North Greenwich station on the Jubilee Line Extension (jle) is the most breathtaking in the collection of inspiring stations. David Bennett talks to project architect Will Alsop of Alsop & Stormer and Robert Benaim of Robert Benaim & Associates, who in a joint venture with Opus (Works) were the engineers on the project.
Conversation with Will Alsop of Alsop & Stormer Architects
David Bennett Should architecture be an integral part of engineering thinking?
Will Alsop When people talk about buildings and structures, they consider architecture as something that is applied to it. That's a terrible development in our vocabulary. Everything is architecture - both good and bad.
DB What's the difference between good and bad architecture then?
WA In my book, good architecture is when there is some element of delight perceived by the people that use a public building or space. In order to create good architecture, certain qualities are needed. One should be educated in structures and how they behave. It is not possible to have a meaningful dialogue with engineers without a grasp of structural form.
I was lucky enough to be introduced to structural design as a student at the aa in a very positive way - there was a whole series of education and learning experiences on pneumatic and tensile structures and folded plates. And sometimes, in order to understand their behaviour and get a feel for the 'tension' and 'compression' forces, we erected human pyramids with our fellow students.
We should consider a design not just as a work of civil engineering but as a piece of architecture as well and be able to have intelligent three- dimensional conversations with engineers using line drawings, cad images and models to better understand what we are trying to do. It is a process of discovery, not about the grand conception nor the architecture or engineering and how it must be; but about throwing up a range of ideas and possibilities.
db If we consider North Greenwich there are a lot of criteria where architecture has no influence - its physical dimensions, excavation depth , structural requirement and so on - and therefore architecture is reduced to just a supporting role, choosing the surface finish and colour scheme.
wa Before we got involved with North Greenwich, there was already a jle briefing document stating that it was to be a 'cut and cover' station with a roof slab. When we teamed up with Robert Benaim, we explored various ideas for the station. I suggested that we should omit the roof slab and open the station to the sky. I developed the idea with sketches, cad images and even models before presenting it to the client , who initially accepted it.
db But you weren't driven by the need to delight, but by the necessity to make the structure work both practically and economically.
wa As an open station we had a 400m long box, 30m high and 25m wide. It contained a sharp edge on the surface which we could put a handrail around, so that, when you looked down into it, you could see the trains running through. There would be a need to put smaller buildings within it for ticket offices and administration and so on - we thought this was a terrific idea.
Economically of course we save money by eliminating a slab on the top for the roof, but the biggest saving comes from eliminating the ventilation ducts and shafts that would be required for an enclosed station. It seemed to me that the quality of the station environment would be better because you could see the sky and there would be greater security for the public using the station because it was full of open spaces.
The way that the scheme developed is that you would descend 5m from ground level down to the ticket office, then walk along a bridge longitudinally, which was de-signed like a terrace garden, before you take the escalator to platform level. It is exactly the organisation of space that we have now built, except that the station has a roof. I remember working with Robert and discussing how we should stop the free- standing side walls from imploding - do we have to have cross beams to strut the walls? Can we design the walls with a sinusoidal profile for greater free-standing rigidity?
db So why did the open station with its garden concourse not get built?
wa Our proposals had to be put to British Gas which owned the land. At the time there was no Millennium Dome, and British Gas was happy to have a station because it would up the value of its land. It saw the station as the centrepiece of the development and went along with our ideas, even wanting to rename the station 'Port Of Greenwich' which was the name of its development site. That's sadly as far as it got. Shortly after that we received instructions from jle that the roof had to go back on. Nevertheless we had gained from our experience, and carried forward the design philosophy of the clean box for the station.
db Those splayed columns that run down the centre of the station - who decided upon them?
wa There are advantages in having fewer columns at platform level and fewer pile-cap bases to build, and close-centred columns from which to suspend the concourse slab and to support the roof. The solution was a logical one and came from engineering efficiency. The visual delight is in the quality of the surface colour and profiled shape of the concrete columns.
db Concrete, as an aesthetic material, has been treated with suspicion by many architects and, judging from examples of your work, you would be among the non-believers .Why is this?
wa It's true that I have not used much exposed concrete, although on North Greenwich I was extremely impressed with the in situ finish of the concrete columns. It was a surprise for me because I did not think the uk construction industry could achieve that quality of finish. French contractors on the other hand can produce very good site-cast concrete.The cruciform legs on the building in Marseilles are just painted site-cast concrete - the quality was superb.
db Undoubtedly the 'Piranesi Blue' columns give the station its dramatic effect, but how can you justify the cost of the highly polished mosaic tiles and terrazzo steps, using cobalt pigments - the most expensive money can buy - and still be on budget?
