The management of smoke in common access areas in flats and maisonettes is vital for the safe escape of occupants. To support the development of Part B of the Building Regulations and keep pace with improving technologies Stewart Miles senior consultant at BRE, has investigated ventilation options.
BRE has recently undertaken a research programme for the Buildings Division of the Office of the Deputy Prime Minister (ODPM). This is to reinforce the development of Part B of the Building Regulations for England and Wales and its supporting guidance. It’s important that this work is undertaken, as it enables the Regulations and guidance to keep pace with new and improved technologies and changing occupational needs. One project this research programme has examined is the smoke ventilation of common access areas of flats and maisonettes, and their escape procedures.
Smoke containment, provided by the provisions for compartmentation and smoke-rated fire doors and coupled with limiting travel distances, play key roles in smoke management strategies for flats and maisonettes. Where smoke does enter the common access areas, smoke ventilation and control measures offer different levels of additional protection. However, there is a current lack of agreement on the requirements for smoke ventilation in common areas, the performance of alternative measures and how the Building Regulations and guidance should be interpreted.
Assisted by project partners Buro Happold and the University of Ulster, and with guidance from a steering group drawn from industry, government and other enforcing bodies, Stewart Miles and Norman Marshall of BRE have studied a range of smoke ventilation measures using computational fluid dynamics (CFD) and reduced-scale physical modelling (see panel opposite). In addition, they were supported by a review of published research and current practice in the UK and overseas. The project has considered the initial stages of a fire, where occupants would be making their escape, and also the later stages, where fire fighting operations may have started. It has also considered when evacuation from other stories of buildings may be needed.
In the majority of the experiments and simulations, a steady fire has been assumed, and the work has been conducted to yield steady-state conditions. Furthermore, a fixed, open door gap from the fire compartment to the adjoining lobby/corridor, and also from the lobby to the stairs, has been assumed. This represents scenarios where fire doors have not closed properly. The experiment provided a suitable means to compare the relative performance of different smoke ventilation designs under smoky conditions. However, it should be noted that this approach, with fixed door gaps or openings, was designed to enable a relative assessment of the performance of alternative schemes under onerous conditions, and not to generate predictions for specific ‘real-life’ scenarios (this may obvious due to a steady fire being assumed). Correct operation of fire doors should limit the amount of smoke transported into the common areas and is an important ‘front-line’ defence, particularly when residents are escaping before the fire-service arrives.
The range of smoke ventilation designs investigated included external wall ventilators, naturally ventilated smoke shafts, mechanical smoke exhaust from the corridor, and pressurisation of the stairway with and without smoke/air relief from the corridor. While the provision for the replacement air was investigated in both the physical and CFD studies, the influence of adverse wind and stack pressures was also studied as part of the CFD analysis.
Physical Scale Modelling
The basic principle of reduced scale modelling is to build a scale model of the geometry, and to undertake experiments that various dimensionless numbers and groups are preserved at reduced and full scale. This means that the results at reduced scale can be extrapolated to full-scale. In Froude Number modelling, the temperature of the gases is the same at both model and full scale, while other parameters such as fire heat release rate, pressure and velocity at one scale can be translated to the full scale.
Over 70 experiments, with a 1/5th scale model of a six storey building containing a fire inside a first floor dwelling and with various arrangements of lobbies, corridors and stairs was conducted. The model was based on one used in a previous project to examine the performance of natural smoke ventilation in fire fighting shafts. Additional features added were residential corridor sections and mechanical smoke extraction shafts.
One arrangement of the 1/5th scale model had a common access lobby ventilated by a natural smoke shaft, and was adjoined by a stair and a fire compartment. In addition to the door gap opening to the lobby, the fire compartment was ventilated with an opening at the rear of the building. This provided sufficient air for combustion to occur and was maintained sufficiently small so that smoke-filled conditions were generated inside the compartment, thus providing a plug of smoke at the door opening.
Temperature, velocity and carbon dioxide concentration (from which an estimate of visibility distance was calculated) measures were recorded. Figure 2 shows an example of smoke travelling along the ceiling of a section of corridor.
The physical scale experiments showed that after a prolonged exposure to smoke, protecting the stair and maintaining good conditions inside the adjoining corridor was not possible by natural smoke ventilation. But by carefully setting the air-flow rate, using mechanical smoke extraction from the corridor and with adequate provisions for replacement air, it was possible to maintain a stratified smoke layer in the corridor, while limiting the amount of smoke entering the stair. By either extracting smoke from the corridor mechanically at a sufficiently high rate, or by using naturally ventilated smoke shafts located in the corridor, it was possible to protect the adjoining stair from smoke ingress, but this was at the expense of smoke filled conditions inside the corridor.
The CFD study supported the findings from the physical modelling study, and extended the scope of the project to examine further smoke-ventilation procedures. These were to use larger fire sizes and door openings (which is akin to a fire fighting situation), assess the effect of adverse wind and investigate stack pressures on natural ventilation. Figure 3 illustrates part of one CFD geometry, containing a fire compartment, corridor and smoke shaft.
Figure 4 illustrates a typical numerical grid. In the corridor or lobby, the typical maximum cell dimension was 0.2 m. This was reduced at critical locations, such as at the narrow door gaps, to 0.025m. Larger cell dimensions were allowed at other locations, notably in the vertical direction away from the fire floor, where values up to approximately 0.5 m were used.
