Stiffness reduction approach for structural steel design

Abstract

Stiffness reduction offers a practical means of considering the detrimental influence of imperfections and the spread of plasticity on the strength and stability of steel structures. In this thesis, a stiffness reduction approach for structural steel design is presented. The proposed method is carried out by reducing the stiffness of steel members through developed stiffness reduction functions and performing Linear Buckling Analysis (LBA-SR) and Geometrically Nonlinear Analysis (GNA-SR). Since the deleterious influence of geometrical imperfections, residual stresses and the spread of plasticity is fully taken into account through the developed stiffness reduction functions, the proposed design approach does not require the use of member design equations, but only requires cross-section checks, thus leading to practical design. Finite element models of steel members and frames are created and validated using experimental results from the literature. Geometrically and Materially Nonlinear Analyses (GMNIA) of the validated finite element models are used to verify the proposed stiffness reduction method in all considered cases. The proposed stiffness reduction method is initially developed for the flexural buckling assessment of columns and the in-plane design of beam-columns, where a stiffness reduction function is derived using the European column buckling curves, providing the same strength predictions as determined through these curves. The proposed method is then extended to the lateral-torsional buckling assessment of steel beams and the flexural-torsional buckling assessment of steel beam-columns. Having established its validity for individual members, the proposed method is applied to steel frames. Both non-redundant and redundant benchmark frames from the literature are considered. It is observed that the proposed stiffness reduction method provides a reliable and accurate design approach.