FIRE ENGINEERING AT THE GLA BUILDING

Introduction

In order to offer clients increased robustness plus some cost savings, Arup Fire is pursuing a performance based approach to structural fire resistance on several major building projects in London. The first of these was the GLA building, designed by Foster and Partners, which was opened last year. Now known as City Hall the building provides offices and committee rooms for the Greater London Authority, as well as public assembly and exhibition spaces.

The use of a fire engineering approach enabled the exposed steel columns to be fire-protected with a thin-film intumescent coating, giving a high quality finish. Natural ventilation was designed to limit the temperature of smoke layers, allowing glazing that did not have a fire rating to be used in compartment elements. Extensive discussion with the building control [London Borough of Southwark] and fire [London Fire and Emergency Planning Authority] authorities, enabled appropriate evacuation procedures to be worked out for the mixed population of public and office staff.

Building description

This is an office building for the Mayor’s administration, with a council chamber for the GLA and several meeting rooms. It is also a public building with general access to the council chamber and meeting rooms, and to the top floor which is a multi-purpose area. Access is at ground and lower ground levels.
The council chamber overlooks the Thames and is the base of an atrium that rises past all the upper floors. A spiral ramp serves every floor. Above chamber level the ramp is inside the atrium. The upper section of ramp is separated from the section serving ground and lower ground, as it passes the chamber level. Ground and lower ground floors are linked by an elliptical atrium. The night photo below shows the lighted upper atrium façade with the ramp visible inside it.

Structural fire engineering

Modern office buildings, retail facilities and the like typically incorporate a high proportion of glazed or non fire-rated elements for facades and atria walls. For this type of construction a performance based approach, allowing for ventilation of heat through sections of façade penetrated by fire, generally shows that the guidance on fire resistance in Approved Document B is conservative. The fire resistance periods can typically be reduced to 60 minutes using a Performance Based approach.

Reducing the fire protection on structural steel members can result in a significant cost saving and other benefits, including:

Traditionally the fire resistance of structural members has been determined in Standard Fire Tests. The time-temperature environment in the Standard Fire Test represents a more severe heating condition compared to that in many typical natural fire compartments. In a well-ventilated compartment the duration and/or the severity of the time-temperature environment is generally less than in a Standard Fire Test. The effect of ventilation and fire load on fire severity is illustrated in Figure 2. Fire tests were conducted in compartments where the fire load and the natural ventilation were varied. The well ventilated compartments experienced lower temperatures and fires of shorter duration. In Figure 2 the numbers identified with each curve indicate the fire load density in kg/m² (ie 60, 30 or 15) and the ventilation area as a proportion of the façade area (ie ½ or ¼).

The compartments used in the tests were small by modern standards but the results are indicative of the influence of fire load and ventilation on the time-temperature environment generated within fire compartments.

Key Factors for Time-Equivalent Analysis

When a fire reaches a stage where there is full involvement of the combustibles within a compartment (known as flashover), the intensity of the heat in the hot smoke layer will cause glazing and non-fire resisting facades to fail, allowing hot gases to escape (see Figure 3). Similarly, openings to atria will also allow hot gases to escape. The temperatures reached in a compartment and the duration of a fire depend on natural ventilation through openings to atria and glazing or non-fire resisting facades that fail in a fire.

Factors that affect the intensity and duration of a fire include: fuel load (quantity and type), geometry of the compartment, the thermal insulation provided by the linings and the natural ventilation following glazing failure.

The principles of the time-equivalent analysis have been understood for many years. Technical papers have been published by Law (1978), and Pettersson (1976). To date, time-equivalent analysis has been used on a limited number of projects to determine the effect of fire on structural members within individual compartments. Publication of the analysis method in Eurocode 1 (ENV 1991-2-2:1995) some years ago, made the approach more accessible to the broader engineering community.
The principle of time equivalence is that the member is exposed to an equivalent heat dose. The equivalent fire severity can be stated more formally as: “the time of exposure to the standard fire test that would result in the same maximum temperature in a protected steel member as would occur in a complete burnout of a fire compartment”.

The thermal inertia of the compartment linings also affects, but to a lesser extent, the intensity and the duration of a fire. Linings that are more insulating, or have a high thermal inertia such as gypsum plaster, slow down heat transfer from the compartment to the walls and ceilings, with the result that the temperatures and the fire duration in a well insulated compartment are greater.

Practical Application of structural fire engineering: the Greater London Authority Building

One of the first applications of the time-equivalent analysis technique to a complete building structure was at the new GLA building. As the illustrations 5, 6, show this is a fully glazed building with the potential for a high heat loss rate to the exterior in case of fire. The time-equivalent analysis method is appropriate for compartments where the fuel load can be characterised, ie offices, retail areas, schools, hospitals, residential apartments and hotels.

Where sprinklers are installed a reduction factor is applied to the calculated value for fire resistance, recognising that sprinklers reduce the intensity of a fire.
A factor of safety for consequence of structural failure is also applied to the calculated value of fire resistance, introducing another degree of conservatism for taller buildings.

When the Structural Eurocode mentioned in 2.2 above was published in the UK a National Application Document [NAD] was included. The NAD gives UK values for parameters in the time-equivalent time analysis, to enable the safety objectives of the Building Regulations to be met. This was the approach used at the GLA building. The use of the analysis method generally results in reductions in the structural fire resistance compared to the values recommended in the Approved Document B, where the building is potentially well ventilated in a fire through a high proportion of unprotected facade.

In the case of the GLA project a 60 minute standard was agreed for the structural elements. The building is over 30m in height and under AD B would have been expected to have 120 minutes. The lower period opened up a wider range of options for fire protection the structural steel framework with architectural benefits as well as economic ones.

Interestingly, when the method is applied to the style of office building common in the first half of the 20th century [when Ingeberg was working], such as figure 7, the fire resistance can be as high as 4 hours. This may be justified by the lower percentage of glazed openings and high thermal inertia of the construction.
Figures 5 and 6, the Greater London Authority building, showing the high proportion of exterior glazing [the spandrel panels are fritted glass].

Temperature control ventilation

Although the floors are constructed as fire resisting compartment floors, the public and private spaces are not separated by fire-rated construction everywhere. The green line on figure 8 represents the boundary between the upper and lower atria. In many places this separation is formed by glazed screens using ordinary laminated glass.

Automatic vents open in the top of the atrium to remove hot smoke, while vents near the bottom admit cooler make-up air from the lower atrium which has a series of automatic opening vents that double either as inlets, in this scenario, or as smoke outlets in the case of a fire in the lower atrium. The complicated geometry of these spaces makes it impossible to illustrate the air paths precisely on a 2D representation.

Means of escape

The public areas of the building are evacuated simultaneously, but the office areas have a phased evacuation arrangement. Depending on the fire location the office levels may not be evacuated in the first instance. A voice alarm system is programmed to provide the appropriate messages.

There are two firefighting pressurised stairs in the core, sized on the basis of use by the top [public] storey, two office storeys and the council chamber [with up to 300 occupants] at the same time. Allowance was made for the time delay of people from the top floor reaching the council chamber level, in assessing peak flows.

A fire in the council chamber should be limited by the fire load. The furniture and fittings were designed to keep the fire load density low.

In the event that a fire on an office level penetrates the atrium enclosure the temperature control ventilation, described above, should preserve the integrity of the atrium enclosure indefinitely, assuming a heat release rate into the atrium of up to 2.5MW.

The building is fitted with sprinklers so it is unlikely that an office fire would grow large enough to penetrate the atrium in the first place.

Conclusions
The use of fire safety engineering