When the contractor building a hospital in Leeds decided on a new structural support system, the cost of the fire protection threatened to spiral. But, writes Alex Smith, a computerised fire-analysis tool took the heat off the specifications team and left the client with money to burn
Everybody, it seems, is happy with the new Nuffield Hospital in Leeds. Doctors are happy because they have some of the most technically advanced equipment in Europe at their disposal: facilities include eight operating theatres, eight intensive care beds and 80 private in-patient beds. Patients will appreciate the 11-storey building's light and airy design, which includes a glazed top-floor restaurant with views over the city. The client is satisfied with its gleaming new building, as it was delivered on time and within budget by design-and-build contractor Shepherd.

But the design and specification of the private hospital was not as straightforward as the smooth handover suggests. The tender issue drawing by architect Carey Jones was for a concrete-frame building, but Shepherd's bid proposed using steel instead. With steel, Shepherd calculated that the hospital could be built within the 90-week schedule desired by the client as opposed to the 100 weeks required for a concrete framed structure.

The drawback of steel is that the capital cost is significantly higher than that of concrete – up to 50% according to Nigel Hiorns, technical manager at Fedra, Buro Happold's fire engineering group. This is partly because steel requires fire protection; concrete is intrinsically fireproof. To make the steel option financially viable, Shepherd employed the services of Fedra. Its task was to value engineer the fire protection strategy and to enable specification of the most appropriate materials.

The use of steel made the project much more complicated for Fedra. It had to take into account the effects on fire performance of the interaction of the steel beams and columns with the thick-slab concrete floors. The use of steel also meant that the floors would be about 30 cm deeper than with concrete, which would have reduced ceiling heights. To get over this problem, Shepherd specified cellular beams, which are I-beams with large holes in them. Instead of services having to be run in suspended ceilings beneath the beams, they could be run through the holes instead.

Fedra used a complex structural analysis model called Vulcan to determine the most cost effective fire protection strategy for Nuffield. For the first time, the effects of fire on cellular beams were assessed by the model. Vulcan was also able to analyse the effects of fire on the junctions between the steel and concrete.

The Building Regulations allow a degree of flexibility in designing a fire strategy. The fire regulations contained in Part B offer prescriptive guidance, which if followed by designers ensures compliance, but it also states that in the case of large and complex buildings, compliance may be achieved by taking the whole safety package into account. This allows trade-offs to be made between elements of fire safety without adding to the risk of fire.

By using fire engineering, Fedra managed to reduce the structure's fire rating from 120 minutes to 60 minutes without increasing the risk. The assessment of the building's design indicated that the fire loads were low and the facade design allowed a significant amount of ventilation. These conditions result in fast burning, short-lived fires, which means fire protection can be reduced without compromising safety. Traditionally, no allowance has been made for the cooling effect of windows breaking because they were usually small and set in solid wall facades. However, in a modern glazed building, the heat that escapes through broken windows is more significant, and can therefore be taken into account when assessing a building's fire performance.

"As Part B is only guidance, it means that building control will accept lower fire ratings if it can be demonstrated there is no extra risk," says Hiorns. Without reducing the materials' fire rating to 60 minutes, protecting the steel would have been complex and expensive.

Hiorns says Fedra helped Shepherd save about £100,000 by taking a fire engineered approach. A large amount of that saving came about because Fedra found, using Vulcan, that 40% of the primary steel beams could be left without fire protection. This was partly as a result of what is called the 3D membrane action, which means beams supported by columns and joined at corners – primary beams – are more likely to sag during fire than secondary beams, which are connected to them. This enabled Fedra to leave many secondary beams without fire protection. It was able to persuade building control that this approach would not result in extra risk because Vulcan's analysis has been proved accurate by BRE's Cardington fire tests.

Once the client decided to use steel frame, Fedra went about helping to specify the most appropriate material to protect the beams and columns. The three options available were cementitious spray, intumescent paint and boarding. Hiorns likens cementitious spray to a Wallace and Gromit porridge machine. "It's not really suitable for a hospital, which you want to keep as clean as possible. It's also a wet trade, which makes it awkward to apply when there are other trades working nearby," he says.

