The American Socnts in Turbomachinery Modeling for

Abstract

Present day turbogenerator installations are statically indeterminate rotor-bearing-foundation systems utilizing nonlinear hydrodynamic bearings. For optimal balancing and diagnostic purposes it is important to be able to correctly predict the system vibration behavior over the operating speed range. Essential aspects of this involve identifying the unbalance state, identifying appropriate dynamic foundation parameters, and identifying the system configuration state (relative location of the support bearings). This paper shows that, provided the system response is periodic at some speeds over the operating range and appropriate rotor and bearing housing motion measurements are made, it is possible, in principle, to satisfactorily achieve the above identifications without relying on the Reynolds equation for evaluating bearing forces. Preliminary results indicate that the identifications achieved promise to be superior to identification approaches that use the Reynolds equation.