wa I was not thinking about the cost but the right colour at the time, and in my defence, I would say that the major cost is surely in the manufacture of terrazzo and tiles, not in the raw ingredients. Remember that the terrazzo surfaces and mosaic tile covering were specified by jle and I was working to the brief. A lot of people say that I am fixed on blue after I designed the Marseilles building, but it was not blue until relatively late in the design development. We had chosen the blue for North Greenwich before Marseilles, and I know that having done so it did influence my decision to go for blue on Marseilles.
db How did you arrive at the choice of blue?
wa In my view it had to be blue, as it was an enclosed box and I was trying to create a feeling of endless space rather than confinement. That's why the under- surface of the ceiling to the roof is not really finalised but there is a hint of it as you see through it, and the blue glass wall beyond. It gives the notion that it goes on into infinity. The reason I like blue is its deep colour, its deadness and the way it absorbs the light like a black hole. It's the same sort of colour tone as a deep copper-beech tree that contrasts with the fresh green of a horse-chestnut tree in May, and this wonderful copper colour sucks all the light in and is not read as a three dimensional form, but as a hole.
db Coming back to abstract ideas and the purpose of your paintings, is it to inspire others or to work within your own thought processes?
wA They are private and personal researches, doodles in paint if you like, which get clearer and stronger as I develop the design. Sometimes it's on a scale where you might have to walk the canvas from one end to the other. So it's the converse of the thumbnail sketch. They are not abstract to me because I can make the connection. In fact some of my teams that have been with me for some years can read my paintings very well.
This process evolved over many years and I draw on my art- school experience. In 1985 we did an urban planning scheme for the City of Hamburg and I thought we should be making illustrations and perspectives of what we were interpreting. What does this plan mean in three-dimensional terms if you are standing at this point in the city and looking at it?
We started to draw perspectives, and I set myself an absurd number as a target, and as time went on and deadlines came nearer and nearer I worked faster and faster and the illustrations became looser and more abstract. Having done them I thought I would take them to show the client. What I noticed with all the councillors present - it was a large formal meeting - was that you had a much better conversation with them on the loose, abstract sketches, than the precise ones. There was enough in the sketches to explain the concept but it was not so precise that they felt inhibited from speaking about them. I felt that this was a better way of expressing my ideas and it would take me far less time to do. It was also a very good device for initiating a conversation with engineers and exploring ideas for yourself. We always talk about teamwork in design, but we have to balance that with private meanderings of our own.
db If I were the client I might be disappointed to be given just a set of abstract paintings and sketches expressing the architecture.
wa You would need the commentary to go with the images and that would explain the approach that I had taken. I would start to build on that, taking in what you were saying about the idea. Then there would be another round of discussions as I began to define the architecture more precisely. The plans and sections of the building or structure would then emerge and gradually I would fill in the more precise detail. All the time I would be carrying the client and design team along with my ideas and making them all feel part of the design process.
db But at this stage you haven't got a contractor on board and you are relying on the design team having a strong grasp of the cost implications and construction logistics?
wa Robert Benaim is one of the few civil engineers that I have come across who can actually sit down and have an intelligent conversation on architecture and structural engineering. He knows what I am talking about and has a deep understanding of the whole process of design and construction, the nitty-gritty of how things are assembled. He is an exceptional engineer. We have worked together on a number of projects but this is the first time that our collaboration has resulted in a finished structure. Cost was an issue on the jle as it has been with our work with them on CrossRail, with virtually the same client. The interesting point to note is that, on North Greenwich, we were appointed six months after Robert Benaim.
db How were you appointed on the jle?
wa We were recommended by Ronald Paoletti and were appointed after making a successful fee tender bid. We did bid rather low, partially through naivete and because we wanted the job. The budget that we had to work with was negligible compared to the civil-engineering cost. As it turned out the architectural budget - not our fee - was 10 per cent of the total contract value and we were within budget.
db You did not do the day-to-day jobbing details as the work progressed, so perhaps a lump-sum fee for the design was appropriate.
wa We did everything up to tender. The overruns and over-spending were more to do with the way jle contracted the project than anything else. It worked on the basis that everything was going to be confrontational and so we all became adversarial.
db How difficult was that to work to?
wa Up to tender stage, while working with Robert Benaim, things went very well. We were to budget and on time - the only problem was in dealing with the services, which had not been finalised. I think that was a disaster, as the services represent a large part of the overall station budget. The services contract was dealt with internally by jle and was not developed to the same level of finished detail as the architecture and structural design when we went out to tender.
If you go backstage at North Greenwich today you will find a number of empty rooms fully decorated and nothing in them but which should have taken services plant. Perhaps some of those rooms are for future capacity such as the possible branch line to the Royal Docks and City Airport. The late appointment of Drake and Scull as design-and- build services contractor meant that any savings in space and capacity could not be recovered. In normal circumstances, if they were part of the design team, we would have caught wind of that and tried to give that volume back to the public areas.
db Now that it's finished, what pleases you most about the station?
wa In the end I feel the architecture is only as good as the client! If Roland Paoletti had not been part of the jle, life would have been impossible for architects. We would not have been appointed and, even if we had, it might have been very difficult to get our ideas accepted. We would consult with Roland to seek his advice on impending problems and he in turn would help our case by arguing it through on the client side. That's why you now see a beautiful structure at North Greenwich and other stations.
