Dr Jim Glockling revisits research into commercial construction methods in the light of building repurposing.
The fire that occurred on the 17th floor of a residential tower block in Aldgate on the afternoon of 17 March 2022 is a further reminder of the challenges of safety management in tall buildings. In this environment, the residents are remote from help, assistance can be delayed, and it is very much up to the fire resisting properties of the building, the inbuilt detection and suppression systems, the attending fire services, and the associated management policies to assure occupant safety. Following incidents like this, the first question asked these days, for obvious reasons, is “what was the cladding?” In this case, the system was a traditional glazed curtain walling system, and indeed there are reports that wooden balconies may have contributed to spread, but that is not to say there is no need to discuss the use of this type of cladding in this type of environment.
Glazed curtain walling systems are synonymous with office blocks. Comprising an aluminium frame (transoms in the horizontal, and mullions in the vertical) attached to concrete floor slab fronts with aluminium brackets, and glazed in clear float glass at room level, and opaque toughened glass spandrel panels across floor slabs; they grace every city centre. Their use in the domestic and residential environment is relatively new, but is increasing as offices are repurposed for use as dwellings, and commercial building techniques and materials are applied to the residential sector. While a completely noncombustible form of cladding, by merit of the use of aluminium and glass, these systems offer little resistance to fire. Contrary to popular belief aluminium does not burn, but it is consumed in a fire, and this can be in large quantities. The working structural temperature limit for aluminium and its alloys range from 200oC to 250oC, and all strength is lost at 350oC, well below fire temperatures (800oC to 1200oC). Similarly glazing can be expected to tolerate 150oC to 200oC, although its performance is linked greatly to temperature differentials across the pane.
The fire performance consideration of glazed curtain walling systems is not therefore one of whether it will contribute and spread fire, but whether it can resist fire well enough to prevent successive floor-to-floor fire spread – fire breaking out of one floor, and then breaking back into the floor above.
For the purposes of calculating estimated maximum loss (EML) values for office type buildings, insurers financed a substantial research programme in 1999 designed to investigate the fire performance of glazed curtain walling systems and the associated failure modes. The research is as relevant today as it ever was. For any insurance policy, EML describes the amount of damage that may reasonably be expected to occur against which to assess premium. In the multi-storey environment, rather than assuming the entire building will be lost in any given fire event, an assumption might be made that assumes, for example, the floor on which the fire starts is lost to fire, two floors above the fire floor sustain smoke damage, and the floor below the fire floor has water damage from firefighting operations. A crude example, but in this case the EML would be four floors of the total number of floors, and the premium would be set accordingly. It is worth noting that many insurers have their own versions of EML, and the detailed meaning of the term may vary amongst them also, but the principles are the same.
Clearly, an EML of less than 100% demands that features of the building design warrant such an assessment. This is the current MMC challenge, where the use of combustible materials in the cladding, insulation, and structure of the building, in combination with assembly challenges such as combustible voids and green walling, can make it very difficult to robustly conclude that a fire will not result in a whole building loss. At the time of the research the typical structures clad in glazed curtail walling would have been quite traditional concrete or steel-framed buildings, and the cladding was viewed as the most likely route for mass vertical fire spread.
The large-scale research programme was conducted at the Loss Prevention Council. A specifically produced steel frame and concrete-floored building allowed the mounting of two and a half storeys of curtain glazed walling system. Typical offices were surveyed for fuel load to determine the size of an appropriate design fire. While wood cribs are traditionally used in fire tests of this type, there was a need to include a 20% component of plastic (by energy) to account for the increasing use of plastics in the office environment – this was shown to be an important factor in replicating flame lengths that could impact storeys above.
The experimental route taken was to substitute all elements of the system in turn with non-combustible alternatives so that the time of failure for each key component could be assessed in isolation. For every test, pig iron hung from the main aluminium grid, ensuring that the loadings of the system were correct for mimicking larger expanses of cladding. Additional tests sought to determine the benefits afforded by sprinkler protection and fire-resistant glazed systems.
The experimental work programme was divided into a number of logical phases.The experimental work programme was divided into a number of logical phases.
Phases
Phase 1 - Development of a stylised fire
Office environments typically possess fuel loadings of around 420 MJ m-2, of which 20% is derived from plastic-based sources (predominantly polypropylene).
Phase 2 - Evaluation of the performance of aluminium brackets to fire
Aluminium brackets, common in curtain wall installations, must support over 30kg per square metre of façade. Failure of the brackets offer the potential for large sections of the façade to fall away from the building during a fire.
