Single-Story Steel Frame Structure in 3D: Sensitivity Study Utilizing FEM Simulations

This paper presents a sensitivity study of a single-story steel frame structure. A basic sensitivity study of linear (Pearson) correlations was conducted to examine the relationships between input parameters and the frame's ultimate resistance. This study provides a basic insight into a problematic of how variations in key input parameters influence the ultimate resistance of the structure across different slenderness levels (various frame heights, keeping the same cross-sections). The input parameters considered in the analysis are treated stochastically and include two material properties (elastic modulus and yield stress) along with four parameters defining the initial geometrical imperfections of the frame: global sway and local bow imperfections in two orthogonal directions. These imperfections affect the structural behavior and stability of steel frames, making their probabilistic assessment crucial. To determine the ultimate resistance based on stochastic data, a combination of the First Order Reliability Method (FORM), the finite element method (FEM), and geometrically and materially nonlinear imperfect analyses (GMNIA) is employed. The FORM-based resistance values are then compared with two deterministic estimates of ultimate resistance obtained following the European standard for steel structure design (EC3). The first deterministic estimate directly utilizes FEM, while the second is based on EC3 buckling provisions, where the critical length is determined using prior eigenvalue buckling analysis of the frame structure. The findings of this study provide valuable insights into the sensitivity of the selected steel frame and demonstrate how probabilistic methods can complement conventional deterministic design approaches. By incorporating both stochastic and deterministic analyses, the research highlights the influence of uncertain parameters on the structural response and enhances the understanding of steel frame behavior under varying conditions. The comparison between probabilistic and deterministic resistance estimates also reinforces the accuracy and reliability of stochastic approaches, which may offer a more refined alternative to traditional design methods. The basic sensitivity study of linear correlations identifies key relationships between input variables and structural resistance, helping to better interpret the influence of different parameters on the overall stability of the frame. These results contribute to improving structural safety assessments and refining design methodologies for steel frames, particularly in cases where uncertainties play a significant role in performance evaluation.