Transverse load-spread of a voided bridge deck

Abstract

Orthotropic bridges can be modelled using a grillage model and an orthotropic plate model. The grillage model has some benefits (i.e.: shorter calculation time and easier to apply pre-stressing) compared to the orthotropic plate model. However, during the Blankenburgverbinding project Ballast Nedam experienced that the transverse load-spread of the grillage model can be lower compared to the orthotropic plate model. This potentially results into a less economical reinforcement design. The aim of this study is to investigate the differences in transverse load-spread of the grillage and orthotropic plate model and to make a judgement on the practical usability of the grillage model based upon the differences in reinforcement design. To test the hypothesis that the grillage model has less transverse load-spread compared to the orthotropic plate model and to quantify the impact on the reinforcement design, a case study on the bridge decks of the Blankenburgverbinding is carried out. The results of a grillage model are compared to an orthotropic plate model. In order to be able to judge whether to orthotropic plate model describes the correct behaviour of the bridge, this model is verified using a 3D plate model. In total, 3 different single-span, simply supported bridge deck layouts were modelled (straight, curved and curved with a skew angle). The results of the 3D plate model are almost equal to the results of the orthotropic plate model. Both these models show the same transverse load-spread. The grillage model shows less transverse load-spread compared to the orthotropic plate model and the 3D plate model. This difference in load-spread increases with increasing eccentricity of the applied load. The torsional and transverse shear stiffness has the biggest impact on the transverse-load spread. When these stiffnesses are higher, the transverse load-spread increases. Lowering the stiffnesses results into less load-spread. In case of the curved bridge deck, the grillage model results in about 7.3% more reinforcement compared to the orthotropic plate model. For the curved and skewed deck the grillage model results into up to 4.7% more reinforcement compared to the orthotropic plate model. The differences in design bending moments (Wood-Armer moments) are relatively small. Although the theoretical difference in amount of reinforcement is significant, the practical difference in reinforcement design between both models can be small. For example, the structural engineer is limited by the available space, detailing requirements and the available bar diameters. These limitations and rounding differences make that for some bridges it is possible to reduce for example 5% reinforcement and for some bridges such a reduction can be very hard to realise. In that case, the theoretical difference between the amount of reinforcement becomes very small or completely vanishes in practice. Eventually, it depends on the situation whether the benefits of the grillage model outweigh the theoretically less economical reinforcement design. When there are a lot of repetitive bridge decks it can be beneficial to use an orthotropic plate model. In that case, the orthotropic plate model has to be created only once and a small reduction in the amount of reinforcement of one deck can add-up to a significant reduction for the whole project. When there is only one bridge deck to model, or all spans are unique, the benefits of the grillage model most likely outweigh the practical difference in the amount of reinforcement.