Dueto various kinds of stator phase arrangements, existing fault diagnosis (FD) methods cannot be applied to different types of multiphase machines. Spurred by the era of big data and artificial intelligence, an improved machine-learning-based FD method with adaptive secondary sampling filtering is proposed for the multiphase drive systems. Experimental results of the proposed method on both five-phase and six-phase motor drive platforms validate its satisfying generalization capability as well as high accuracy and strong robustness.

The voltage source inverter (VSI) with pulse width modulation (PWM) is widely used in motor drive occasions owing to the technical maturity and simplicity. However, the inverter operating in the discrete and impulse state generates the high-frequency common-mode voltage (CMV) in the neutral point terminal of stator winding which induces the negative effects. This paper proposes a CMV elimination modulation scheme for the dual two-level inverter fed asymmetrical six-phase permanent magnet synchronous motor (ASPMSM) which has two sets of three-phase windings spatially shifted by 30° electrical degrees. The modulation algorithm shifts the PWM signal with the designed end-to-end cyclic sequence which can retain the zero voltage vectors for each sub-CMV of dual three-phase windings and theoretically eliminate the total CMV for the motor drive system. The validation of the proposal is verified by simulations and experiments.

This paper introduces a novel pulse width modulation (PWM) scheme for an OW-PMSM driven by dual two-level three-phase inverter with common dc bus which can effectively deal with the inherent zero-sequence current (ZSC) problem. Based on conventional symmetrical unipolar double frequency SPWM scheme with appropriate phase-shift, the common mode voltage (CMV) of two inverters can keep the same and cancel out each other to eliminate the modulated zero sequence voltage (ZSV) disturbance source. In this case, the double frequency effect can be retained to reduce the ac side current ripple and suppress both the corresponding motor vibration and acoustic noise which is advantageous to improve the synthetic performance of motor. The DC bus voltage utilization of the novel PWM scheme is proved to reach the maximum value as same as the conventional modulated ZSV elimination scheme. Meanwhile, a zero-sequence controller is designed to suppress ZSC by further adjusting the two CMVs to counteract other zero-sequence disturbance sources. To verify the analysis, the proposed PWM technique associated with the control method is implemented in an OW-PMSM experimental setup to validate the superiority of proposed method.

Determination of high-frequency circulating current is important for coupled inductor design of paralleled voltage source inverters (VSIs). A time-domain model describing the circulating current of coupled inductor within a switching period is presented in this study. Using this model, the peak or RMS values of circulating current can be calculated for different PWM methods. Both the conventional 180o interleaved PWM and zero-CM (common-mode voltage) PWM algorithms are studied here. Considering the saturation nature of the magnetic cores in determining the effective inductance of coupled inductors, the variable switching frequency PWM (VSFPWM) is proposed to better utilize the coupled inductor's saturation limit, which is determined by high-frequency circulating current peak value. With the proposed strategy, both switching losses and EMI are improved with respect to the constant switching frequency PWM (CSFPWM). Moreover, the detailed power losses analysis of paralleled VSIs has been presented and the dynamic performance of proposed VSFPWM has been well investigated. Finally, the proposed method has been extended to the system with 2N paralleled inverters. Detailed simulation and experimental results are provided to validate the proposed method.

The high fault-tolerant ability of multiphase drives is favored in safety-critical applications. Under open-phase faults (OPFs), to guarantee ripple-free torque, stator current references should be revised. In this paper, a global fault-tolerant control strategy based on an online current optimization algorithm (OCOA) is proposed for symmetrical multiphase machines (SMMs) to achieve both the maximal torque production range (TPR) and minimal stator winding losses under all possible OPFs. The OCOA can calculate the minimum losses current references covering the full TPR online, regardless of the number and locations of faulty phases. With the online-optimized references, ripple-free torque and minimum losses in full TPR can be achieved under faulty conditions. Additionally, the proposed method can be adopted to SMMs with arbitrary phase numbers. Simulation and experiments demonstrate the online optimization can be completed in a short time, meanwhile, achieves the same TPR and loss reduction compared with existing offline strategy based on pre-stored look-up tables. Due to the flexible online calculation ability and excellent extensibility, the proposed method is especially favored in modular multiphase converters and machines with high phase numbers, where converters can adapt to machines with different phase numbers or the number of possible faulty conditions can be extremely large.