Abstract: The key of stable suspension of bearingless permanent magnet synchronous motor is the closed-loop

control of rotor radial position displacement. Usually，the high performance suspension operation of the motor relies on displacement sensor，which destroys the robustness of the motor structure and hinders the low-cost practicability of the motor. Therefore，a new displacement estimation strategy is needed to replace the traditional mechanical displacement sensor. By establishing the state equations of flux linkage，voltage and current of suspension and torque windings of the motor，the rotor displacement estimation model based on least square method was constructed，and a hybrid weighted least square method based on the common least square method and the forgetting factor least square method was proposed. By sampling the voltage and current of two sets of windings，the hybrid weighted least square algorithm was used to identify the rotor displacement online. Simulation and experimental results validate that this method can achieve effective estimation of the motor rotor displacement under sensorless condition，the motor suspension and rotation system has good performance.

Key words: bearingless permanent magnet synchronous motor；least square method；rotor displacement

estimation