Progressive Collapse of Steel Frames in Fire Using Hybrid Models with Explicit Formation of Connections

Failure of beam-column connections is one of the main reasons for progressive collapse of steel structures under fire. The computation cost in a high-fidelity model is becoming an issue for accurately considering the effect of connections on structural collapse. A hybrid model of steel frame that integrates solid elements for simulating top-and-seat-angle with double web-angle (TSDW) connections and beam elements for other components is presented in this study, and its performance efficiency is confirmed. The collapse behavior of steel frames simulated by the hybrid model is investigated by addressing the effect of load ratios, connection forms, number of TSDW connections and fire scenarios. It is found that the proposed hybrid modeling method can accurately and efficiently predict the collapse mode and collapse temperature of structures in fire. The collapse temperature of structures decreases in a range of 11% to 45% with the increase of load ratio by an interval of 0.3. The form of connections has a great impact on the collapse behavior of steel frames. The collapse modes of steel frames significantly depend on the fire-exposed area, and it is necessary to define a set of real fire scenarios for accurately predicting realistic collapse behavior of steel frames in fire.