Jämförelse av FE-modeller för lastspridning i tvärled : Parameterstudie för dimensionering av betongfarbana på parallella stållådbalkar

Sammanfattning: A great deal of the design of bridges today is carried out using finite element models and analyses. These models are usually generalized versions of the actual bridge, as these kinds of models often is able to produce results similar to reality. In this thesis, a study regarding how the considerations of some of the details of a bridge affect the results using a FE analysis is performed. The analyses are focused on the change of transverse load distribution in the bridge slab. The thesis also addresses the use of a parametric design approach in FE modeling and analyses. The models are created using Python code with the possibility of choosing which properties that will be active or not before each analysis. The study is focused on the modeling of composite bridges with parallel box-girders. The bridge Kollektivtrafikbron, which is a new bridge currently in design as part of the Hisingsbridge project in Gothenburg, is used as a case study. The study has shown that a great deal of simplifications is reasonable to carry out in the modeling of a composite bridge, as long as the designer is aware of how these simplifications affect the results obtained. For the type of bridge investigated, the study has shown that the presence of transverse stiffening beams in the model as well as how the material properties of the edge beams are specified have a large impact on the transverse bending moment. The study has also shown that the use of shell or beam elements in the modeling of the main beams have a large impact on the shear force in the slab. Furthermore, the choice of element types and how constraints are specified have been shown to entail greater differences in the results than the detailing level of the model. These aspects should therefore be thoroughly considered in all modeling work. Additional aspects treated in the study is the impact of radius, the recess in the slab for the track as well as the size of the surface specified when applying traffic loads. The parametric design approach implemented in this study have provided the possibility of carrying out more analyses and investigating more details than what would have been possible if each model had been created manually. The greatest benefit of using a parametric design approach have been the possibility of carrying out analyzes automatically without monitoring or adjusting, which has made it possible to utilize the time provided for the study in a highly effective way. The design approach has been shown to be most effective when using simplified FE models, as these can be generated automatically fairly quickly. Therefore, it could be very beneficial to implement this kind of design approach in the early stages of design.