Due to the thermophysical properties of cast steel material the build-up of imperfections, such as shrinkage pores or cavities, is unavoidable. These defects may significantly influence the local fatigue strength and have to be considered at early stages of the design process. The aim of this master thesis is to optimize cast steel specimens during construction phase to obtain a defined porosity distribution within these parts. From these cast samples, specimens for fatigue testing will be manufactured to investigate the influence of pores on the local fatigue strength experimentally. The first part of the thesis deals with the evaluation of varying simulation parameter settings and their influence on the general porosity. These preliminary studies act as foundation for more detailed specimen based examinations. In the main study, a variation study of casting process dependent parameters is used to determine the effect of individual geometries on porosity. Furthermore, the gating and feeding systems are evaluated and modelled according to the industrial manufacturing requirements. The specimen geometry is specified in such a way that gating system and global dimensions are not separately modified, while the variation of single local parameters affects shape and size of the inhomogeneity in terms of porosity probability. Four parameter sets result as specimen geometries from this main study, which show different levels of porosity formation. In addition, another geometry to examine the cast surface is defined, which exhibits the same manufacturing boundary conditions as the previously described specimens but exhibits almost no interior defects. Finally, a test specimen is simulated which includes a feeder containing a comparably huge defect. By the aid of this specimen, the numerical simulation procedure and the real casting process is compared. It is shown that the tendency for cavity formation is estimated quite well and the simulation methodology set-up in this master thesis is appropriate applicable for further work. To conclude, utilizing the introduced casting simulation technique, the local porosity probability in cast parts can be simulated depended on the geometry and manufacturing process parameters. Further validation of simulation results is planed as outlook based on radiography, computed tomography (CT) as well as visual examination of the developed cast specimen geometries.