IET Power Electronics最新文献

This paper presents a non-invasive fault detection and diagnosis (FDD) methodology for permanent magnet synchronous machine (PMSM) drives, using low-frequency phase current signals. Specifically, this work focuses on the detection and diagnosis of power electronics-related inverter faults, which are a common source of system failures. The proposed framework introduces a pairwise feature fusion technique to enhance class separability and employs a three-stage selection process to distil a compact, discriminative feature set from Clarke-transformed current data. Diagnosis is performed by a hybrid machine learning model that ensembles the predictions of random forest, histogram-based gradient boosting, and k-nearest neighbours classifiers via a late-fusion strategy. The performance of the proposed method is evaluated on a publicly available experimental dataset containing nine operational states (one healthy and eight distinct inverter faults). The proposed method achieves an overall accuracy of 93.3% and a macro F1-score of 95.91%. The results demonstrate that the proposed approach can accurately diagnose multiple inverter faults without requiring high-frequency data acquisition or additional sensors, offering a cost-effective solution for enhancing the reliability of PMSM drives.

While most grid-tied inverters are grid-forming (GFL), it is foreseeable from the current development of grid-following (GFM) techniques that the modern grid will face a mix of both GFL and GFM inverters. This work addresses coordinated secondary control of an islanded microgrid that contain both GFM and GFL inverters; such a system possesses complex dynamics due to its different operating characteristics. A unified control framework is developed that coordinates both GFM and GFL, and a set of distributed secondary control methods is proposed to provide secondary restoration of the system's operating frequency and voltage and to enforce proportional sharing of active and reactive power across all GFM and GFL. Compared with the conventional control that focuses on the steady state, the proposed method explicitly bounds over the dynamics of power fluctuations induced by large load disturbance, which resulted in enhanced system dynamic performance. The scheme does not require precise system parameters, and could be applied to a wide range of grid topologies without additional communication needs. The Lyapunov stability of the controller has been proved, and its performance has been extensively analysed in both steady and dynamic states. A controller-hardware-in-the-loop testbed for an islanded microgrid with two GFL and two GFM inverters have been developed and the performance of the proposed coordinated control has been validated under various scenarios.

The flexible power point tracking (FPPT) strategy is designed to address frequency regulation resource shortages in modern power systems. However, it introduces power quality challenges in single-stage photovoltaic generation systems (PVPGS) operating on the left side of the P–V curve. Specifically, non-smooth AC power responses characterised by reverse power spikes and steady-state ripple occur during dynamic operation. This work develops a small-signal model of the PVPGS topology to identify stable operating regions through eigenvalue analysis, revealing an inverse power relationship between PV arrays and DC capacitors during left-side FPPT. Leveraging these insights, a parameter adaptation method is proposed to enhance AC output smoothness without hardware modifications or FPPT algorithm alterations. Simulation and experimental results verify the proposed method's effectiveness in suppressing AC power reverse spikes, demonstrating a minimum 40% reduction in peak reverse power magnitude and 59% attenuation in average spike amplitude while maintaining grid-compliant steady-state operation.

With the increasing penetration of power electronic devices in distribution networks, wideband harmonic resonance poses a serious threat to system stability. Conventional improvements based on a single-inverter control strategy are insufficient to achieve wideband suppression, while additional devices introduce extra cost and power losses. To overcome these limitations, this paper proposes a damping-reuse-based inverter topology reconfiguration method for wideband resonance suppression. First, a unified representation based on RLC branches is established. It reveals the general influence of the auxiliary damping device on port impedance. Based on this insight, an impedance reshaping method using parallel RLC branches is proposed to enhance the port capacitance. However, passive damping devices inherently introduce energy dissipation. To address this limitation, an active module is designed to achieve an equivalent port impedance. The proposed module is integrated into existing inverter structures. Topology reconfiguration and parameter design are then performed. The proposed method effectively suppresses multimode resonance peaks. It also mitigates newly introduced resonance points associated with the auxiliary branches. As a result, wideband harmonic resonance is effectively suppressed over the entire frequency range. Theoretical analysis shows that the proposed method effectively disrupts the original resonance network. It significantly enhances system damping. Meanwhile, it avoids additional hardware cost and active power losses. Simulation and experimental results further confirm the effectiveness and general applicability of the proposed method for wideband harmonic resonance suppression.

In this article, a novel active power filtering (APF) control method is proposed to suppress ripple power in the DC-link of railway traction systems, including both the inherent second-order ripple of single-phase rectifiers and additional frequency ripples introduced by the pantograph–catenary arc. First, according to the transmission process of ripple power in the traction system, a ripple power decoupling model is established. Then, the model predictive control (MPC) scheme of the APF circuit with low switching frequency is introduced to primarily suppress the second-order ripple voltage. Furthermore, to enhance steady-state performance, these residual ripple voltages on the DC-link are suppressed by compensating the capacitor current reference. Finally, experimental results demonstrate that the proposed control method has higher steady-state accuracy and faster dynamic response compared to the traditional control method at low switching frequency, making it suitable for railway traction systems.

