Heat treatment of steels is a widespread technology to adjust the properties of materials. Throughout the heat treatment process of big shafts damage can occur leading to failure of those parts. During this process hydrogen in combination with eigenstresses is of decisive importance. Furthermore the microstructure has a great inuence on the occurrence of so-called hydrogen induced stress corrosion cracks. Especially hard phases react susceptibly to this kind of cracking mechanism. Those hard phases can form due to segregated areas in the shaft. Segregated areas are known to show delayed phase transition compared to the matrix material. It is the aim of this thesis to estimate and reduce the damage potential of big shafts during heat treatment processes with the finite element package ABAQUS. For this the computation of residual stresses of 1st and 2nd order is needed. Residual stresses of 1st order are computed in a macroscopic model that delivers results on continuum length scale. Those results are transferred to a microscopic scale model that computes 2nd order residual stresses. Based on the results of those simulations an alternative process design is derived that leads to a reduction of the damage probability during the heat treatment process. A promising concept to avoid cracks during isothermal annealing is to increase the cooling rate in the present standard process. This leads to a prevention of di_usion caused enrichment of hydrogen in segregated areas and therefore lowers the probability of damage occurring during the process. In contrast to the isothermal annealing a reduction of the cooling rate of the quenching process is recommended due to the hard phase fraction being the major cause of damage. Therefore media giving lower cooling rates than water should be considered. In addition the stress position and magnitude can be shifted in shafts with smaller diameters by an appropriately chosen heat treatment coe_cient so that areas with a tendency of damage can be relieved.