Approach for early-warning collapse of double-span steel portal frames induced by fire

The unexpected collapse of burning buildings has posed a great threat to firefighters. Hence, early-warning methods for fire-induced collapse are urgently needed to avoid secondary casualties. This paper proposes an early-warning approach for predicting the collapse of double-span steel portal frames based on real-time measurement of displacements and displacement velocities of the burning frame. Firstly, numerical models are established to simulate the collapse behavior of double-span steel portal frames under fire, and six collapse modes of the frames are summarized through parametric analysis. The displacements and displacement velocities of the apex, eaves, and mid-span of rafters, defined as the key monitoring physical parameters (KMPPs), are found to have a close relationship with the collapse mode and time of the burning frames. Secondly, by exploring the rules of the KMPP-time curves, the characterized points that can be used for early warning of the collapse of the frame are extracted. Then, the early-warning approach applicable to six collapse modes is proposed based on the emergence of various characterized points. For universalizing the collapse prediction, early-warning time ratios are introduced and determined according to the reliability theory. Finally, the practicability and accuracy of the proposed approach are validated by an existing fire test.