251 - Occupant loads in Car Parks with Mike Spearpoint

Wojciech: 0:00

Hello, everybody. Welcome to the "Fire Science" show, episode 251, 25% into I've set a goal of 1,000 at some point. Let's keep going towards that. In today's episode, more fire science coming your way, and I have a special guest, Dr. Mike Spearpoint from OFR. And we will be uncovering a, a piece of research that, uh, been partially involved in. Uh, the part that we discuss today is, is purely Mike's Uh, we were looking into car parks. We were looking into figuring out a number of adequate period for a car park. That's a story for perhaps a different episode. In today's episode, we focus on one nuance that was a part this study, and that was figuring out the occupation load of a car park that we are investigating. So a really simple question, how many people could be a car park? And, uh, how reasonable that number is in the broader, uh, safety risk analysis or fire safety assessment. You could argue, "Yeah, of course we know the number. You just put two people per parking space, and you're done it." Well, why not five? Why not one? Uh, how do we justify that number? Are people always in the car park? If not, how long they are there? How many people can be concurrently in the car park, you Once you start digging deeper, you get into very spaces, and as you can imagine, Mike, uh, got into very spaces that we will very much discuss in this podcast When you listen to this conversation, I would ask you to not just look at the question of the occupant load or the question of how many people are there in the car park. Yes, this is important. That's the point of the study. But I would like you to follow the methodology or, or kind the way of thinking that Mike presents in here, how we dig how we uncover more of the question at hand. Like how we go into this rabbit hole and study phenomena at a level which really allows us to, to get more complete answer to the question, and that's something I think we should in fire science. I mean, not the occupant load studies. Perhaps we need more of those, but more rational analysis, understanding how buildings work, and more linking of- features of the building to the real-world, uh, use case in which the fires in those buildings may happen. Um, I hope I built a little bit of curiosity on, of what find in this episode. Uh, I had a blast talking with Mike, like always, and I can recommend you sit with us and enjoy the conversation. Let's spin the intro and jump into the episode. The Fire Science Show podcast is brought to you in with OFR Consultants. OFR is the UK's leading independent multi-award winning fire engineering consultancy with a reputation for delivering safety driven solutions. we've been on this journey together for three years so far, and here it begins the fourth year of collaboration between the Science Show and the OFR. So far, we've brought you more than 150 episodes, Which into nearly 150 hours of educational content available, free, accessible, all over the planet without any paywalls or hidden agendas. This makes me very proud and I am super thankful to OFR for long lasting partnership. I'm extremely happy that we've just started the year four, I hope there will be many years after that to come So big OFR for your support to the Fire Science show and the to the fire safety community at large that we can And for you, the listener, if you would like to learn more or perhaps even become a part of OR, they always have awaiting. Check their website@orconsultants.com And now to the episode. Hello, everybody. I am joined today by Mike Spearpoint once again. Hey, Mike. Welcome back to the podcast.

Mike Spearpoint: 4:49

Uh, thanks for having me back again.

Wojciech: 4:50

You're always welcome, Mike. Your ep-episodes are breaking the popularity rankings. I should just do episodes with you, I think. There's no comment on that.

Mike Spearpoint: 5:02

I'm sure people don't wanna hear too much from me.

Wojciech: 5:05

Well, uh, Mike, we have done a project, or perhaps rephrase uh, WIFOR has done a project and invited us to be a part as the ATB on, uh, providing the adequate fire for open-sided car parks, uh, for, for guiding, uh, the changes in, in, the ADB, I believe. and you've done an very, very interesting and in-in-depth of that, uh, work, which was, uh, evaluating the occupancy in the car parks, and I would love to talk about this today, both in the sense of, uh, how we fill up the car parks with humans and how many people could be in a car park and what actually changes when this number changes. But also, you know, in a kind of philosophical way of how do we evaluate that in fire safety engineering and how big part of fire safety engineering it is and what it should be. Wanna start up with the philosophical part, perhaps?

Mike Spearpoint: 6:07

yeah, I think that's a, a reasonable place to then that will lead us on to the, the specific, part of this project, the, the specific question we're covering in the, in the project. So, I mean, when we think about doing some kind of, analysis, I suppose that's the kind of classic way we might use, um, this question of how many people do we expect to be in a building, which we wanna put into calculation, into our model, then there's sort of how, how do we approach that? And I guess there's, you know, there's a couple of doing it. One way might be do some kind of survey or something of an existing building. But I suspect for most people, the approach is to go their document of choice. It could be the SFP handbook or their local codes or and then there is often a table of some sort Um, density, floor space factor or, or something like and it will have a certain number of people per unit or a unit area per person, and you would identify each space, activity space, pick the number off the table put it into your design and populate your building in that way. I mean, there's quite a few questions about where numbers come from, what do they represent in terms of some sort of statistical representation. I think the numbers come from probably when we started to industrialize society and we had questions about, much ventilation should we provide for offices and how much lighting we should require. And my guess is that some of those numbers then got adopted in fire safety engineering. But clearly those numbers must represent some sort of, um, part of a distribution. Are they average numbers? Yeah. But- Are they some percentile?

Wojciech: 8:03

That- that's interesting Difficult to know like, i- in we also have a table with exactly what you said, uh, square per person in a building, and, uh, those are written directly in our technical requirements for buildings. So it's kind of an act of law that you have followed. It's not voluntary. And th- those would be very high numbers. So like for a shopping mall, for example, I would have four meter per person of a shopping mall. And that, you know, creates a quite an interesting, uh, because suddenly I have four square meter per person in shop, and at the same, you know, four square meter per person in the whole mall area, four square meter per person on stair, four square meter per person in every toilet corridor. You know, every meter weighs the same in, in the building, and it's not very uncommon for me to end up twenty, twenty-five thousand people, uh, inside the mall for my evacuation analysis, and this obviously guides widths of escape routes, escape doors, et cetera. And then I think, okay, that's probably unreasonable. I, I have no idea what it represents, like you said. I-- Just last week, I had, uh, a promo at the local mall where they were giving like every shop had like second item for percent price. And I believe that number was there present in the building. Like there was definitely four square meter per person on that day at that hour of the sale. But the year has 365 days, and, uh, that sale happens once a year. Is this really the point for which I should do all my fire Could you call it like the worst credible scenario? I, I think people like that name in, in- Yeah fire safety.

