Ten years on, BAA is extending London Stansted Airport. Would the construction team be able to respect Norman Foster's ground-breaking design yet still deliver substantial cost savings?
Excited passengers jostle for position IN front of the budget airline counters. Indicator boards flash out flight codes, beckoning them to the waiting aircraft that will fly them to Berlin or Barcelona for the price of an Indian mean for two.

In their eagerness to get the check-in over with, most passengers do not register the solid partition that stands, temporarily, in place of the glazed facade that once graced the west elevation of Foster and Partners' award-winning terminal building. But the partition separates them from a busy construction site where work is under way on a £57.5m extension.

Foster and Partners' seminal piece of airport architecture, which opened in 1991, features a light, uncluttered roof, held aloft on steel trees like a sail billowing over the single-level concourse. Buried below the ground, in an enormous undercroft, are all the services and access roads. Part of the original brief was to design the terminal in such a way that it could be extended once passenger traffic had increased enough to warrant it. In fact, planning permission for the original terminal included outline planning for its future extension. It is this extension that is now being built.

"The extension will respect Fosters' original concept," says Russell Batchelor, BAA's head of projects at Stansted, "although we have chosen not to proceed with Foster as architect, because he is not one of BAA's framework suppliers." Instead, the project architect is BAA-vetted Pascall + Watson, which will be working to Foster's original design concept for the extension.

The extension will add two more 36 m wide bays to the five existing bays. This will increase the concourse area by 11 500 m2, the number of check-in desks by 18 and will double the amount of retail space. It is part of the phase-two development of Stansted, a £200m package to increase passenger capacity at the airport. Other elements of the project include a new aircraft taxiway, additional car parks and a satellite terminal.

This is an important project for BAA. Using the original as a benchmark, the scheme will give the airport operator the opportunity to assess whether its drive for leaner construction and framework agreements will result in a terminal that looks like the one built 10 years earlier, but is put together more efficiently – a principle the budget airlines would surely approve of.

The project is also a test flight for BAA's revamped project process, explains Batchelor. The operator's original project process, which he describes as being about "consolidating and using best construction practice at the time", was defined in 1995. In 1999, BAA revised the scheme to incorporate lessons from the previous four years so that "it is more aligned to a manufacturing process". This is the first time the new process has been tested and will probably be the last opportunity before the much larger Heathrow Terminal 5 project kicks off.

The revamped project process coincides with E E BAA's reorganisation of its framework suppliers – designers and contractors BAA has agreed to work with over a 10-year period on the understanding that they share information and ideas with other members of the construction team to improve the construction process.

The suppliers were originally grouped into airport teams. However, this generated what Batchelor refers to as "silo thinking", with inconsistencies developing in the way construction was procured at different airports, making it hard for suppliers to standardise. The reorganised framework now transcends airport boundaries, although the airport teams remain site-oriented.

As a further incentive to drive down costs, BAA has introduced a project reward scheme. "We pay everybody what the work costs: we fix the overhead and profit associated with an agreed target figure," says Batchelor. "If a supplier can deliver a cost improvement, they increase their margin, and the benefit for us is less cost." At Stansted, if the scheme is delivered for less than the £57.5m target figure, and on time, BAA will share any saving 50:50 with the supply chain.

"It's about encouraging suppliers to work together," says Batchelor.

Enabling works started in January 2000. More than £1m was spent on separating the construction site from the passengers. Construction started in earnest in April.

The revamped project process appears to be working. The method of roof assembly, for example, has changed, although its appearance and form of construction remain true to the original. The steel trees have square trunks of vertical tubular steel members and horizontal ties, and they rise up from foundation level through the undercroft to a point 4 m above the main concourse. Here, four interconnected tubular steel columns spread out to connect with the 18 m2 grid of the roof structure.

Tubeworkers, the steelwork contractor for the original terminal building, went into liquidation not long after the project was completed.

When Rowan was awarded the contract for the terminal extension, the first thing it did was assemble a truncated prototype tree at its Nottinghamshire works. "You shouldn't practise on a construction site – get the guys who are going to build the thing to practise somewhere else," says David Coulson of construction manager Mace. Following the mock-up, Rowan changed some of the detailing and then produced a series of jigs to ensure that each of the dozen trees needed for the extension was assembled exactly.

Rowan also built the diagrid, the 18 × 18 m structural dome that supports the steel roof deck. The tree and roof structure was constructed using Tubeworkers' original fabrication drawings, to the same specification, except for the supporting columns below the concourse where the tubular steel of the original trees has been replaced by I-section columns to cut costs.

The diagrids were assembled on the airfield, where they had been delivered in large sections. Once the structure was complete, cladding was installed on the unit's square-shaped centre section along with the insulation and waterproofing, followed by the triangular rooflights. Finally, the steel roof deck was attached and the units were driven across the airfield to be craned into position.

"The main difference in the construction of the roof for the extension has been installing the cladding and glazing to the diagrids before lifting the units into position," says Coulson.

"We also ensured that more of the decking was fitted before the structures were lifted. Compared to the original scheme, we did a lot more of the assembly work on the ground. It saved more than 500 man-days with a subsequent increase in productivity and safety." However, waterproofing the junctions of the diagrid still had to be installed once the domes were in their elevated position on top of the trees.

BAA's project process also resulted in a complete rethink of the concourse floor structure. The original building had an 800 mm deep concrete coffered floor construction supported from reinforced concrete columns. The new extension will be supported on steel columns with a composite metal deck and concrete screed supported by a steel frame.

"The starting point for the change was the need to provide a floor," explains Batchelor. "So we had a brainstorming session with the consultants and suppliers on how we could do it." This generated a raft of solutions, "75% of which weren't worth pursuing". BAA's performance specification for floors gave values for factors such as loading and finish, but because BAA had not yet finalised the baggage handling system, the floor had to be capable of having holes knocked through it in the future to accommodate new risers.

"This narrowed the form of construction down to three or four possible solutions," says Batchelor. The design team looked at the impact of each option on the programme as well as cost; they also looked at the possibility of prefabrication.

A scoring matrix was produced from which the favoured solution emerged, explains Batchelor. This was a steel-frame system, and it saved eight weeks' programme time for the trades that followed.

Completion of the terminal was originally scheduled for Easter 2002. Coulson says this date has been revised because of improvements to the programme and that the terminal should be open in time for the Christmas rush in 2001. By this date, the cladding and handrail that were removed from the original terminal will have found a new home on the extension, along with the bow trusses that supported the original glazed facade. The glazing, however, will not be reused: BAA's performance requirements have changed, so laminated glass units will now replace the two layers of toughened glass of the original facade.

Batchelor estimates that by Christmas 2001, passenger numbers will have topped 13.5 million, just 1.5 million short of the terminal's maximum capacity. "We are only just keeping ahead of demand," he says. BAA is "already in discussion with the local authority to expand the airport's capacity beyond 15 million with a new terminal building," he says.