Major City of London office developments always pose plenty of problems for the designers and contractors involved. Few are the sites which are not close to listed buildings. Foundations and basements are usually complicated, and there is always the possibility of significant archaeological finds during preliminary excavations.
All these and more were faced by the team responsible for the £78M Lloyds Register of Shipping HQ in Fenchurch Street. But for specialist concrete sub-contractor O'Rourke there was a very special challenge - the building's architect is Sir Richard Rogers, famous for his innovative use of exposed structural concrete and the very high standards he demands.
On this project columns, downstands and floor soffits are exposed and will be highly visible both inside and outside the fully glazed building, which steps up to 14 storeys. O'Rourke's £8.7M subcontract included the congested two level basement but from the outset it was the structural frame that absorbed most attention.
'The specification for the finishes alone runs to three pages,' reports O'Rourke chief engineer Colin Banks. 'And we had to build a two bay single storey prototype at our Grays HQ to prove we could meet all the requirements of the client and the architect, especially on consistency of finish.'
The prototype was completed around Christmas 1996. And to the dismay of contractor and subcontractor alike, it was rejected by the architect.
According to structural engineer Anthony Hunt Associates project director Bjorn Watson, the problem stemmed from the mixture of precast and insitu construction initially chosen by main contractor Sir Robert McAlpine. 'It was impossible to match the precast raking columns, vault units and ribs with the insitu columns, stability ribs and spine beams,' he says (see diagram).
With the main contractor already committed on price and programme O'Rourke had to come up with a potentially more consistent alternative in a hurry. A switch to a fully precast frame might have seemed the obvious solution, given the quality of finishes required, but the building at first sight was not the ideal design for precast concrete construction.
Columns are at 6m centres in one direction, but in the other centres vary from 10m to 8m to accommodate the tapering floor plan. Stability in one plane comes from the distinctive raking columns, but in the other the frame is unbraced, relying on stiffened monolithic joints. All fixings and service holes would have to be pre-located. And on a more practical level, the tower cranes for the site were already ordered, which put a strict 7.6t limit on the size of any precast units.
But there were very strong arguments against going for insitu construction as well. 'The column diameter is the same as the spine beam width, which would pose major alignment problems for insitu construction,' points out O'Rourke project manager Tony Backler.
'And, more importantly, there is a current shortage of highly skilled carpenters in London, which would make it difficult to achieve the quality of finishes needed.'
O'Rourke eventually came up with a hybrid solution which promised to solve all the problems - but which pushed hybrid construction techniques into new areas. Raking columns and ribs would be of conventional precast concrete, vault units would also be precast but slimmed right down to minimise weight. Everything else would be hollow precast shells filled with insitu concrete on site.
Main reinforcement would be within the thick shell walls - up to 100mm thick on the columns. Continuity steel would be within the insitu concrete only, a design feature that originally shocked the Anthony Hunt team.
'We consulted both the British Cement Association and the Reinforced Concrete Council and they both said it wouldn't work,' Watson reports. 'As far as we could see nothing like this had ever been done before.'
An alternative insitu design was progressed alongside the hybrid concept for a few months, until work at Imperial College, London, helped develop a theory of thick shell hybrid design which finally convinced the structural engineers that O'Rourke's proposals were sound. Meanwhile, O'Rourke was also putting a lot of effort into getting the concrete mixes right.
Banks explains: 'Basically we're using a C50 mix throughout for both pre-cast and insitu. After a series of tests we decided on normal OPC, pink limestone coarse aggregate and white limestone fines for the precast.
'The insitu was more complicated, especially the mix for the cores of the precast shells, which contains an expansive admixture to minimise short term differential shrinkage, and a superplasticiser. Because volumes are small and the admixtures have a short working life we decided to batch this mix on site using pre-bagged aggregates.'
Long term differential shrinkage between the column shells and the insitu cores should not be a problem, as the expansion of the shells under load should take up any shrinkage of the core. Actually producing the column shells did turn out to be something of a problem, however, despite their convenient lack of projecting bars.
'Forming the internal cores was the difficult bit. We even tried inflatable formers, without much success. Now we're using slightly tapered collapsing steel tubes,' says Banks.
Moulds for the spine beam shells took some time to get right as well, thanks to the thin shell walls. But at the end of the day O'Rourke had come up with an alternative design which was no more expensive than the original and which could even be built slightly more quickly. Time did not permit the luxury of a second prototype, only the erection of trial units on site.
The first production units arrived in early February this year. By mid- May construction was up to the fifth floor. Erection of the internal bays is relatively straightforward, with the column shells placed, located and plumbed, continuity bars added and core concrete skipped into place before the spine beams, ribs and vaults follow. Floor toppings, 95mm thick above the crown of the faceted vault units, are ready mixed C50 concrete with 10mm limestone aggregate, supplied by Pioneer Concrete and pumped into place.
At the bay ends a triangular support frame - variously dubbed 'the 'rocket- launcher' or 'the Scud' - is used to support and manoeuvre the raking columns. These have to be socketed into the column boots and the stability ribs after the latter are positioned but before they are filled with concrete. Macalloy bars running through the raking columns are then stressed before the columns and boots are filled in one operation.
More than 2000 precast units will be supplied from O'Rourke's Grays precast yard before erection is completed in early September this year. The team has already had time to benefit from hindsight and identify what they would do differently a second time. Backler and Banks both agree that the insitu splices are too constricted and the steel fixers have had a difficult time as a result.
Backler adds: 'Getting a cable to the temperature sensors fitted to each spine beam has involved us in a very tricky detail as well.' Watson also agrees on the problem with the splices. 'Next time I would go for braced rather than unbraced frames,' he says. 'And I would seriously consider post-tensioning the floors to cut down the reinforcement even further.'
Otherwise all involved are happy with the basic thick shell hybrid construction concept. 'The only problems have come because we had to stick to the original geometry of the insitu solution,' Banks comments. 'If we had started with a clean sheet of paper we could have avoided the reinforcement congestion and taken full advantage of the concept's real potential benefits.'