Conversation with Robert Benaim of Robert Benaim & Associates
db It must have been a delight for you to be working with Will Alsop.
Robert Benaim It was a very creative collaboration but it was also a complete coincidence that we were appointed along with Will Alsop on North Greenwich. The engineer and the arch- itect appointments were made separately. We were appointed several months before Alsop.
db You had to bid for the design up to tender stage. Was it very simple?
rb It was a very structured contract in which we were treated as contractors and not as consultants. We had to bid for the four stages of the work - stage one was the preliminary design, stage two was detailed design, stage three was tender documentation and stage four was site supervision. We actually had to give a lump-sum bid which covered four years' work, but we were not allowed to front-load the bid for cash-flow reasons. We had to decide very early on if we were doing this job just for the money or really to show how skilfully we could design the station. We chose the latter, as we put the quality of our design as very high priority.
db Who did you report to in the jle? And did the client have its own consultants?
rb We reported to jle civil-engineering managers who actually did no design work and were only monitoring the design submission. We were given a preliminary concept design of the station to work with in our tender brief which set out what the station had to provide and what its shape and size was.
In the end, with all the variations - there were close to a hundred - and last-minute changes made by jle after we had fully designed the station and had gone to tender, it left the client very vulnerable to contractor claims. This is the down side of the design-and-tender approach. We had to make our own claims for the delays and the extra work, because suddenly at the end of the four-year period we were faced with a number of major design changes which we had to incorporate in a design that we had already completed.
db Perhaps design and build would have been better? It seems to have worked well on the Hong Kong Mass Transit.
rb Yes, a current project that has worked well is the Singapore mrt. It is an impressive contract, highly structured to integrate the system-wide aspects of the m&e work and the civil engineering. That's one of the biggest problems of any underground station - integrating its architecture to its civil engineering and services engineering.
db But the argument against design and build is that there wouldn't have been the singular role for architecture that there has been on the jle stations?
rb That's probably true. You could not have achieved the quality of architecture of North Greenwich if you went down the design-and-build route.
db So if you do want architectural quality you have to spend more on the design fee and actually appoint an architect?
rb I think the jle stations are of exceptional quality - I don't think there are any stations like them in the world. There was such freedom given to the architect and engineer to think through the station from first principles. We were looking for ways to economise on materials and to explore different ways to express the structure.
The least expressive thing you could do is to have three horizontal slabs - a roof, concourse and track slab - spanning between side walls that are cut off visually from each other. This is typical of so many underground stations. It is simple to build, but it can be wasteful in materials. Stations designed this way have a huge amount of wasted space in the concourse area. When we were thinking about the station with Will Alsop we decided that we did not need a complete concourse floor slab and that leaving it out made structural savings. Of course there were problems with the side walls which would span further. We increased the rebar in the walls to cater for this without significantly increasing the wall thickness and so were able to capitalise on the savings we had made.
db What about the idea of not having a roof slab?
rb We went one step further and asked ourselves why we needed a roof. The trains are 25m down below, but removing the roof would not be a problem. We would need to prop the walls and would use struts. It was Will Alsop's idea that the station be opened up completely in this way.
db What are the real benefits to emerge from your design partnership with Alsop?
rb Architects are tuned to think three-dimensionally about space and light, while we engineers tend to think two- dimensionally about structural efficiency. It is probably true that, initially, it would not have occurred to us to think about omitting the roof. Will Alsop's prompting was valuable. Having taken the roof off we would get rid of all the fire protection, smoke extractors and ventilator structures above ground, but there were other problems to deal with. One of which is preventing children from throwing things into the station, but we would find solutions to this without having to put the roof back on.
db How early on did these ideas come together?
rb We evolved the open-plan station in stage one of the design brief within six weeks of working together. We then put the scheme design to jle to be vetted, commented upon and approved before we could go on to stage two. We had prepared drawings of the open station with the side walls held apart by flying struts, escalators threading their way between the struts - it was such an exciting station concept. In the end the jle wanted the flexibility of keeping the station roof - it's possible it may have just been informed of the need for a bus station.
db Do you find architects are better than engineers at communicating ideas to the client?
rb Yes they are. I think in a people environment they are very good because that's what they are trained to do. They are trained to think about space and the movement of people. At North Greenwich, we were really designing an underground building with a large civil- engineering content. I feel that all the architects have made such a positive contribution on all of the jle stations.
