Improvement of Sensorless Rotor Angle Estimation Using Zero-Sequence Voltage Measurement

Sensorless rotor angle estimation methods are gaining importance due to the increasing demands on electric motors. However, their use is often limited by insufficient rotor angle accuracy and associated additional energy losses. Under dynamic constraints, this paper shows how to reduce the estimated rotor angle of the zero-sequence voltage measurement using high-frequency square-wave injection. This is achieved by including higher order anisotropies in the motor model and estimating its parameters by injecting voltages in specific directions. With the improved model, the estimated angle becomes more accurate. First, the mathematical model and numerical simulations for this new approach are presented in this paper. Measurements on a permanent magnet machine validate the applicability of this novel approach and address concerns regarding accuracy and energy efficiency. As a result, the maximum rotor angle estimation error can be reduced from 12 ° to 4 ° for the considered motor.