Boarding around the cellular steel beams and services would have been a "bit of a pig" according to Hiorns, who says the complexity of the work would have made the labour costs prohibitive. However, the project team did decide to protect the steel columns with boarding because it could be done by the subcontractor that installed the partitioning. This integration of trades meant that Shepherd could put out a single package for the tendering of the partitions and boarding.

Intumescent paint was specified for the cellular steel beams, which was applied on site for greater quality control. The 2 mm coating meant that it did not interfere with the services running through the beams' circular holes.

At the project's end, Fedra was as happy with its achievements as the doctors were with their new building. And, it would seem that the fire engineer's satisfaction was not misplaced: "At the last seminar I went to five out of the six speakers there mentioned Nuffield in Leeds," he says with a smile.

Taking the heat out of fire protection costs

Architect Carey Jones had an important role in making the fire protection cost efficient. It devised the overall fire strategy for the 11-storey building and by thinking laterally was able to come up with a number of cost-effective fire safety solutions. Negotiations with Leeds council and West Yorkshire Fire Service to devise the fire strategy commenced in June 2000. The fire strategy was carried out in accordance with the Hospital Technical Memorandum 81, a code of practice that satisfies Part B of the Building Regulations. Information on factors such as fire separation lines, travel distances to escapes and fire loads was fed into Vulcan to assess how much fire protection the steel structure required. Each floor was divided into three or four fire compartments with some then further divided to create a minimum of two alternative protected escape zones. Circulation corridors and holding areas were introduced to allow beds to accumulate before descending lifts. This would allow for progressive horizontal evacuation. Total evacuation is always a last resort option because of the risks to patients who may be undergoing operations or be on life support. Carey Jones said other rooms, such as bedrooms, did not require the 30-minute fire protection set out in HTM81 because part of the document referred to geriatric/ psychiatric wards rather than ordinary wards. The practice also successfully argued that the 30-minute fire-rated doors and plasterboard provided a high level of fire protection anyway. This meant that it did not have to incorporate fire dampers in ventilation ductwork passing through these walls, and therefore these rooms were given a nominal 30-minute fire rating. The narrow width of the hospital along with the large size of the central lift area presented a tricky specification challenge. HTM81 usually requires the inclusion of a lobby to separate lifts from the corridors connecting fire compartments, but in this case the lift doors opened directly onto the corridor to save space. Carey Jones was able to overcome this potential conflict with building control by specifying smoke seals for the lift doors. Otis, the lift maker, does not provide smoke seals as standard. So, with the help of Nuffield’s fire adviser, Carey Jones came across Sematic, an Italian manufacturer of smoke seals, and convinced Otis that it could fit them to its lift doors. Performance specifications for the fire detection systems were provided by M&E engineer Waterman Gore at tender stage and developed further by Shepherd Engineering Services. All building components were specified on an “or similar” basis, which meant Shepherd was not limited to a similar material during the tendering process. “It’s frowned upon for us to specify products because contractors don’t think we can get the prices,” says Carey Jones associate Richard Barlow. On occasion, Shepherd made alternative suggestions to Carey Jones, which the architect accepted. “For the fire doors we specified cherrywood but Shepherd suggested white oak, which was much cheaper and looked fine,” says Barlow.

Vulcan: A futuristic fire-protection tool

Vulcan is a 3D structural analysis tool that is designed to predict the performance of steel in fires. It was developed by Sheffield University, and it is continually being modified using data gleaned from real-life situations. Vulcan can take such factors into consideration as the influence of the fire load, the effect of column protection and connection stiffness. By using Vulcan, Fedra was able to make an accurate assessment of where fire protection for the steel was required. As well as being able to analyse the performance of cellular beams, Vulcan was also able to take into account the effect of the 300 mm concrete floor slab. This provides the steel structure with rigidity and determines its acoustic performance. The interaction between the concrete and steel can provide the structure with additional strength in the event of a fire. “Previously the industry would have tested a single steel beam. Vulcan allows you to take into account the composite action of concrete and steel,” explains Hiorns. With the aid of Vulcan, which was used to analyse every room in the hospital, Fedra was able to ascertain that 40% of the steel beams required no form of fire protection.

Fire proofing