In this paper, an improved deadbeat predictive current control (DPCC) method with parameters identification for surface-mounted permanent magnet synchronous machines (SPMSMs) is proposed. With the proposed DPCC method, zero steady-state current error and deadbeat dynamic current response could be achieved, even with inaccurate initial motor parameters. On basis of the conventional DPCC method, a novel parameters identification for the stator resistance and inductance is developed, which is the main contribution of this paper. The proposed parameters identification method works based on a reconstructed characteristic vector from the disturbance observer with current injection. Compared with traditional recursive-least-square (RLS) methods, the proposed method can be implemented with greatly reduced computation burden. Additionally, since the design is established based on the fully-discretized model, the effectiveness will be guaranteed on both low-frequency and high-frequency motors, which is a significant advantage of the proposed method.

In this paper, a position sensorless drive and online parameter estimation method for surface-mounted PMSM based on adaptive full-state feedback current control is proposed. The position sensorless drive is established by the detection of the back-EMF in the ${\gamma\delta}$ synchronous reference frame, which is effective at the medium-speed and high-speed range. Besides, accurate estimation of the winding resistance, the stator inductance and the flux linkage of the PM is achieved independently. Compared with the traditional recursive-least-square (RLS) methods, the proposed parameter identification method can be easily implemented because of the significantly reduced execution time. With the help of the parameter identification, the precise position estimation can be achieved by the proposed sensorless control method regardless of the parameter variation during the operation. The stability of the proposed method is proved by the Lyapunov-function method. Finally, the effectiveness of the proposed method is validated by the simulation and experimental results.

This paper proposes an analytical model for the prediction of airgap magnetic field distribution (MFD) for axial flux permanent magnet (AFPM) machine with partial magnet demagnetization. The AFPM machine geometry is first converted to a polar representation. Consequently, the sub-domain model based on current sheet technique is developed. Then current sheet representation of PM is derived to consider the partial demagnetization using superposition principle. The back electro-magnetic forces and cogging torque are obtained accordingly based on Maxwell's equations. The results show that the results of proposed approach agrees with that of finite element (FE) method. The model is further validated by experiments under both healthy and demagnetized conditions, which can validate the proposed approach. Main contribution of the work is to consider the partial irreversible demagnetization Moreover, the proposed method is applicable for both AFPM and radial flux permanent magnet (RFPM) machine.

Back–electromotive force estimation-based rotor position estimation methods are usually adopted for high-speed surface-mounted permanent magnet synchronous machines (SPMSM) drive to deal with the limitations of position sensors in high-speed applications. Considering inductance mismatch is the main cause of the position estimation error, this article proposes a novel pulsewidth modulation (PWM) current ripple-based online inductance identification method to improve the position estimation accuracy. The proposed method utilizes the inherent PWM current ripple of the voltage source inverter for inductance identification. In addition, the transient circuit equations on the estimated $\gamma \delta$ frame are adopted as the identification model, and the recursive least squares method is used to obtain the estimated inductance. Compared with the traditional parameter identification methods, the proposed method does not need to inject additional signals or fix parameters. During the sensorless drive of high-speed SPMSM, the proposed method can accurately identify the inductance while not affecting the normal operation of PMSM at all, which is the main contribution of this article. With the proposed method, the estimated inductance can rapidly converge to the accurate value, and then the position estimation error will be eliminated. Finally, the effectiveness of the proposed method is verified by simulations and experiments.

This paper proposes an analytical model for the prediction of airgap magnetic field distribution (MFD) for axial flux permanent magnet (AFPM) machine with various types of misalignments. The AFPM machine geometry is first transformed to a polar representation. Consequently, the sub-domain model based on current sheet technique is developed. Then the stator coordinate system is chosen as reference coordinate to consider both static/dynamic angular and axis misalignments. The back electro-magnetic forces and cogging torque are obtained accordingly based on Maxwell's equations. The results show that the proposed approach agrees with the finite element (FE) method. The model is further validated by experiments under healthy, dynamic angular and axis misalignment conditions, which can validate the proposed approach. It turns out that the proposed approach can predict the performance of AFPM machines with types of misalignment quickly and effectively, which is a great significant for the further fault detection.

In this article, a novel dynamic-decoupled active damping current controller is proposed for an LCL -equipped high-speed permanent magnet synchronous machine. Compared with the conventional stationary current-control method for the LCL -type system, the proposed method is established in the synchronous rotating frame for improving the current transient performance. When taking the controller into the synchronous coordinate, there are two following challenges: first, the synchronous resonance frequency varying in a wide range because of the synchronous coordinate transformation, and second, eliminating the coupling between the ${dq}$ coordinate. To address these issues, an improved synchronous capacitor-current-feedback active damping method is designed based on arbitrary pole assignment and is significantly effective for the LCL resonance within the Nyquist frequency. Moreover, a novel dynamic-decoupled motor-current controller is proposed to eliminate the coupling between the ${dq}$ -axis motor current. The gain selection method is discussed to acquire sufficient phase margin and gain margin. Finally, the effectiveness of the proposed method is verified by driving the tested motor to 72 kr/min.