Structural Engineer Mr. DEVRAJ talks about his inspirational design projects

Well thank you very much for inviting me. I am very pleased to be here. I have lived in Sydney for 20 years recently returned to London and am missing the place

already so Perth is nearly Sydney it is near enough for me anyway. I was told I had to talk about the water cube so I am afraid the first, first half of the

talk will be a bit about the water cube but it is I think the project where we did the

most intense collaboration between multiple disciplines including architecture and various

branches of engineering. And I'll do my best to try and describe why we took the decisions we took.

It was a competition that we entered in 2003 so the work I am going to describe happened

10 years ago. It is strange for me to think it was that long ago but it was 10 years ago. And a group of designers came together from PCW Architects from Arup Engineers but also

a firm in China called CCDI who flew their Architects down from Beijing to Sydney to

join the design team for the competition so it was a 10 week competition.

And we did something quite unusual which I am surprised to be able to say to you I've never done before nor since.

I can't work out why I've never done it again but what we did is we lined up the 10 lead

engineers leading their disciplines to say on day one what would success look like to

them. What do they think a swimming centre should be from their discipline perspective?

So the mechanical engineers said look it's about 14 degrees on average in Beijing during the year but inside a swimming a centre we would probably want water temperatures of

about 28 especially in a leisure pool to keep people comfortable. Therefore it is a heating problem and the best way of heating something is simply to

absorb as much solar energy as possible make use of it and stop it getting out. So they wanted an insulated green house.

The lighting electrical engineers said well from a lighting perspective that would be great too but only if we diffuse the daylight because actually you get real glare problems

if we get direct sunlight reflecting off swimming pools and it becomes a safety issue.

The next people who are highly influential and often are not even in the room at the beginning of a project were the Acousticians.

They said well that's all very well chaps but if we make it out of glass it'll sound

awful. Tiles you need around swimming pools for hygiene reasons reflects' sound water reflects sound,

if we make the roof and the walls also reflect sound it will sound appalling and it will be just an unpleasant place to be.

And it was actually the Acousticians that said well look there is this plastic stuff called ETFE Ethylene Tetrafluoroethylene recently used on the Eden Centre in the UK and it lets

sound through. It doesn't reflect it except at the very highest frequency. So if you make your greenhouse out of ETFE it'll sound good too.

There is by the way a by-product of that it lets sound in but you know I will let you

think about that. It also sounds a bit like a drum when it rains but we sort of by-passed that bit.

As a structural engineer on the team I said I don't really care what we do with the structure

too much but I would rather it wasn't outside because it was likely to be a long span and it's likely, therefore to be steel for a cost and strength ratio and in China it is quite

an aggressive polluted atmosphere in Beijing and maintenance is not necessarily their strong suit while they are busy building more cities than the rest of the world and therefore I

would rather it wasn't outside. But at the same time inside is also the wrong place to put steel work in a swimming pool

because the chlorine used to disinfect the pool water is pretty aggressive and turns into sulphuric acid or hydrochloric acid sorry and attacks, attacks the steel work.

I don't care what we do structurally I just don't want it outside nor inside.

But of course architects also come in different flavours. It's not just the engineering disciplines that split ourselves up so the, what the Americans

call Program Architects and I don't think there is quite an equivalent in Australia or in the UK lexicon but those who worry about how a building should be laid out, which functions

goes where, they worked out that really the key to a successful Olympic swimming centre

is very simple. It's to have a use after the Olympic. Not so much about the actual competition pool hall which is very specified.

But it's putting in as much leisure facilities as possible for the people of Beijing to use it afterwards so they simply maximise the footprint of the building to fill the whole

site which happens to be square and so we had a square building.

The urban designers the people who care about how the building fits into the environment around them, they said look we've got this brilliant opportunity because the Olympic

site (this is a google map of Beijing) is right on the north south axis of Beijing that

goes right through the middle of the forbidden city and is actually the primary cultural access of Beijing.

Down to the south is the temple of heaven one of the most important cultural sites in

Beijing the east west axis is the commercial axis of Beijing. We've got this opportunity from the master plan that was already proposed that the stadium

and the water, the aquatic centre became the gateway from the city on this major access

into the site for the rest of the Olympics. So what they said is how these two respond to each other that counts, now these are all

very well we put all these things up on day one and four weeks later we are still arguing about what we are meant to do or what we are going to do.

And the Chinese architects were drawing icebergs I think if I remember rightly with mist curtains

around them. The Australian architects were drawing wave forms and they both agreed that they wanted

something that was absolutely a sort of a physical manifestation of water if you like.

They thought that was the way to win an international design competition but we didn't actually know what to do.

