Early-warning approach for fire-induced collapse of steel portal frame based on overall deformation

Existing early-warning approaches for fire-induced collapse are complicated and challenging to implement on site due to the difficulty in measuring key parameters. To address this issue, this paper presents the development of an applicable early-warning system for the fire-induced collapse of steel portal frames, based on numerical analysis. An analytical method for assessing the overall deformation of steel portal frames in fire scenarios forms the basis of this system. The results reveal that the explicit dynamic algorithm can effectively predict deformation and collapse modes under fire conditions. Factors such as vertical load, wind load, restraint, fire scenarios, and fire protection significantly influence the collapse mode, failure temperature, and collapse time. Characteristic deformation curves and feature points for different collapse modes are identified, leading to the proposal of a four-stage early-warning framework. This framework utilizes vertical and horizontal displacements to provide timely evacuation warnings at various stages. Earlywarning ratios are derived from a series of numerical analyses. Additionally, an analytical geometry method is introduced to estimate the mid-span rafter displacement from column top displacements, demonstrating acceptable accuracy for practical implementation. This method facilitates the real-world application of this early-warning system, providing valuable guidance for firefighters on site.