Experimental and Numerical Study on Early-Warning Approach for Fire-Induced Collapse of Steel Portal Frame Based on Rotational Angles

The existing early-warning methods primarily rely on detecting structural displacements which are often challenging to measure accurately in real fire scenarios. To develop innovative early-warning strategies, this paper experimentally and numerically investigates the fire-induced collapse of an 8 m x 6 m steel portal frame assembly. Detailed thermo-structural responses of the frame were measured and presented, including the displacements and rotations. The results revealed that the vertical mid-span displacement and horizontal displacement at the rafter end are key to developing an effective early-warning system. Structural rotations seem sensitive to structural deformation and emerges as a valuable safety indicator for structural systems. Furthermore, parametric analyses were carried out in order to investigate the effect of load ratio, fire protection and heating curve on key parameters of the structure subjected to fires. It is discovered that the increased load ratio can reduce the peak value of vertical displacement at the mid-span of the rafter. A rotational angle of 6 degrees in the steel beams is optimal for predicting the collapse of steel portal frames in fire conditions. Based on the parametric studies, an innovative early-warning approach using rotational angles is proposed and validated against the test frame, demonstrating significant applicability and reliability. The rotation-based early-warning approach works in two distinct stages, being activated respectively by the maximum and zero rotational angles at the end of rafter. The early-time ratios for the respective warning stages are 0.65 and 0.88. For better precision and practical reliability, it is further recommended to combine the rotation-based and displacement-based approaches for the on-site early-warning of fire-induced collapse of portal frames.