6. Steelwork Fire Resistance

 

Fire resistance is measured by the length of time an element can survive in a standard fire without exceeding specified performance limits.

 

Figure 27. BS476 Part 20 standard time-temperature relationship for fire tests.

 

Fire Resistance is usually expressed in terms of compliance with a test regime outlined in BS476 Part 20 and 21 (11). It is a measure of the time taken before an element of construction exceeds specified limits for load carrying capacity, insulation and integrity. These limits are clearly defined in the standard. The characteristics of the time-temperature relationship for the test fire from BS476 are shown in (Figure 27).

 

Although the strength of steel at 550oC is about 60% of its room temperature strength, the fire resistance of a simple steel element depends on other factors such as its temperature profile and the applied load.

 

 

Figure 28. Steel strength decreases with temperature.

 

All materials become weaker when they get hot. The strength of steel at high temperature has been defined in great detail and it is known that at a temperature of 550oC structural steel will retain 60% of its room temperature strength (see Figure 28). This is important because, before the introduction of limit state design concepts, when permissible stress was used as a basis for design, the maximum stress allowed in a member was about 60% of its room temperature strength. This led to the commonly held assumption that 550oC was the highest or "critical" temperature that a steel structure would withstand before collapse. 

Recent international research has shown however that the limiting (failure) temperature of a structural steel member is not fixed at 550oC but varies according to two factors, the temperature profile and the load.

 

6.1 Effect of Temperature Profile

 

The temperature profile is an important factor.

A joint test programme by Corus and the Fire Research Station has shown that the temperature profile through the cross-section of a steel structural member has a marked effect on its performance.

 

The effect of temperature on yield strength is characterised by results obtained from small-scale tensile tests.

The basic high temperature strength curve shown in Figure 28 has been generated by testing a series of small samples of steel in the laboratory, where the whole of each test sample is at a uniform temperature and is axially loaded.

 

The effect of a non-uniform temperature profile is to allow redistribution of stresses from hot parts of the cross-section to cooler areas.

When these conditions are repeated in full-scale member tests, e.g. unprotected axially loaded columns, then failure does indeed occur at 550oC. But if a member is not uniformly heated then, when the hotter part of the section reaches its limiting temperature, it will yield plastically and transfer load to cooler regions of the section, which will still act elastically. As the temperature rises further more load is transferred from the hot region by plastic yielding until eventually the load in the cool regions becomes so high that they too becomes plastic and the member fails.

 

The top flange of a beam supporting a slab remains cooler than the web and lower flange, and this results in an increase in critical temperature to about 620oC.

The most common situation in which temperature gradients have a significant effect on the fire resistance of structural steel is where beams support concrete slabs. The effect of the slab is both to protect the upper surface of the top flange of the beam from the fire and to act as a heat sink. This induces temperature differences of up to 200oC between the upper and lower flanges in standard fire tests. Test data shows that the limiting (lower flange) temperature of fully loaded beams carrying concrete slabs is about 620oC. This compares with 550oC for beams exposed on all four sides.

 

6.2 Effect of Load

 

Beams fail when the applied bending moment is equal to the plastic moment of resistance, reduced to account for temperature. Lower loads or stronger beams will therefore result in increased failure temperatures.

It is known from full scale fire tests that a simply supported beam carrying a concrete floor slab and 60% of its cold load bearing capacity will become plastic at about 620oC. It is also known that if it carries a lower load then plasticity will occur at a higher temperature. Thus, at low loads fire resistance is increased.

In BS5950 Part 8 (12) (See Section 7) load is expressed in terms of the 'Load Ratio' where 

Load Ratio = the load at the fire limit state/ the load capacity at 20oC

The load at the fire limit state is calculated using load factors given in BS5950 Part 8. A fully loaded beam in bending would normally have a load ratio of about 0.50 - 0.6. It is known from the research data that, with a load ratio of 0.25, for example, failure in simply supported beams carrying concrete slabs will not occur until the steel reaches 750oC, an increase of 130oC on the limiting temperature in the fully loaded case.