< Back to previous page

Project

Multi-scale analysis of lean duplex stainless steel structures.

Stainless steel combines high mechanical properties with excellent corrosion resistance, making it an appealing choice for load-bearing elements in civil engineering, greatly reducing maintenance for structures in aggressive environments, such as coastal areas. The nonlinear stress strain curve, high strain hardening at relatively low strains and the differences in residual stresses for welded specimens compared to carbon steel necessitate specific design treatment for structural stainless steel. This thesis focusses on the improvement of the design rules for stainless steel beams suffering lateral torsional buckling. In addition, improvements on the shear buckling resistance and the strength of stainless steel fillet welds were investigated.

Firstly, an extensive experimental programme is presented on 13 beams suffering lateral torsional buckling made out of two lean duplex grades, EN 1.4062 and EN 1.4162, and one austenitic grade, EN 1.4404. Additionally, one carbon steel beam (S235) and one stainless steel beam (EN 1.4062) suffering shear buckling were tested, as well as 24 fillet welds, made of three stainless steel grades (EN 1.4062, EN 1.4404 and EN 1.4307) and two welding processes (GMAW and GTAW) under tension and shear. During these experiments, LVDTs, load cells, inclinometers and digital image correlation (DIC) were used to measure the displacement field and applied loads. In addition, DIC was also used for the 3D measurements of the initial geometric imperfections of the beams and for the measurement of the fracture area of the fillet welds.

Secondly, geometrically and materially nonlinear numerical models were validated against the lateral torsional buckling experiments, together with experiments collated in the literature. This was followed by a parametric study on the fundamental case of lateral torsional buckling: a beam supported by fork supports loaded by a constant moment. This parametric study covered 30 cross-section geometries, with beam slendernesses ranging between 0.3 and 1.95, and three stainless steel grades, one from each stainless steel family. The results showed that the current design rules slightly overestimate the lateral torsional buckling strengths in the lower slenderness range for ferritic and austenitic stainless steel, and that, for beams with higher slendernesses, undue conservatism is present. Based on a reliability analysis of the numerical results, improvements to the stainless steel design rules were proposed, including the derivation of individual imperfection factors for each stainless steel family. Those improvements were inspired by the recent new design proposal by Taras and Greiner for carbon steel beams suffering lateral torsional buckling. It improves the safety of the design rules in the lower slenderness range, while greatly reducing the conservatism for higher slendernesses resulting in a more efficient, more consistent and safer design.

Thirdly, for shear buckling, the current design rules were assessed using a parametric study based on geometrically and materially nonlinear numerical models, which were initially validated on the performed shear buckling experiment and experiments found in the literature. A sensitivity analysis investigated the effect of all design parameters on the predictions of the shear buckling strengths. Based on the numerical results and the observed failure modes for the shear buckling experiments, an extra term was included in the current shear buckling design equation, taking into account the stiffness of the non-rigid end post. Although more research is needed to further validate this proposal, more efficient predictions were achieved which is very promising.

Last, the strength of stainless steel fillet welds was investigated. It is a complex subject which therefore should rely on a large experimental basis due to the many variables influencing the weld behaviour. By analysing the experiments carried out in the frame of this work and comparing them to results found in the literature, large scatter between the different testing programmes was noticed. We concluded that a uniform measurement method of the fracture area is crucial to get consistent predictions. Nevertheless, improvements to the correlation factor βw could still be proposed based on a reliability analysis of all experiments, allowing again a more efficient design of welds.

Date:19 Nov 2014 →  18 Sep 2020
Keywords:lateral torsional buckling, shear buckling, strength of welds, Multi-scale analysis, lean duplex stainless steel structures
Disciplines:Metallurgical engineering
Project type:PhD project