In recent years, the increased need for renewable energy has put more interest in wind energy and therefore in making wind turbines more efficient and building them as efficiently and cheap as possible. Delft Offshore Turbines (DOT) is developing a drivetrain that replaces the heavy and expensive parts in the nacelle of a wind turbine and makes these parts more accessible and aims to improve the reliability of the wind turbines by replacing the parts that are most prone to failure.  Due to the innovative design of their drivetrain, several unknowns still have to be answered since this has never been tried before. One of the most important parts of this drivetrain requires rolling contacts that endure high contact stresses and have to operate for a long time in harsh conditions. These rolling contacts have to be tested and for this, DOT created the Roller Test Bench (RTB) to test different materials and material combinations until failure and to see what type of failure is occurring and what the lifetime of these rollers is. The rollers are a lot smaller compared to the rollers found in the drivetrain to reduce cost and testing time. The scaling of the dimensions has been done several times in different fields of research but it is unclear how this affects the results and how the results translate to the original rollers. In literature some papers have stated that the results scale in a certain way but this has not been proven or the statements have not been argumented in a clear way such that these statements hold no ground.  A rolling contact has several ways that it can fail; wear, deformation, corrosion or fatigue are most commonly found. From the literature study performed in this thesis it has been noted that the main mode of failure in the RTB is rolling contact fatigue since the rollers are smooth, have no lubrication applied, are free of external contaminants and have no significant amount of friction on the surface. With the stress applied and the material used it is also noted that the fatigue will occur after a high amount of cycles and in literature this is also called high-cycle rolling contact fatigue. Experimental testing of such problems takes a very long time, even if the rollers are smaller and are run faster through the load cycles.  With both problems stated, the lack of information on the influence of scaling on fatigue life and the testing of high cycle fatigue, the goal of this thesis was found and it was to find a method to predict the lifetime and wear mechanisms of rolling contacts in the DOT Direct Drive Pump. This method could then be used to figure out how the scaling affects the fatigue life for rolling contacts and make a possible statement of its influence.  From the literature study a numerical modelling method was found that incorporated the microstructure and included both crack initiation and propagation and was able to model high cycle fatigue. This model suited the problem the best and was proven to provide results that agree well with previously performed experiments and their experimental results. This method implemented a Voronoi tessellation to reproduce the microstructure and applied continuum damage mechanics to calculate the damage evolution and apply the damage to the microstructure. A jump-in-cycle method was applied to improve the efficiency of the model and greatly reduce the time to model until failure. The created model is able to reproduce subsurface initiated fatigue and with that produce a scatter and L10 life comparable with previous models and experimental results. The model was validated by comparing the stress distribution, crack formation and the initiation and total lives with other models and data. The model was off the required range since the models uses an homogeneous isotropic material which causes the scatter to be lower.  With the model created and validated, it was applied to several different half-widths to test how the scaling affects the fatigue life. From the results it was noted that for a larger roller size the fatigue life decreases but it has to be noted that further tests, experimental and numerical, have to be done and improvements to the model have to be made such that at some point a correlation can be made between the scaling and the fatigue life. Further recommendations were provided on improving the model and with that creating more reliable results such that a better prediction can be made on how the fatigue life is affected but also to predict the life of a rolling contact.