Mike Spearpoint: 9:55

I mean, that's one way to look at it. And so, so, so there's a couple of things there. Firstly, depending on which code or which philosophy using, you might decide that you said there you have in the stairs, people in the shops, people in the Now, some would say you wouldn't have all those things f-full at the same time, that the people in the shops are be, gonna be the same people going to the right? So some codes will allow you to say you don't need to occupy the toilets or the corridors or something you're expecting those people to either be in one or the other and not be concurrently occupied. But you're right, there's a temporal question there buildings, through the day, through the week, depending on the type of building, through the year, will in the number of people. So these sort of questions of, of sort of... I mean, th-these, and these are not new, new questions that, that, you know, somehow I-I'm claiming I've of thought myself. There's... I mean, there's some work. I mean, look at office. So, so you look at offices, myself and Charlie Hopkin probably about 10 years ago started looking at this a little bit, but there's work by Jim Milky and Caro, think it is, and, uh, on offices, and there's, I NFPA or, or you can find documents from the 1920s or a bit earlier where it's got some numbers for offices. I've got this 100 square feet a person or something that. So these things are, um, are not, you know, not kind uh, new questions. And the same with the, the shopping mall question. Uh, again, myself and Charlie, looked into that. Um, there's the work by Gianluca, Gianluca De in Switzerland who- Mm who did some work on that as And we did some simulations using, um, my little model. There's some work that we use from, Hawken Francis and his student, So, you know, th-this is not the only... You know, we're not the only people, uh, who are looking at this question for various types of occupancies. As I said, some of this comes when I was writing my, what it's worth, network model, and I'm wanting to people, um, in terms of putting a probability or a n- a number distribution- Mm within each space, then thinking how those people are, are going to be in different spaces. So kind of that leads on to car parks, right? So when we think about a building which have a car it might be a mixed-use building, it might be Well, we're talking about open-sided car pa- a car park next to the building, some retail on the ground It might be a residential building, and that car might be somewhat for residents and somewhat for using the retail. Then either you might think, "Well, those people are in the car park using their car, or off doing else." Yeah. So even with the building and the car park for the you wouldn't populate the car park with the The res- the- with the residents because they're either one place or another. And there was other work that, again, that Charlie did as part, uh, related to his PhD as well, where, we looked at, uh, numbers of people in residential and we used the English Housing Survey, looked at types of, single family dwellings, multi-family, of flats. And you can either do it per bedroom s- sort of or you can do it per square meter. So, so the car park stuff is kind of an extension of the office work that people have done, the retail work myself and Charlie and- Mm independently Jean Luc has done, and then there's the, uh, residential type

Wojciech: 13:40

W- w- wait. I'll, I'll stop you before we got there because I'll drop one more interesting, uh, twist when we're talking about loads, and that's for me a railroad. Railroad tunnels, railroad stations. I'm designing currently a large ra- railroad station, and one thing that's expected by the client is that we take by far largest train that is allowed to travel through a tunnel, is apparently 400 meters long. That's pretty long passenger train, if you ask me. And they want Because i- of the interoperability directive of the European Union, you technically should allow any train in Europe on that tracks in that tunnel. Like, you cannot forbid,"Oh, yeah, this, this train doesn't look like it can go through." So if it's allowed as a train in France, it should be able to access Poland. So we end up with some crazy déjà vu. I think it's Alstom Model M, if I'm not wrong. It's basically a double-decker train, which is 400 It's like two trains next to each other, and that one has people 1,460 people in it. And I'm, you know, supposed to design my evacuation like, one meter wide, pathways in a tunnel because you really make larger because that you know, doubles the size If I wanted to, to have, like, five minutes evacuation, I have to build, like, three meter wide, passages next to it for a train that I don't even know it will ever enter Poland, for the number of occupants that can potentially maximally it. we were designing a metro system, uh, that was not, uh, the double-decker, it was just a metro, uh, line, but there was 800 people in one train. There was a station, so there was one train on one platform, second train on the second platform, and also 1,600 people on the platform itself, uh, waiting for next two trains. So suddenly I have 3,000 people in the station. Like, you get into really ridiculous scenarios if you just to go the simplest way. Oh, what's the maximum biggest number we can put in that And I think in a lot of engineering tasks, that's where heading with this evacuation modeling. Like, what's the biggest n- number we can justify, and let's just cross-check for that one.

Mike Spearpoint: 16:03

and you know, I m- and, and that might be the case and, and you know, there's an element of what I was what's the reasonably worst case? I mean, we could think about the, the maximum number people, and I think in something I wrote somewhere, the maximum number is some... You can know the world record for the number of squashed into a telephone box or a Mini or Yeah. could go there you know, that's, that, that is the... But it's not a credible maximum.

Wojciech: 16:27

Yeah, yeah.

Mike Spearpoint: 16:28

That's an incredible maximum. So then the question is, what's the credible And someone might, for your railway example, might say,"Well, it's a very popular sports f- game, and it is The train is full with people." It could be. It might be standing room, you know.

Wojciech: 16:43

I, I'm, I'm not saying it cannot be. It is likely it's gonna be, but- Yeah.

Mike Spearpoint: 16:46

So, so they are quite relevant, discussions to have what, what is that credible maximum, um, for the occupancies you, you might be dealing with.

Wojciech: 17:00

car park is easy. How many parking spaces you have? 100. That's 100 vehicles. How many people fit into a vehicle on average? Five. Here you go, 500 people.

Mike Spearpoint: 17:10

That's right. That's one way you could do it. I mean, if you look at the current ADB guidance, it says you use two people per parking space. So that basically assumes that you've got every space with a vehicle in it, and concurrently two people in each vehicle, and that would be your of people you design to. it just feels like Quite a big number. And when I say that, it's quite... That's for any car park, right? So that's not... When you think about car parks, they're associated an activity.