In this paper, a modified isolated Z-source AC–AC converter with inherent commutation and reduced polarity changer switches is proposed. This converter enables flexible output voltage adjustment, determined by two degrees of freedom — namely, the high-frequency transformer turns ratio (n) and the duty cycle (D). Its circuit structure comprises four unidirectional switches — two high-frequency switches responsible for adjusting the output voltage amplitude and two low-frequency polarity changer switches to provide output states in inverting and non-inverting configurations relative to the input. One of the advantages of the proposed converter is its ability to resolve the commutation challenge without the need for snubber circuits, dedicated safe commutation techniques with intentional dead-times implementation. In addition, it provides inherent commutation advantages without requiring the circuit to be separated into positive and negative switching units or incorporating series diodes at the input side. As a result, the converter maintains a simplified switching pattern and control while reducing the total number of circuit components. Additionally, all magnetic components remain active throughout each half-cycle. The converter also maintains a continuous input current waveform and allows step-changes in output frequency. Moreover, since only one high-frequency switch operates per half-cycle, switching losses are minimised, leading to improved overall efficiency compared to conventional converters. In terms of switch voltage and current stress, switching device power and power density, the proposed converter is cost-effective and lightweight. After explaining the operating principle, a laboratory prototype with 110 Vrms is developed and evaluated under three frequencies — 25, 50 and 100 Hz — to validate the theoretical relationships.

Switching devices, the most vulnerable component within converters, underscores the critical need to enhance their reliability and prolong the power converter's lifetime. This paper proposes a multi-state reliability model for insulated gate bipolar transistors (IGBTs) that departs from conventional two-stage models with constant failure rate. Moreover, it quantifies the effect of DC-link voltage on IGBT reliability for switching frequencies below 5 kHz and derives an operating voltage ratio envelope of about 60% of the rated voltage. Operating within this band maximises lifetime and reduces maintenance cost, providing a practical voltage reference for control strategies. Finally, Monte Carlo simulations across multiple cases verify the feasibility and robustness of the proposed model and comparative hardware experiments support the underlying thermal assumptions of the proposed model.

The parasitic inductances of power loops in Silicon Carbide (SiC) power modules are critical parameters affecting dynamic current sharing, and their model can provide theoretical guidance for the design of dynamic current balancing. However, the coupled parasitic inductance matrix (CPIM) involved in the traditional model contains self-inductances and complex coupled mutual inductances, which hinder direct quantitative evaluation of the parasitic inductance differences in the power loops of paralleled chips. Based on the circuit equivalence principle, this paper proposes a decoupling calculation method for the CPIM, which realises the solution and modelling of equivalent parasitic inductances (EPIs) by matrix diagonalisation. Combined with the switching states of chips, the current distribution characteristics in dynamic current sharing are clarified. Then, according to the concepts of partial self- and mutual inductance, the coupled parasitic inductance network model (CPINM) is developed. Based on the identical V–I characteristics of model circuits before and after decoupling, the CPIM is diagonalised to calculate the EPIs. Finally, the EPI models for actual 4-chip and 6-chip paralleled power modules are developed, and the accuracy of the models is verified by theoretical and experimental analysis.

The virtual synchronous generator (VSG), as a representative grid-forming control method, has become a key technology in distributed renewable energy systems. However, there exists an inherent trade-off between fast power reference tracking and high virtual inertia support in conventional VSG control. To solve it, this paper proposes a novel coordinated control strategy, whose first step is a power reference splitting and feedforward module. The key idea is to split the power reference into high-frequency and low-frequency components, where the high-frequency component can be fed forward to the current loop reference of the VSG control, thereby improving the dynamic response. Meanwhile, the low-frequency component serves as the mechanical power to the VSG swing equation, ensuring adequate virtual inertia support without compromising the power tracking speed. However, the direct injection of the high-frequency component into the point of common coupling causes a disturbance to the electromagnetic power of the VSG. To mitigate this disturbance, a high-frequency power compensation module is further proposed. The entire coordinated control strategy, comprising the above two control modules, is seamlessly compatible with conventional VSG control circuits, enabling a fast power reference response in grid-connected mode without accurate line impedance knowledge while ensuring sufficient inertia support in standalone or weak grid conditions. Finally, simulation and experimental results validate the effectiveness of the proposed control strategy.

This paper presents an optimization-free PWM control method for a single-phase 9-level flying-capacitor (FC) multicell active neutral-point-clamped (A-NPC) inverter. The controller measures the output current and the FC voltages and compares them to their references, then converts the comparison results into logic variables. These variables drive a set of logic equations that simultaneously (i) regulate each FC voltage to its target value and (ii) select the appropriate switching state to synthesize the requested multilevel output through PWM. Because the gating decisions are produced by direct logical evaluation—without cost functions, iterative search, or computationally heavy optimization—the method is fast and simple to implement. Experimental results verify reliable voltage balancing and proper operation during abrupt changes in DC-link voltage, modulation index, and output frequency.