152 - Why we need good standards with Björn Sundström

Transcript

WEBVTT

00:00:00.699 --> 00:00:02.402
Hello everybody, welcome to the Fire Science Show.

00:00:02.402 --> 00:00:16.496
In the podcast we are trying to talk a lot about some aspects of fire safety engineering profession that perhaps are not that much known to the engineers and from my perspective, that's definitely the world of fire testing and standardization.

00:00:16.496 --> 00:00:25.774
Not many of us engineers are involved in those processes and yet they are highly influential over everything that we do as fire safety professionals.

00:00:25.774 --> 00:00:32.112
This is the reason why I'm bringing guests from the world of laboratories and standardization into the show.

00:00:32.112 --> 00:00:39.628
I've had my good colleague Piotr Turkowski in the podcast where we've talked about the fire resistance and how fire testing looks like.

00:00:39.628 --> 00:00:42.149
We've had Birgit Messerschmidt.

00:00:42.149 --> 00:01:02.887
I've recently had Rudolf Van Mielo All of them important episodes showing you the kitchen of how standards and classes and the entire system that allows product certification and placing products on the common market in the European Union works like, and this episode is no different.

00:01:02.887 --> 00:01:10.674
I have invited Bjorn Sundström from LIA, but Bjorn has most of his life spent at RISE or SPPAC.

00:01:10.674 --> 00:01:24.007
He was also director at Fire Research AS in Trondheim, which was then acquired by SP, and he's also involved with Brandfusk, so entire career spent in a fire testing laboratory.

00:01:24.007 --> 00:01:29.384
So he definitely has great insights over how stuff works from this perspective.

00:01:29.384 --> 00:01:47.594
In this episode we try to go into what makes a good standard, why we need standards and I try to do it from perspective of laboratory manufacturer and also the end user or the society and we also try to understand why it takes so much time and why we really need good standards after all.

00:01:47.594 --> 00:01:51.129
So enough of talking important episode.

00:01:51.129 --> 00:01:51.870
I hope you enjoy it.

00:01:51.870 --> 00:01:53.966
Let's spin the intro and jump into the episode.

00:01:58.301 --> 00:01:59.846
Welcome to the Firesize Show.

00:01:59.846 --> 00:02:03.402
My name is Wojciech Wigrzyński and I will be your host.

00:02:03.402 --> 00:02:22.895
This podcast is brought to you in collaboration with OFR Consultants.

00:02:22.895 --> 00:02:25.837
Ofr is the UK's leading fire risk consultancy.

00:02:25.837 --> 00:02:36.655
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.

00:02:36.655 --> 00:02:52.471
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 a hundred professionals.

00:02:52.471 --> 00:03:04.127
Colleagues are on a mission to continually explore the challenges that fire creates for clients and society, applying the best research experience and diligence for effective, tailored fire safety solutions.

00:03:04.127 --> 00:03:17.768
In 2024, ofr will grow its team once more and is always keen to hear from industry professionals who would like to collaborate on fire safety futures this year, get in touch at ofrconsultantscom.

00:03:17.768 --> 00:03:24.740
Hello everybody, I'm here today with Professor Bjorn Sturmström from Lulea University.

00:03:24.740 --> 00:03:26.308
Hello, bjorn, good to have you in the podcast, bjorn.

00:03:26.329 --> 00:03:26.550
Sturmström.

00:03:26.550 --> 00:03:26.992
Hello, thank you.

00:03:27.120 --> 00:03:29.847
Thank you for being here, sturmström, I'm very happy to have you in the show.

00:03:29.847 --> 00:03:34.651
I've met you like a decade ago when we had a funny conference in Poland.

00:03:34.651 --> 00:03:55.865
That was good times and I know you were very involved back then running a large testing laboratory and, throughout your career, very involved in development of fire safety from, let's say, the standardization testing house laboratory side.

00:03:55.865 --> 00:03:57.407
So I would love to talk to you about the standardization in fire safety.

00:03:57.407 --> 00:04:05.169
I had some talks in the podcast so far where mostly we were criticizing it, if anything kind of of interesting discussions.

00:04:05.169 --> 00:04:16.725
But I would love to openly talk standardization with you today and my first question would be like why do we need standardization related to fire safety products?

00:04:16.725 --> 00:04:24.148
How does the world where we don't have one compares to the world in which we have products that are standardized?

00:04:24.889 --> 00:04:26.232
Oh yeah, that will be.

00:04:26.232 --> 00:04:30.000
For one thing, you will not be able to sell products between countries.

00:04:30.000 --> 00:04:36.733
This is one important issue, that is, it takes away trade barriers.

00:04:36.733 --> 00:04:47.047
If you look at the European system, the standards there allows you to sell products in 30 countries or whatever, which is very important.

00:04:47.047 --> 00:04:49.250
And the other thing is about safety.

00:04:49.250 --> 00:05:02.031
You need to be able to declare the properties of a product, because that is what you need when you're trying to make something at a certain safety level and you need a declaration procedure for that and that's a standard.

00:05:02.031 --> 00:05:03.084
So you need the standards very badly.

00:05:03.084 --> 00:05:03.889
You can compare with cars, for example.

00:05:03.889 --> 00:05:05.237
They get one tooth standard, so you need the standards very badly.

00:05:05.237 --> 00:05:17.206
You can compare with cars, for example they get one tooth and you fall five stars and they are tested according to a scenario where they crash into walls and being crashed from the side and whatever.

00:05:17.206 --> 00:05:18.670
Take the passengers.

00:05:18.670 --> 00:05:22.730
So standards are extremely important and you cannot live without them.

00:05:23.172 --> 00:05:33.591
I like the example of cars, because every time I see an advertisement of a new car it always has five stars, Like if no other measure existed for the manufacturer.

00:05:33.591 --> 00:05:43.187
But in the end I feel the stars are in some way an ambiguous summary of safety of the vehicle against crash.

00:05:43.187 --> 00:05:43.928
That's my feeling.

00:05:43.928 --> 00:05:45.382
I don't know the details how it works.

00:05:45.382 --> 00:05:51.982
That's my view and, as a consumer, that this car will deliver me safety because it has five stars.

00:05:51.982 --> 00:05:54.548
How about like safety products?

00:05:54.548 --> 00:05:59.100
How do people perceive like you talked about, market manufacturers?