Wojciech: 17:43

Yeah.

Mike Spearpoint: 17:43

An airport, a sports stadium, a shopping mall. Office. An office. So one might ask the question, okay, but does that for the different types of occupancies? and should, could we... I say should we. Could we, provide numbers that might be a bit more for different types of occupancies? And so that was part of this work, this question. and then, then you, then you get to the, the question such as, okay, so, how many people per vehicle? We might come back... We'll come back to that in a minute. And how many vehicles? So some of this question of, well, let's assume single parking space is full, um, relates back to work, um, that, Zaheer Tohir did for his PhD when we looking at design fires and looking at how many spaces might be full. And this, this is kind of relevant to the wider fire project in the sense of fire spread between Is every parking space full? Uh, what size vehicle is in each parking space? Are there spaces in between, uh, that mean not every parking space is full? So you might start ask those questions and talk from... If you want to do a Monte Carlo s-s-sampling, one would... Well, there's a couple of questions. How many vehicles are in the car park? Mm-hmm. And how many people are in each vehicle? and then that would allow you to sort of come up with some sort of distribution of if you know how many are in the car park as a function of time, and you know how many people are in the, i-in their vehicle, you can of that distribution that just assumes probability. You, you, you get data, right? So you can get data, uh, somewhat from the parking about how many vehicles are in a car park as a of time, and you can use that to sort of model the, number of vehicles in. You can get... There's quite a bit of information sort of in the places in the literature on how many people per There's data on, I found various papers where, uh, none do with fire. They're looking at, um, how many people are in vehicles for different trip activities. So if people are going to the office, people are out, are going out, um, to the shopping mall or

Wojciech: 20:08

how do they get the data?

Mike Spearpoint: 20:10

Uh, how do they get it?

Wojciech: 20:11

Yeah.

Mike Spearpoint: 20:12

I, I'd have to go back and read the reports, right?

Wojciech: 20:16

Are they spying on people?

Mike Spearpoint: 20:17

Who are they spying on people?

Wojciech: 20:18

Like you on the balconies.

Mike Spearpoint: 20:20

Well, well, I've done s- I will come back to me doing spying on car parks in a minute. So I, I, I have to say right now, I don't know how they got that data. But what you find is, and kind of not surprising, if look at people going to work, to the office, you you have one or two people per vehicle. Mm-hmm. And if you look at people going to some sort of family event or something, the number's bigger. So, so you can get an average number of people per Now, using averages for this sort of analysis, you might go, "Well, we want some sort of, you know, more numbers." So there is, um, a little bit of data out I found some data from, uh, Transport for London, they, they looked at the number of, uh, people in although they didn't break it by activity, but you get a distribution. Yeah. Obviously, you've got the driver, and you've got

Wojciech: 21:09

you've got- There's a, there, there's a quote in your I can read it out. Uh, the quote is, "Around six in 10 car driver trips are made alone without any passengers." So six in 10 without"A quarter include one passenger, and the remainder have or more passengers." So that's the TfL stat. Yeah.

Mike Spearpoint: 21:29

So from that you can w- And with using the averages using that, you can create some sort of approximate for different types of trip activities- Mm-hmm and put in some sort of distribution, you might say. So now, the first, one of the parts of the project was to then start combining the, the numbers of expect to be in the car park with the number of people, and run it as a Monte Carlo. Mm-hmm. a s- a distribution, and then you can pick percentile or a 95th percentile, whatever percentile want, and that would give you something, that you say is a, is starting to give you a sort of credible number of people in the car park on that

Wojciech: 22:12

But, but, but here you concurrently assume that the people in the cars which are in the car park.

Mike Spearpoint: 22:17

Yes. I mean, that's, so it's a conservative approach.

Wojciech: 22:20

that's- For a, for a car cinema, yes.

Mike Spearpoint: 22:25

Other activities. But it's a s- you know, it's a place to start, right?

Wojciech: 22:28

Yeah.

Mike Spearpoint: 22:28

So then I suppose, you know, that you might look at and go, "Well, that's all, that might seem a little conservative." so the question might be, well, when a'Cause we're asking the question here is, if a fire how many people are in the car park at the, when the starts who are potentially going to be, uh, put at Mm-hmm or, or, or subject to the hazard from the, from the car fire? And obviously they, they'll evacuate. So, so the next thing is just gonna go, "Well, I if I can at least get some sense of how many people have been in car parks when real fires happen." So can go and find Images of the, of the Luton car park Mm the Liverpool Echo fire, the Stavanger fire, and got pictures of the first car on fire, and I can tell you now that there isn't, two people per vehicle out of their cars- Mm at the same time, right? Now, they're airports. Obviously, there's a, you know, a transition there. The Liverpool Echo one was, related to an entertainment type, um, uh, venue, but there is not... I mean, the car parks from the pic- pictures are full, but there's hardly anyone there.

Wojciech: 23:39

Mm.

Mike Spearpoint: 23:39

So it either means there wasn't two pe- let's say two people per car, or if they were, they've all got out got away before the fires got very big.

Wojciech: 23:48

But, but, but this is very interesting. You bring up, uh, Liverpool. I would say Liverpool was a lucky shot because If you assume that the fire could have happened with the same probability any time, if it happened, like the moment the, the event has finished, you could have a l- awfully lot of people- Yeah in the car park.

Mike Spearpoint: 24:06

Yeah. Yeah.

Wojciech: 24:07

So I, I wonder, average value is one thing, but the curve at which the car park fills out and empties, like have you into that?

Mike Spearpoint: 24:16

So that, that's, that's what we'll, we'll, we can on to. So that's-- So we call that the, what we've been about so far, I call it the s- the static analysis, assuming that, that somewhat there's a fixed...

Wojciech: 24:30

and- I mean, this is a, a more, an engineered value with a better proof than just value of two from, uh, a guidance,

Mike Spearpoint: 24:38

Yeah. The, it's a, it's a first step to, to, to doing that. And, and again, uh, you know, we appreciate that, like you're saying, with some different activities, where an entertainment activity, where there's a crowd of at the end of an event, that car park might get quite full of people.