And then fortunately our neighbour, the bird's nest stadium scheme was published, so they

were having a competition at the same time but a little bit ahead of us. And this winning scheme by Herzog & de Meuron and also Arup as Engineers came out and was

put in the press and we sat there, and I'll never forget it, 20 people sat down around a table and we said ok, so theirs' is red and round so we will do a blue box ok?

And I know it sounds sort of trivial but it was honestly like that and it was sent me a note "we wanted ying and yang" so red in Chinese cultural sort of mythology is masculine

and blue is feminine, circular things speak to the heavens, square things speak to the earth.

One is internally focused and one is externally focused. And at the same we brought all those ideas from the first day which had been sitting

in people's head but it hadn't actually been put forward properly yet to the table and we made it out of ETFE and we made it an insulated green house and we made it have its natural

light and we made it sound good so it was again, the structure, my bit was undetermined

at this stage. But we did know we needed two layers of these ETFE pillows because to make ETFE which is

a nought point two millimetres thick plastic, it's really just a polythene bag, into something

that can actually do anything you have to inflate it with air and make it pressurised. So you make these pillows and we knew we need two layers of them to keep the heat in at

night otherwise all the heat we gained during the day was just going to go out again during the night. So at that point I looked at this and said aha that means there is a space between the

inside pillow and the outside pillow, where I could put my structure that is neither inside nor outside.

And then I thought well how does that structure want to inhabit that space?

It is not a question we normally ask. We normally ask how the space, how the structure enclose space for human habitation.

How do we build walls and roofs and things? I suddenly thought how does structure inhabit space and I realised I'd never questioned,

never written before, I'd never heard anything at University or afterwards that answered

that. And then I saw this photograph which I took to be some piece of biology, an organic material,

and I said aha that looks to me like structure inhabiting space, that's what we are going

to do. That was a great aha moment then follow that with a mmm.

But what is it? And how do you define it? And how do I work it out?

And then the magic 2003 was really the early days of the internet and I found of course the World Wide Web can help you in all these matters.

So I went out and did some research and I found out that Lord Kelvin in the 1890's had

asked the question "how do you sub-divide three-dimensional space in the optimum fashion

into equal volumes" and he came up with this thing called Kelvin foam, which is a truncated

opti-heaven as the optimum way of doing it. And all his friends and colleagues had a look at it and they all tried as well and none

of them could do any better so they sort of gave up and for 100 years you know that was the answer as to how you sub-divide three-dimensional space.

But it didn't look very exciting to me and it didn't look very like my photograph either.

I then found a Belgium physicist at the same time a fellow called Plato who was looking into soap bubbles and actually an infinite array of soap bubbles will also be the answer

to "how do you sub-divide three-dimensional space" because the surface tension in the bubble will try minimise the surface area which means it's the same challenge.

And he'd observed, Plato had observed that soap bubbles always come together, three films at 120 degrees and at the end of a line, four lines come in and they are at a tetrahedron

angle of a 100 and, if I can remember it is either 104.9 or 109.4.

I think it is 104.9. And so I said that's the answer isn't it? All you have to do is draw those lines and keep drawing them you know in a 3D CAD model

and you will come up with the answer. Well of course it doesn't work like that otherwise somebody would have worked it out a long time

earlier. And lastly, I found that a fellow called Professor Weaire and his PHD student Doctor Phelan in

the Trinity College, Dublin in 1993 about 100 years after Kelvin had come up with a

formation of geometry which was more efficient than Kelvin foam. And it's made of these two polyhedral that you see in front of you collected together

to form the molecule on the right which then repeats. It's mainly made out of a pentagon with the odd hexagon in it.

And what I found is that if I took a very large block of Weaire-Phelan foam and then

cut it at a skew angle, and it was cutting it at the skew angle that was the clever bit,

or the odd bit, so if you go across the top here from left to right and then the middle from right to left and at the bottom you will see how we did it.

We took a block of the foam at a skew angle, we chopped out a fifty metre height of it,

then on plan in the middle we chopped out a 200 metre square box and then on the bottom

we hollowed out the spaces, the volumes that we wanted to occupy, the actual halls inside

the centre itself and you end up with this, a wonderful structure.

And this is you know, this is sort of really new to us because all that was done inside a machine.

It wasn't done on the back of a napkin. It wasn't done with paper and pencil and it came out with 24 no 22,000 elements joined

together at 12,000 nodes or intersections. And we said that's it that's what we want to do and it's going to be a structure like

no other, it produces a wonderfully random pattern for the pillows to go on, on the surfaces

which is just what the architect liked. The only question now was how do we show the judges that this is what we are putting forward

so we said that's easy we'll just make a model, like a physical model and actually as you

might realise by now this isn't going to be an easy thing to make a physical model of so we said that's, that's ok because these are these things called rapid prototyping.

