Characterization of the Full Non-Linear Behaviour up to Failure of the Sheared Panel Zone under Monotonic Loading Conditions

Nowadays, modern building codes, such as the Eurocodes, require from the civil engineers to ensure an appropriate robustness to any structure. In steel and composite structures, this request for robustness mainly focuses on the joints between the structural members, which are seen as “weak” elements in the structure. To meet this request, it is recommended to provide sufficient ductility to the joints that would allow them to deform significantly without breaking in the case of an unforeseen exceptional event. However, the analytical method currently available in the part 1-8 of Eurocode 3 (EN 1993-1-8), i.e. the component method, does not allow to predict the rotation capacity of the joint under large deformations.
In this context, a large research project was launched at the University of Liège with the aim of extending the component method towards the large deformation field, and under complex loading conditions (impact, fire, explosion, earthquake…). The present thesis is the first outcome of this project and focuses on the behaviour of the panel zone (PZ), which is known to provide a significant reserve of ductility to the joint, when activated and appropriately designed. The main objective of the thesis thereby consists in providing a new sophisticated analytical model, which can predict the full non-linear behaviour of this component up to failure, under monotonic loading conditions.
To achieve this objective, an extensive literature review of existing scientific models was first conducted and set the theoretical background of the thesis. The performances of these models were assessed through comparisons with many experimental results carefully selected from the scientific literature. These comparisons revealed that none of them was able to accurately capture the complete non-linear behaviour of the PZ up to failure. Even for the prediction of the plastic resistance, all the investigated models also failed in providing consistent results. These observations validated the need to develop a more sophisticated constitutive model for the PZ.
The prediction of the plastic shear resistance was considered first, in the context of simple welded joints. The finite element (FE) approach was used to gain insight into the complex phenomena governing the PZ behaviour. This approach included the development and validation of a FE model and the use of this model to perform large parametric studies on various joint configurations. Based on the careful analysis of the FE results, the key geometric and mechanical parameters governing the resistance of the PZ could be identified and introduced in a new complex analytical model. After a validation step, this model proved to work well and to outperform existing analytical models.
Based on the knowledge acquired in the plastic field, the model could be further extended to the large deformation field. Again, the FE approach was used to identify the key parameters governing the deformability and the failure of the PZ. The resulting full-range model, encompassing those key parameters, was extensively validated against numerical and experimental results, where it proved to predict the PZ ultimate resistance and ultimate deformation capacity with reasonable accuracy.
The case of bolted joints was eventually tackled, considering the prediction of the plastic resistance first. Some adjustments in the new analytical expression as well as in the assembly procedure of the component method were suggested and validated through comparisons with experimental results. Secondly, the prediction of the deformation capacity was addressed, through preliminary comparisons with experimental results. From these comparisons, some limitations in the approach were highlighted and perspectives of improvement were discussed.
An additional outcome of the work consists of a set of new simplified design criteria for the prediction of the PZ initial stiffness, plastic resistance and deformation capacity. The first two expressions outperform the current EN 1993-1-8 criteria, proved to be unsafe in many cases, while the third expression fills a gap in the EN 1993-1-8 norm where no criterion is currently available for the prediction of the PZ deformation capacity. This new set was proposed for integration in the forthcoming prEN 1993-1-8 pre-normative document.