Tracking the Polar Principal Axis for Magnetic Bearing Control and Flywheel Balancing

Controllers for flywheel magnetic bearing designed by the usual “position direct feedback” approach are prone to system imbalance and sensor error problems. Remedies for such problems with that approach are complex and unreliable. In the long run, fatigue cycles would eventually change the flywheel system characteristics from their initial values on which the remedy parameters were designed. In this paper, a simple and reliable alternative design approach will be presented to help avoid most flywheel control problems. This design approach is based on an adaptive Kalman-Bucy state estimator that in spite of various system uncertainties can accurately track the motion of the wheel polar principal axis for feedback to the magnetic bearing controls. An implementation of this design on a Digital Signal Processor was tested for robustness, energy efficiency, and versatility. This controller prototype is easy to use for nearly any high-speed flywheel. Compact, analog implementation of this controller is feasible, which would reduce costs incurred by electronics and digital components. Potential applications of this Polar Principal Axis tracking technique for the design of voice-coil dampers for repulsive-pole bearings, and for wheel balancing, alignment, and health monitoring are also presented.