00:05:59.100 --> 00:06:06.350
What about just users, I'd designers, people who are not interested in putting products in the market?

00:06:06.350 --> 00:06:09.456
They have to use the products in their design.

00:06:09.456 --> 00:06:17.721
How is the regulatory system of fire safety providing them with this overarching information of what safety is Well?

00:06:18.101 --> 00:06:22.709
the standard is the declaration of a property of a product, a property of a product you know.

00:06:22.709 --> 00:06:25.673
You take, for example, non-combustibility.

00:06:25.673 --> 00:06:37.947
That means that your product, in real light situation, will not continue to fire to any extent that it's increasing the danger for normal use, right?

00:06:37.947 --> 00:06:39.446
That's one thing you know.

00:06:39.446 --> 00:06:49.375
And the other end of these European classes is products that could cause flash over in a small room for sort of 12 seconds or something like that.

00:06:49.375 --> 00:06:50.704
So you have information there.

00:06:50.704 --> 00:06:58.786
And then the other aspect is, of course, the regulation, because the regulation asks for certain properties than our products.

00:06:58.786 --> 00:07:10.632
And if the regulations ask for non-combustible construction, then you have these standards for non-combustible construction, then you have these standards for non-combustibility and you have this product that declare that property.

00:07:10.632 --> 00:07:22.190
There's a C mark on them saying this is Euroglass A1 or A2, which means that it had been tested and verified to be non-combustible.

00:07:22.190 --> 00:07:24.420
So then you can select those products.

00:07:24.420 --> 00:07:28.629
And if you don't have that, you can select, you know something else.

00:07:29.399 --> 00:07:39.420
So, between the bodies, markets, the regulations and the user, do you think for whom the standardization has the biggest value?

00:07:39.420 --> 00:07:42.670
You know how valuable it is to have standardization.

00:07:42.670 --> 00:07:47.202
Is it more important for the manufacturers because of the market regulation?

00:07:47.202 --> 00:07:50.329
Is it key for regulators?

00:07:50.329 --> 00:07:53.483
Could regulators regulate without having standardized tests?

00:07:53.483 --> 00:07:55.209
Perhaps that's a better question.

00:07:56.562 --> 00:07:58.048
That's a very interesting question.

00:07:58.048 --> 00:08:03.565
Can you regulate without knowing the properties of the products that you're regulating?

00:08:03.565 --> 00:08:05.346
That's quite difficult.

00:08:05.346 --> 00:08:08.987
You see, you need to be able to declare properties.

00:08:08.987 --> 00:08:18.146
Whatever that is, you know it can be scenario-based standards that we have now, or it can be some sort of fundamental properties.

00:08:18.146 --> 00:08:33.620
If you're building a bridge, you don't have to, you know, test the bridge because you can test the properties of the steel, and then you have models that allow you to build a bridge that will have certain strengths, but you're still declaring properties.

00:08:33.620 --> 00:08:39.567
You have to do that, or you could say we have no regulations on fire.

00:08:39.567 --> 00:08:40.971
See what happens.

00:08:42.240 --> 00:08:50.764
This is again a very interesting comparison because in fire safety let's now think about something very complex like a facade system.

00:08:50.764 --> 00:08:57.466
You know we, you technically, could test all the components and get some outcome of that.

00:08:57.466 --> 00:09:14.788
Let's say, do material testing on the components, run cone calorimeter or non-combustibility test on every material, and then you have a large facade mock-up, like your Swedish test or the newly designed European test, where you test a massive sample in the full scale.

00:09:14.788 --> 00:09:29.687
And I wonder to what extent there is even a chance that we can predict performance of a product from just knowing the properties of the subcomponents or little things that go inside.

00:09:30.360 --> 00:09:32.427
Well, that's again it's a very interesting question.

00:09:32.427 --> 00:09:47.841
It's extremely complex thing because you have to predict the behavior of the system, which is not only the behavior of the incoming materials and their burning behavior, ignition and pyrolysis rate and whatever.

00:09:47.841 --> 00:10:01.389
You also have to predict the flow that's going on there inside the air gaps and so on, and the mechanical properties of the construction when it is exposed to heat.

00:10:01.389 --> 00:10:29.797
So this is, of course, a thing that would be very nice to do, but I think it's also extremely complex Because you know, on a facade, if you have a fire that's going on inside a flat, you have a big flame going out through the window exposing the construction around the window, for example, the appearance of whatever is there, so you don't have a fire going on inside the window, for example, the appearance of whatever is there, so you don't have a fire going on inside the facade.

00:10:29.876 --> 00:10:41.441
And then you have perhaps barriers, heat barriers between the floors to stop the flame spread going on, and you know mechanical properties, things falling off the facade and everything.

00:10:41.441 --> 00:10:48.308
You have to predict all that in comparison to making a test, and the test, of course, will not show you everything either.

00:10:48.308 --> 00:10:50.687
So these things are very complex.

00:10:50.687 --> 00:11:09.027
I mean, there is one thing you can do on materials test, and that is if you prescribe that your facade should be non-combustible, which is a material test, you will have steel and brick, concrete and whatever, and that facade is not going to burn.

00:11:09.440 --> 00:11:10.907
That's the ultimate prescription.

00:11:10.907 --> 00:11:27.986
I wonder, from your very long career in one of the biggest European laboratories, were ever such efforts undertaken to try and get complex system assessment from a collection of smaller tests, and to what extent were they successful?

00:11:29.139 --> 00:11:32.567
Yeah, well, it was done with the cold calorimeter.

00:11:32.567 --> 00:11:46.613
You know, to predict the room fire in the small room standardized rule corner test procedure and you could use the cold calorimeter metadata there to have some quite good predictions.

00:11:46.613 --> 00:11:53.451
But you know, then again that's not the building, so the road fire is the reference scenario in the SPI test.

00:11:53.530 --> 00:11:57.270
Now it can also be the reference scenario easily to the cold calorie.

00:11:57.270 --> 00:12:01.250
So there is this superposition models that do that.

00:12:01.250 --> 00:12:07.702
I think also you find it in the SDS, that kind of water you could put in colorimetry data there.

00:12:07.702 --> 00:12:15.726
So that's what it is around, and then I don't really know what is going on there.

