Design of Stainless Steel Homogeneous and Hybrid I-Sections Exhibiting Strong Flange Behavior

Abstract Stainless steel built-up I-profile are frequently designed with webs significantly more slender than their flanges and greater material strength in the flanges in case of hybrid sections. Given the fundamental load case of pure compression or bending, the compressed flange can develop significant postlocal-buckling strength, approaching its fully plastic capacity, whereas the web reaches only its local buckling resistance, a long-overlooked failure mechanism not currently captured by current design approaches. The effect is particularly pronounced in stainless steel, whose nonlinear stress–strain response and pronounced strain hardening further enhance this reserve capacity. In this paper, experiments are first used to validate a finite element model of such behavior. Following validation, a parametric study is conducted to evaluate the influence of geometric and material parameters on the ultimate strengths. The study includes both homogeneous and hybrid built-up sections, where the homogeneous sections are made of four grades (EN 1.4062, EN 1.4462, EN 1.4307, and EN 1.4404), and the hybrid sections are made using four hybrid combinations covering a wide range of cross-sectional slenderness. A total of 1,325 homogeneous and 955 hybrid FE models are developed, focusing solely on strong flange behavior. The performance of the codified Effective Width Method (EWM), Direct Strength Method (DSM), and Continuous Strength Method (CSM) is evaluated against the numerical results. The comparison highlights that element interaction and local buckling strength reductions cause I-sections with slender webs and stocky flanges to exceed the codified strength predictions. A modified Direct Strength Method (mDSM) approach is then proposed. By introducing two new parameters, Ω and α , to distinguish between cross-sectional behaviors and better account for strong flange effects, we address the limitations of traditional “whole section” methods while retaining their inherent simplicity, providing greater accuracy and less scatter than any of the codified methods. A reliability assessment provides evidence that the new equation achieves the codified target for acceptable probability of failure when combined with the current resistance factor.