So what we will do we will make one of these wonderful models using a computer driven rapid prototyping machine never been used in the architectural engineering world before but

common place in the car and aerospace industries but no one had ever made a model with 22,000

different bits in it. They were used to making one piece that was a complex shape.

It took us four weeks eventually to get our data in the right format to get the model to work and we made this model 24 hours before the deadline for the competition.

And we shipped it from Melbourne to Beijing overnight where the Chinese had hand made

the surfaces which they then glued onto the model. And I'll never forget one of them I got a phone call saying the north wall doesn't fit

ok what'll we do? And for some reason, I have no idea why, I said what happens if you turn it around ok

and luckily that was the answer and it did fit.

And we made the model and submitted it and fortunately won the competition which is, which is great because we won from the jury which is the official judges but also the

Chinese had put it on the internet for public voting and sometimes that had happened before

and I'd normally cheated a bit and asked everybody in Arup you know to vote for the Arup scheme.

In this case we won by a million votes so I don't think the Arup voting system would

have made much better. Then you get this incredible euphoria when you win a design competition but of course

to win a design competition you normally push the boat quite a long way away from the shore and then the next day you wake up and you think aha.

Now we have to make it work ok? So from a structural point of view I built then an analytical model.

You might think I should have done this before but we were so busy making the physical model to convince the judges that we hadn't thought about the analytical model.

I mean it looked like a structure, you know, it felt like a structure, if you like it smelt like a structure but was it one?

And the first time we built an analytical model I couldn't make it stand up.

If the elements were put in rather large you know to make them strong enough they were also too heavy and if they were put in smaller than they were too weak and the thing fell

down. So again in this sort of brand new digital age we invented a whole new analytical process

that sorted it out for us and what that meant is what we did was simple, it seems trivial but it you know it was one of those things, one of those breakthroughs.

You put every element in rather small, you analyse it, you find out the stresses in every

element and then compare it with literally with the code of practice of China and not some arbitrary number but the actual code of practice of China, if you didn't meet the

code then make it bigger by one small increment yeah and if some reason it seemed to be larger

than it had to be make it smaller by one small increment. Don't try and guess the actual answer just move it in the right direction slowly.

And what we found after a lot of experiment was it would converge as you can see here after about 20, 20 cycles it converged on a stable answer which also I am glad to say

proved it was also an efficient structure at about 5,000 tonnes of steel.

This picture shows you slightly shady different colour there and one of ten different sizes. We only use ten different sizes in the whole model and you may be able to work out now

why I couldn't guess which one went where. Actually the thing is so highly redundant for the structural engineers in the room that

you could come up with an infinite number of solutions that also works. It didn't matter which one you used as long as it would then produce a set of calculations

that would satisfy the Chinese authorities within the linear elastic state it would comply

with the Chinese code. And every time you ran it you got a different answer which was rather nice you know so every

day we got a different structure and so we just said one day ok we'll have that one.

Ok -- moving on. This shows its dynamic properties. It is a completely framed structure; there is no triangulation in it so it relies entirely

on the moments at the end of each element. And what's really strange about this is not, these are the first four motor vibrations

so the top two, left and right, not surprisingly are very similar and have a frequency of about

one hertz and because it is a square thing in the end you know so it is doing roughly the same thing.

But what amazed me is the bottom two, the vertical and the torsional are the next two modes and they're also around one hertz.

I still don't know why, I can't tell you why, but what it meant was in an earth quake and

earth quakes are the major dominant loads in Beijing, every single element played its

part in absorbing energy or could play its part in absorbing energy from the earth quake. So it actually turned out to be the most earth quake resistant building in the world probably.

I can't prove that I can tell you it takes more than one gravity sideways and it's still

standing up quite easily. It just forms plastic hinge after plastic hinge after plastic hinge after plastic hinge

until it gets to 44,000 of them before it gives up.

Having done all the sizing of the elements inside the computer we then wrote a piece

of logic that said well we will make the nodes spherical because that's an easy way of elements

coming together and we will size the spheres according to the size of the elements coming into them, and we'll write a rule and therefore we'll develop all the spheres as well and

then all the elements get cut back until they meet their spheres and then we have you know what we now call a bin model.

We have a fully defined piece of structure. So far it hasn't been touched by anybody.

It hasn't been meddled with manually if you like and we got it down to we could re-do the whole structural process if, for example, we wanted to make the building smaller, which

we had to do several times or add a doorway or do something to it in three days.

So we could go three days from changing it to having a new set of construction documents.

The construction documents could be portrayed in lots of different ways. Up till now I have always said there should only be one set of contract information for

fear of there being a contradiction so it was normally the 2-dimensional drawings with the contract information.

Here we said, well actually, as this is just a set of data that hasn't been touched by anybody then we can represent it any way we want.

