Robustness of Adaptive Unbalance Control of Rotors with Magnetic Bearings

Rotor unbalance in the primary cause of unacceptable vibration in rotating machinery. Over the last decade, researchers have explored different methods of taking advantage of the active nature of magnetic bearings to attenuate unbalance response including both feedback and adaptive open loop methods. An important issue in the application of this technology to industrial machines is the robustness of the unbalance control algorithm. The stability and performance robustness of a promising adaptive open loop control algorithm is examined. Expressions are derived for a number of unstructured uncertainties. Experimental results are then presented, which evaluate the algorithm's robustness with respect to three variations: gain schedule errors, random additive errors, and feedback loop gain. The robustness exhibited in these tests was quite good and, along with the excellent vibration attenuation obtained, recommend the algorithm for further testing and industrial application. The experimental results indicate that the theoretical robustness expressions do provide an upper bound on actual performance, however this bound is not tight. Although the conservatism in the results is partly due to the variations considered and the worst-case nature of the performance robustness guarantees, the results also indicate that further research is needed on unstructured performance robustness for this method of rotor vibration control.