Design of Sensorless Bearingless Motors for Displacement Estimation Using Mutual Inductance

Bearingless motors can produce electromagnetic torque and force using a single machine. The forces are often used to actively levitate the rotor which requires high-bandwidth displacement sensor feedback. These sensors add undesired cost, bulk, and cabling to the drive system. To mitigate these issues, the motor itself can be used as the rotor displacement sensor in a 'sensorless' or 'self-sensing' approach. Based on electromagnetic variation due to rotor eccentricity, real-time self-sensing control algorithms can generate estimates of rotor displacement which can be used for levitation feedback signals. This work proposes a framework for motor design specifically to optimize for self-sensing capability. The main goal is to understand and maximize the displacement self-sensing sensitivity and linearity. A multi-objective optimization is used to explore the design space and quantify performance trade-offs in power conversion, force actuation, and self-sensing. Leveraging the optimization results, a prototype sensorless bearingless motor is fabricated and tested. Its eccentric-rotor inductance is measured statically versus frequency and rotor angle, showing less than 10% mutual inductance ripple. Real-time online self-sensing results are provided which demonstrate 15 μm resolution displacement estimates at 800 Hz bandwidth during transient motion.