So we could represent it as you see on the left here as a wire frame model with a lot of tags on it and a whole set of data saying what each tag meant or we could represent

it on the right hand side as a set of 2D information or on the bottom of the right hand side, you

can barely see it I am sorry, a bit washed out, but 3-dimensional views of the actual connection details for example.

And you can visualise it so this isn't what the architects normally call a CGI, you know

an artificial if you like of visualisation of something where a lot of work has been done in photo shop to make it look nice.

This is actually just a view of the construction documents and it was the first time we'd done

this so this is we are now just into 2004. But there was some magic in the original geometry and that is that all the, even though the

tube sizes were different throughout the building the geometry repeated because it would come from that piece of Weaire-Phelan foam that repeated and so we said that the way to build

it was to make a jig which would always be a fixed jig and you make up these little four

legged elements and then on site you bolt them together and as long as you bolt the right ones together in the right order you get the structure you first thought of.

But at the time we said that looks like a bit of hard work, why don't we just make lots of tubes and lots of balls and a big scaffold and we will weld it together?

And they did. And I have no idea whether they welded the right one to the right ones.

None whatsoever but because actually and I say that flippantly but because of the huge

redundancy in the structure it doesn't actually matter. You formed plastic hinges wherever you needed to and it would re-distribute you know, the

strength was there somewhere if you will. Actually I think they probably got it right, they took it very seriously.

And then eventually a year later they took the scaffold out and we'd be able to see what we had constructed.

This is the inside of the leisure hall that you didn't see during the Olympics it was

still not fitted out and then it was finished. So I went, I was invited to the official opening on Australia Day in 2008.

I got the invitation at 5:00 pm on a Friday that the official was going to be at 9:00 am on a Monday but fortunately I managed to get there.

One of the other things we did is to get the heat and light balance right. All the ETFE pillows were covered with a silver dot pattern but we varied the density of the

dot pattern according to how much light and how much heat we wanted into that particular space in the building.

You can see here that when the lights behind the building you see through it you get some

transparency and here again the same thing. But as the sun comes round on to the face of the building it goes this incredibly metallic

colour instead so now all you are seeing is the silver dots. They are reflected back at you so it's a bit like the Sydney Opera House; it changes its

mood and its feeling according to the lighting conditions and what you get before it. Which wasn't something I must say we totally predicted but it was still quite nice.

And then we went inside and inside the main pool hall we were actually only getting 5%

daylight because this television company wouldn't let us let in more than that and they said

if it was any more than that we would black it out anyway and use our flood lights because they're worried about controlling lights for television.

A bit like our camera at the back whose spotlight me to make sure I can be filmed which is a

similar sort of thing but actually 5% daylight is tonnes if you're inside it will make you

feel as though it were in natural light and bright light. This I think was my favourite space this is just a café in one corner but it had a real

sort of serene calm feeling, almost spiritual and this is Mr Phelps winning his 7th Gold

Medal. Forty-five world records were broken in the swimming centre in Beijing and I to this day,

I will uphold the idea that it was because it was such a beautiful place to be but of

course the athletes performed so well and it was nothing to do with the shark-skin suits that they were wearing.

And then this is the bit you didn't see at the Olympics this is the leisure centre so this is the legacy, this is what the people of Beijing get to use every day now.

At one time the Beijing government said actually we don't want to make it into a leisure centre,

we'd rather make it into a posh restaurant for because we think it's so good it should

be a very expensive to go into and we pleaded with them that we would much rather that it were for the people of Beijing and they acquiesced which is great.

So there we have it, the two form this gateway, I think, if you think about the Olympics I

don't think you would remember any other buildings really than these two. There is actually a building right next door the main convention centre the main indoor

arena which is larger than either of these but you probably can't remember what it was.

So I got three slides in amongst these that talks about projects that sort of pointing out some, I guess some of my approaches to life or approaches to design so the first

one is, at Arup we say we shape a better world. That's our strap line if you will but what do we mean by that?

And for a long time I thought well it actually just means that we've got to make the world sustainable. Now I don't back down from that at all we do have to make the world sustainable but

sustainability is a lot about reducing impact, you know reducing environmental impact, reducing

financial costs, reducing social impact. That could produce many different worlds actually that comply with that and which one do we

actually want? So what about the benefits? What are we actually trying to do as well as being sustainable?

Mind you, it is about improving the world for people you know and there is various,

there is tonnes of different levels you can look at that, you can make people safer make them healthier, give them more functionality, more amenities but ultimately it's also to

inspire them. We want to enjoy life surely we want out built environment pleasurable.

To be delightful so in the end I short circuit all of that and say delightful efficiency.

Delightful maximise the benefits, efficiency minimise the costs whatever they might be.

