Model Reduction of Large Rotor-Bearing-Foundation Systems

Finite element models of mechanical or structural systems contain too many degrees of freedom (dofs) for dynamic analysis, particularly if the system is nonlinear. This paper presents an investigation on the performance of nonlinear rotor-bearing-foundation systems with reduced rotor models obtained by the Guyan and the mode superposition methods. A simple unbalanced rotor-bearing-foundation system wherein the rotor is supported via four hydrodynamic journal bearings; i.e. a nonlinear statically indeterminate system, was selected for numerical evaluation purposes. The steady state responses obtained using the reduced rotor models were compared with the responses of the original unreduced system. The results show that whereas the Guyan rotor model provides good accuracy in low frequency range, the mode superposition rotor model is capable of good accuracy in the high frequency range as well. It is concluded that the mode superposition method is preferable to the Guyan reduction method for mechanical systems, such as turbo generator units, that operate over a wide speed range.