OUT OF THE FLAMES
They've been burning a new eight storey office block in one of the gigantic airship hangars at Cardington. Over and over again. The fires have included the biggest test fire ever lit. Computer models are still analysing the data, so the scientists are staying officially cautious until early next year when the full data comes out on cd-rom.
But they know that the results have been sensational. Millions of pounds in unnecessary fire protection could be saved - the live tests have indicated clearly that current assumptions about safety margins for composite metal- deck steel buildings are too pessimistic and that it may no longer be necessary to fire-protect all the structural beams in buildings for 60 minutes fire resistance, which covers 80 per cent of all multi-storey buildings in the uk. A hope too is that fire engineering will become integrated into structural engineering rather than remaining a tiresome add-on.
British Steel has contributed to the Cardington programme which is supported by the European Coal and Steel Community and the Department of the Environment, Transport and the Regions in collaboration with European institutes tno (Netherlands), cticm (France), the Steel Construction Institute and Sheffield University. The structure, designed by engineer Peter Brett Associates, is an eight-storey braced non-sway frame with steel-deck floor slabs typical of current multi-storey steel construction. Beam spans were 6m and 9m with partial-depth end plate beam/column connections and fin-plate secondary beam fixings. Imposed loads were those of a typical office. External cladding was also typical: a spandrel wall of blockwork with glazing above.
The tests began by comparing the performance of a single restrained 9m- long beam with that of a similar beam built into the test frame. The latter performed significantly better. This confirmed the scientists' belief that, as had been observed in the 1990 Broadgate fire, it was unrealistic to design on the basis of the performance of isolated test beams. By the end of the tests it was clear that the floors of multi-storey buildings incorporating composite floors comprising steel beams, permanent steel shuttering and integrated reinforced-concrete slabs, have a major impact on the structure's performance in fires. However once a column on a lower floor goes it has a partial domino effect on the columns above. So columns still need to be protected.
Beams are another story. In very high temperatures up to 1200degreesC the test beams noticeably deflected by as much as 0.5m in 9m. But although there was no strength left in the beams themselves, the floors retained their integrity, probably because of the mesh and permanent steel shuttering. And there was no transmission of fire up through floors for the same reason.
Although with hindsight you might say it was not entirely surprising to find that structural members rigidly connected also interacted together under high stresses, nobody before had done the live research which showed how this might work out in real life. In fact the tests were conducted to verify computer models which had suggested the results. British Steel's Jef Robinson, metallurgist and manager of market development for British Steel, Sections, Plates & Commercial Steels says: 'It's always been assumed that 550degreesC is the temperature at which steel will collapse. The Cardington tests suggest that that assumption can no longer be sustained. The new question is: how does steel behave in real structures? At the moment nobody knows what the real safety limits are. They may be very small. The probability is, however, that they are large and that our existing margins of safety are unreasonable.
'But,' he warns, 'we have now to get the results out to the scientific and engineering community and then take the information and their observations and produce an outline guidance.'
Although unprotected beams and composite floors remained safe it is likely that new guidance will call for protection to long-span beams while allowing secondary beams to remain unprotected.