00:12:15.726 --> 00:12:32.312
That is something I think you would like to have that you are being able to use the huge amount of data that is generated through the standards and being able to use that for predictions.

00:12:32.312 --> 00:12:40.884
That will be one thing that is very nice to do, and otherwise it's, of course, to develop test methods again or standards again.

00:12:40.884 --> 00:12:49.349
That would provide you with useful data for using the SDS and these kind of all these works are pretty things.

00:12:49.940 --> 00:12:54.231
Yeah, but still cone would be like still some sort of material properties.

00:12:54.231 --> 00:12:58.410
How about, if we can use the example of facade again?

00:12:58.410 --> 00:13:00.626
You have a facade with a cavity.

00:13:00.626 --> 00:13:05.552
You can have a cavity width of five centimeters, six, seven, twelve, fifteen.

00:13:05.552 --> 00:13:10.113
Perhaps each of them will behave differently in in different fires.

00:13:10.653 --> 00:13:42.389
You may have hundreds types of render on the facade, you may have the different types of insulation in your facade and also because of architectural function of that facade, it is kind of built for a purpose of a specific building, so it will have specific architectural details around the windows, specific corners, connections, perhaps those tons of tiny, tiny things that we know very well that influence the outcome of the final test.

00:13:42.389 --> 00:14:04.783
So I wonder if there exists any approach or were any developments in like testing the details and from this test medium or small scale test extrapolate to large samples, like testing different I don't know closings of cups or the detailing around windows, you know, instead of of gaps or the detailing around windows.

00:14:04.783 --> 00:14:16.355
You know, instead of building 10 meter tall facades to just check the frame around the window, could we perhaps check you know 100 frames and then say, ok, these 30 are good and the other ones are bad.

00:14:19.220 --> 00:14:19.782
I don't know any such work.

00:14:19.782 --> 00:14:22.831
Maybe you know better than me if there's anything like that going on.

00:14:22.831 --> 00:14:23.352
I don't know, I don't know.

00:14:23.352 --> 00:14:23.854
That's extremely complex.

00:14:23.854 --> 00:14:25.197
You know better than me if there's anything like that going on.

00:14:25.197 --> 00:14:25.779
I don't know, I don't know.

00:14:25.779 --> 00:14:28.427
That's extremely complex, you know.

00:14:28.427 --> 00:14:41.365
If you go back to the cars again, it would be like instead of running a crash test as they do, you send in a drawing of your car and you send in the material properties the yield and the products and then you calculate the crash.

00:14:42.831 --> 00:14:43.785
I don't know they don't do that.

00:14:43.785 --> 00:14:44.716
I don't know they don't do that.

00:14:44.716 --> 00:14:48.509
I know, and your example here of a facade is extremely complex.

00:14:48.509 --> 00:14:50.847
So I mean, you would tell the test.

00:14:50.847 --> 00:14:55.527
And the test is not going to be enough either, because, as you say, there are variants here.

00:14:55.527 --> 00:14:59.779
Do they really build it exactly as you have tested, and so forth.

00:15:01.205 --> 00:15:02.408
But I will go for the test.

00:15:03.160 --> 00:15:23.350
I'm building this example, example, you know to challenge you because, uh, a standard, as we said before, or as you you've explained me, at the beginning there's like market, there's regulators, there's user, but in the end it's the, the market, the manufacturers who pay for the test, who obtain the certification, who put their products into the market.

00:15:23.811 --> 00:15:48.246
I always wondered is there a big interest for a manufacturer to actually do tests that is, building-related or element-related, like more complex than just the product characteristic they're selling, more going outside of the box that standard defines as the minimum stuff you need to declare to sell the product you know, and and perhaps the amount of tests.

00:15:48.246 --> 00:16:07.014
Or perhaps the reason why we don't have such things is because maybe the manufacturers do not have any reason or good good enough reason to actually pursue that, because the regulatory system that is created allows them to obtain their goal, which is selling the products.

00:16:07.014 --> 00:16:24.731
I wonder to what extent we're limited by the fact that it's the manufacturers responsible for obtaining the certification and characteristics of the products, and when they do that, it relates to their own goals, right, perhaps not for all manufacturers, but I would assume for most their own goals right.

00:16:24.772 --> 00:16:27.200
Perhaps not for all manufacturers, but I would assume for most.

00:16:27.200 --> 00:16:33.451
Well, there are manufacturers that will go beyond the test to prove that they have a very excellent product to gain market.

00:16:33.451 --> 00:16:43.153
Of course there are, but you know, the general picture is, of course, that you fulfill your S and then you could put a C mark on your product.

00:16:43.153 --> 00:16:44.667
You can sell it off the market.

00:16:44.667 --> 00:16:49.071
This is a or should I call this a core.

00:16:49.539 --> 00:16:53.490
The regulators play a big role here.

00:16:53.490 --> 00:16:56.448
That is the demand of what do they want?

00:16:56.448 --> 00:17:10.771
And it's not only the regulators, it's the, for example, insurance companies that could further and say we see now a lot of insurance cases where there are fires here and we want something better than this regulation.

00:17:10.771 --> 00:17:12.734
And then something happens.

00:17:12.734 --> 00:17:18.833
But there somewhere must be also a demand, or you see a catastrophe, fire.

00:17:18.833 --> 00:17:35.432
You know, I guess, the regulation in the UK they are developing quickly now due to the Grenfell case and, for example, in shipping, the fire in the Scandinavian Star many years ago that started the development of the IMO requirements and rules for ships.

00:17:35.432 --> 00:17:38.288
So they go hand in hand.

00:17:38.288 --> 00:17:50.605
You have the manufacturer, you have the market and you have the regulator and you have the fire safety, engineering, people, it sort of all acts together pushing forward to better fire safety.

00:17:51.363 --> 00:17:56.103
Another thing that comes to my mind when thinking about the entire fire safety system.

00:17:56.103 --> 00:18:10.071
The characteristic is like the final stage of the process and there's usually a pathway leading to that characteristic pathway that sometimes is a simple success.

00:18:10.071 --> 00:18:15.290
Someone you know develops a product, obtains the characteristic they've hoped for and they're done.

00:18:15.290 --> 00:18:20.906
Sometimes it's a bumpy road where the product failed in many ways that they did not expect.

