Holding up under fire
It is really refreshing to be presented with research that has been thoroughly explored and which has not fallen prey to the expediency of hurried publication. It is all the more surprising given that the research in question offers fundamental benefits to safety, costs and time.
Corus Construction Centre, in conjunction with the Building Research Establishment (BRE) and Steel Construction Institute (SCI), has been carrying out a test programme since 1994 to examine the real effects of fire on structural steelwork. Even though the Building Regulations and Eurocodes stipulate fire resistance standards (which dictate the amount of insulation applied to given components), observations of the actual aftermath of fires in steel buildings suggested to the researchers that the amount of fire protection being applied to some steel elements was excessive - and in some cases - unnecessary.
In 1990, a fire occurred in a partiallycompleted 14-storey office block in Broadgate in London and reached a temperature in excess of 1,000degreesC. The columns had deformed but the structure showed no signs of collapse, even though there was no f ire protec t ion to much of the steelwork and the sprinkler system was not yet operational.Other fires, such as the Churchill Plaza Building in Basingstoke and others around the world, offered similar evidence.What was it that was holding these structures up against all regulatory expectations?
At the BRE's test facility in Cardington in Bedfordshire, an eightstorey full-sized test building was built to replicate real conditions. The steel frame building comprised grid layout columns and beams with intermediate beams supporting a profiled metal former with a reinforced insitu concrete floor above.The external skin consisted of traditional brick/block cavity walls with large areas of glazing. The floors were packed with sandbags to reproduce actual loadings and combustible material was positioned around the floor plate. The building was then set on fire within one of the compartments.
The first fascinating piece of evidence indicated that, even though the building had aluminium windows (notoriously poor performers in fire), while the windows remained closed there was insufficient air flow to encourage combustion and the fire rapidly died out. This happened several times and raises questions about the number of fires that occur in the UK that never become sufficiently fierce to pose a problem. It also raises important questions about naturally ventilated building technology, which unmanaged, could possibly fan the flames more than air conditioned building management systems.
After smashing a window however, the test recommenced and the fire raged to a peak of 1,100degreesC and, as predicted, the building did not fail structurally. The subsequent research into the state of the remaining structure has been quantified in a new guidance publication, Fire Safe Design: a new approach to multi-storey steel framed buildings, which describes the new concept of 'fire engineered steelwork'. The years since the tests were carried out have been spent in developing a theoretical appraisal and producing scientific tabulations of the empirical measurements.
The principles of the research reveal the conditions for its successful application. It is imperative that designers do not get carried away and think that these test results apply universally; they apply to multistorey steel-framed buildings with composite decks. This is the basis of the test research and other structures, which may show similar patterns of inherent fire resistance, cannot yet be vouchsafed by this research method. Provisos, which are spelled out fully in the publication, must be adhered to.
One pre-requisite is that the main structure comprises a steel column grid supporting perimeter beams (perimeter to each compartment under consideration).These provide the lateral stability so there can be no reduction in standard fire protection to these members.The fire protection for columns and primary steel must extend around connections with secondary beams and continue up to the soffit.
The composite floor above must comprise profiled metal decking with reinforced concrete over it. The research stipulates that the reinforcing mesh be positioned 15mm to 40mm above the deck for trapezoidal decking profiles.
Importantly the guidance does not apply to precast concrete slab floors and fabricated beams or to lattice beams.
The secondary beams - that is the beams spanning between the perimeter beams and not connected to columns - can, however, forego fire insulation. These are the beams which break up the span of the floor above. The document explains how the omission of the insulation can be assessed satisfactorily.
The reason is straightforward. As the fire takes hold, the main structure's fire resistance performs its task and maintains its integrity and stability. The unprotected steelwork, however, loses its structural integrity reasonably quickly and buckles away from the soffit.The concrete is then left spanning the full area and, since it has not been designed as a self-supporting slab, bows in the middle.
However, as the slab is still supported at the perimeter and is suitably reinforced, it does not collapse. This is known as 'membrane action'.
The major costs incurred in fires in buildings of this type are internal fittings and fixtures and the structure tends to be written off when fires reach the level of intensity reproduced in these tests. The key factor to consider, therefore, is the suitability of the structure to withstand collapse - to minimise hazard to the occupants and the facilitation of safe escape.
As Jef Robinson of Corus Construction Centre says: 'We have built in a factor of safety.Used correctly, this system does not erode fire safety at all.' Dr David Moore of the BRE confirms that there has been considerable liaison with Building Control throughout the testing process (such that they will accept it as an alternative compliance with Approved Document Part E).
He says that insurers have also shown a healthy acceptance of the principle.
With calculation sheets of its own, Fire Safe Design has the potential to supersede some of the fire resistance calculations in the bible of plasterboard design, Gyproc's 'White Book'. It is usually assumed that beams designed in accordance with BS 5950-3 support the factored loads in normal design. The fire engineering calculations allow the designer to work out the residual bending resistance of the secondary beams in fire, relative to the utilisation of the beams for normal design. By reading across the tables, one can find out whether the designed beam and slab is adequate or whether one or other needs to be upsized. The mesh size and location can be changed to give greater tensile strength as the concrete bows, or the beam can be over-designed by a factor given in the tables. A sample table is shown.
Other calculations are necessary to test different meshes and mesh locations. It is also important to realise that as the secondary beams collapse, additional loads are applied to the perimeter beams. These must all be factored in to perimeter beam designs.
In essence then, omitting fire protection can make significant savings and this can only be a good thing. However, Mark Lawson, of the Steel Construction Institute points out that, rather than this being seen simply as a straightforward cost saving, the monies should preferably be reallocated towards a good sprinkler system.
It is well documented that the sprinkler system is one of the best mechanisms to prevent a fire from taking a hold and, if it is satisfactorily zoned, it can cause minimal water damage.
The research embodied in these calculations will enable considerable cost and time efficiencies to be made on schemes and/or the reallocation of resources into prevention rather than cure. It is an opportunity too good to miss.
The tests were sponsored by the DETR and European Coal and Steel Community, with further input from the TNO (Netherlands), CTICM (France), SCI (Steel Construction Industry, UK), Sheffield University
It is important that this system is understood fully. The BRE recommended that architects and engineers attend briefing meetings to recognise the benefits, raise any queries and become acquainted with the methods.
Workshops will be held on the following dates:
16 November Manchester 29 November Glasgow 6 December Birmingham 17 January London Additional dates are expected and full details are available on 01344 872776.