

00:36:48.380 --> 00:36:59.117
So the value of extinction coefficient that we have in our modeling is very artificial and highly, highly influenced by the person who's carrying the calculations.

00:36:59.117 --> 00:37:04.068
I'm not sure if everyone realizes how powerful your impact is on that.

00:37:04.068 --> 00:37:17.135
Actually, we had a paper with Gabriele Vigne on the sooth-healed factor and how you can use higher sooth-healed factors to reduce the uncertainty that you bring into the model Just because of the exponential nature of the Lambert Spear low.

00:37:17.135 --> 00:37:19.981
You're recommended to read that paper.

00:37:19.981 --> 00:37:29.807
However, you must realize that we have huge influence over this parameter and this in the end hugely influences the visibility in smoke.

00:37:30.514 --> 00:37:36.568
So, to wrap up the episode, I think the Gene's experiments were really brilliant.

00:37:36.568 --> 00:37:41.266
He was really interested in safety of people.

00:37:41.266 --> 00:37:45.646
They've noticed that people perish in fires because of loss of visibility.

00:37:45.646 --> 00:37:51.360
They lose the ability to escape and this is an important aspect of the visibility in smoke as sustainability criterion.

00:37:51.360 --> 00:38:02.740
It does not kill you, it does not damage you that much the visibility itself but it makes you lose the orientation, it makes you unable to escape and other things in the fire may be fatal to you.

00:38:02.740 --> 00:38:24.686
He also noticed that development of mathematical model of visibility in smoke based on the physical optical properties of the smoke and other phenomena may be attractive for researchers, but it could be very complicated, and he even uses a sentence it must be very complicated and intense to be out of practical use.

00:38:24.686 --> 00:38:27.719
I think that's the only part where gin was very wrong.

00:38:27.719 --> 00:38:40.320
I think this has enormous practical use if we had a good visibility in smoke model and I hope the model that we are developing with Lucas Arnold will be a thing like that and it will actually find a practical use.

00:38:40.762 --> 00:38:44.115
There's a lot of beauty in the research of gin and there's a lot of joy in studying this.

00:38:44.115 --> 00:38:44.545
For me, ancient experiments.

00:38:44.545 --> 00:38:45.293
I've learned a lot of beauty in research of gin and there's a lot of joy in studying this.

00:38:45.293 --> 00:38:47.922
For me, ancient experiments.

00:38:47.922 --> 00:38:59.668
I've learned a lot when going through the original material, much more than from the limited remarks that are left in the handbook and his plenary paper.

00:38:59.668 --> 00:39:07.795
I think a lot of things, a lot of important details are, and I don't think that they were lost because of someone's bad intentions.

00:39:07.795 --> 00:39:29.822
Actually, 1997, when the plenary lecture was given, it's still the visibility was already an important part of modeling because it was using zone models, but perhaps still it was not recognized how impactful it is over the built environment, and Gene was so fascinated by the human behavior in smoke that he spent the rest of his professional career studying that.

00:39:29.822 --> 00:39:40.059
There's much more space given to those experiments rather than the first experiments of the visibility in smoke that we've discussed in today's episode.

00:39:40.059 --> 00:39:48.623
However, I truly believe that those first experiments from 1970 and 1971 were experiments that have changed the fire science.

00:39:49.315 --> 00:39:53.826
Let me know what you think, let me know if you like this episode.

00:39:53.826 --> 00:39:56.043
I have a very intimate relation with this.

00:39:56.043 --> 00:39:57.960
Modeling is very important to me.

00:39:57.960 --> 00:40:02.947
I would love to hear your observations, your ideas about this research.

00:40:02.947 --> 00:40:04.500
How do you value it?

00:40:04.500 --> 00:40:06.021
What's important in it for you?

00:40:06.021 --> 00:40:11.347
Maybe you have different conclusions that I have, and I would be very, very happy to hear about them.

00:40:11.347 --> 00:40:23.474
And if you've enjoyed it, I'm very happy to cover papers three to five the ones that correlate the smoke density to human behavior in another episode of experiments that have changed the fire science.

00:40:23.635 --> 00:40:28.322
I'm not sure if you're ready for that much visibility in smoke, so I need some feedback in here.

00:40:28.322 --> 00:40:33.099
If you like it, I'll do it, and that would be it for today's episode of the Fire Science Show.

00:40:33.099 --> 00:40:35.744
I hope you've enjoyed this one.

00:40:35.744 --> 00:40:41.702
I had a huge pleasure to go once again through the jeans papers and record this to you.

00:40:41.702 --> 00:40:58.179
I hope you have a good week, nice weekend, summer is here, the summer in Poland is beautiful, so I'm going to enjoy that and I hope, wherever you are, the summer or, if you're in Southern Hemisphere, your winter are fantastic and you're having a good time.

00:40:58.179 --> 00:41:02.641
Thanks for being here with me and see you here again in the next FireScienceShow episode.

00:41:02.641 --> 00:41:26.085
Cheers, bye, bye.

162 - Experiments that changed fire science pt. 9 - Jin's experiment on visibility in smoke

Listen On

Experiments that changed fire science Episodes

Smoke and visibility Episodes

Recent Episodes