Fibre reinforced polymer is a new material that is more and more used in civil engineering works. This lightweight and high-strength material comes with a lot of advantages. The application of this lightweight material in bridge structures can result in a reduction of the out-of-service time of railways. On the other hand, the low stiffness of this material offers a big challenge. The stiffness requirements of railway bridges are strict to ensure safe and comfortable passing of trains. This thesis looks at the application of a FRP deck in railway bridges. The literature study describes (1) the material FRP, (2) state-of-the-art in (FRP) bridges and (3) a dynamic model for (railway) bridges.
The main properties of FRP materials are listed including common manufacturing processes. According to a state-of-the-art study two designs are presented: (1) A sandwich deck, commonly used in FRP bridges, and (2) a through deck, commonly used in railway bridges. Furthermore, the requirements and loadings for railway bridges are listed. The dynamic model for railway bridges is based on a mass-spring rigid body system with Rayleigh damping. This model is used to verify the finite element models. A material study is performed to determine the most suitable material properties for FRP railway bridges. Glass fibres and polyester resin are choses based on individual material properties, market prices and common use in civil engineering structures. Foam has no structural integrity but is used as permanent form work. The lay-up of fibres is based on the literature study and differ for flanges and webs. The lay-up for flanges is:
[12.5% - 90° | 12.5% - ±45° | 62.5% - 0° | 12.5% - ±45°]s
The lay-up for webs is:

[25% - 90° | 20% - ±45° | 35% - 0° | 20% - ±45°]s

The volume fraction of both laminates is 50%. These lay ups in combination with the conversion factors and safety factors result in a FRP material suitable for FRP railway bridges. The design study describes the sandwich and through design with the materials from the material study. Two finite element model are made using Sofistik. These finite element models are verified using analytical models. A static and dynamic analysis is performed on both designs. Based on the results from the static analysis the maximum deflection and maximum stresses of the Sandwich Design do not exceed the limits, and the maximum deflection and maximum stresses of the Through Design do exceed the limits. Based on the dynamic analysis both designs do exceed the limits. Furthermore, in consultation with a FRP manufacturer, the laminate thickness must be reduced to ensure both designs are manufacturable. Based on the results of the design study and the requirements presented in the literature study a new design is created. This design includes the following improvements: (1) reduction of the span by transferring the forces in transverse direction, (2) extra height of the design, (3) more webs, (4) extra supports, (5) reduction of the laminate thickness to 30 mm and (6) the use of pre-camber. This final design is transformed into a finite element model. A static and dynamic analysis is performed. Based on the results of the static analysis the maximum deflection and maximum stresses do not exceed the limits. The dynamic analysis results in maximum acceleration of 19.48 m/s2, which exceeds the limit and results in a unity check of 5.57. The final design doesn’t meet the requirements based on the results of the dynamic analysis. An important remark on the dynamic results is made in the discussion. The dynamic load model, consisting of point loads, is applied directly on the bridge deck, omitting the dynamic behaviour of the ballast bed. Taking into account the ballast bed in the dynamic analysis might result in a significant reduction of the maximum acceleration of the bridge deck due to the spreading and damping effect of the ballast bed. In conclusion, according to the results of the static analysis of the final design it is possible to apply a FRP deck in railway bridges. Improvement of the dynamic numerical model is needed before drawing conclusions on the dynamic behaviour of the final design.