Modeling of Frequency-Dependent Damping for Fast Vibration Prediction in Permanent Magnet Synchronous Machines

Vibration prediction enhancement in permanent magnet synchronous machines (PMSMs) through frequency-dependent damping is proposed in this work. Rayleigh damping coefficients are used to implement frequency-dependent damping. Mass- and stiffness-dependent Rayleigh damping coefficients are determined by utilizing frequency response functions (FRFs) from the impact hammer test for two prototype 12-slot 10-pole (12s10p) PMSM designs and are used during vibration prediction rather than the assumption of constant damping. Predicted vibration spectra are experimentally validated through run-up tests of two prototypes. A damping coefficient prediction strategy is proposed to enable fast vibration prediction of future builds and to reduce the dependency of accurate vibration prediction on the availability of a prototype. The damping prediction method uses the change in mass, natural frequency, and stiffness from one design to another to predict the Rayleigh damping coefficients of a future build. The proposed damping prediction method is experimentally validated using a third prototype.