When an early Brunel experiment in cast iron bridge building was discovered this spring next to Paddington Station in west London (NCE 1 May), the nation sat up and, briefly, took note of the local landscape. The modest Brunel structure was revealed, cocooned within the ugly brick span of Bishop's Bridge, a mish mash of late 19th and early 20th century structures that gracelessly shamble over the Regent's Canal and 15 railway lines.
This August bank holiday, with demolition of the old brick elements of Bishop's Bridge complete, contractor Hochtief and specialist steelwork subcontractor Cleveland Bridge will hoist the old structure's Parker steel truss into the air.
In its place a sleek new five/six lane steel girder bridge is to be built, providing twice as many lanes for traffic, which for years had a tendency to jam up either side of Bishop's Bridge during rush hours (NCE 21/28 August 2003).
Client Westminster City Council hopes to banish congestion around the station by removing the bottleneck.
The bridge replacement is just one £24M component of a far wider £62M project to improve traffic flows. Westminster is bearing most of the cost itself, but has some support from Network Rail and BAA.
The new Bishop's Bridge is reasonably long - it will consist of three spans, one of 70m across the canal, and the second two, with a combined span of 101m, crossing the railway lines. It is crossing the lines that poses the greatest challenge and makes this job particularly interesting, says Hochtief project manager Sally Sunderland.
Hochtief is carrying out all work over and either side of the railway lines during Network Rail's planned night time maintenance possession. With Cleveland Bridge it has devised a way of removing the existing Parker truss without needing to buy extra possession time.
While the Parker truss is suspended the new bridge will be slid across from the canal side of the tracks to the station side, underneath the old structure. The Parker truss will then be lowered onto trailers and hauled off to be broken up. In this way Hochtief can be absolutely sure that no demolition debris will fall onto the tracks.
'The methodology grew out of the Ladbroke Grove rail crash and the need to keep signal sightlines at Paddington clear - there was no clearance to put scaffolding and a debris deck underneath [the old bridge], ' Sunderland reports.
To get the Parker truss to the point where it can be lifted clear of its piers it has been necessary to build four temporary lifting towers, one at each corner. These are founded on 20m deep, 4.3m diameter concrete caissons - one pair on Paddington's platform 10/11, and the other to the south of the railway lines.
These foundations will serve a double role for, as soon as the Parker truss is lifted, reinforced concrete piers for the new bridge will be constructed between the towers.
The towers themselves are heavy duty structures, standing 26m tall and weighing 50t each. Their tops are fitted with 400t strand jacks, which will be used to raise the truss 9m into the air.
'On this part of the job, the big issue was positioning of the towers - fitting them around the railway tracks, ' says Cleveland Bridge project manager Raj Soni. And because the new bridge is a good deal wider than the old, the towers have had to be offset by around 5m at platform 10/11. The new bridge will be launched between them. It has been impossible to lift the bridge directly at each corner, so the truss has been extended with add-on 'handles'.
'We've added new steel to extend the truss to where the towers are, ' explains Soni. 'At platform 10/11 we have a huge plate girder spanning between the towers, and the strand jacks will actually lift the girder, not the bridge itself.
Extension posts cantilever out from the ends of the bridge to connect with the plate girder.
At the canal end the lifting system is more like a bracket.'
Connections are bolted, with 'cheese plates' - steel plates cut with holes - used to pack between the riveted truss and the temporary works.
Cleveland Bridge has carried out extensive analysis of the Parker truss's steelwork to ensure it is not stressed by lifting in ways it was not designed for. To strengthen the old structure, corroded areas have been plated.
'The bridge will be up in the air for 15-18 months while we assemble and launch the new deck, ' says Sunderland.
'We have tried to maintain existing forces within most members to ensure the change in the means of support doesn't change the behaviour of the bridge. Any additional forces will go to new members, not old.'
In preparation for the lift the old structure's concrete deck and its heavy steel parapets have been stripped off, cutting 250t from its total weight. 'It's down to its bare bones, ' Sunderland says.
Trial lifts were taking place as NCE went to press to find out whether there was any physical connection between the truss and its bearings and to gauge whether it will behave as predicted. If there are any problems Cleveland Bridge has just enough time to come up with a remedial solution, says Soni.
Once the bridge is up in the air, Hochtief will be able to knock down the Parker truss's brick supports and get on with construction of the new piers, while Cleveland Bridge will be lifting the canal span into place:
this is a conventional steelconcrete composite structure.
Then the really interesting bit begins, says Soni.
The rail spans have a complex geometry, tapering from six lanes wide over the canal to five lanes wide at the south abutment. There is a height difference, with the north abutment around 5m higher than the south. The deck twists over its length. And because it will be supported at mid-point by the platform 10/11 pier, its section is some 1.3m deeper there than at either end. As the bridge is pushed out from the canal span abutment it will want to dip, slew, roll and rear up.
'The central [platform 10/11] pier and the canal pier will have sets of rollers which need to be adjusted in almost every direction. We'll be steering the deck in three dimensions, ' Soni says.
To accommodate the changes in deck thickness, rollers on the platform 10/11 pier will be mounted on rack and pinion type climbing jacks which can be lowered and raised as the deck advances. These will also help compensate for the expected 2m of longitudinal deflection in the bridge as it is launched.
The first 70m of the rail spans is scheduled to roll out over the tracks in April next year. There will then be a short pause while the rear 60m is added. Launching is due for completion in July next year.