00:18:20.906 --> 00:18:28.548
They have to fix stuff, you know, change stuff and then obtain their characteristic, but the end user has no idea about it.

00:18:28.548 --> 00:18:30.945
The end user also has no idea.

00:18:31.086 --> 00:18:42.741
If it barely made the test, or it made it easily, with confidence, it should work for 30 minutes, work for 50, right, and that's a huge.

00:18:42.741 --> 00:18:45.894
Now, is there any way because that's a huge?

00:18:45.894 --> 00:18:46.419
Uh, now, is there any way?

00:18:46.419 --> 00:19:02.250
Because there's a ton of knowledge in there, a ton of knowledge related to the, to the behavior of products and and a ton of stuff that we could extrapolate, perhaps the entire buildings, if we understand failures and the reasons why things do not pass the tests.

00:19:02.250 --> 00:19:11.343
I wonder if, uh, from your experience now, now to tap from your experience now, now to tap from your experience, maybe in the forum, because you've also dealt with a ton of different laboratories.

00:19:11.343 --> 00:19:15.132
Are we even able to tap into this knowledge ever, do you think?

00:19:15.132 --> 00:19:16.001
What do you mean?

00:19:16.442 --> 00:19:25.270
Having our data published, or what I understand, you cannot just publish the data because it's copyrighted, it's, it's owned by the manufacturer.

00:19:25.270 --> 00:19:31.240
It perhaps you know, it perhaps has some secret information in it that you cannot share.

00:19:31.240 --> 00:19:33.164
But it's not just data, it's knowledge.

00:19:33.164 --> 00:19:39.207
You know, if you understand how systems fail, you perhaps can make better regulation.

00:19:39.207 --> 00:19:43.387
You perhaps can educate the user better, you perhaps can build something nice.

00:19:43.387 --> 00:19:53.426
But this knowledge I'm also working in a test lab and I understand like how much we learn as the test lab employees from failed experiments, from failed uh tests, yeah.

00:19:53.426 --> 00:20:03.060
So I wonder if there's any way that we could, you know, somehow share this with the community, because it's it's knowledge, it's, it's, yeah, it's something that could be useful.

00:20:03.060 --> 00:20:04.521
Yeah, it could be very useful.

00:20:04.823 --> 00:20:05.864
I mean laboratories.

00:20:05.864 --> 00:20:06.704
They have knowledge.

00:20:06.704 --> 00:20:10.810
You're testing something and you see what the failure mechanism is.

00:20:10.810 --> 00:20:14.355
You can also see how you can improve that kind of a problem.

00:20:14.355 --> 00:20:26.365
You see things like that, I guess and this has to go through research, like Brandsforge, who I'm also representing, who are funding research products.

00:20:26.365 --> 00:20:28.588
That will be the way to go.

00:20:28.588 --> 00:20:34.536
You know, to some products describing failure mechanisms of product.

00:20:34.536 --> 00:20:40.561
That's the way to go.

00:20:40.561 --> 00:20:49.009
I think you could use that and then you could use also the experience from laboratories and put them together with universities and so on and so forth to produce that kind of a knowledge.

00:20:49.009 --> 00:20:52.567
I think that will be very, very helpful if you could go on in that route.

00:20:54.364 --> 00:21:03.471
I think, of course, obtaining new funding for projects like that is a very interesting way and definitely something could be done.

00:21:03.471 --> 00:21:18.931
I thought more about the already existing, you know, thousands and thousands of test reports and plots of temperature in function of time and and other figures and small grass that are getting dust in laboratories over the world.

00:21:18.931 --> 00:21:31.912
I'm not sure if there still needs to protect data that was collected 20 years ago and perhaps we could today with big data and AI, perhaps we could still learn something from that collection.

00:21:32.031 --> 00:21:35.548
Yeah, I think that would be very, very good if you could do that.

00:21:35.548 --> 00:21:45.426
You need to change the law or something like that because it's proprietary data, but I mean, if you could get hold of that, that would be extremely valuable.

00:21:45.700 --> 00:21:59.446
I know you know, in the development of the European system we made 1,500 tests, something like that which is wow, that's not a lot compared to what you would like to have if you want to go into that direction.

00:21:59.446 --> 00:22:11.827
There are, say, 60 laboratories that are notified to the European Commission now producing documents that are legal in the entire European Union and they are testing, you know, every day.

00:22:11.827 --> 00:22:16.872
I don't know how many hundreds of thousands test units last 25 years that are around.

00:22:16.872 --> 00:22:21.449
So that will be big data and that will be very helpful if you can get a hold of some of that.

00:22:22.279 --> 00:22:24.528
I'm going to be picking brains of people.

00:22:24.528 --> 00:22:35.330
Maybe one day we will find a way in which you can do it safely and not cause damage to anyone's property, but perhaps build something nice for everyone around.

00:22:35.330 --> 00:22:59.000
Another question that relates kind of to that, if I wanted to dig into data from standardized tests, the standardized test also, you know, contains specific performance objectives or defines how to pass that test-based file criteria, and in some way those things drive the design of the products.

00:22:59.000 --> 00:23:08.880
I wonder if you share the same opinion that the standards in some way shape the products as well, because people design them to pass the tests right.

00:23:09.421 --> 00:23:12.467
Yeah, it does, of course, because that's one of the aims of the standard.

00:23:12.467 --> 00:23:15.653
You know that is to influence the product's behavior.

00:23:15.653 --> 00:23:19.423
But it also goes hand in hand with the regulations.

00:23:19.423 --> 00:23:19.742
You know.

00:23:19.742 --> 00:23:23.586
You know you ask for certain performance, high performance.

00:23:23.586 --> 00:23:30.452
You will push the products being put on the market in your country to fulfill that requirement.

00:23:30.452 --> 00:23:34.717
So it influences the products, of course it does.

00:23:39.799 --> 00:23:41.583
How does it influence innovation, growth development?

00:23:41.583 --> 00:23:57.634
Because if there's a well-established standard, it doesn't mean that there are no other ways to obtain a characteristic that would be more feasible, like instead of characteristic A, someone could come with another one that could be more representative for a specific end-use condition.

00:23:57.634 --> 00:24:07.747
But it may be challenging for them to get that done because everyone standardizes or declares class within the other testing regime.

00:24:07.747 --> 00:24:09.123
It's very difficult to you know.