Wojciech: 24:55

Mm.

Mike Spearpoint: 24:56

And, but that will be different from an o- an office or different from an airport. So, so what we're trying to say is there's an element of, uh, of just maybe looking at the different of the car park or the associated functions and saying maybe, uh, there's a-- we can sharpen our pencil a bit and not use a single value for all different types of... or occupancy types or activity types. Maybe there's a bit more engineering analysis we can do. It might be, in doing this, we conclude that two per space is not enough for some car parks. You know, it doesn't mean we're always gonna go We might find for an entertainment value, uh, venue, when everyone leaves at the same time, and it's groups, that four pe- it's full of cars, and maybe people are getting in each car, and maybe that number's not high enough. So it's not about just making numbers lower. It is about making numbers, um, appropriate for the

Wojciech: 25:56

Yeah, yeah. So, so, uh, given those distributions of how many vehicles are, uh, how many person per vehicle, per trip, per specific occupancy, w-what did you get in terms of, of the percentiles? You said that you can get percentiles out of Monte Carlo.

Mike Spearpoint: 26:11

Yeah. So I, I looked at eightieth, ninety-fifth, and percentile, just, just making some choices. Uh, and I'm not gonna tell you all the numbers, so we'll look at the ninety-ninth percentile. So I broke it into, uh, three activities because it was more-- it wasn't possible to break it into others. So I broke it into commuting and business. Mm-hmm. In that case, I got one point one three, so that's per vehicle. For education and other, it was close to the two. I got one point nine, and for shopping, I got one five eight. So they're, they're the numbers that you end up with Well, residential... Well, there wasn't, um, it was difficult to say what you do with residential.

Wojciech: 26:54

Okay. because it's kind of a starting point of your travel

Mike Spearpoint: 26:58

starting point, right? Because it might be people are going to the office or they're, or they're coming back- Okay or they're going out. so I mean, I put-- I think I, I got an average on an which is about one point five two, and that might be a way of saying that might represent, uh, sort of a because the, the activity that people are coming or to is, is either... Is it might be one of these three other types. And, and it wasn't, it wasn't easy to, to sort of, um, uh, separate that out in terms of the analysis.

Wojciech: 27:29

Yeah. But you also did, uh, compare that with some, uh, real data, right?

Mike Spearpoint: 27:34

well, yeah, so the real world data. Um, we-we'll come back to that. So, so, so that's assuming that all of the vehicles kind of concurrently occupied, and there's a of, of vehicles in the car park. I mean, obviously the other way to do it would be to all the spaces are occupied, and you have a of people. Um, I think I did that as well. I'll have to go back and look through my tables. But the other way to think about a car park is, is to experiment. Imagine you've got a car park that's empty, and what is you've obviously got vehicles that come in. So a vehicle will come in, and it will have a certain number of people in it, and it will go and find a space. The people in that or person will get out and walk to the stairs or whatever. And so they have a sort of transition time. They, uh, there's like three phases. There's while they're in the car parking, while getting out and doing whatever they need to do to out of their vehicle, and then the transition vehicle to the place of relative safety. Could be outside, but in this case, if we're thinking

Wojciech: 28:40

Kind of the target destination.

Mike Spearpoint: 28:42

target destination. And obviously if people are coming back in, they do reverse. Mm-hmm. They walk to the car, they do what they're gonna do, then they leave. So, so then you can imagine that obviously in that car park there's gonna be a flow of vehicles coming in- at some rate with people, people in the vehicle. And obviously they're potentially, uh, if a fire I mean, they're in the car park. There's gonna be people doing their activity, putting, getting out their stuff, uh, put it in or And there's people walking to or from the, the target. So then you can, you can model that. each vehicle coming in.

Wojciech: 29:22

Yeah.

Mike Spearpoint: 29:22

Um, the time it takes for each person or group of to do their activity around the vehicle, and the it takes them to- Mm-hmm reach, reach the, the exit or the, the place of safety. In ways that's, the last bit is kind of the easiest, you might say, because you can use walking speeds

Wojciech: 29:39

kind of like evacuation analysis.

Mike Spearpoint: 29:41

It's an evacuation analysis, but people aren't They're, they're, they're doing their, doing their, daily activity. And of course, at some point a fire starts. So, so what we're doing is we're, we're just modeling the activity of the car park at the point when a fire starts. And at that instant we can say, "Right, how many people are in the car park?" Mm-hmm. Some of those people are gonna be in the car, some them are gonna be doing their getting in and getting activity, some are gonna be walking to or from the, the, the target, the, the stairs or the outside. And I say the, the s- the walking bit It's quite easy. I mean, we've got walking speeds, classic, you know, the literature. We could pick a reasonably slow walking speed to We can pick a distance they need to, need to car parks have certain sizes, and so you might you know, some sort of average or some kind of walking distance.'Cause obviously, some cars are gonna be near the some are gonna be far away, and you don't really want to start trying to model e-exactly each parking space. You could do, but, you know, there's only so many hours in the day to do analysis. So then the question of the other two is, is the coming in, the vehicles going out bit, and then the around the vehicle. So, so you can find data for the, the transition time of vehicles, um, how long on average it takes to travel in and out. So this, this is something of interest to parking operators. So there is data available that allows you to, to, to do that. And you... And I found at least one paper or two papers where the end they almost have a l- equation that says this is the time it takes And it de- and, and that depends on the, um, how full the car park is. So obviously, the first car that comes in, um, imagine your empty car park, doesn't take long for, to find parking space- Yes right? But as the car park gets more and more full, it takes longer to find a parking space. So you can relate the time to park to the utilization of the car park.

Wojciech: 31:46

the size of it, how many floors it is- Yeah how many spaces.

Mike Spearpoint: 31:48

the size it is, that sort of thing. So, so that, that's a function you can put into your'cause if you c- if you keep a track, sort of a running track of how many vehicles are in and out, then you y-you can use the equation. And so that gives you a time basis. Now, I've assumed that the time it takes for the car get in and find a space, and the time it takes for a to leave and get out are the same. Some might argue, well, maybe it's quicker to get out than it is to get in, but then there might be a barrier that people need to pay a ticket at and all that sort of thing. But again, a little conservatism says, "I, I'm gonna it takes just as long to get out." There might be a of a queue, right? Mm-hmm. So, so, so we can use that to, to then look at the it takes.

