This paper presents a delay-suppressed sliding mode observer (SMO) to observe the real-time rotor position of a permanent magnet synchronous machine (PMSM) controlled by vector control (VC) algorithms. Firstly, in order to solve the low pass filter (LPF) delay problem existing in the traditional signum function-based sliding mode observer (SMO), a brand-new hyperbolic function is initially selected as the switching function. Because a hyperbolic function with proper boundary layer is capable of reducing the chattering phenomenon of a SMO, it is not necessary to re-employ LPFs to eliminate the adverse impacts of chattering on the position estimation accuracy. In order to ensure the reachability and stability of the hyperbolic function-based SMO, the observer gain is calculated by the means of a Lyapunov function in this paper. Secondly, to solve the problem of calculation delay caused by digital computation, a current pre-compensation scheme based on dual-sampling strategy in one switching period is proposed. After compensating the calculation delay, the accuracy of position estimation as well as the motor control performance can be improved. Finally, the proposed SMOs with and without delay compensation are verified by both simulation and experiments which are conducted on a three-phase 1.5kW PMSM drive prototype.

The motor drive system plays a significant role in the safety of electric vehicles as a bridge for power transmission. Meanwhile, to enhance the efficiency and stability of the drive system, more and more studies based on AI technology are devoted to the fault detection and diagnosis (FDD) of the motor drive system. This article reviews the application of AI techniques in motor FDD in recent years. AI-based FDD is divided into two main steps: feature extraction and fault classification. The application of different signal processing methods in feature extraction is discussed. In particular, the application of traditional machine learning and deep learning algorithms for fault classification is presented in detail. In addition, the characteristics of all techniques reviewed are summarized. Finally, the latest developments, research gaps, and future challenges in fault monitoring and diagnosis of motor faults are discussed.

Time delay of the finite control set model predictive current control caused by a large amount of calculations will degrade the control performance of the permanent magnet synchronous machine system. This letter first demonstrates the delay issue at length. Then, a new direct compensation method, by predicting the current variation within the delay time, is proposed. In comparison with the traditional two-step prediction strategy, the novel method is also easy to implement and can single out the best switching state in each period for optimal control. Simulation and experimental results are presented to verify the effectiveness of the proposed method.

Three-phase interior permanent magnet synchronous motor (IPMSM) drive systems have gotten widely used in various applications due to their speed regulation characteristics and high power density. Compared with classical field-oriented control (FOC)and flux-weakening control, model predictive control (MPC) technique is more efficient in achieving excellent performance without complicated controller tuning but with explicit constraints. Yet, the optimum application of MPC algorithm in machine control system is in the exploratory stage.In order to lower the current and torque ripples, this paper retains a modulator, but replaces all PI-control loops which are contained in the conventional FOC systems with a single MPC controller. Moreover, we adopt a brand-new linearization approach to tackle the strong coupled nonlinear IPMSM mathematical model, obtaining an improved linear plant model (ILPM) which is suitable for the motor with constant load torque. When it comes to flux-weakening control,the required d-axis current is calculated, after which it is used as the input of the proposed MPC controller, abandoning the previous natural field weakening method. Finally, for the purpose of lowering the dynamic speed and current overshoot, a further constraint in the change rate of manipulated variables is discussed. The improved control algorithm has been verified in simulation and experiment.

This paper proposes a new finite control set model predictive control (FCS-MPC) strategy that simultaneously evaluates two targeting control objectives (TCOs) including speed and currents in a single cost function, achieving high-performance single-closed-loop control structure. Besides, aiming at the calculation delay problem of the MPC controllers, a novel calculation delay compensation method by predicting the current variation within the delay time is proposed. In this paper, an improved machine model that is specially designed for the multi-objective FCS-MPC operation is illustrated at first. Then, a new cost function which can evaluate the tracking performance of speed and d-axis current and steady-state performance of q-axis current is developed. Compared to the conventional FCS-MPC approaches, extra speed controllers are not needed so that the proposed control topology becomes simpler. Then, in order to tune the weighting factors for the speed and currents in the cost function, an efficient handling strategy containing two implementation procedures, state variable normalization (SVN) and balance of state sensitivity to voltage alteration, is developed. Finally, a brand-new computation delay estimation and compensation technique based on dual sampling within a switching period is proposed to reduce the current and torque ripples during control process. Comparative simulation and experiments are conducted on a three-phase 1.5kW PMSM drive system.

This letter proposes a novel sliding mode (SM) disturbance observer-based technique to diagnose demagnetization fault of interior permanent magnet (IPM) motors with stator parameter mismatch impacts eliminated. First, the IPM motor model incorporating the disturbances caused by the PM demagnetization and stator parameter mismatch is established. Then, an SM disturbance observer is constructed to identify the overall disturbance caused by all parameters, with its stability discussed by using the Lyapunov function. Third, a current-analysis-based method is developed to extract the disturbance only caused by flux linkage mismatch from the overall disturbance. Third, the extracted disturbance is adopted to calculate the real flux linkage, achieving demagnetization fault judgment and demagnetization degree calculation. Finally, experiment is conducted on two IPM motors to validate the proposed flux linkage estimation and fault diagnosis methods.