So this building I have already introduced to you. This is the Bird's Nest Stadium in Beijing. When I look at that I have no doubt at all that it's the most beautiful stadium ever

built in the world. But unfortunately it was also quite expensive.

It uses an awful lot of steel to make it work because it is a sculpture with a structure

to hold it up if you will. It is not really made by intense collaboration between engineer and architect trying to work

it out how to do both at once. So I sat down with my best friend Philip Cox and did this competition for a stadium in

Durbin for the world cup, for the soccer world cup and you might notice that it never got built.

It wasn't part of the world cup. I still haven't got over losing this one so I am going to tell you about it instead.

So this is about how an engineer and architect can try and come together and trust each other

enough to try and work out how a form will work both as structure and architecture.

It's a pretty intense and trusting sort of arrangement because, particularly for Philip,

he has to trust me that when he comes up with something he likes the look of and I say that doesn't really work as a structure then I am right and then I am acting in his best

interest and genuinely trying to find something that will work as a structure and look the

way he would like it to look. So anyway after a while we came up with what we thought was a really you know a really

beautiful building that also worked highly efficiently and I'll just for a moment describe

how it works. So if we were to work like the Birds Nest Stadium which are really just these big deep

trusses spanning across it this is a cross-section and it shows something like a nine metre deep

truss cantilevering 70 metres. It shows you have got tension in the top cord and comprehension in the bottom cord under

a vertically downward lobe. Well when you do that in 3-dimension you have to put them sort of 12 metres apart or thereabouts,

you have to put spanning elements between them to hold up the cladding. You have to keep the compression cord stable laterally so you add another layer of structure

at the bottom. It's now looking pretty nasty so you probably add a ceiling underneath that can be the bird's

nest house and you end up with a lot of steel and a bit of a jungle.

If instead you take exactly the same principles and smear them out into 3-dimensions and now

you offset the bottom cords from the top cords you can now have a structure to where all

the structure is carrying the planning. Where the structure is also auto stable, that is, the difference in curvature and the offset

keeps everything from buckling and you literally take out 60% of the weight by going from the

truss structure to this one. And in our case the brief was for a stadium that could hold 70,000 for the World Cup and

then go down to 50,000 afterwards and we could put the 20,000 extra seats up into those humps

in the roof that wouldn't have been available to you if you had done a truss with a ceiling on it.

So I am still, oh and the last bit is of course, is the inwards curvature of the wall, which is part of what we wanted it to look like also gave you exactly the right amount of

floor place of each level for that particular stadium brief. Often in a stadium you get a problem that the floor plates get bigger and bigger as

you come down but actually there is no more facilities that you need at the bottom than you need at the top.

But as I said unfortunately we lost that one. But instead we did the baby brother.

This is AAIMI Park in Melbourne. I love this photograph of AAIMI Park because it looks like a CGI but it's not, it's the

real thing. This is what we built. The principles the same, it's a much more modest design.

It's only a 25,000 seat stadium with a small roof. But we did do is use parametric modelling to help us make it as efficient as possible

so on the left there you see a program called Catia and the red line links the leading edge

of the roof together. And all the white lines are rule based and follow the red line so we set up a whole lot

of geometrical rules that operates in the red line so we adjust the red line a bit and

we mil, the structure model pops out we size it automatically like we did at Beijing using the same piece of software and the architect gets to look at it and says do I like it or

do I not like it, does it look better? And we also looked at the panelization of the façade, the cladding, are we getting

more or less repetition, is it more efficient to clad it with this shape or the other shape?

In theory you can link all these together and use some kind of genetic algorithm to sort out the most optimum solution.

In reality I find it's better to do it by hand actually so adjust the red line, look

at the different outcomes and then think about how else you should you adjust it to make it better.

You know, which is more important, making it look good, making it repetitious, making it less steel and eventually we decide on something.

The other advantage as I said before in a way is that this is a CGI this is a rendering

you know we made of the building before we built it and this is the building after it was constructed so you get a pretty good idea now what you're going to get before you go

into the physical world. Here we see it being used and the same idea of the seats and the humps and in this case

we likened it to a club you know where you might sit in your booth at the back, you might have your favourite piece that you always go to that you occupy.

It breaks up, the inside of stadiums can be quite characterless, they all look the same once you're in them you know, we make them different from the outside, not necessarily

from the inside. This again is one of my favourite shots because again this is reality.

This is it sitting on the river in Melbourne and then lit up at night and another thing

Melbourne is fantastic because you can, you'll be able to hear as well, you can see your stadium from the city, you can see what is happening and you can decide you know, am

I going to go and watch a match tonight as you come out of work, or am I just going to have a drink there or am I going to walk around the park - great place.

So here is my second little sort of philosophic piece and that's about design.

When I was first made an Arup Fellow, 10 years ago, the first question I was asked is ok how do you do it?