Wojciech: 32:31

Well, th-this is actually quite interesting that you mention queue because, two cases from, uh, from Poland, from, uh, world. One was a shopping mall, again, a different shopping mall, perhaps the biggest one in Warsaw, where there was, again, some event, uh, ridiculous amount of people in the mall. it was related to the marathon. They were giving out the marathon starting packages for Marathon, some sort of event like that. So you suddenly have, you know, thousands of people coming to the mall at the same time. And- They blocked not just the mall, but the streets the mall to the point where they locked out. Like, there are roundabouts, and the roundabouts Like, there was no way for cars to leave the round- and police had to clean out the whole district. Like, block the access to the district, try get the cars out just so the cars can exit the car park in which they were because they could not exit because the streets were locked the amount of cars that were trying to get in and out of the car. So that was a very interesting case because now you high utilization of a car park with the people actually stuck in their vehicles in the car park.

Mike Spearpoint: 33:41

Yes.

Wojciech: 33:41

So th- that's actually a case... But this is, like, once in a very, very long time event. I, I- Yeah uh, but I, I have another one, uh, which is quite common. Uh, we have some streets in the Warsaw, um, commercial offices, where basically 5:00 PM everyone tries to get out of their office with a car, and they all, you know, leave into street. And this inevitably creates a traffic jam where people spend for a long time in their vehicle, in that, in, in, in in that space. So now this travel time is kind of tied to the time of a day and the traffic outside. So if you, you know, go out at 4:00 PM or 6:00 PM, you're driving, like, as fast as you entered. But if you're going out 5:00 PM when everyone else leaves, a different trip.

Mike Spearpoint: 34:33

Yes. Uh, and remember here what, what we're trying to this particular calculation to is, is like a risk question, right? Yeah, yeah. So, so we can come up with a, a s- a case of I r- I a, a case, a car park where this, this thing So, so in that case there, that specific design- Mm but, but, you know, that, that would mean any of th- any the codes might fail if you can provide me a case

Wojciech: 34:58

says- No, I, I, I mean, We, we, we, with, um, anecdotal like this, what I want to say is that it's very, very specific. And if you understand what's the point of the occupancy, can start seeing the scenario. So if you, for example, give me two people per car, uh, per, parking spot in a residential building, ah, ah, that, probably never gonna happen. But in an office or a, or a shopping mall, I've actually seen that happen. Mm. So it's... I, I mean, and as soon as you realize what the building is to do and what the humans in the building are supposed to do, it kind of makes sense, and I think we We don't do a s- amount of rational analysis in fire safety engineering a, I don't want to go there.

Mike Spearpoint: 35:57

that's another discussion. Yeah. Um, so but anyway, let's go back to the- Yeah the, the

Wojciech: 36:02

at hand. You've done case studies.

Mike Spearpoint: 36:03

Well, well, here you go. So, so we've got the, we've got the walking in and

Wojciech: 36:07

Yeah.

Mike Spearpoint: 36:08

We've got a vehicle in and out. Obviously there's vehicles coming in and going out, how quickly is the turnover of vehicles? And you find things, like there's a, the IStructE says, I think it, it's a quarter, a quarter... Oh, no. A quarter of the car park vehicles transition every minutes or something like that. Hmm, Anyway, there's a, there's a transition rate'Cause obviously the question of, you know, this of vehicles coming in and vehicles going out, what's sort of the rate of vehicles going in and out? What's that, what's that, that, that turnover of And so you'll get various, numbers. And I say there is a sort of guidance value, where I it's, it's 25% every 15 minutes. so, so we can get that. So then, then the question comes up with, okay, so, so we've got that third stage with the people getting of their vehicle, and they might have activities. So they might be getting stuff out of the boot. They might have children. They might have their dog with them or whatever. And again, it partly depends on the, um, the type park, the activity, right? So again, you can imagine an office, people might be in and out quite reasonably quickly compared with a type of activity where there might be more complex going on. So I thought, well, my... You know, the, someone must have published some this. it might not be a fire, uh, but someone would've it. I spent quite a while digging through the and I couldn't find anything. Doesn't mean it doesn't exist, but at some point you go, "I, I, I just don't think there's anything that's gonna be useful." So what do you do? You think, "Well, I'll just have to measure m- measure this myself." And so then you think, "Well, how am I

Wojciech: 37:49

Oh. You went spying on car parks.

Mike Spearpoint: 37:52

Well, that's it, right? So then you realize that there's car parks with 24-hour cameras in it, and you think, "Well, that would be So I can get the camera, and I can watch the car The problem here is, is when you go to in- indoor car parks, the camera doesn't show you much, right? Because obviously- Yeah the car park's quite low, and you can- can't see much.

Wojciech: 38:14

it. Okay.

Mike Spearpoint: 38:15

But I found a car park, uh, outside and in a seaside in Norfolk, which I, I... It's written Hunstanton But I think it's, I think pronounced Hunstanton. Some- someone told me that it's there. So, so this car park looks over, or this There's a that looks over a car park with something like 25 as I remember. Um, now it is outside, right? So someone would say, "Well, anyway, you're using

Wojciech: 38:44

But- Well, human not car park is the problem.

Mike Spearpoint: 38:47

Yeah. So but the advantage of this is, firstly, you a few spaces to look at. The other advantage is it's because it's on the seaside, people are turning up to do various activities, right? Which means you're gonna And those activities are to be, you know, sort of, um, recreational. Well, they're not in a hurry. They bring family members, they bring Often there was lots of pictures of people with them, and they're their dog in and out. um Well, you get a relatively diverse range of and you get activities that include families They get where, where one of the parents is getting pram or a pushchair out. Families with elderly people, which might have a frame or something else. So you're getting some nu- some data for range, ranges of activities. So basically what I did was I would log in. I'd, I, for, I did it for about two months. Most mornings I'd log in, and I'd watch it, and I'd a timer going. And every f- every car that came in- I would watch, and I'd have a timer going, and I would count the number people getting out, and I'd s- count how long it took them to do what they're doing, and I'd note it down. And of- and some of them, right, so the first few would turn up at, at 7:30 in the morning or whatever, were people clearly going off to work, right?'Cause someone would park their car, literally get close the door, and they were off, and it's like 10

Wojciech: 40:11

Yeah.

