The paper introduces a comprehensive investigation in end winding inductances of large two-pole turbo-generators. With the aid of an analytic-numeric approach, where Neumann's formula is applied, the influence of geometric characteristics of double-layer stator end windings with involute shape is analysed. This parameter study results in approximation formulas for the stator self and mutual inductances at stand level as well as for the common used end winding leakage inductance. In order to consider field affecting components as pressure plate, flux shield, rotor shaft and rotor retaining ring, finite elements models for two machines (250 MVA and 1150 MVA) are created and computed. The results are integrated in the developed approximation formulas. Finally the simulation results of machine 1 are compared to the data of two different measurements. All approaches introduced in this paper show good correlation. The high speed of the analytic-numeric calculation is combined with the accuracy and opportunity to consider field affecting components within the extensive finite element computation successfully.

Turbogenerator coil retaining rings are shrunk-fitted onto the rotor over the coils, in order to restrain them against the centrifugal force. They are typically subjected to low cycle fatigue, with a cycle being completed at every machine switch-on and switch-off. The subject of this paper consists in the determination of the failure probability of a coil retaining ring. The failure mode of the ring cracking, when it swells in tension, due to the centrifugal force is here considered. The reliability assessment is preceded by the study of the input variables affecting the low-cycle fatigue load and of their stochastic distributions. This question is tackled by the experimental determination of the static, cyclic and fatigue curves of the involved material and by the application of a statistical model to compute related parameters and their standard deviations. Upon the determination of variable distributions, the probability of failure is estimated in the form of a cumulative distribution function by a computationally efficient methodology, based on the Advanced Mean Value approach. The obtained results account for the material response and the local stressstrain states at the most loaded coil retaining ring region. The determined probability at the end of the machine life, in the order of 10^-12, is compatible with reference values for structures under fatigue in the mechanical and aeronautical fields.