FIRE DESIGN OF STEEL MEMBERS - Civil and Natural Resources ...

Results from experimental tests show that more complex structural arrangements often
have more fire resistance than single members. Results from series of tests from
Cardington and articles such as O’Connor and Martin, (1998) who reported the results
of tests studied by British Steel indicate that unprotected steel frames behave
significantly better than indicated in single member tests.
1.4 NZS 3404:PART 1:1997 SECTION 11:
The New Zealand Steel code, (SNZ, 1997) includes a section devoted to the fire safety
of essential steel elements of a structure. The section includes formulas to estimate the
maximum temperature that an element can reach before it will no longer be able to
carry the design fire load and therefore fail, and the time until this temperature is
reached. These formulas are based on the member ends being simply supported, so no
redistribution of loads is allowed for. This gives a conservative design of elements as
a lower temperature of the element will result in failure than if the element was ‘built
in’ to a frame with axial restraint and moment redistribution possible.
Other design tools for fire included in the section are formulas that estimate the
variation in mechanical properties of steel with temperature. These are the Yield
Stress and Modulus of Elasticity of steel, whose variation with temperature dictates the
strength of the steel at a particular temperature
The time until the limiting temperature is reached, the variation of temperature across
the cross section and the rate at which the temperature increases, depends on the
element cross section, the perimeter exposed to the fire and the protection applied to
the member. This is accounted for in the code equations by a ratio of the heated
perimeter of the member to the cross sectional area of the steel member, known as the
section factor, or H p /A (m -1 ). This value is the inverse of the average thickness of the
steel member so gives an indication to the rate of temperature rise.
In NZS 3404:1997, there are formulas provided for three and four sided exposure to
the fire for unprotected members. For protected members the temperature of the
element with time is based on a single experimental test exposed to a standard fire,
7