From the beginning we considered a number of different materials. However, as the structure evolved around issues of cost, programme, risk reduction, corrosion and weight, aluminium offered significant advantages.
Early discussions with manufacturers indicated that to achieve undistorted curvature in the cladding panels required 3-4mm-thick aluminium. Using the skin structurally as a monocoque provided a more elegant solution.
Our computer model of the structure was built using a three-dimensional finite element package. Ribs were modelled as stick elements, and all were connected together with a 'membrane stiffness' representing the skin. The limited curvature of the pod would not enable the pod to work as a pure shell, although when combined with the additional bending stiffness of the aluminium ribs this enabled the pod to work as a semi-monocoque. This form also allowed some leeway for sculptural refinement.
Given the free-form geometry of the building, the number of different load-case conditions within the model posed a huge processing challenge. We developed a post-processing programme to quickly and effectively assess structural performance. Information was interpreted graphically on screen, enabling 'hot spots' to be checked manually and the model adjusted.
Having a clear understanding of the movement of the building changed the way in which we considered aluminium as a structural element. Aluminium is an extremely supple material, and the daily deflections in the building are relatively high. Thermal movement was the other big issue. By designing connections which allow the whole building to swell, thermal movement was limited to 10-15mm rather than the 30-40mm you would expect for a 38m-long building.
Aluminium is a fundamentally different structural material to steel. When heated, it returns to its natural strength. It also expands rapidly, making it very difficult to weld. We took the decision to design the pod for its parent properties, enabling the boat-builder to use a common pliable grade of aluminium. Therefore it is not the high-shine, high- strength aluminium associated with aeroplane construction which would be riveted and not welded in order to maintain strength. This was never an aesthetic issue since it was always planned to spray-paint the exterior like a car body and not to have the highly polished surface associated with aeroplanes.
Arups has many links with the yacht-building industry, with a number of engineers being involved with the design of carbon fibre hulls for the Americas Cup. Pendennis Shipyard in Falmouth was contacted at an early stage of the project since it is very experienced in working in aluminium and doubly curving surfaces.
Pendennis, like most boat-builders, operated from a high-technology base, understanding how to use a computerised three-dimensional form. It also appreciated the value of prototyping and mock-ups for one-off custom design. Liasing with a Dutch firm, Pendennis used a specialist programme to 'fair' or smooth the surface of the model, abstract the flat cutting patterns of the curved surfaces for computerised plasma cutting and marking of the elements to provide a precise reference in three-dimensional space.
This programme also marked out the rolling patterns to enable the flat plate to be two-dimensionally curved. Each plate was coded with the unique directions, wheel-sets and number of passes it should make through a two-dimensional roller to construct a specific double curvature. Tolerances - depending on the size of the plate - were within 2-3 mm.
Fire protection, thermal insulation, and acoustics were achieved within one layer to minimise the loss of internal space. We achieved this by using one-hour fire-rated Rockwool, which had some insulation properties, topped up with normal thermal insulation before applying the foil protection. Fire certificates were obtained from both the oil and boat-building industries, who have been using these materials and practices for more than 15 years. The services are integrated into the depth of the finned skin structure in the same way they would be placed in the hull of a boat.
The mezzanine floor hangs from the roof, while the 9m-high inclined glass wall rests in the floor, but supported by the mezzanine. In this way, as the roof moves up and down the glass remains stationary. This is relatively straightforward if you have vertical glass, but this glass is inclined: as the mezzanine floor moves down it also pushes the glass out. The result is a very complicated movement arrangement between the glass, the mezzanine and the roof. But the beauty of the fixings is that they appear incredibly simple but perfectly engineered for the job.
The pod sits on two concrete grp-clad legs which incorporate the lifts, risers for the media cabling, fire escape and access stairs to the existing stands. At mid-height they stabilise the existing stand and go up to form the concrete ring-beam which accepts the floor of the aluminium pod while the lift shaft over-runs support the roof. Space was limited due to the need for accress for fire engines and people around the legs, and there was a very limited choice of location. The deep pile foundations are designed to take the permanent over-turn load as the building's centre of gravity lies over the stand where the writers' and tv rooms are.
The flexibility and precision of the analytical process made this project buildable. These tools have created the possibility of working directly with craftspeople from other fields. Complex geometries like the Media Centre which we modelled analytically and visually can now also be modelled physically in a matter of minutes. This means we can simultaneously address the differing concerns of client, architect, engineer and contractor to achieve an optimal solution.