Permanent magnet synchronous machines (PMSMs) torque production capability depends on the permanent magnets (PMs) magnetization state, which can be affected by PMs temperature and of the current flowing throughout the stator windings; knowledge of the PMs temperature can be therefore of great importance both for control and monitoring purposes. PMs temperature can be measured or estimated; PM temperature measurement is not easy and is not normally implemented in commercial drives. PM temperature estimation methods can be divided into thermal models based, BEMF based and signal injection based methods. Existing HF signal injection methods estimate the PM temperature from the measured stator HF resistance. Unfortunately the resistance is also affected by magnetoresistive effect, which can limit the accuracy of the estimates. This paper proposes the use of the stator d-axis HF inductance for PM temperature estimation. This makes temperature estimation insensitive to magnetoresistive effect. In addition, it allows the use of higher frequencies, reducing the adverse impact of the injected signal on machine performance.

Demagnetization in permanent magnet synchronous machines results in a decrease in the motor average torque and could also increase the torque ripple and consequently vibration and acoustic noise, degrading the overall performance and reliability of the drive. Demagnetization detection can be therefore of tremendous importance. Use of Hall-effect sensors has proven to be a viable option for demagnetization detection. Results reported using this method were based on the analysis of the flux density complex vector; the accuracy of this implementation strongly depends on the angle of the Hall-effect sensor, being also sensitive to implementation issues such as offsets and gain variations of the sensors, magnet temperature, manufacturing tolerances, and stator current injection. This paper proposes the use of the zero-sequence component of the magnetic flux density for demagnetization detection. The use of this signal will improve the accuracy and reduce the sensitivity to implementation issues compared to the use of the flux complex vector.

Negative d-axis current injection used in interior permanent magnet synchronous machines (PMSMs) weakens the permanent magnets' (PM) flux, eventually increasing the PMs losses and reduces therefore the machine efficiency. Several studies have reported the increase in the airgap flux harmonic content as the primary reason for the PM losses increase in flux weakening operation, disregarding PM properties variation with the magnetization state (MS) due to PM magnetoresistive effect. This paper analyzes the variation of PM losses with the MS and flux weakening current, the target being threefold: a better understanding of PM losses production mechanisms; to analyze the impact of MS on the thermal behavior of the machine; use this knowledge to improve the machine design

Torque measurement/estimation in surface permanent magnet synchronous machines (SPMSMs) is needed in many applications. Torque measurement systems are expensive, require extra room, add weight, can introduce resonances into the system, and their performance can be compromised due to electromagnetic interference. Alternatively, torque can be estimated, accurate knowledge of machine parameters being critical in this case. Injection of a high frequency (HF) signal in the stator via inverter has been shown to be a viable option for online identification of machine parameters identification required by torque estimation methods for interior PMSMs (IPMSMs). However, the reported methods require the use of two pulsating HF currents, which introduces concerns both on the adverse effects on machine's performance (additional noise, vibration, additional, …) as well as to the additional computational burden, as HF current controllers are required. This paper proposes a torque estimation method for SPMSMs using a single HF signal; further more, a HF voltage signal can be used, avoiding the use of HF current controllers, and therefore simplifying the implementation.

Torque measurement/estimation in permanent magnet synchronous machines (PMSMs) is required in a large variety of applications. Direct torque measurement by means of torque transducers is typically limited to laboratory applications due to cost, reliability, and room concerns. Alternatively, torque can be estimated. Torque estimation methods require accurate knowledge of machine parameters. Injection of a high frequency (HF) signal in the stator via inverter has been shown to be a viable option for online machine parameters identification in PMSMs. However, this introduces concerns on the performance of the machine due to the noise, vibration, and additional losses produced by the HF signal. This article proposes a torque estimation method for PMSMs using low cost Hall-effect sensors. The method is based on the dependency of the leakage flux complex vector due to interaction between permanent magnet flux and stator coil flux and does not interfere with the drive operation of the machine.

Torque pulsations are one of the major concerns in permanent magnet synchronous machines (PMSMs). Torque pulsations can be mitigated by proper machine design and/or by control. Independently of the means being used, precise measurement of torque pulsation is highly desirable for the validation and improvement of machine designs and control strategies. This paper proposes a non-invasive, wireless vibration measurement system aimed to provide precise estimates of torque pulsations. The system is mounted on the rotor shaft of the PMSM, i.e. there is no mechanical coupling, and allows on-line measurement without interfering with the normal operation of the machine.