A switching power amplifier (SPA) provides the driving current for an active magnetic bearing (AMB) to achieve magnetic suspension (MS) control. The performances of the current bandwidth and current ripple should be significantly considered in the SPA of the AMB. The SPA can obtain a wide bandwidth by increasing the dc-link voltage. However, a high dc-link voltage will cause a high current ripple. To achieve a wide bandwidth and low current ripple, a new GaN-device-based SPA for the AMB of a flywheel energy storage system (FESS) is proposed in this article. The GaN SPA can operate at a high switching frequency and high dc-link voltage. Thus, the GaN SPA can improve the performance of the current bandwidth and current ripple compared to the conventional SPA. According to the characteristics of the GaN SPA, a low-loss PWM method and an optimal dead time strategy is proposed to increase the efficiency of the SPA. Furthermore, a high-response control approach of dead-beat model predictive control is proposed to further improve the current bandwidth. The proposed GaN SPA is applied in a 500-kW MS FESS prototype. The effectiveness of the SPA is verified experimentally.
Extended Kalman filter (EKF) is widely applied in the permanent magnet synchronous motor (PMSM) position estimation method. However, the estimation accuracy will be degraded, when measurement noise is not 0-mean random noise. To solve this problem, this paper proposes an EKF algorithm with moving horizon estimation (MHE) to estimate the rotor position of PMSM more accurately. The proposed MHE EKF algorithm uses the concept of a moving time domain window to estimate the motor operating status by integrating the window information of the N moments. By establishing a cost function and adding random noise to replace the measurement error, the prediction problem is transformed into an optimization problem. Simulation and experiment results show that this algorithm can effectively improve the accuracy of estimation position.
In this paper, a planar flux-switching permanent magnet (PFSPM) motor, which is the first of its kind for planar motors, is designed and analyzed. First, the topology structure, modeling, and operating principle of the PFSPM motor are clarified. Then, the initial dimension of the PFSPM motor is determined in terms of its topology structure. Based on the initial dimension, the width of the iron core and the length of the permanent magnet are optimized to increase the thrust force and reduce the cogging force. In addition, a 3-D finite-element analysis (FEA) is performed. Finally, the experiments are conducted based on a developed prototype of the PFSPM motor with the optimized dimension. The results demonstrate that the thrust force is increased by 21%, and the cogging force is reduced by 56% via FEA; the thrust force of the experimental measurement is largely consistent with that determined via the FEA.
DC-link voltage control needs to be achieved for flywheel energy storage systems (FESSs) during discharge. However, load disturbances and model nonlinearity affect the voltage control performance. Therefore, this paper proposes a load-current-compensation-based robust DC-link voltage control method for FESSs. In the proposed method, the model is linearized via load current feedforward compensation and dq-axis current-to-DC-current conversion. The uncertainty of the linear model is analyzed and an H∞ robust control method is applied to overcome the uncertainty. Furthermore, experiments involving the proposed method are conducted on a 1.2 kWh magnetic suspended FESS prototype. Compared with the general proportional integral control method, the proposed method can increase the voltage response speed by 37.1% and reduce the voltage fluctuations by 29.5%. The effectiveness of the proposed method is verified experimentally.
This paper presents a DSP-based sliding mode controller of the planar switched reluctance motor (PSRM) for precision positioning applications. The mechanical structure, control scheme, and the design of sliding mode controller are presented and analyzed, respectively. The DSP-based sliding mode controller is implemented for the PSRM experimentally. The steady-state positioning errors of x- and y-axes are 0.016 and 0.029 mm, respectively. The effectiveness of the DSP-based sliding mode controller for the PSRM is verified through the experimental results.
16 April 2025 in IEEE Transactions on Industry Applications
Model predictive position control (MPPC) is a promising control method for planar switched reluctance motors (PSRMs) to achieve high-precision trajectory tracking. However, the control performance of MPPC may deteriorate due to a mismatch between the system model and the actual system. To address this issue, online parameter identification can be used as an alternative method to model refinement. In this paper, an MPPC method based on recursive extended least squares (RELS) is proposed to improve the trajectory tracking performance of a laboratory-developed PSRM system. The conventional dynamic model and RELS-based parameter identification dynamic model for the PSRM are established. The predictive model is developed using the parameter identification dynamic model, with its parameters updated online. The control law is derived from the developed predictive model and the defined cost function, and the system stability is subsequently analyzed. The simulation and experimental results show improvement in the trajectorytracking performance of the PSRM system. The effectiveness of the proposed method is verified.