Large structural components of energy-generating units strongly influence the weight of such facilities. These components are typically made of ductile iron. The dimensioning of such parts is commonly dominated by guidelines and codes which are very conservative. The scope of this thesis is to illustrate the potential for lightweight design by utilizing local material properties for the fatigue design of these facilities. The first part of the present work deals with the verification of casting simulation at the example of a wind turbine hub made of EN-GJS-400-18-LT. Microsections as well as specimen for tensile tests have been take from different locations within the hub. Results showed, that the microstructure is in good correlation with the simulation after calibration. The static material properties could not be estimated properly. The technological size effect of ductile iron has been thoroughly investigated within the present work. It has been shown, that the static material properties exhibit a less pronounced technological size effect then the cyclic mechanical properties. Fatigue limits of 151,4 MPa to 211,6 MPa have been tested for a fully ferritic matrix. An empirical material model, as well as a fracture mechanical model based on the number of graphite nodules per square millimeter have been derived to describe the microstructure dependent fatigue properties. To quantify the effect of larger highly stressed volumes, specimen of different sizes have been tested. Special focus was laid to characterize the mutual operating influences isolated from the technological size effect. Therefore, the influence of torsional load, the mean stress influence, the effect of stress gradients as well as the influence of low temperatures on the fatigue behavior have been systematically investigated. The influence of dross on the fatigue properties of ductile iron has been evaluated for different microstructures. It has been shown, that dross significantly decreases the fatigue strength independent of the microstructure. The amount of dross influences the remaining fatigue properties, large amounts of dross yield in lower fatigue strengths. In the last section of this thesis the generated material models are combined into an optimization based on local material properties. The classical optimization loop is enhanced by the process simulation as well as the fatigue post processor. Therefore, components can be optimized using damage calculated with local material properties as an objective function. Some sample optimizations are used to illustrate the differences compared to the classical optimization.