20 June 2013

Mersey Gateway preferred bidder announced

A preferred bidder has been announced for the £2bn Mersey Gateway scheme in North West England. The Merseylink consortium (Kier, FCC, Samsung) now enters final financial negotiations with the promoter, Halton Borough Council, leading to contract award later this year.
 

While this is good news for anyone who like big bridges and to see people employed welding together lumps of steel and pouring tons of concrete into holes, I am more interested in what it means for the design of the bridge.

The original design was by Gifford (now Ramboll) with Knight Architects, and was used to secure planning consent prior to inviting construction tenders. Their design featured three mono-tower masts supporting the deck via "harp"-layout cable-stays, i.e. an arrangement of parallel cables. This is visually very attractive, but the low slope of cables near the bridge towers mean the deck is not well-supported in these areas. The design also allows for essentially a single plane of cables along the bridge deck centreline (actually, two planes very close to one another), which is visually very open and legible from all perspectives. This is the original design:


For multi-span cable-stayed bridges (with more than two towers), a key challenge is how to stabilise the spans, as they are far less stiff than a bridge with only one or two towers. The original Mersey Gateway design addressed this issue using a stiff steel truss deck, pictured in cross-section below (an excerpt from the planning drawings), which was conveniently deep enough to also accommodate future light rail tracks below the road deck (an idea later dropped from the scheme requirements).


A deck of this type would be lighter than a concrete alternative, and subject to reduced wind load. However, the cost of steel fabrication and site assembly would be high.

I mentioned the bridge in May 2012, noting that the promoter was sensibly varying the planning conditions, to give tendering contractors flexibility to reduce costs. I commented that "the structure which will be built is unlikely to retain either the harp cable layout or the truss deck, although I imagine the single plane of cables has a fair chance of surviving."

It's always nice to be proven right, as this is precisely what has happened in the preferred bidder's design, which I understand is by Flint and Neill. The cable arrangement is replaced with a single-plane "fan"-layout which is structurally much stiffer, and the single plane will be less expensive to install than two closely-spaced planes. This is the proposed design:



In place of the steel truss, the new design uses a conventional post-tensioned concrete box girder, which judging from the video will be cast in-situ in sections. The image below is taken from the project's new publicity video and shows the use of internal steel frames to transfer load between the girder webs and the cable-stays:


I imagine the deck will be shallower overall than the original truss design, which will make up for any heaviness of the solid concrete. I also expect that it makes the transition to approach viaducts very straightforward, as the same box girder cross-section can be applied consistently.

There is one significant visual disadvantage to the preferred bidder's proposal, which comes in the relationship between the tower above and below deck. This is a common problem with large cable-stayed bridges, where a slender tower above deck often gives way to an extremely stocky tower below deck. The Dartford Crossing is a particularly egregious example of how bad this can look. The Millau Viaduct shows how it can be done well.


In the Knight / Gifford design, the truss deck could be pierced, so that the bridge mast and piers form one continuous column passing through the roadway. A similar approach was used on Stonecutters Bridge in Hong Kong, to pick just one example. This is clearly impossible with a concrete box girder, and so the contractor's alternative design features piers below-deck which are much bulkier than the tower above, and appearing doubly so because of the large crossheads used on the two taller piers to provide torsional restraint to the whole deck.

Despite this, I think the promoter's decision to allow a certain range of variation from the original design was clearly the right approach. We will have to wait until later in the year to see whether it has realised the cost savings they expected.

No comments: