EFFECTIVE LENGTHS OF CONCRETE-FILLED STEEL SQUARE HOLLOW SECTIONS IN FIRE.

Abstract

Current fire design to the Eurocodes allows the effective length of concrete-filled steel hollow-section columns, in braced multistorey buildings, to be taken as 0·7 or 0·5 times the system length of the column, depending on whether the column is continuous at one or both ends. In the case where the column is continuous at both ends, numerical analyses presented in this paper show that, provided local buckling does not occur, an effective length of 0·55 can be assumed. However, observations from tests on isolated members suggest that local buckling generally occurs at the top of heated concrete-filled columns. This is caused by differential thermal expansion, where the expansion of the steel shell is greater than that of the concrete core. This differential expansion causes the steel shell to support the applied load, which, due to loss of strength as its temperature rises, will buckle locally, with the load being redirected back to the concrete core. The analyses showed that if the effects of local buckling were included, the effective length of columns in fire was increased to a value of 0·75. It is recommended that this value should be used in design. Analyses were also conducted on columns that were continuous at one end. These indicated that a value of 0·8 times the system length should be used in design, if the column is continuous at its base. If an idealized-pinned foundation is assumed, the effective length should be taken as the system length.