I shrugged my shoulders and said I don't know, that's what I do, I am not an academic, I

don't think very much in those sorts of ways, I just you know do what I do. And so it made me think you know and so three or four years later I decided that this is

the way I tend to go about a project you know. The first thing is to think about what so you want to achieve, not as a structural engineer

but what should the project achieve, what outcomes does your client really want, it's

often not written in the brief actually, you often have to go and talk to people to find out what they really want.

And also of course, stakeholders, it's not just about what your clients want but what do the end users want, what does the community want, what do people overall want?

The second one is a sort of what I think of as a design actually being is to challenge the status quo, to not accept what has been done before is the best that can be done.

It may be, but you don't accept it initially, you've got to say, can it be done better. It doesn't matter what it is.

It could be a spreadsheet, it could be a project management plan, it could be how you write

an email, it doesn't matter, all of those things when you do them, can they be done

better? The third one as I said before is trying to improve people's lives, and in particular

I think to get an emotional response on a human scale. I remember I worked for a wonderful engineer called Peter Rice, when I was young, who unfortunately

died young like George I guess. Peter said to me though, he said Tristram, it doesn't matter if your structure is wizardry

and wonderful, what matters is how people will respond to it.

How will they think you know what emotional response will you get from the people who visit it. Then there is as I have already said you make it as efficient as you can, you make it as

delightful as you can, and lastly and most importantly for us in our industry you have to deliver it.

It's no good just ending up like Durban did you know in the ether, it has to come out

and become real and that's the hardest challenge of the lot. It has to be built.

For this project I was asked to include some more international projects so these are projects

I have been working on recently that haven't always been completed. So this is about an airport in Kuwait with Norman Foster as the architect.

The question here is what actually are you trying to do in Kuwait? What's Kuwait about? Kuwait is about hot heat and dirt or dust, desert storm, it's about solidity and trying

to find shade and the feeling of shade even when it's not, or the feeling of cool rather

that comes from shade and mass and masonry.

This is the plan of the airport super imposed off Hong Kong's harbour so it wasn't a small

challenge that we were faced with. This airport is 1.2 kilometres from tip to tip with a triangular element.

Yet, Fosters in their wisdom decided it should be made of concrete.

So we were faced with a challenge of, this is just one arm of it, of a concrete roof

probably ten times larger than any other concrete roof that's ever been built. With up to a 60 metre cantilever around all its edges and a 48 metre span between it and

then how do you do this efficiently? So we came up really with a schema you see in front of you that the bottom is the set

of arches and ribs out of post tension concrete so if you like the opera house construction

or bridge, big segmental bridge construction and on top of it the idea of this light weight concrete shell, but the light weight concrete shell is not easy.

In in situ concrete it will be 300 millimetres thick which could actually weigh an awful lot. You'd have, every single panel is different, I should have said all the shells are different.

So the formwork would have been horrendous and then it would have fallen apart under thermal load because it's so big.

So in the end we came up with this idea of using fruit boxes, we call them, so we call this fruit box engineering and it was simply the idea that if we make panels of, and I

called these concrete panels for Foster's sake but for me they are hybrid panels because they are steel frame with a concrete base.

All the steel is vertical, vertically set in real life and flat and later cut so as

to get them as accurate as possible, four pieces. Then we put them on a formwork and we cast the piece of concrete in the bottom.

And then as you see at the bottom there you produce these cruciform elements also from laser cut steel which you then just bolt everything together and in theory it'll all work as long

as everything is made accurately. And this is an image therefore of the central piece, probably the simplest bit of how it

will all go together each piece is about 4 metres square, the fruit boxes to give you a scale and the concrete is only 130 millimetres thick.

That shows you the more complex piece around the edge with the 60 metre cantilever but it was applying a ruthless piece of logic if you like, a piece of process across an

analogue idea. Here's a fruit box and this is its clever bit.

The clever bit is that concrete will carry quite a lot of compression as I think we all know and therefore in compression it all sort of forces itself together if it gets hot.

But if it pulls, starts pulling itself apart the concrete won't take much tension so what happens those steel side plates flex and bend and allow it go away so it was all about balancing

stiffnesses. How do we get the right amount of in plane stiffness so the shelves worked the right

amount of bending stiffness so they stayed stable and didn't buckle in compression but the least amount of tension stiffness so we could release the tension forces out of the

concrete. It was quite an ask and we ended up doing full dynamic, not dynamic, but using a dynamic

non-linear program to analyse as you can see it, at this level, every bar of re-enforcement, all the welds, all the steel and in a non-linear fashion pulling it and pushing it and bending

it just to make sure that all those stiffness parameters were actually what we were going to get.

