Over the past few decades, the Netherlands has built numerous steel railway bridges to improve its infrastructure. The repeated loading from railway traffic makes steel bridges susceptible to fatigue damage. The stress ranges at fatigue-prone locations are generally higher in railway bridges than in road traffic bridges. Due to the limited space available due to road and rail alignment, the connection between the main girder and cross-girder is often identified as a critical fatigue location.

The main- to cross-girder connection can be designed for fatigue using different design principles. Designing the connection with a certain rotational stiffness leads to higher fatigue stresses at this location. Meeting the fatigue requirements at the connection can often be achieved by local adjustments, such as welding extra steel plates to the fatigue-induced location to evenly distribute stresses from the cross-girder to the main girder. This can be a costly solution. Designing a connection that is flexible could reduce stresses at the main- to cross-girder connection. Solving fatigue issues can be done by making global adaptations to the structure. The question arises: which aspects can be adjusted such that slight changes can notably improve fatigue resistance in a cost-effective manner?

A literature study and a finite element investigation of the main- to cross-girder connection are performed. A reference model is made after which parameters are altered to investigate their impact on the fatigue response of the connection. The reference model is based on the design of the bridges from project Oostertoegang. However, instead of considering a connection with a certain rotational stiffness, the design is made more flexible for this study. Ansys 2022 R2 is used to create a finite element model of the bridge with shell elements. The hot spot stress method is used to conduct the fatigue assessment. In total four critical fatigue locations are researched for six parameters. The parameters researched are: the center-to-center distance between the cross-girders, the height of the cross-girder and main girder, the thickness of the inner web plate of the main girder, the diaphragm, and the steel deck plate.

From the analysis, it can be concluded that for one detail (M3) local measures should be applied to meet fatigue criteria. The three other details can satisfy requirements within feasible limits. The most cost-effective and realistic way is to increase the thickness of the inner web of the main girder. Other cost-effective but less feasible solutions to optimize the flexible design for fatigue are: decreasing the thickness of the diaphragm and increasing the height of the cross-girder. These parameters show the best ratio between the costs needed to alter the parameter and the total fatigue damage change of the critical detail.

Furthermore, it is determined that the fatigue assessment using finite element analysis with shell elements can be optimized by using the hot spot stress method in combination with modeling the weld using an increased thickness.