Influence of Ductility in the Design of (High Strength) Steel Bridges

In connections of steel structures stress concentrations occur due to the rapid geometric changes in the cross section. If the material has sufficient ductility then, due to plastic deformations, forces are redistributed so there will be equilibrium in the internal forces. In design codes this is accounted for by imposing certain requirements with respect to the tensile/yield strength ratio and minimal strain at fracture. This problem is of particular importance when high strength steels are used as questions are raised related to their ductility capacity. Over the past decades, high strength steel has gained significant ground in the steel structures market. A steel grade such as S355, which was considered to be a high strength steel 20 years ago, is now one of the predominant grades used for steel construction. In practice the occurrence of stress concentrations raises questions on how to deal with them and how to ensure a certain amount of ductility to avoid brittle fracture. In most cases an inelastic finite element analysis in shell elements and based on the nonlinear behaviour of the material is required in order to calculate the strains at the notch-tips and prove redistribution of stresses takes place. However this is a time consuming and costly procedure. This paper addresses such issues and, based on Neuber’s formula for nonlinear material behaviour, a new method is developed to calculate the strains at the locations of the stress concentrations. This new approach, entitled Stefanescu Method (SM), is based only on the results of the linear finite element analysis to estimate the value of the strains. The procedure is first developed in theory and applied to a simple case of a plate with a hole in tension because in this case the stress concentration factors are already known. The results of the SM are checked with those of an inelastic finite element analysis. Based on the simple case of a plate with a hole in tension, the SM approach gives good estimates of the strains up to a limit load value of approximately 90% of the smallest force that would cause yielding in any nominal section of the plate. The applicability of this method is extended to a bridge connection from a real life project (the A1/A6 Diemen-Almere Havendreef steel railway arch bridge developed by Iv-Infra). The method is first studied on a simplified geometry of the gusset plate from the bridge connection. As the method again yields good results it is extended to the more complex geometry of the bridge connection. Comparing the results of the SM with those of an IFEA, the method gives good estimates of the strains. Based on the models used it can be concluded that the SM method can be applied to estimate strains at the location of the stress concentrations under the conditions that the loads are known and these are below 90% of the smallest force that would cause yielding in any section of the elements subject to the stress concentrations. Also, this paper addresses the issue of the ultimate to yield strength ratio and to which extent this influences the ductility capacity of the material with respect to stress concentrations. For this purpose three different ratios were studied and their influence on the strains at the notch-tips is compared. With respect to stress concentrations and the local demand for ductility to redistribute stresses, the presence of strain hardening significantly increases the ductility capacity. The lack of strain hardening leads to low deformation capacities of the plates studied. The fu/fy parameter together with a minimum elongation at fracture can lead to very high ductility capacities of the elements.