Modeling of a High Speed Rotor Test Rig With Active Magnetic Bearings

Abstract

This paper reports on the modeling and experimental identification of a high speed rotor-magnetic bearing test rig. An accurate nominal model and an uncertainty representation are developed for robust controller synthesis and analysis. A combination of analytical modeling, model updating, and identification is employed for each system component and for the system as a whole. This approach takes advantage of both the behavior modeling and input/output modeling methods. Analytical models of the rotor and the magnetic bearings are first developed from physical laws and refined by comparison with the experimental data. The substructure model is directly identified from the experimental data by a structured identification approach. Models of the electronic systems, such as the filters, amplifiers, sensors, and digital controller, are developed through experimental identification. These component models are then assembled to obtain the overall system model. Closed-loop tests are conducted to identify parameters in the model. Advanced control techniques based on H∞ control and μ synthesis are developed and successfully implemented on the test rig, which further validates the model.