Mike Spearpoint: 40:11

Then as the day went on, and particularly the you would start to get family groups turning up. And I'd get all excited when I'd think, "Oh, look, a, this one's gonna have five people in it, and one of them's gonna have to get something out of the boot."

Wojciech: 40:25

As the 22nd.

Mike Spearpoint: 40:26

Yeah. Tw- well, some of them took minutes, right? Some pe- but then, but then you can create a of results, right? Yeah. So I get a number of people per each vehicle, the they did, how long they took to do it, and some notes and whatever. And of course, the other thing I could do was look at the number of vehicles going in and out. Now, it was a, it, it was a bit tricky, you know, when it got quite busy,'cause you'd have vehicles coming vehicles going out. So I'd have several timers running at the same time. Now, my eye would be flicking between them. Now and again, you'd miss someone getting in or out because you, because, you know, they were a bit away. now and again, because it was on the seaside, would get in with their fish and chips, and they would be in their car for 30 minutes or something really long as they'd be sitting eating their food. So at some point I had to made a judgment that, that at some point it's so long it's unlikely to represent an activity of someone getting, you know, a transitional activity. They would be getting in their car to look at the and, and that's a different activity. But you can take all that data, and you can process and you can get a distribution for the activity time. I did find another car park. This was in, um, in the US.

Wojciech: 41:47

Shreveport, Louisiana.

Mike Spearpoint: 41:48

Shreveport, Louis- Louisiana. This was outside a, a, um, a restaurant or something A mall.

Wojciech: 41:52

A restaurant, okay.

Mike Spearpoint: 41:53

Um, that one, the data was less interesting A lot of I think, I think it was a, they had takeaway. So a lot of it was people get... A person would get out of their car Mm-hmm go and pick up, and go and leave again. So, so the data there was, was obviously a different country, different location, whatever. But it, it wasn't, it didn't have that range of that the, um, the seaside one was. So, so I, I, I, I used the seaside one to allow me to do my statistical So I could put that into my model. So now I've got my model, my transition time of going in and out, my transition time of people with activities, and my walking transition time, and from I could then create this more, complicated or if we wanna be really hopeful about it, more approach to, to the analysis that says this represents analysis- Mm because it's all happening

Wojciech: 42:53

can- Yeah. And how this, this compare to the static one?

Mike Spearpoint: 42:56

from that, you can then create a table, and again, you can look at different percentiles. You can again split it s- by, uh, activity, um, in terms of the, um- Mm the trip activity. and now what I've done is applied the same seaside Um, I, I can apply the number of people in the vehicle, uh, but I can apply then the seaside data to all of So again, it's a little bit conservative 'cause so offices, you know, uh, the distribution is gonna be- a fairly long tail. So it more represents the education sort of But from when you do that, you end up with, kind of surprisingly A, a s-, a less on people per, per because you're, you're, you're expecting not all the

Wojciech: 43:45

at the same time There, there, you missed the simultaneous of it.

Mike Spearpoint: 43:47

Yes. So if we took the two persons per space approach for um, where all vehicles are occupied concurrently, drops to 0.6 people per- Mm space. Now, I've done it per space, not per vehicle, so I've basically averaged out and said to kind of match with the various codes and that, I've taken my analysis and then said, "But I know how many spaces there are," so I've, I've reported on a per space even though I've been working out per vehicle. Mm.'Cause that includes the fact that some are gonna be either, uh, vehicles occupying a not. So we had two people kind of similar, well, same ADB for holiday stroke, uh, education. That drops to 0.6. For, commuting and business, we went from about 1.2 to 0.4, and then for the other, which I called shopping, leisure, and, and actually kind of all car parks you go from 1.6 to 0.5, and they are 99th percentile. Obviously, if you took 98th, 95th percentile or No those number would pr- change again. So obviously from a, a, an AHJ sort of approach, they could, they could choose if they wanted to, do they a 99th percentile or do they want a 95th percentile or whatever, because you've got your cumulative curves, and you can pick any percentile you want. Obviously you can't, you know, up to 99.9 or 99.99 whatever number you, if you wanted to. It's- So, so you, you, you can, you know, uh, look at those numbers. Now, and obviously they are quite a bit lower than our original starting point. What was interesting was I, because I had the of vehicles going in and out, um, and how long they there, the, the guidance, the ISTRUCT-EVAT guidance what I got were, were quite close. I mean, that, that, that was quite interesting to see, um, that I was getting the same result for the, for turnover of vehicles, or the maximum turnover.'Cause obviously in that case, the car park in the early in the morning, would be quite empty, and car would turn up and someone would get out. So when you were look... And, and then during the day, that car park, the during lunchtime and all that would got really busy, there was always vehicles going in and out as people came and went. So that gave you a sort of the, the, the turnover of car park reached the numbers that you saw, I, I got from the, from, uh, a particular guidance document.

Wojciech: 46:18

Beautiful. the listeners, if you go to the report, which is linked in show notes, if you go to, to, uh, Appendix D, which is work WP3A Um, you will find all what Mike, uh, listed, including the table with, uh, what the, uh, people on the beach, uh, did and how long it took them to exit the vehicle. So it's, it's, it's, uh, quite an interesting read. But Mike, um, we're, we're running out of time, but I have, uh, uh, another question about the utilization because of course what you just said now was to get a much better informed number on what is the amount of people per parking space, uh, for purposes. But from the utilization, I know that this value was also used later in the fire spread modeling and in the structure and I know it was a significant factor on all, all of can you comment on, on, on, on how utilization translates to the, the, to the design fire perspective in a, in a short...