00:24:09.123 --> 00:24:15.480
Once this is established and once this is pretty much ruling the market, it's very difficult to break out of that.

00:24:15.480 --> 00:24:19.451
How do you see developments in that related to innovation?

00:24:20.500 --> 00:24:21.842
Well, you need to.

00:24:21.842 --> 00:24:32.296
If you go back and look at a standard, it has, as it goes from a fire scenario that you select, that fire scenario has to be reasonable.

00:24:32.296 --> 00:24:43.432
You know, if you are saying I'm going to do a test here for testing linings, classifying them, I don't put the candle towards the wall.

00:24:43.432 --> 00:24:46.135
You know that would give me no results.

00:24:46.135 --> 00:24:56.673
You have to have a fire scenario that's reasonable, and that's extremely difficult because in the real world there are millions of scenarios.

00:24:56.673 --> 00:25:08.778
You have to pick something that is reasonable, useful, and then you have to develop your standard that has to be reproducible, useful, and then you have to develop your standard that has to be reproducible, and so on and so forth.

00:25:08.778 --> 00:25:13.162
Some sort of validation you need.

00:25:13.182 --> 00:25:14.265
But that can be extended, or should, I think, be extended.

00:25:14.265 --> 00:25:15.750
So the system you have now should develop to be extended.

00:25:15.750 --> 00:25:23.170
Applications of, say, there are standards for that as well, and so you should allow and you should develop the system.

00:25:23.170 --> 00:25:25.153
But you still can do that now.

00:25:25.153 --> 00:25:33.134
I mean you could produce a product that is being proven through fire safety engineering to be very useful for application.

00:25:33.134 --> 00:25:37.391
Then you can say that happens in the shipping area, of course.

00:25:37.391 --> 00:25:40.368
So yeah, I mean you can do that already today.

00:25:40.368 --> 00:26:02.807
That depends, then, on special niche products that you can use in that way, of course, but I think for the bulk of the products in Europe you need to extend the application of the declarations of all the products and build what's more generalized properties perhaps?

00:26:03.592 --> 00:26:06.805
And how about applying standards for specific scenarios?

00:26:06.805 --> 00:26:17.482
The thing most perhaps vivid example is the use of SBI or Euroclass system for performance of facades.

00:26:17.482 --> 00:26:33.096
Again, you know that's that's very far from the original intention of SBI and I also had Rudolf Amirlo in here who told me that facade system sorry, that facade was the next one to be regulated after Euroclasses.

00:26:33.096 --> 00:26:35.153
It just didn't happen at that time.

00:26:35.153 --> 00:26:38.054
But still we have examples like that.

00:26:38.054 --> 00:26:55.537
Perhaps the way we use it goes a little bit too far from the original scenario, from this original representative, fhir there, when the standard was created here, do you see need to rapidly develop new standards?

00:26:55.537 --> 00:27:24.617
Is it more a need for regulations and the way what we allow to be used, the way what we allow to be used, or perhaps instead of that, we need to use fire engineering and engineering judgment or modeling to, to extrapolate, because again, in the world of rapid innovation, we face these things all the time and and they come quicker and quicker new solutions that escape existing regulatory regimes yeah, I, I mean you're right.

00:27:24.805 --> 00:27:28.756
You're rudely perfectly right about the facades and the SPI test.

00:27:28.825 --> 00:27:43.794
The SPI test was developed for hires in rooms and was validating against the reference scenario and the reference scenarios in Turnaland was deemed to be good enough to represent real buildings, so you could go that way.

00:27:43.794 --> 00:27:47.204
Facades was seen during that time as another scenario to be good enough to represent real buildings, so you could go that way.

00:27:47.204 --> 00:27:52.277
Facades was seen during that time as another scenario that SPI was not useful for and it's not useful for that.

00:27:52.277 --> 00:27:56.635
So if you use it for facades, you really don't know what you're doing.

00:27:56.635 --> 00:28:03.172
So I guess that's why you're getting a facade test now, sort of 20 years too late.

00:28:03.172 --> 00:28:15.525
But it's important I think to it's about education as well to explain the systems, what is it good for and what are the limits, and nothing of this is absolute.

00:28:15.525 --> 00:28:21.048
Fire safety, you know, is a moving target, as you're saying, and you should.

00:28:21.048 --> 00:28:23.999
I think, as you say, it's a very good idea.

00:28:23.999 --> 00:28:35.807
If you can put together modeling and testing the standards of the modeling world and amalgamate them, that will be very fruitful.

00:28:35.867 --> 00:28:40.038
I think, Now there are different worlds to a large extent, which is not too good.

00:28:40.944 --> 00:28:49.085
You've mentioned validation multiple times, so perhaps let's talk about what are the features of a really good, successful standard.

00:28:49.085 --> 00:28:59.857
If you had to define what makes a standard successful, what would be the key characteristics of a good standard?

00:29:01.200 --> 00:29:02.580
Right, that's a long list.

00:29:02.580 --> 00:29:13.930
Talking, then, about product standards RH5, for the European system, and I worked a lot with that, you know, together with the regulators and the markets and everything.

00:29:13.930 --> 00:29:26.566
And the first thing you know is there must be a demand for the standard, there must be a regulatory demand, there must be a market demand, something like that Regulatories.

00:29:26.586 --> 00:29:30.257
There are regulators, but there are also insurance companies and so on.

00:29:30.257 --> 00:29:37.333
The UL has a number of standards, for example SM as well, and SM, by the way, they are doing this.

00:29:37.333 --> 00:29:46.097
They are using modeling, together with their testing, to handle their product, which is the insurance, which is very clever.

00:29:46.097 --> 00:29:47.339
So you need that.

00:29:47.339 --> 00:30:03.513
And then you say, okay, then I'm going to make kind of standard and I need to validate that standard, because you know there are thousands of standards out there and some of them, a fair amount of them, are not validated.

00:30:03.513 --> 00:30:11.613
They are just sort of, you know, you put a booms and booms into something and then you say 30% must be bad, 30% must be good and then we have new jobs.

00:30:13.788 --> 00:30:19.571
And you move to evaluate, which is, you know, a science by its own because the real life is extremely complicated.

00:30:19.571 --> 00:30:37.199
So how it's done for linings and pipe insulation and cables in the European system and floorings as well, the European system correction, I used to say I have this reference scenario which, for the SBI, is the rule of fire.

