Reduced order models are required for controller design purposes to allow practical realization and implementation in real-time applications. In the rotor dynamics field, the controllers associated with magnetic bearings are classic examples of such a requirement. A technique is introduced to reduce the order of a finite element rotor model to one involving only the required rectilinear displacements in sensor and controller nodal planes. The reduced order model is shown to predict a number of natural frequencies and responses to step changes in unbalance with considerable accuracy. A simulation of the rotor making contact with a rigid retainer bearing is also established using a constrained Lagrangian multiplier. Again, the differences between the full and reduced order model rotor orbits are shown to be small.

This content is only available via PDF.
You do not currently have access to this content.