Railway transition zones represent a significant challenge due to abrupt variations in substructure properties, leading to differences in vertical strain. These differences cause the wheel-rail dynamic interaction to get excited, which can cause passenger discomfort, track deterioration, and derailment. Consequently, proactive design strategies are crucial to mitigate these stresses. In previous research, an optimal geometry of railway sleepers was designed, however it was assumed that every sleeper would have the same geometry. This renders the design to be susceptible to inherent uncertainties. To address this limitation, a new design should be made incorporating uncertainties. However, railways involve numerous parameters
influencing performance. Employing all parameters within an optimization framework is computationally not feasible. Therefore, this thesis aims to identify the most influential design variables within a railway transition zone. This leads in the following research question:

How can the most influential design variables on the performance of a railway transition zone be determined using a static model while taking uncertainties into account?

In this research, the design variables have been limited to only railway sleepers. To answer this question, a comprehensive review on how uncertainty is implemented into designs was done. Leveraging this knowledge, Optimized Latin Hypercube Sampling (OLHS) was employed to generate samples defining the length and width of each sleeper within the transition zone. These samples were then used as input variables within a static Ansys model, modified to mimic dynamic factors. This model subsequently calculates the stress and settlement in the ballast and soil layer for several locations across the transition zone. This process is repeated for ten different samples created with OLHS. The location exhibiting the highest settlement and stress for each sample was identified, resulting in a dataset for comparison between the
samples with the most favorable and least favorable performance.

One pattern that can be seen in the results is that higher stresses occur when there is a higher
disparity between adjacent railway sleepers. This may be the cause of differences in bearing surface area, where longer sleepers spread the load from the rails over a larger area than shorter sleepers. This causes uneven settlement, and consequently higher stress concentrations. If this hypothesis holds true, incorporating gradual transitions in sleeper length throughout the transition zone becomes a critical design consideration for future optimization efforts. However, further research is warranted to definitively validate this observation.