00:30:37.199 --> 00:30:48.192
So I make large gauge experiments in my rule of fire and I say reference scenario represents the real world, which of course is not true, but it's you know, it's good enough.

00:30:48.192 --> 00:30:50.057
So you say that is good enough.

00:30:50.057 --> 00:30:53.372
And then I validate my test against that.

00:30:53.372 --> 00:31:02.012
And you know, if I have an ignition test, I say my scenario, my fire scenario, is, you know, the cattle standing under a cut.

00:31:02.012 --> 00:31:08.430
That's my risk scenario, so I'll make a test for that, and so on and so forth.

00:31:09.184 --> 00:31:20.285
So validation in here would be making sure that the standard really represents what it's supposed to represent, be it a real-life scenario or some specific objective.

00:31:20.625 --> 00:31:24.455
Yeah, you need to be able to explain what you're trying to do.

00:31:25.065 --> 00:31:35.515
This is also important, you know, for preventing the things like using the standard in an incorrect way, because if you had it well validated, it should not be used in an incorrect scenario.

00:31:35.515 --> 00:31:36.837
Exactly, exactly.

00:31:36.998 --> 00:31:38.047
Exactly so.

00:31:38.047 --> 00:31:38.951
It's very important.

00:31:38.951 --> 00:31:43.757
And I mean, if you take the facade, it's easy to see what you're talking about here.

00:31:43.757 --> 00:31:53.333
The validation is you build up the facade and then you test it, and it's not going to be perfect either, because there are variants of all things, but hopefully it's good enough.

00:31:53.333 --> 00:31:55.230
So that's your validation.

00:31:55.230 --> 00:32:03.953
And then you have this thing when this stand-up test doesn't have to work in the real world, which means it must be repeatable or reproducible.

00:32:04.744 --> 00:32:20.375
If you take the SPI, then there are what did I say about 60 laboratories that are notified and hundreds of them that are testing and putting out documents on the market in more than 30 countries all the time.

00:32:20.375 --> 00:32:28.969
Now, if your test is not repeatable or reproducible enough, the system will collapse very quickly and then you can do this.

00:32:28.969 --> 00:32:33.342
I mean, there are again standards for measuring repeatability and reproducibility.

00:32:33.342 --> 00:32:36.769
So you do that, you do round robins, a lot of work on that.

00:32:37.211 --> 00:32:52.309
The SPI number two round robins, I think, was extremely Is it something that's an ongoing effort, like it happens also also, or you do it just when you develop the standard and then you say, okay, it's fine, or you just you know repeat the round robins every five years or so.

00:32:52.309 --> 00:32:54.895
I think eagles is doing that right yeah, you do.

00:32:55.056 --> 00:33:01.476
You do yes, because your accreditation requires you to do this kind of work constantly.

00:33:01.476 --> 00:33:04.229
And also, you know, do the things like?

00:33:04.229 --> 00:33:10.133
For example, something called the Group of Multified Bodies under the commission that meets.

00:33:10.133 --> 00:33:19.413
You know the system that meets to talk about how do we run this system and what is the problem that has occurred now and how do we handle that on a European basis.

00:33:19.413 --> 00:33:32.855
And then there is another one that is EGOR, a laboratories organization that do technical work with the standards all the time, improve the standard, and then the standards come back and you improve them, and so on and so forth.

00:33:32.855 --> 00:33:34.888
You look at the SPI standard, for example.

00:33:34.888 --> 00:33:38.636
It's 100 pages of instructions how to test.

00:33:38.636 --> 00:33:58.055
You have to do work all the time there, because that is extremely important, because if a standard is not reproducible, you know people will discover that you get different results in different cultures and different laboratories and things will be problematic, but I mean it must be difficult to do it in a perfect way.

00:33:59.165 --> 00:34:02.214
How do you accommodate for little differences that can be?

00:34:02.214 --> 00:34:09.927
Is there a range of, are there reference scenarios and there's a range of outcomes in the calibration test, for example, that you can have?

00:34:09.927 --> 00:34:18.996
That tells you okay, this apparatus built in the new laboratory represents the 60 apparatus that already exist across Europe.

00:34:18.996 --> 00:34:25.851
Or how does one building a new laboratory, assure that they become one, Because they were not part of a round-robin, for example?

00:34:26.414 --> 00:34:26.815
Oh, they do.

00:34:26.815 --> 00:34:36.452
But first of all, the standard is extremely detailed on how the apparatus should look alike, and there are manufacturers in Europe.

00:34:36.452 --> 00:34:41.257
It's just a few of them two, three or whatever that delivers.

00:34:41.257 --> 00:34:45.896
So you have some sort of ability there that you get the same similar apparatus.

00:34:45.896 --> 00:34:55.929
At the beginning of the SBI there was one manufacturer doing apparatuses for those laboratories, and how do you do it?

00:34:57.951 --> 00:35:15.396
Well, during your accreditation procedure, you have to show that you are able to do this test according to the standard, that you have a quality assurance system that could include things like that reference testing within other laboratories, and so on and so forth.

00:35:15.784 --> 00:35:39.338
There is the accreditation procedure that will go in the level of details that allows to recreation of the test method, unambiguity in the description of the procedure, the detailing that allows someone to take this standard and, from start to the end, just follow the instructions there and obtain the test.

00:35:39.338 --> 00:35:41.690
I guess that's also a characteristic of a good standard.

00:35:41.690 --> 00:35:47.753
It should provide all of that, because if it does not, then it's a challenge for this reproducibility right.

00:35:48.275 --> 00:35:49.077
No, that's right.

00:35:49.077 --> 00:36:00.896
But that's also why you need organizations like EGOR that are continuously working with these things, because they are developing, there are new products, there are new programs, there are new, you know.

00:36:00.896 --> 00:36:12.512
So you have to have a system that is continuously working, improving, changing the standard, doing new round robins and so on, have to sort of take care of the system.

00:36:13.867 --> 00:36:16.215
How about the market impact?

00:36:16.215 --> 00:36:28.115
Because I also that's something I took from the interview with Rudolf he was mentioning there was a political will to do that and we were discussing the SBI and Euroclass system.

00:36:28.115 --> 00:36:37.934
I was wondering, like, to what extent the political will should be, you know, something shaping a technical system, but I guess the impact on the market.

