Abstract
Levitation control of bearingless motors requires rotor displacement sensors. Recent research has explored self-sensing (sensorless) control which enables removing these sensors, but only for separated windings. Combined windings—where each coil can create torque and force—have been proven to be a critical component for realizing high performance bearingless motors for industrial applications. These applications are often cost-sensitive and require bearingless solutions with minimal overhead compared to traditional motors. Self-sensing control with combined windings removes barriers for wider adoption of bearingless technology—this paper forms the analytic basis for showing this is possible. The primary contribution is a unified modeling framework which encompasses several popular combined winding types found in the literature. For each winding type, winding functions are defined—each type with differing winding coil circuit arrangements and drive interfaces. The terminal inductance matrix and back-EMF is computed to include the effects of a non-centered rotor, and then transformed into decoupled torque and suspension force components. Through the presented transformations, all combined winding types support radial self-sensing, albeit with additional control complexity. Finite element simulations and hardware experiments validate the modeling approach. Future self-sensing research can build upon these models to design control algorithms and machines which work for any combined winding type.
Funding Information
  • Wisconsin Alumni Research Foundation
  • Accelerator Challenge