A three-dimensional magnetic field analytical model of axial-flux permanent magnet couplers is presented to calculate the eddy current loss, and the prediction of the copper plate temperature under various loads is analyzed. The magnetic field distribution is calculated, and then the eddy current loss is obtained, with the magnetic field analytical model established in cylindrical coordinate. The influence of various loads on eddy current loss is analyzed. Furthermore, a thermal model of axial-flux permanent magnet couplers is established by taking the eddy current loss as the heat source, using the electromagnetic-thermal coupled method. With the help of the thermal model, the influence of various loads on copper plate temperature rise is also analyzed. The calculated results are compared with the results of finite element method and measurement. The comparison results confirm the validity of the magnetic field analytical model and thermal model.
Rotational permanent magnet eddy current couplers are promising devices for torque and speed transmission without any mechanical contact. In this study, flux-concentration disk-type permanent magnet eddy current couplers with double conductor rotor are investigated. Given the drawback of the accurate three-dimensional finite element method, this paper proposes a mixed two-dimensional analytical modeling approach. Based on this approach, the closed-form expressions of magnetic field, eddy current, electromagnetic force and torque for such devices are obtained. Finally, a three-dimensional finite element method is employed to validate the analytical results. Besides, a prototype is manufactured and tested for the torque-speed characteristic.
An improved structure for axial-flux eddy current couplings with combined rectangle-shaped magnets is proposed to minimize cogging torque without sacrificing torque density. Then, a novel analytical electromagnetic-thermal model, which takes the change of material electromagnetic and thermal characteristics under temperature, is presented to calculate cogging torque, electromagnetic torque, eddy current loss and temperature. The model is valid for the whole working area. The magnetic field distribution is calculated by the quasi 3-D electromagnetic field analytical model. Then cogging torque, electromagnetic torque and eddy current loss can be obtained. Furthermore, an equivalent thermal resistance network model is established. The thermal model takes the previously obtained eddy current losses as the heat source, which is the eddy current loss of copper and copper back iron. With the help of the electromagnetic-thermal model, the characteristics of slip on electromagnetic torque, eddy current loss, copper and permanent magnet temperature rise is also analyzed under the whole working range. The validity of the electromagnetic–thermal model is confirmed with finite element method (FEM) and measurement.
To address the challenges of diminished back-EMF, high noise interference, and reduced accuracy in traditional low-speed sensorless control, this study proposes a rotor position estimation method based on structural vibration characteristics. The coupling mechanism between air-gap electromagnetic force density and stator structural vibration modes is analyzed. This analysis reveals that rotor spatial information is embedded within specific modal phases, establishing the physical basis for utilizing vibration phase as a position carrier. Accordingly, a workflow encompassing signal acquisition, modal selection, and phase calculation is developed and integrated into a sensorless control system. Simulation results demonstrate that the proposed method achieves stable estimation even under strong noise. The estimation error shows clear performance advantages over conventional back-EMF-based methods in the low-speed region, validating its effectiveness and robustness at low speeds. This research provides a new approach that introduces non-electrical structural information as a complementary channel to overcome the inherent limitations of electrical-signal-based position estimation at low speeds.