And there you see the final, an image of the final product. This is only out to tender as we speak so hasn't yet been proven to work, this is uhm,

but it has been fully designed. Second to last is this Singapore sports hub so this is a stadium for Singapore.

Singaporeans are not known for being sports fans but they wanted a national stadium for

their national identity. It had to be a multi-purpose stadium so one of its attributes is it has an open and closing

roof so you can use it in the rain and also so you can keep people cool in the heat of

the Singaporean summer. The schema is Arup's architects and engineers so the scheme we came up with is this enormous

dome that career spans across the seats that provide a moving piece at the top, a solid

piece at the edge of the seats and then an open louvered piece at the perimeter.

And again to give you scale this is a 300 metre diameter dome, I am told it's the worlds' largest dome, I haven't done my own research but I am told it's the worlds' largest dome.

And of course it can be open, so you can use it either in close mode or you can open the roof as you see here which, just remember that shot because if, the structure I am going

to talk about later is highly dependent on this problem that you've got a big weight that can be in the centre or it can be separated and then these naturally vented spaces at

either side. So to keep people cool in summer we provide under-seat air conditioner if you like but

we were not going to try and air condition the whole volume its' enormous, we are just going to produce local cooling.

To do that though and at the bottom there I've got a picture of a CFD, Computational Fluid Dynamics, of all the people in their seats on the left, then the middle image is

the bottom tier, next to the right is the middle tier and the far right is the top tier.

What you can see in the far right is its beginning to get warmer and the reason it's beginning to get warmer is there is radiant heat coming in through the roof so the critical part for

the fixed roof is to insulate it as well as possible and to reflect things but the critical

for the moving roof, the top bit, is to make it as light as possible, is because it's this

moving load that will crucify the fixed roof otherwise. So it's this delicate balancing of everybody in the room talking about what do we want

out of each component, how do we make it work? The moving roof is actually clad in ETFE like Beijing to keep it as light as possible and

then the outside is louvers just for shade for the external concourses.

And this is a rendering of the outside so you can see these components so you can see the pillowy shape is the ETFE clad moving panel then the triangular shape if you like

with the lines across it is the fixed roof which is kalzip aluminium cladding with a lot of insulation underneath it and then the louvers down below for the external concourse.

And this is the primary structure but it's not the whole trick this is part of the primary structure so it's trusses.

The trusses there vary in depth and width from the centre of the building where they're deepest to the perimeter of the building where they're shallowest and this comes back to

this human reaction bit which I'll talk a bit about later. But there is also these bit diagonal trusses and you are looking at it and probably thinking

why are they there? Well at the bottom of all these trusses is a ring beam that holds it all together but

all the steer is in compression the ring beam is in tension. But the ring beam cannot cope without a balanced load.

It can only cope with a uniform pressure pushing outwards on it and what the diagonals do then

is as this moving load changes it re-directs the moving load outwards so when the roof

is open it pushes towards the end and when the roof is closed it allows the stresses

to come down in the middle. That's quite complicated and I haven't got a decent image to show you, but I hope you'll

get it, then across that is a set of secondary trusses and a set of louver trusses, lots of different geometry being used in different places to get different responses or different

impacts. And of course here you see that moving load, that's the moving roof in its open position.

The secondary trusses aren't shown otherwise it would look just like a jumble of lines.

And this shows you the deflections of the roof and without the moving roof shown in some of the images and with it on the top right, so you see the top left is the dead

weight deflection of the whole roof without the moving panel. The top right is what the moving panel does to it when it's closed, the bottom left is

what the moving panel does to it when it's open and the sections in the bottom right show you they're different so it goes down in the middle when it's closed and it goes

down on the sides when it's open and that was the thing that was really challenging about this roof.

And here you see the louvers now complete, the louvered zone down at the concourse, now I think, I hope you see the benefit of tapering the trusses they're now only 2 metres deep

at the perimeter, they're 5 metres deep for buckling reasons up at the top centre to carry

the moving panel. But this brings them down to something that's more of a human scale where people get to

meet them and interact with them. And here you see the whole structure, last week the fixed roof was completed and the

prop so we've now got the worlds' largest dome now actually exists.

This is my last slide which is about the impact of IT in the Built Environment you might have

noticed that I talked a lot about the virtual world. I think this revolution, even though it has been going on all my life, my professional

life, has still got a long way to run. The impact that the digital world has on the physical world through our ability to now

design things absolutely perfectly before they are constructed. It used to be that every building was its own prototype you only tested it when it was

made and in those days the car industry made physical prototypes and crash tested them

against solid bits of concrete, nowadays both of us can do it in the virtual world without

resorting to the physical world. And we know what we want before we build it. I would then suggest that construction will become more manufacturing where we make things

in factories all over the world and bring them to site and assemble them because we know what we want we can predetermine it.