00:36:37.934 --> 00:36:42.548
The market must need the standard and the standard must not destroy the market at the same time.

00:36:42.548 --> 00:36:45.594
Right, it's some sort of symbiosis.

00:36:46.266 --> 00:36:47.391
Yeah, it's a symbiosis.

00:36:47.391 --> 00:36:55.853
I learned when I was working with this, creating the system, that you need three partners at least.

00:36:55.853 --> 00:36:59.235
You need the regulators, because they will ask for it.

00:36:59.235 --> 00:37:01.353
That will require the demand.

00:37:01.353 --> 00:37:02.027
We need this.

00:37:02.027 --> 00:37:03.431
You know which was.

00:37:03.431 --> 00:37:04.775
You know the commission and the demand.

00:37:04.775 --> 00:37:04.804
We need this.

00:37:04.804 --> 00:37:05.014
You know which was.

00:37:05.014 --> 00:37:06.389
You know commission and the sky and the regulators.

00:37:06.389 --> 00:37:08.447
And you need the industry.

00:37:08.447 --> 00:37:29.186
You know, because they have, they know the rules of their products and also and they have also technical capabilities, they know their products and you need the technical people, the laboratories, and if you want to create a large system that will be successful, I think you need the three of these partners At least.

00:37:29.186 --> 00:37:37.056
You need, I should add, also you need the science, you need the modeling, you need that sort of thing, the understanding of high dynamics, of course.

00:37:37.056 --> 00:37:39.692
So not alone.

00:37:39.692 --> 00:37:42.289
You cannot be alone here representing the product.

00:37:42.289 --> 00:37:43.692
You cannot be alone here representing the world.

00:37:43.711 --> 00:37:46.275
Yeah, I guess it would be a challenging pathway.

00:37:46.275 --> 00:37:59.414
If I stand here today, draw a new way of testing, let's say facades on a piece of paper and then say to the world this is the new standard, that would be quite challenging to build the support with.

00:37:59.414 --> 00:38:16.811
I observed the development of European standards for facades and how much work that is to actually create a new regulatory regime standard that has all the features that you mentioned that will be widely accepted.

00:38:16.811 --> 00:38:20.570
And actually I worry it will not be widely accepted.

00:38:20.570 --> 00:38:24.646
That's a worry for me because I think that it will be too costly.

00:38:24.646 --> 00:38:33.367
Like, okay, my perspective as a research laboratory, I I would take it that's a that's a good business opportunity, right.

00:38:33.367 --> 00:38:40.809
So I'm I'm happy to deliver tests of this magnitude, complexity and at this budgets that we look at.

00:38:41.329 --> 00:38:52.311
But I also know that clients will perhaps do everything in their power, you know, to not do that test more often than they really absolutely necessarily need, because it's just going to be super expensive.

00:38:52.311 --> 00:38:58.288
So I wonder, like is this something that also should be considered when you develop a standard?

00:38:58.288 --> 00:39:01.351
Like, what will be the cost?

00:39:01.351 --> 00:39:07.561
Like, will the cost justify the gains you get from the regulatory system?

00:39:07.561 --> 00:39:11.516
Of course, if it is your ticket to sell on the market, then that's the only way.

00:39:11.516 --> 00:39:16.295
It's not that you have a choice, but I wonder if we should also consider that.

00:39:16.335 --> 00:39:22.981
Yeah, you have to, and it has to be acceptable by the authorities and it has to be acceptable by the market, by the authorities, and it has to be acceptable by the market.

00:39:22.981 --> 00:39:29.867
As you say, if the market says we will not have this large scale test, we can't have it, and so on and so forth.

00:39:29.867 --> 00:39:47.393
You need to either have a very strong regulator, like the European Union that you know you make mandatory, everything mandatory, and then the politicians agree, the European agree to it, and part of that is the market and you need to have acceptability there.

00:39:47.393 --> 00:40:07.898
It cannot be too complicated and too expensive either, but it has to do a job and if you look at it in general, the test also must be able to test all the building products according to those test score, which is another thing that is quite difficult.

00:40:07.898 --> 00:40:13.561
And what about the time factor, the time characteristic, how long it takes to build a new standard from scratch?

00:40:13.561 --> 00:40:16.402
You said it starts with the demand.

00:40:16.402 --> 00:40:19.715
So let's say there is an indication that demand will be there.

00:40:19.715 --> 00:40:21.632
We need a new standard on X, Y, Z.

00:40:21.632 --> 00:40:26.916
How long is the process to get it done realistically?

00:40:28.045 --> 00:40:29.065
That's a good question.

00:40:29.065 --> 00:40:31.128
I should take a complicated thing.

00:40:31.128 --> 00:40:43.001
I guess the work on the European system for reaction to fire was on for, let's say, 10 years before there was a decision by the commission.

00:40:43.001 --> 00:40:47.626
This is how it's going to be and they're going to work with it, so it can be.

00:40:47.626 --> 00:40:53.952
It can be a very long process Depends on, you know, the standardization organizations.

00:40:53.952 --> 00:41:05.869
Then can be very slow, Depends what the funding is, what the demand is and what the you know your discussion with industry can be very lengthy, so it's a long process.

00:41:05.869 --> 00:41:09.835
There are a lot of countries that had to agree, so it's not quick.

00:41:11.748 --> 00:41:13.855
And what happens in between you know.

00:41:13.855 --> 00:41:24.358
So if you define there's a demand, it means there's a real need, that we need some specific characteristics to design fire solutions.

00:41:24.358 --> 00:41:28.670
Right, and standard is not yet there and it perhaps will be out there.

00:41:28.670 --> 00:41:33.949
If I we're really pushing the facades of people, sorry guys, but it's.

00:41:33.949 --> 00:41:40.186
It's a nice example how long we are talking about the facade standard, at least since grenfell.

00:41:40.186 --> 00:41:40.847
You know it's very.

00:41:40.847 --> 00:41:48.003
It happened very shortly after Grenfell that we started looking into European regulatory regime and Grenfell is already what?

00:41:48.262 --> 00:41:53.574
eight years ago, seven years ago long time ago and and the standard is still way ahead.

00:41:53.574 --> 00:41:56.286
And yet every day we need to design new facades.