It was a brave decision to give architects on the jle the freedom to express their ideas and a great deal of credit must go to Roland Paoletti for having the vision and courage to appoint relatively new architectural practices with no experience of underground station design.
db What other innovative engineering solutions did you come up with on North Greenwich?
rb By removing the weight of the concourse slab, the station box became more buoyant. It's always a problem with underground stations where the water uplift may be greater than the weight of the station. In a conventional station with the weight of all the floor slabs and roof in place it's more or less balanced. Once the concourse slab was removed there was a net uplift on the station.
Our skill is in using our analytical abilities to find economical and thoroughly well-engineered solutions to overcoming problems. The imagination comes in not thinking in conventional ways - we consider solutions from first principles, we think things through rigorously, whilst always maintaining the integrity and durability of the structure. Engineers design conservatively if they are not confident of their designs. When we are confident of the actions of the soil forces, the water pressures and buoyancy forces and we know how the structure will behave in resisting those forces, we can design efficiently and imaginatively.
db But it is very complex to predict the variability of the soil and to get an accurate picture of the soil structure interaction - is there a need to generalise and to 'fudge' things a bit?
rb We find the right expertise and we employ specialist geo-technical engineers who really understand what the soil pressures are and how they act on underground structures. We take into account the flex- ibility of the structure itself and the soil behaviour - how it may arch to miss out the weak points and pick up the strong points. If you really understand how the forces can change and vary you can design tightly.
db Given the same set of background information and soil-investigation reports, is it possible that another engineering consultant would have designed a fairly bulky structure with enclosed floor slabs?
rb Yes it is possible. On the basis of flotation for example, we have found that the most expensive way of holding the station down is by increasing the mass of the station externally. If you have a base slab that is 2m thick and floats, the most expensive way of preventing it doing so is to make the base slab 3m thick - but that is the usual way it is done.
The cheapest way in our view is by mobilising friction in the soil to prevent uplift. That needs a lot of careful thinking about how to deal with the temporary works holding the hole open and the backfilling material. The other cheap method which we used at North Greenwich is to design an external toe to the base which will be held down by the soil mass above it, the friction on the faces of the side walls and the pull-out strength of the contiguous piles.
All of these issues were analysed and that is why we were able to make such a big difference to the costs of construction. Another solution is to use tension piles, but we have found that to be expensive.
db You also haunched the roof slab to reduce material weight, and tapered the deep side walls taking away mass again.
rb If you can thin down the concrete volume and sensibly reduce the amount of reinforcement it is a good thing. To haunch the roof slab is very economical as the deeper haunch section attracts moment and that is where you concentrate reinforcement. We have found that haunched slabs in tunnel roofs are very efficient. It is also easy to build if there is a lot of repetition as we had in North Greenwich. We have a good grasp of construction cost as we do a lot of alternative design bids for our contractor clients on bridges and long road-tunnel work. We halved the reinforcement quantity in the roof slab in our alternative design for the Limehouse Link tunnel.
db What is the function of those wonderful columns running down the centre of the station? Are they propping up the roof slab?
rb They are doing two things - they are resisting water pressure acting on the base slab and supporting the roof. They were highly constrained in their size by the space requirements as they had to be sandwiched between the escalators and the platform clearance.
So there was the initial problem of trying to make the columns sufficiently slender to squeeze through the space available and we had to think about the progressive collapse of the columns from impact by trains. It was one of the design criteria of the jle that if a train was to demolish one of the columns the station structure should remain stable.
The splayed or V-shaped columns are very efficient because, if you had a series of straight columns and you were to remove one column, this would double the span of the beams spanning longitudinally between them. This does not happen with the V-shaped column arrangement so that we do not have to increase the beam depths or thicken the roof slab. The columns we designed were very highly stressed and had to be minimum in size, so we decided to specify 80N concrete. High-strength concrete was well proven but we thought the strength high at the time, as the columns were to be cast in 1993. We could have gone higher but we did not want to make it such a special mix that it could lead to problems letting the contract. We thought 80N concrete was within the competence of a large number of contractors. We also thought it was better to precast the columns.
db But the contractor chose not to precast them in the end.
rb I was delighted they were cast in-situ, but at the time we thought that to cast such very tall columns would need specialist plant that might not be available on site. Despite our initial concerns, the contractor did an extremely good job.
db When you were thinking about the construction of the station and decided upon the position of struts and props for the deep excavation, did you assume the working space and sequence of operation that the contractor would anticipate?
rb What we normally do in situations such as this, is to think through one way of building the station - how the hole could be dug and the structure could be built - but we can't impose it on the contractor. We let the contractor know how we approached the construction. On Canada Water Station we had to impose 'top-down' construction on part of the work. What we propose is usually adopted by the contractor because we have had so much experience in this kind of work.
db Having designed the station and gone out to tender, do you have no further involvement with the project?