One of the reasons we pour concrete into timber forms on site is to get it right on the spot

you know to alter things as we go to make the holes required for the services to pass through the structure. But all of that can now be predetermined we can coordinate it all in the digital world.

I can talk more and more about this but I won't do too much because I am probably out

of time. And in the end it's the internet of things out there in the physical environment itself

you can have actuators and sensors so you can get feedback from what people are actually doing and you can change the physical environment in real time.

This is most readily seen in smart infrastructure so the idea that you can control traffic systems according to traffic flows, you can control energy supply according to energy demand etc.

And lastly if we were clever designers we would take all the feedback and use it to

make our designs better and make our plans better. At the moment I am ashamed to say most designers when they've finished a building walk away

from it. They very rarely go back and find out how it's actually used or whether its occupants

actually like it because the myth is almost always better than the reality but we stick

with our ideas, our concept of what we want to have happen rather than finding out what really happened.

This is the last building, last project, back in Australia, this is in Brisbane, this is 111 Eagle Street, and it's the tower in the middle that I am going to talk about.

We put it between two very, very good, very beautiful buildings by Harry Siedler, two

modernists' buildings on the left and the right. This is on plan the two grey blocks are the two existing buildings and our site was right

in the middle but it was sitting right on top of the existing car park and right over

the existing loading dock which you can just see with a car in it right in the middle of this slide.

And so we said well either we just you know we sterilise the whole basement in sight while

we carve it up, build a new basement you know and build our building on top or we make use of what we've got.

We leave the loading dock where it is, we put our columns where we can for the building, we put our new core where we can, which is offset to one side, because of the existing

loading dock and then we make a virtual out of it and see what we can do. Unfortunately, when the core is off centre like that it no longer works for a 40 storey

building in a cyclone zone and to take the wind loads alone so we decided we also then had to have a perimeter structure to make the building work against wind load.

So what we did is we said that if we start with a random position of columns, in other words we only put them where we can then why don't we just grow them randomly up the building

and let's see what that looks like and actually it occurred to me that if you incline columns

very slightly they'll actually probably work to carry the wind load as well. And so that is precisely what we did, we wrote a computer program that would generate random

numbers and produce random sets of columns. It's not quite as simple as that. What we said is that from every level let's generate a new set of columns and then let's

test it. Let's say, do we have as many columns leading to the left as to the right, in other words is it balanced under vertical load rather than wanting to fall over one way or the other.

Is it balanced front to back, is it balanced torsionally, more importantly does the floor

above get a decent amount of support or do the edge beams get too long you know and we said we'd limited that and we said we want only an 8 metre edge beam.

And then we played with the amount of randomness as you went up the building so what happens is each time it gets to a floor it generates a set of columns it says pass or fail, every

about, only one in ten thousand would pass times and then it goes and generates another lot, another lot, another lot and it might get stuck part way of the building and say

actually I can't find the new set of patterns of columns, it takes the next lift, in which case just go back to the bottom and start again.

It's incredibly profligate you know only one in a million attempts comes out with success

but it doesn't matter because it only takes minutes you know and then you say ok that's

one, there's one I'll put it on the shelf and when you get twenty of them on the shelf, you go down to see your friendly architect again, Philip Cox in this case and you say

"Do you like any of them Philip? Do they look any good?" And he normally says no, and so you come back again and you re-write the program to some

slightly different logic you know and as you can see this is the final one and what we've done is make the columns get more random, more chaotic, more excited as you get to the

top of the building because they are carrying less load by then so you can. Down the bottom they are quite strict, they're quite vertical, they're quite you know, they've

got to have a five degree average lean to take the wind load, after that they can be you know more upright.

This is the Cox drawing of the same building I am glad to say, here they coloured coded the columns according to their size so you can see in a way that it does actually work.

The columns are behaving the way we'd expect them to. And for the structural engineers this is a, on the left the vertical load diagram, on

the right the wind load diagram showing the columns are doing what they are meant to do. They are carrying more load as you get towards the bottom relatively evenly and on the right

hand one, they are barely going into tension on one face and they are taking the compression on the other. This is my favourite drawing I think because this is 2008.

On the right is our revit model and on the left is the architect's revit model winded

and for once they are the same thing ok? Only time in my life so far, we'll get there.

And here it is built. Here it is in Brisbane, it was the only high rise building in Australia to not stop construction

in 2009. DPT, our client had the courage to continue, and by the way we were open with them and

didn't rely on the competition, we said we think this is a building you will like, it will cost you $6m more you know and they said ok we'll have it, we'll do it.

They've done it, it was fully tenanted when it was opened so they had the courage of their

convictions to continue construction through the recession with a modest hiccup we had in Australia. I think now we might be having a recession unfortunately.

And there it is lit up at night. So there, ladies and gentleman, thank you very much.