Reliability based design method for buckling of steel columns in fire

Abstract

Purpose This paper aims to reliability analysis of axially loaded steel columns at elevated temperatures considering the probabilistic features of fire. Design/methodology/approach The response function used in the reliability analysis is based on the non-linear FEM calculations. The stochastic variability of temperature is integrated with the procedure similar to the parameters of loading and material properties. Direct Monte Carlo simulations (MCSs) are implemented for probabilistic analysis. Computational costs are reduced by polynomial approximation of the response function of the column. Findings A design method for practical applications in the common Eurocode format is proposed. The proposed method can be used to estimate the failure probability of a steel column in fire conditions. If standard reliability criteria are applied, the results of the steel column buckling capacity in the fire according to the proposed procedure deviate from the Eurocode results in certain parameter ranges. Originality/value The proposed method for design calculations makes use of the advantages of MCS results, while the need for the tedious amount of calculations for the end user are avoided as the predefined factors are implemented in the procedure of Eurocode format. The proposed method allows better differentiation of the fire probability in the capacity assessment compared to the existing design methods.