5. Conclusions
In this paper, a large experimental campaign on CFS columns at both ambient temperature and simulated fire conditions with restrained thermal elongation is reported. The developed finite element model is also described and the final results regarding the calibration process are presented. In the buckling tests at ambient temperature, the advantage of using built-up members was clear since the increase in the buckling load was significant. In terms of the design predictions according to EN 1993-1-3 (2004), it was observed that increasing the number of profiles the design predictions become unsafe.
In the fire resistance tests, it was found that the interaction between the initial applied load and the imposed level of restraint to thermal elongation significantly influence the behavior of isolated CFS columns under fire conditions. When some level of restraint exists, additional forces are generated, which may lead to premature collapse and consequently to lower critical temperatures. It seems that increasing the level of thermal restraint the failure of the columns may be controlled by the generated axial restraining forces, whereas for lower levels of thermal restraint, the failure is controlled by temperature increase and consequent degradation of the mechanical properties of the S280GD þ Z steel. It seems that the higher reductions on critical temperature, because of restraint, occur for lower values of ak.