Related resources
Full-text held externally
Search for item elsewhere
University researcher(s)
REALISTIC MODELLING OF THERMAL AND STRUCTURAL BEHAVIOUR OF UNPROTECTED CONCRETE FILLED TUBULAR COLUMNS IN FIRE
Journal of Constructional Steel Research. 2008;(64):1086-1102.
Access to files
Full-text and supplementary files are not available from Manchester eScholar. Full-text is available externally using the following links:
Full-text held externally
Abstract
This paper employs the commercial finite element analysis package ANSYS to model the thermal and structural behaviour of isolated CFT columns in fire. Although CFT columns have been numerically analysed by many researchers, this paper presents details of a number of features which have often been neglected by many researchers, including the influence of an air gap and slip at the steel/concrete interface on CFT column temperatures and structural behaviour, the sensitivity of CFT fire resistance to concrete tensile behaviour and CFT column initial imperfections. The finite element model is validated by comparing the simulation results against experimental results of standard fire resistance tests on 34 CFT columns with different structural boundary and loading conditions. A numerical parametric study is then performed to investigate the sensitivity of simulation results to different assumptions introduced in the finite element model. The results of these numerical studies show that whether or not including slip between the steel tube and concrete core in the numerical model has minor influence on the calculated column fire resistance time. The fire resistance of CFT columns with an air gap is generally slightly higher than that without an air gap. However, including slip gives better prediction of column deflection behaviour. Using different tensile strength or tangent stiffness of concrete has minor effect on the calculated column fire resistance. Different amounts of column initial deflection have some influence on column fire resistance times. Nevertheless, the influence is relatively small so that it is acceptable to use a maximum initial deflection of L/1000 as commonly assumed by other researchers.