Experimental Low-Speed Performance Evaluation and Comparison of Sensorless Passive Algorithms for SPMSM

Sensorless algorithms for PMSM have achieved an increasing interest in the technical literature. They can be divided into active methods and passive methods. Active methods exploit rotor anisotropy by injecting high voltage frequency signals, whereas passive methods are based on observers. Even if passive methods do not have the drawbacks of active methods - acoustic noise, torque ripple and harmonic distortion - their effectiveness is reduced in the low speed region. For this reason, active methods are usually used at low speed and the motor control switches to passive methods in the high-speed region. Since high frequency injection is associated to many drawbacks, the aim of passive methods is to work at lower speed as possible or, ideally, to be able to start directly the motor. That is particularly true for Surface-Mounted PMSM: since they have a low anisotropy, high frequency injection should be particularly elevate to achieve sufficient signal-to-noise ratio. It is difficult to compare the performance of passive algorithms proposed in the technical literature since experimental tests have been done with different conditions: motors, inverter features, measurements, type of tests, etc. In this paper, five different sensorless passive algorithms for Surface-Mounted PMSM, selected among the most promising algorithms available in the technical literature, are compared, performing the same tests on the same experimental setup, to achieve a fair comparison and establishing the most performing one.