Mike Spearpoint: 47:23

so, uh, yes. So we had the question of the f- of the, uh, for the analysis in terms of the design fire and how fire would spread from vehicle to vehicle. Um, there was various ways that we did that.

Wojciech: 47:37

talk about. There will be a separate podcast episode on that- So- I'm

Mike Spearpoint: 47:39

I, I mean, I, I put together a, what you might call a hand calculation approach and compared it to a number of the sort of well-known car park fires we've about. Then, second way that we approached it with a, was a cellular automata sort of approach, where you can uh, you have a grid that represents your car park, and

Wojciech: 48:00

bits- It hops from car to car.

Mike Spearpoint: 48:01

H- yeah, hops from car to car. And that was, uh, uh, and Danny Hopkin, uh, sort of focused on that bit. And then we also had a, a, a CFD analysis that y- Wojciech, did. And we could compare the three, um, um- approaches. And, and, and so we could look at that in terms o- of rates of spread. Uh, we looked a little bit about how wind might spread in a, in a sort of predominant direction. And we looked at the question of if you've got uh, empty spaces, how that affects car park, um, fire development.'Cause, 'cause clearly, and again, this sort of back to Saher's work, if you have big e- far enough between vehicles Then it's more difficult to fire, to spread from one vehicle to another compared with they're- Mm all sort of packed together. So, that, that's, that's kind of the other part of it, the, the other part of, of the analysis. I mean, maybe you sort of see it as an A set, R set of thing. We've got the occupant side of it, and then we've the, the fire side of it, and obvious- uh, and, and you know, with the specific question about the f- what resistance do we, um... fire res- What's adequate fire resistance. Uh, that included a cost-benefit analysis, in terms of lifetime costs, lifetime benefits. So there was a lot more to that study, um, outside of this element. Now, we, we could have, you know, we could have done project by just adopting the A, the, the, uh, a value such as ADB, but we felt... I mean, you called it a bet- I'm not sure it's a

Wojciech: 49:39

A more complete.

Mike Spearpoint: 49:40

It's a more... It, yeah, it, it might be an improved analysis. Whether it's better or not, that's for- Yeah that's for the audience to decide.

Wojciech: 49:47

Well, o-one aspect of this utilization, because obviously what we found in our, uh, cellular automata or what, uh, what and his team found in cellular automata, uh, was that the utilization you had, the, the, the higher the, adequate fire resistance period is, which kind of translates into larger and higher fire load, if I may use those terms, uh, i-in this case, which is kind of obvious. More vehicles leads to, to larger fire. But o-one thing, I, I don't think we, we considered that, that could... That's an interesting research follow-up question. Like, if you understand the role of the building and the way how a car park works, you may have a car park where people to park at the nearest to the exit. If I'm going to a shopping mall, I always have a, a spot as to the exit as I can. Even if it just improves my travel time. I'm, I'm willing to spend five minutes in a car looking for a spot to save 20 seconds of walking time.

Mike Spearpoint: 50:46

see, I'm not. You are. See, I'm a negative. I'm the opposite of that. Well- I hate that. I... If I'm in a vehicle, I'm with someone, I say, just go to the top floor, and we park." I don't wanna spend 20 minutes trying to find a parking space to 20 seconds of walking. So I'm completely opposite. Right? Go as far away as possible, and I'll walk.

Wojciech: 51:04

I leave that to the listeners, which is more common in, society. But you're right. But, but they would clump, you know, closer to the exits. But when you have like a residential, when I know my parking spot is parking spot number 67, which drives my children for some reason I don't completely comprehend, I always park on that spot No matter if the spot next to me is free or not, you know? So this utilization of, let's say, 40% in a shopping mall lead to completely different, you know, spread of the within the car park versus utilization of 40% in a residential.

Mike Spearpoint: 51:38

Uh, yes. I mean, I, I remember this was something that Zaheer many years ago debated. Uh, in the end, it, you know, it's, it's another level of complexity, and I suppose, I mean, what I-- what done here, you might say, is a, um, you know, it's it's a certain level of complexity. Um, I'm sure within the, vehicle management parking there'll be some very highly complex parking model you could, you could, you know, simulate each and the probability it mo- it, it's either you in the vehicle or me in parks and different places. And obviously car parks have, you know, different They have disabled spots. They have places, like you say, nearer exits and places, if you're in a shopping mall, it might be where you've put your shopping cart. Um, and some people might be wanna get the, you know, get the shopping out and move that. So there's all sort of behavioral elements that, that will not idealize the way that the vehicles might themselves around the car park. Um, but that's another, yet another level of, of that you could add into this analysis if you had the, you know, the time and the energy, the money, whatever, to do that.

Wojciech: 52:47

I mean, I mean, the, the purposes of this episode were One is to get to the numbers which you, we've revealed in f- grand finale of your analysis, but also, you know, to, to way of thinking. If you want to s- if we want to, you know, switch from the worst case scenario that could potentially happen of its probability and design always for that. Some buildings, yeah, that's probably adequate, but many like in my tunnels or my railroads, railroad stations, I get to paradoxes of design which creates spaces for which I really deliver safety in those conditions. And yet at the same time, I know that the combination of of, you know, fire happening to the most, populated the worst location of a train tunnel and blocking the worst, uh, evacuation exit at the same time, that's quite low. I, I really appreciate it when we are able to switch into methods and talk about risks rather than, uh, those type of scenarios. And, you know, on the example of how many people can be in the car park, we, we've shown or you have shown, um, the way of thinking that leads you to, uh, perhaps a more complete

Mike Spearpoint: 54:12

or not- Yeah. I mean, it, it gives you a more informed answer to a, have a technical discussion with whoever That says, you know, you can provide that and then that becomes a societal, a risk-based question that says you to adopt any of the numbers I've come up with. You, if you don't, you know... They are there. There's a method, there are numbers, and that allows to table those in a discussion. It might be that w- a different decision is made, but it's an informed decision, and that's, uh, that, that can't be a bad thing in my eyes.

