Non-Linear Assessment of a Concrete Bridge Slab Loaded to Failure

Sammanfattning: This thesis covers an investigation regarding the failure in the bridge slab of Gruvvägsbron, which was the result of the full scale test that the bridge was subjected to prior to demolition. Using the non-linear finite element software ATENA 3D, a model of the bridge was assembled, with the purpose to attempt to reenact the test procedure and realistically capture the failure load and behaviour. This in order to be able to conclude what kind of failure that occurred. The initial part of this thesis presents a summary of a conducted literature study, which aims to give deeper knowledge regarding the linear shear and punching shear phenomena and their respective failure mechanisms, and how they are applied on bridge slabs. Furthermore, the shear capacity of the bridge was calculated according to current design codes. A parameter study was conducted on the model, which initially showed a over-stiff response. The aim of this was to study the influence of key variables on the outcome of the analysis, and hopefully get closer to the failure load acquired inthe experiment. From the studied parameters, it was observed that a combined reduction of the tensile strength and fracture energy, together with a low fixed crack coefficient had the largest influence on the outcome of the analysis. It was also observed that the location of the failure and failing load was dependant on how the loading was applied to the model, i.e. via load control or deformation control. The final model failed at a load which surpassed the actual failure load by 10.5%. The mode of failure obtained in all the analyses were the result of a large shear crack propagating from the edges of the loading plate, through the slab to the slab/girder-intersection. This indicates that the type of failure that occurred was primarily due to a linear shear mechanism with a secondary punching effect. The design values calculated by keeping with the current codes resulted in too conservative values when compared to the obtained failure load from the experiment. This proves the difficulty in regarding the internal force distribution in slab struc-tures as well as the shear carrying width, which from the analysis were found to be larger than that obtained from the code.