Performance-based fire protection of clustered individual members within a steel frame structure

Abstract

Fire resistance rating (FRR) in the existing codes and literature sets an overall criterion for buildings by different fire resistance levels with undifferentiated fire protection. However, the performance diversity of individual members within a building is not considered. Therefore, a performance-based fire protection method for clustered individual members within a steel frame structure is established. First, parametric models of a steel frame structure exposed to two typical fires are built for data acquisition. Three features considering the mechanical properties of steel members and their impact on the entire structure are selected. Then, K-means clustering is adopted for the classification and ranking of individual members to assign fire resistance ratings and conduct clustered fire protection design. A comparison between the undifferentiated fire protection strategy and the clustered fire protection strategy is conducted. The quantitative evaluation results indicate the feasibility of the proposed method for clustered fire protection design, with the consideration of the performance diversity of individual members within the steel structure.