Progressive collapse analysis of 3D steel frames with concrete slabs exposed to localized fire

This paper numerically investigates progressive collapse resistance of three-dimensional steel frames with reinforced concrete slabs exposed to localized fire using LS-DYNA. An explicit dynamic analysis is carried out and the real fire time is scaled down to save the computational cost. The prototype of the model is based on the eight-storey building in Cardington tests. The scenario of heating individual columns on the ground floor is first studied followed by a sever case of simultaneously heating four columns in one compartment. The collapse modes and load redistribution scheme of the frame subjected to different load ratios and fire locations are investigated. The modelling parameters such as mesh size, initial imperfection, timescale are first studied by validating against the fire test data. It is found that the mesh size has little influence on the responses of members at elevated temperature. The initial imperfection has significant effect on the load-bearing capacity of columns. The quasi-static behaviour of structures under a fire duration of hours can be simulated in a dynamic analysis by being scaled to seconds without causing oscillation of responses. The numerical results of the 3D frame show that the frame does not collapse in the case of single column heated for a fire design load (load ratio of 0.25 for columns). By increasing the load ratio to 0.5 as for the ambient design, progressive collapse occurs. For a fire of four columns heated, the frame collapses for the corner bay fire and long edge bay fire but withstands for the internal and short edge bay fires. The collapse modes are dominated by the uneven load redistribution in the two horizontal directions and the fire location, which cannot be simulated by a 2D model. The loads previously sustained by the buckled heated column are transferred more along the short span than the long span. The critical temperature of a column in a frame is significantly lower than that given in EC3, due to the fact that the translational and rotational restraint increase its load ratio and reduce its effective length, respectively. The critical temperature for the global collapse of the frame is about 50-100 °C higher than that of individual heated columns.