Fire protection: full-frame structural fire engineering

Sensible precautions

27 October 2017

Full-frame structural fire engineering offers a practical approach to the analysis of steel-frame and composite concrete slab structures, according to Cate Clooney and Mark Jones


Full-frame structural fire engineering analysis is carried out to enhance the passive fire protection of steel-framed buildings with concrete composite slabs. The technique can help significantly reduce the need for passive fire protection on the secondary steel beams, and in many cases mean that none is needed at all.

Testing context

Normally, fire tests are only carried out on individual beams or columns because of the limited sizes of the furnaces where testing takes place. Tests are carried out to well-defined standards, and are the basis of certification for passive protection products applied to structures so that they can meet the fire resistance periods specified in the Building Regulations guidance.

However, when connected together into a frame, the structure behaves differently – usually better than the individual elements, because there are different load paths in the composite slab and protected primary structure around it, as tests such as the Cardington series have demonstrated. This redistribution of load is commonly referred to as “membrane action”.

Conventional fire design for steel-framed buildings usually takes the following approach:

  • design the structure in such a way that it achieves the required performance at ambient temperature
  • apply insulation to the structure so that, if a fire occurs, the steel should not be heated beyond the point where the structure is still adequate.

No analysis is normally carried out of the structural behaviour at high temperatures – it is usually assumed that the approach detailed above examines enough safety factors to ensure that the building is unlikely to fail. While very few modern buildings collapse due to fire, it does not necessarily mean this is the most efficient design approach, however.

The methodology of Vulcan – developed at the University of Sheffield – and similar software is different in that it aims to analyse the complete structural frame or subframe to assess how it would behave in the event of various fire scenarios. By understanding the frame’s behaviour when subjected to a fire, it is possible to identify and apply protection only to the steel members that need to retain their strength and rigidity to ensure the building’s overall structural stability.

Exova Warringtonfire uses Vulcan to allow experienced fire engineers to model the way that the entire frame would behave in a fire. This enables more flexibility in the design approaches, and can often reduce costs significantly.

Load-bearing legislation

Requirement B3 of the Building Regulations 2010 states: “The building shall be designed and constructed so that, in the event of a fire, its stability will be maintained for a reasonable period.”

The aim of the structural fire engineering analysis is to show that this requirement can be met by targeting protection effectively in the building structure – without reducing the level of safety or the performance of the building in a fire.

This approach simply directs fire protection to those parts of the structure where it will perform a useful purpose. Any structure that only supports a roof would not require fire protection, for example, as Approved Document B or BS 9999 recommend.

Both of these recognise that alternative fire-engineered approaches may be used in order to comply with the Building Regulations, which provides an opportunity for a performance-based design approach. According to BS 9999, beams and columns should maintain the load-bearing capacity for the required fire resistance period when tested to BS 476, part 20 of which states that a column shall be deemed to have failed when it cannot support the load. This can be checked by the finite element analysis, because the model stops running once any single structural element loses its load-bearing capacity.

BS 476 part 20 stipulates that a beam shall also be deemed to have failed if it can no longer support the load, and limits the maximum deflection of a beam to L/20, where L signifies the clear span in metres. This limit is also applied to steel beams without external fire protection.

BS 9999 requires floors to maintain load-bearing capacity, integrity and insulation for the required fire resistance period. As with the beams and columns, a slab shall be deemed to have failed if it is no longer able to support the load, with a maximum allowed deflection of L/20 under BS 476.

Part 20 of BS 476 explains that failure to maintain integrity shall be deemed to have occurred when a floor slab collapses, there is sustained flaming on the unexposed face, or the criteria for impermeability are exceeded; impermeability is normally satisfied for composite slabs. Should a slab collapse, the finite element model will then stop. For composite slabs, insulation is not normally a problem.

Due to the nature of composite steel-framed buildings, localised failure – that is when a limited number of structural elements lose stability – is unlikely to cause the whole frame to collapse. However, should the model stop before the required fire resistance period, for example 90 minutes, the structural frame is not considered to meet the fire resistance requirements.

The required fire resistance period can be obtained from guidance documents such as Approved Document B or BS 9999. The latter also allows the designer to take account of available ventilation when considering the severity of a fire and the subsequent fire resistance period required, which draws on the “time-equivalent” value, as further detailed in BSI Published Documents PD 7974-3 or PD 6688-1-2, outside the scope of this article.

Modelling considerations

This analysis is usually carried out on buildings that have sprinklers fitted or are otherwise compartmented floor by floor. For a building that has sprinklers throughout, it is assumed to be effectively compartmented on a floor-by-floor basis; that is, a fire is unlikely to spread to more than one floor.

Due to the nature of composite steel-framed buildings, localised failure – that is when a limited number of structural elements lose stability – is unlikely to cause the whole frame to collapse

Finite element modelling has been carried out using Vulcan software to show that the steelwork in each compartment – columns, beams and the slabs above – will not fail when exposed to the conditions for the required fire resistance period under BS 476; say, for 90 minutes.

The finite element model should be developed through close consultation with the structural engineers to ensure that it accurately represents all of the following features:

  • concrete slab details, including cross-section shape, location and size of rebar and concrete grade
  • steel sizes; each beam can be entered into Vulcan in bespoke fashion, enabling inclusion of hollow and parallel flange channel sections
  • type of connection between beams and columns
  • boundary conditions
  • loading at the fire limit state; this refers to exceptional circumstances when a structure is experiencing an extreme accidental event of low probability, and to model the state slab, perimeter beam and column loadings are multiplied by the appropriate load factors derived from Eurocode or BS 5950-8, if the building has been designed to that standard
  • any non-composite elements of the structure.

The method of modelling cellular beams, within the limitations of the program, should also be justified.

For each steel beam and column, it will be necessary to apply a bespoke time–temperature curve that reflects either the response of the unprotected steel member to the standard fire curve, or a protected response in which the temperature of the protected member is restrained to its limiting temperature throughout the course of the analysis. BS EN 1993-1-2 and BS EN 1994-1-2 both contain guidance on these different temperature responses.

Even if the building has sprinklers fitted, it is assumed that a fire may involve a whole floor or compartment, ignoring any non-fire-rated subdivision, but will not spread to other floors or compartments. Therefore, it is reasonable to model one floor at a time. Each subframe will be exposed to BS 476-standard fire for the required period of resistance. This is a conservative assumption, because the provision of the sprinklers will isolate the fire to a far smaller area than those used in the models.

Because this is a computational technique, it is appropriate for the designers to carry out sensitivity studies concerning the model resolution to ensure that the results have converged, without necessarily using an overrefined model mesh.

Interpretation of results

Results should generally report the profile observed for the most severe slab and beam deflections, with reference to the acceptance criteria of BS 476 part 20. The model runs will normally involve several iterations to determine the optimum configuration of unprotected structures. The structural drawings can then be marked up and supplied to the design team to indicate the extent of unprotected secondary beam structure that is permitted.

Where floorplates are similar across different levels of a building, it may be appropriate to consider the level that has the highest loading and apply the results of that model to the similar floors elsewhere, without necessarily having to model each floor individually. Any changes to the structural design following the modelling would, however, necessitate remodelling to confirm or update the original recommendations.

Cate Cooney and Mark Jones are Principal Consultants at Exova Warringtonfire

Further information