For the majority of the simulations, the fire compartment was located on the second floor of a five storey building. Additional simulations were performed with the fire compartment on the second floor of 10 storey and 20 storey buildings. In all of these scenarios, the fire compartment was ventilated to the outside through a small, low-level opening. This ensured that air was available for combustion, and provided for a completely smoke-logged compartment.
A steady fire size of 0.25 MW, combined with a 0.1m door gap to the adjoining corridor (extending the full height of the 2m door) was used in the main series of simulations – this provided a smoke filled compartment. Different scenarios to explore simulations for larger fire sizes of 1MW and 2.5MW (which had the compartment door opened to 0.5 and 0.78m) were conducted. This provided conditions of greater thermal load, akin to those possibly encountered during fire fighting operations.
Each CFD simulation illustrated predictions for the distribution of smoke and temperature throughout the modelled building. Other information, such as airflow speeds and pressure distributions, was examined. Figure 5 illustrates a typical output from the CFD model; in this example there is some smoke reaching the stairwell.
Main Findings
The findings were taken from an assessment of all three components of the project, that is the physical scale and CFD modelling studies, and the literature review and assessment of current practice.
The project has demonstrated that, if exposed to smoke from a dwelling fire for more than a short duration, the adjoining common corridor/lobby can be expected, to become smoke filled. In the absence of appropriate smoke management measures, neighbouring corridors, lobbies and stairwells can also become smoke filled. For this reason, smoke containment, provided by the provisions for compartmentation and by smoke-rated fire doors, and coupled with limiting travel distances, plays a key role in the overall smoke management strategy for flats and maisonettes. Furthermore, the additional benefit to life safety provided by two escape paths (and when present, two escape stairs) has been reinforced by the project findings.
While external wall vents to the lobby/corridor may then, in some circumstances, maintain tenable conditions inside these spaces, an engineered mechanical solution would be required. This could be achieved by direct pressurisation of the corridors and lobbies (generally in combination with pressurising the stairs) with air/smoke relief from the dwellings, or by smoke extraction directly from the corridor and lobby by a specially designed system. The latter could include distributed ceiling exhaust vents or cross-corridor ‘smoke dispersal’ by means of supply and extract at opposite ends. However, it must be noted that these smoke extraction methods require careful fire safety engineering design.
If the remit of the smoke ventilation scheme is primarily to offer additional smoke protection to the common stair, then a range of relatively straightforward natural and mechanical approaches are suitable. These include:
– Natural smoke venting. Arguably external wall ventilators provide the simplest and toughest approach. These are located in the adjoining common lobby/corridor. However, some smoke can be expected to enter the stairs if the door from the common corridor is not closed when not in use.
In the absence of adverse wind or building stack effects, natural venting into a smoke shaft located in the adjoining lobby or corridor protects the stairs very well, albeit at the expense of leaving the former smoke filled if exposed for more than a short duration. This approach depressurises the corridor next to the adjoining stairway. While it can offer better protection to the common stairs than an external wall ventilator, a smoke shaft arguably requires greater design effort and post-occupation maintenance.
– Suitably designed mechanical ventilation systems. These can provide protection to the stairs that is, in theory, resilient to adverse wind and building-stack effect pressures. This can be achieved by two methods: either by depressurising the corridor or lobby relative to the adjoining the stair (extracting air/smoke from the former) or by directly pressurising the stair with air/smoke relief from the corridor or lobby. Conditions inside the adjoining corridor/lobby are not, in general, maintained. By an informed choice of extraction rate, however, a depressurised corridor scheme may be able to prevent smoke migration to the stairs, while at the same time maintaining some degree of smoke stratification in the corridor.
The ‘cold-flow’ (non-fire) performance requirement of the mechanical ventilation systems for either depressurising the corridor or for pressurising the staircase can be defined in terms of appropriate closed door pressure differentials and open door airflow speeds. These are broadly compatible with those in BS 5588-4:1998. Indeed, the numerical simulations indicate that in non-firefighting situations, there could be scope to reduce the design criteria. However, it should be noted that the ‘cold-flow’ CFD simulations did not precisely replicate the conditions required according to BS 5588-4:1998, and that any design criteria should compensate for adverse wind andstack pressures.
Figures 1 and 2. An example of the physical scale modelling with a smoke layer travelling along a common corridor (bottom photo) from the fire compartment.
Figure 3. Computational Fluid Dynamics (CFD) geometry showing fire compartment corridor and smoke shaft.
Figure 4. Typical CFD numerical grid.
Figure 5. Contour plots of temperature for a corridor with adjoining natural smoke shaft.
A final report, giving details of the whole project, including the modelling studies and review, is available from the BRE website at: www.bre.co.uk/adb
References
Harrison R and Miles S (2002).
Smoke Shafts Protecting Fire-fighting Shafts:
Their Performance and Design. BRE Project Report No. 79204.
Harrison R. and Miles S (2003).
Smoke ventilation of fire-fighting shafts.
Fire Safety Engineering January 2003 pp. 16-17.
[…] the other hand, in flats and maisonettes, smoke containment in communal areas and escape routes will be significantly important, especially […]