Adaptive compensation of sensor runout and mass unbalance in magnetic bearings

Active magnetic bearings (ANBs) have increasingly become the choice for high-speed, high-performance rotating machinery because they provide the scope for contactless and frictionless operation. Since magnetic bearings are open-loop unstable, they require careful control system design. Although general feedback control techniques have been proposed for precise shaft levitation, the problem of sensor runout (SRO) has been largely overlooked due to its similarities with mass unbalance in creating periodic disturbances. Furthermore, the important problem of synchronous SRO and unbalance compensation has not been adequately investigated.; To improve the accuracy of magnetically levitated rotors, we propose for the first time an adaptive control framework that can compensate SRO and unbalance, both individually and simultaneously, while providing shaft stabilization about the geometric center. In our approach, bias currents in the magnetic coils are periodically perturbed to create persistency of excitation that guarantees individual identification of the harmonic components of the synchronous disturbances. Through feed-forward cancellation of the disturbances and careful control system design, the algorithm provides geometric center stabilization that is robust to uncertainty in plant parameter values. While Lyapunov stability theory and its derived passivity formalism provide a solid theoretical framework for the algorithm, corroborating experimental results establish the simplicity of the design and implementation procedure. The algorithm applies to both SISO and MIMO systems involving a rigid rotor and future studies are expected to broaden its applicability to flexible rotor models.