00:41:56.286 --> 00:42:01.838
So how can engineers cope in the absence of the design of the standard?

00:42:01.838 --> 00:42:04.204
And in this intermediate period, right, because the market exists, the users exist, it's just the standard.

00:42:04.204 --> 00:42:08.911
And in this intermediate period, right, because the market exists, the users exist, it's just the standard that it's not there yet.

00:42:10.086 --> 00:42:11.331
Now, right, what you need to?

00:42:11.331 --> 00:42:17.831
You need to use things like you know, fire safety engineering, the national test that you have.

00:42:17.831 --> 00:42:27.152
You have a national test in Poland, you have a national test in the Nordic countries, you have a national test in the UKic countries and you have a national test in the UK since many years on croissants.

00:42:27.152 --> 00:42:31.434
So the demand is also a driving force.

00:42:31.434 --> 00:42:44.148
I mean, if the demand is very high, like you know again for ships the fire in the Scandinavian Star your regulations can leak it because people realize we need something here now.

00:42:44.148 --> 00:42:54.054
But you know you really need the fire science and the fire safety engineering to go in there, work together with standards.

00:42:55.405 --> 00:43:33.125
And another feature of standard that comes to my mind if eventually we figure out that some part of the standard is not really working as intended, how much effort did this to fix a standard, you know, to change a substantial thing in the standard To, I don't know, change the criteria of assessment or to respond to a new need, you know, because it's also like you cannot expect that there's a standard done in like 1970 and expect that in 2024, it will still, you know, be representative for the demand that's created in the 70s at the first hand.

00:43:33.125 --> 00:43:40.918
So how do those standards evolve and respond to the new needs that appear while they're already, you know, in power?

00:43:40.938 --> 00:43:42.802
already in power.

00:43:42.802 --> 00:43:55.150
The system is the standardization process, and you could add things like the group of modified bodies could produce a document saying for this kind of product, this is the way we are going to test it, and so on.

00:43:55.150 --> 00:44:02.670
There is a process, and sometimes very slow, and sometimes it can be fast.

00:44:02.670 --> 00:44:04.813
Then again, it depends on the demand.

00:44:04.813 --> 00:44:13.659
It's a negotiation between 30 countries when you do something like that, which can be extremely slow or fast?

00:44:13.719 --> 00:44:18.054
can it define slow or fast in years, like three years would be slow or fast?

00:44:19.215 --> 00:44:20.766
I don't know, I wouldn't say I would.

00:44:20.766 --> 00:44:33.434
I mean, I mean if, if your problem is very large, I mean this can go, you know, within a year or so the problem is, you know, not extremely large.

00:44:33.434 --> 00:44:35.871
It could take forever you have everything there.

00:44:36.766 --> 00:44:41.652
The last thing I wanted to ask is how does this work with performance-based design?

00:44:41.652 --> 00:44:43.735
How does this work with performance-based design?

00:44:43.735 --> 00:45:00.811
So I assume that in the prescriptive world, the role of the regulator would be to use those characteristics of a product in order to say, okay, only product of this specific characteristic is allowed and anything less is not allowed.

00:45:00.811 --> 00:45:04.771
Right, it's just like a gate In performance-based design.

00:45:04.771 --> 00:45:10.173
If I want to design something, I would perhaps need a different sort of data.

00:45:10.173 --> 00:45:14.070
Perhaps I don't need a class, Maybe I need heat of combustion.

00:45:14.070 --> 00:45:19.052
You know, Maybe I'm not interested in information if it's incombustible or not.

00:45:19.052 --> 00:45:35.211
I perhaps would just like the value, because it allows me to engineer with and for any other product, for any other characteristic, there's a continuity of stuff that you measure and then there are brackets in which you put the product based on the results of the test.

00:45:35.211 --> 00:45:45.846
How do you feel this works with the performance-based design, this standardization regime and use of the standards tests?

00:45:46.967 --> 00:45:49.432
Well, the standard is not the regulation.

00:45:49.432 --> 00:45:51.777
You're building code.

00:45:51.777 --> 00:45:59.344
In your country you're building regulation can be performance-based, like the Swedish-speaking building regulations.

00:45:59.344 --> 00:46:01.550
So you're not obliged.

00:46:01.550 --> 00:46:04.436
There are no really gates.

00:46:04.436 --> 00:46:13.159
In that way you could do whatever you like, provided that you sort of can stand for it and do the proper fire safety engineering work and so on.

00:46:13.159 --> 00:46:20.458
So you have that mechanism and then the standards can be helpful because they declare a certain property of the product.

00:46:20.458 --> 00:46:22.338
If it goes into that context, it's going to be helpful.

00:46:22.338 --> 00:46:24.639
Property of the product If it goes into that context, it's going to be helpful.

00:46:24.639 --> 00:46:42.972
If you have a prescriptive regulation, you tend to be in trouble for the developer because then you have to add these gates all the time and cannot build a spherical building because your building code assumes that buildings are not spherical, that sort of thing.

00:46:43.744 --> 00:46:53.896
Is this something that, when designing a new standard, you look into Like in what kind of regulatory systems it will be used and how will it work with those regulatory systems?

00:46:53.896 --> 00:47:01.746
Or is it something you don't care about and it's up to the regulators to figure out a way to incorporate the standard in their own?

00:47:01.786 --> 00:47:03.409
system.

00:47:03.409 --> 00:47:11.675
Ideally you don't care about how it's going to be used, provided that it's validated and you can explain what good it makes.

00:47:11.675 --> 00:47:18.452
And then it's up to the regulator to use the standard or not, like the ISO standards that are voluntary.

00:47:18.452 --> 00:47:26.414
So people that use these ISO standards, they sort of pick them some of them and use them, like IMO is doing for shipping.

00:47:26.414 --> 00:47:47.257
And on the other hand, when you're making the European system, you have regulations around Europe that were very different, very different tests giving very different results, and so you have to have a dialogue with the regulator because he would need something that he can use.

00:47:47.257 --> 00:47:47.579
You know.

00:47:47.579 --> 00:47:55.190
But on the other hand, you're not constructing a standard to fit a regulation that will not work very well.

00:47:55.190 --> 00:48:02.409

152 - Why we need good standards with Björn Sundström

Listen On

Fire science, engineering and education Episodes

Recent Episodes