Wojciech: 54:41

a-and one more aspect of that which we have not discussed, uh, once you get into this more complete answer and you have sources, it's much easier to trace the uncertainty of the Because when you have, uh, two people per car park, like, the uncertainty if we don't even know where the number Of course, there, there must be someone who knows, but it's to track.

Mike Spearpoint: 55:04

Yeah, I think I tracked down a particular reference that two people. I mean, when you look up-- when you look at the data number of people per vehicle, uh, I've also found data that tracks it somewhat historically, and the of people per vehicle has gone down historically. Okay.

Wojciech: 55:24

Vehicles are more popular.

Mike Spearpoint: 55:25

Yeah, because it... I mean, this was-- I think this was, uh, I think it was a report done, uh, to do with, uh, done in Europe the EU to do with, vehicle exhaust carbon and that sort of thing. And so, so it was environmental, um, um, type thing. And, and I think they were saying in that report, have to dig it out, that it's gone down because, because people has, have got more cars per, per person. But, you know, so, so the number of people has, has down. Um, I found a, I found a really interesting paper way back, I don't know when it was. I don't know if it was even the 1950s or 1960s, of how many people were in vehicles going to US national parks. Okay. Um, and that was, that was actually quite a high, that was a quite a high number of, uh, people per, per You can imagine that people are going on holiday, on vacation and whatever, so you're gonna have a, a with everything i-in a... And you can imagine the, in those days, the v- the you can picture the vehicles in the 1950s and '60s are

Wojciech: 56:29

relatively. Our, our me- our memories are that you could fit much more into very small vehicle, uh, than today's. My parents' vehicle were probably half the size of mine, and we fit comfortably.

Mike Spearpoint: 56:41

Yeah, that's right. I mean, that, that is something when you look at data, of course, is the, the types of vehicles that have, uh, access to has changed in terms of their, um, yeah, their physical size, the number of seats, expectation. So, so you have to be, you know, somewhat careful with any data. And again, it might be somewhat regional, um, you know, international. Mm. That vehicles that, uh, that are used in one country or one area region might be different in terms of their I mean, again, you've got this idea, um, of the um, particularly in the US and, and North America in in the '60s and that, you've got the big Cadillacs Buicks and that sort of thing. I can imagine here in the UK, um, well, I know somewhat the vehicles that were, were, were, were a different of vehicle in terms of size, and therefore maybe the of people you could put in each of those would be So, so you would have to be careful just the data that you might get from both in terms of time and location will have, um, uh, nuances that, uh, that might be quite important. It might... I mean, we could move on to the... And again, not inside this discussion, we've got about the fire load in car parks and- Yeah.

Wojciech: 58:04

day. I mean, there's, uh, I think the good, good, uh, data for is the NFPA reports, uh, on modern vehicle fires and papers. Uh, uh, the recent one was by a team of Jonathan Hodges, I It, it, it's in the show notes most likely, and I've just had Jonathan on the, on the podcast on the design fires where that. So yeah.

Mike Spearpoint: 58:27

there's various people who've been, uh, and Of course. and, uh, he has done, and there's various bits of

Wojciech: 58:35

there and my dream is, Mike, to get all those people on one and one GitHub, uh, study- On one which we'll discuss, uh, after we stop recording, which is approximately now. Thank you, Mike, for, for coming, uh, to this podcast episode

Mike Spearpoint: 58:51

That's fine.

Wojciech: 58:54

know that.

Mike Spearpoint: 58:54

last thing would be, uh, I, I'll, I'll be talking this work hopefully... Well, the plan is at the Human Behavior and Fire Yeah that Enrico and that is organizing.

Wojciech: 59:04

Campus Katzenburg. It's beautiful. I've been there and it's, it's, uh, it looks like a really conference. It's when? It's in September?

Mike Spearpoint: 59:12

September. So if you wanna come and argue with me about the, then, then, uh, then, then, then is your chance.

Wojciech: 59:20

Come with more than three people in the car and break Mike's study with, uh, one anecdotal data point of you coming in the vehicle. Oh, that's, that's good. A link to the conference is also in the show notes. Thank you, Mike, for coming here again.

Mike Spearpoint: 59:34

Thank you.

Wojciech: 59:36

And that's it. Thank you for listening. So we've started with the baseline rule of two people per space. Then we've accounted that normal-- not all of them will be up with people. Then we've accounted that not all vehicles will have two in the vehicle. Then we've accounted that not all vehicles have people in the vehicle all of the time. They have to get in, do their stuff, park the vehicle, get come back to the vehicle, do their stuff, get out with the itself. It's all transient phenomena, and this number, you and drops from two to one and a half, one, eventually even fractions, zero point six, zero point three per parking space. Very interesting study that Mike done for a kind of very design guidance that exists out there. You know, a magic number that just sits there, and we just nice to apply to our everyday engineering without having ability to question it. If you try to question that on a real-world project, that, perhaps a little bit challenging. I know something about that. But, hmm, look how many numbers like that we have in fire It's, it's perhaps ridiculous how much trust and power those old constraints that once upon a time have been put on our discipline and have not been questioned at all with, uh, new knowledge, new data coming up, with, uh, completely types of structures, completely different built in this case, completely different vehicles even. And, uh, yet we still keep designing that. I love to say that, uh, you know, looking at structural fire resistance based on standard, curve and fire tests, we the twenty-first century buildings based on nineteenth data on how buildings burn down, because that's how far you can trace the origins of the standard fire testing. Um, the point of this episode was, one, to give you the It's ki- kind of interesting that we have a, a more number right now. But also, Mike was able to show you how it's done. And I think, you know, when you have a chance, it's really It's kind of, you know, going down the rabbit hole of fire engineering to find a better number and justify the better and perhaps define the uncertainty of the better number. I highly recommend you, if you find such a magic number, dig deeper, and if you find something interesting, uh, and I'm you will, let me know and let's record a podcast episode about it. That would be it, uh, for today's episode of the Fire Science Show. I hope you've enjoyed this episode on the occupant loads in the car parks and, uh, we will have more fire science and coming your way next Wednesday, and the Wednesday after the one after it, and hopefully we'll be doing it for a time. Thanks for being here with me. Cheers. Bye.