Phase 3 - Evaluation of the performance of the aluminium framework to fire
The aluminium framework can react to a fire in a number of ways, all of which have an impact upon the overall performance of the system:
- Softening of the mullions at elevated temperatures may cause “slippage” at the pin-joint with the brackets.
- Distortion of the framework has the potential to cause the glazed units to break.
- If conditions are such that the fire consumes parts of the framework, then support for the glazed units will be lost.
- Distortion of the framework may cause gaps in the fire-stopping, enabling spread of hot gases and flames to the floor above.
Phase 4 - Evaluation of the performance of the window and spandrel units to fire
It is likely that the choice of materials for the window and spandrel units will have the greatest impact upon the reaction of a façade to fire. They have the potential to fail catastrophically, resulting in:
- sudden ventilation of the fire which will be expected to increase in size resulting in increased enclosure temperatures
- the formation of a pathway for smoke, gases, and heat to escape from the fire which may form a tall fire plume capable of threatening the floors above
- sudden loss of the fire-stopping material at the front of the floor slab enabling passage of the fire to the floor above behind the façade.
Phase 5 - Evaluation of active fire protection (Sprinklers)
A sprinkler system offers the possibility of reducing the fire threat and exposure potential to non-threatening levels.
Phase 6 - Evaluation of the passive fire protection
If all elements of the façade can be made to withstand the evolved fire temperatures for a suitable period of time, then break out and subsequent upward external fire spread is unlikely. This can be achieved by: (a) constructing the façade out of materials that may withstand the elevated fire temperatures, or (b) protecting a non-fire-resisting façade by applying cooling (window drenchers).
Results
Brackets
To evaluate the performance of the aluminium bracket sets, a totally fire-resisting façade was created (aside from the brackets) using mild steel mullion sections, and calcium silicate board in place of the window and spandrel sections. When the brackets reached temperatures of around 500oC at 20 minutes into the test, the first signs of slippage of the façade were measured. At 28 minutes the brackets were observed to fail at a temperature of 650oC. Inspection of the bracket sets after the test clearly showed the mullion pins to have pulled through the front plates under gravity.
Aluminum grid
To evaluate the performance of the aluminium façade frame, the frame was attached to the rig with mild steel bracket sets, and the window and spandrel panels were replaced with fire-resistant calcium silicate board. At around 24 minutes, when gas temperatures in the enclosure had reached 900oC, the mock window and spandrel panels fell away from the building due to large-scale consumption of the aluminium façade frame by the fire.
Glazing
A completely glazed façade system was installed onto the rig, comprising clear float glass windows (double glazed) and toughened glass spandrel panels (double glazed). Four minutes into the test the first pane of glass was observed to crack, but not fail. At around 13 minutes one of the spandrel panels fractured catastrophically. Now ventilated, the fire size increased rapidly from 3 MW to 6 MW. Destruction of the remainder of the façade thereafter was rapid. Flame lengths in excess of 4 metres were observed.
Active and passive protection measures
Sprinkler systems were demonstrated to maintain gas temperatures to levels that would not threaten the cladding system. The Pilkington Pyrodor fire-resisting glazing system in association with steel frame and bracketing system also showed itself to be capable of surviving the fire event.
Summary
For the stylised fire (which must be treated as conservative in size and development rate) the chronology of events was shown to be:
- 13 minutes - Failure of the glazed window and spandrel panel units.
- 24 minutes - Failure of the aluminium façade frames (subject to the windows staying intact).
- 28 minutes - Failure of the aluminium fixing brackets (subject to the windows and frame staying intact).
Most marked in these experiments is the rapidity with which large scale destruction occurs, and the uninhibited manner in which it may proceed to upper floors by internal (through the fire-stopping) and external (by the issuing fire plume) routes.
A key takeaway from this work was that, without sprinklers, there were no reasons why uncontrolled vertical fire spread would not be able to continue unchallenged. At the time, office buildings of this type would have been sprinkler protected. Now that this cladding is becoming commonplace in the residential environment, where sprinklers might not be mandated, perhaps it is the time to consider whether a review of safety measures is required, especially given the added challenges of sleeping risk and a less well-regulated environment for fire. Perhaps deployment of cladding systems that have little resistance to fire should be wedded to a requirement for sprinkler protection.
This article is based on a paper presented at the 2nd International Tall Building Fire Safety Conference, University of Greenwich, London. A study of external fire spread on multi-storey buildings and some prevention techniques by Dr Jim Glockling.
Fire & Risk Management is the UK’s market leading fire safety journal, published 10 times a year, and is available exclusively to FPA members in digital and print format depending on your requirements. You can find out more about our membership scheme here.
Dr Jim Glockling is Technical Director of the FPA and Director of RISCAuthority.