IET Power Electronics最新文献

This paper presents a novel design approach for an LLC resonant DC-DC converter tailored for electric vehicle battery chargers. The design methodology considers the non-linear load, ensuring a narrow switching frequency range and zero-voltage-switching (ZVS) conditions for optimal soft-switching. A hybrid design procedure is proposed with a strong emphasis on high efficiency and power density. The resonance frequency and an accurate range for the magnetising inductor are calculated by time-domain analysis using the efficiency formula. The resonant capacitor, magnetising inductor and resonant inductor are then calculated by frequency-domain analysis, considering ZVS conditions and the non-linear load's profile. In the time-domain analysis, a 3D surface map of efficiency versus resonance frequency and magnetising inductor provides an accurate resonance frequency and the range for the magnetising inductor. Furthermore, by drawing a surface map of load considerations versus the calculated resonance frequency and the defined magnetising inductance range, based on the load's profile and ZVS conditions, the optimal design parameters are identified by applying the maximum switching frequency and inductance ratio. Finally, the step-by-step design methodology is validated through experiments on a prototype converting 400 V from the input to an output voltage range of 320 to 420 V at 3.3 kW with a peak efficiency of 97.9%.

Boost converters are one of the most commonly used converters in power electronics due to their advantages, such as simple structure and low-cost implementation. Due to these issues, many papers have been published about the control method of this topology. This paper proposes a reinforcement learning (RL) controller based on the soft actor-critic (SAC) algorithm for boost converters. This approach aims at decreasing the limitation of conventional controllers, such as long settling times, overshoot, undershoot and instability. It also removes the need for a system model, which is often necessary in traditional control methods. Furthermore, the proposed method addresses the challenges of control under dynamic conditions, which is a problem that has not been sufficiently explored in prior reinforcement learning (RL) based controllers. The proposed approach has demonstrated promising performance in both steady-state and dynamic regulation across a wide range of operational variations. Simulation and experimental results show the superiority of the SAC algorithm over the conventional ones without relying on the system model.

The integrated-gate commutated thyristor (IGCT), known for its low conduction losses and high reliability, is increasingly favoured in high-voltage and high-power applications. The junction temperature is a critical determinant of reliability, making the rapid and precise analysis of the transient temperature field indispensable. An analytical-numerical hybrid solution for the transient thermal analysis of the outer gate IGCT considering electrical-thermal coupling is proposed. Validation against FEA confirms the accuracy of the proposed analytical solution, while its computation time is reduced by 94.9%, which underscores its superiority in computational efficiency. Based on this, the proposed analytical-numerical hybrid solution is integrated into the VSC-HVDC simulation model, utilising the calculated maximum junction temperature feedback to optimise the switching frequency of VSC. Simulation results demonstrate that controlling the switching frequency based on the maximum temperature calculated can reduce the device's maximum junction temperature during VSC overload conditions, thereby enhancing the overload capacity of the converter. The analytical-numerical hybrid solution proposed in this article can be used for system simulation and has the potential to be applied to active thermal management of converters.

Current source inverters (CSIs) offer an alternative to voltage source inverters (VSIs) for electric motor applications. CSIs have some advantageous characteristics, such as the ability to inherently boost voltage, the absence of electrolytic capacitors, and reduced voltage stress on the machine. However, the increased component count and the need for a pre-stage to control the inductor current have limited the adoption of CSIs. For applications that must operate above a base speed, such as more electric aircraft (MEA), single-stage CSI converters are a promising solution. The ability to directly measure the terminal voltages allows for a simple and robust sensorless control approach to be used with this power converter. This approach based on a phase-locked loop (PLL) directly connected to the voltages with a feedforward compensation that depends on voltage drops on Rs and L0 is compared against a back-EMF (BEMF) observer. If a machine with an anisotropic rotor is used and the impedance is small, the feedforward compensation can be neglected leading to the classical PLL. Experimental results on a CSI7 converter driving a permanent magnet synchronous machine (PMSM) report that this kind of sensorless control can be a suitable alternative in motor drive applications.

An LLC resonant converter employing a novel reconfigurable rectifier (RR) for applications requiring a wide output voltage range is proposed in this paper. By integrating multiple rectifier structures—such as centre-tapped, full-bridge, and voltage-doubler rectifiers—into a single reconfigurable unit, the proposed RR can dynamically switch among different operating modes according to the output voltage requirement. As a result, the output voltage range is extended by a factor of 2 to 4 compared to conventional LLC resonant converters. Consequently, the normalised voltage gain range of the resonant tank is significantly narrowed, leading to reduced conduction and switching losses and improved overall conversion efficiency. The operational principles and mode-switching control strategy of the converter are analysed in detail. A 500 W prototype was built and tested to verify the effectiveness and feasibility of the proposed method.

Permanent magnet synchronous motor (PMSM) is prone to irreversible demagnetisation faults in complex operating conditions, threatening system reliability. To address this, a novel fault-tolerant control strategy is proposed, with its key contributions being: First, the introduction of a nonsingular fast terminal super-twisting sliding mode observer (NFTSTSMO) for high-fidelity, real-time flux linkage observation and demagnetisation fault reconstruction; Second, the development of a dual-vector finite control set model predictive fault-tolerant control (DV-FCS-MPFTC) scheme that explicitly incorporates the observed flux linkage to compensate for prediction errors. Experimental results confirm that the proposed method significantly enhances the fault tolerance and robustness of PMSMs under demagnetisation conditions, outperforming conventional approaches.

Permanent magnet synchronous motor (PMSM) is prone to irreversible demagnetisation faults in complex operating conditions, threatening system reliability. To address this, a novel fault-tolerant control strategy is proposed, with its key contributions being: First, the introduction of a nonsingular fast terminal super-twisting sliding mode observer (NFTSTSMO) for high-fidelity, real-time flux linkage observation and demagnetisation fault reconstruction; Second, the development of a dual-vector finite control set model predictive fault-tolerant control (DV-FCS-MPFTC) scheme that explicitly incorporates the observed flux linkage to compensate for prediction errors. Experimental results confirm that the proposed method significantly enhances the fault tolerance and robustness of PMSMs under demagnetisation conditions, outperforming conventional approaches.

Grid-connected inverters (GCIs) operating in grid-following (GFL) control may suffer from instability when connected to a weak grid characterised by a low short-circuit ratio (SCR). Conversely, GCIs operating in grid-forming (GFM) control may be unstable under a strong grid. Furthermore, there is a certain contradiction between the stability and performance of the GCIs. To improve the dynamic performance of the GCIs while ensuring stability, a neural network-based multiple-model adaptive control (MMAC) strategy is proposed in this paper. First, a small-signal stability analysis is conducted for different control modes, and a control-switching scheme is introduced to ensure stable GCI operation across a wide SCR range. The proposed method employs the virtual dynamic short-circuit ratio for GCI control design, enabling a unified stability margin while maximising dynamic performance. Then, the online application of multiple-model adaptive control is realised by the neural network, which can adapt to the fluctuations of grid impedance and power. The experimental results verify the effectiveness of the proposed method.

Grid-connected inverters (GCIs) operating in grid-following (GFL) control may suffer from instability when connected to a weak grid characterised by a low short-circuit ratio (SCR). Conversely, GCIs operating in grid-forming (GFM) control may be unstable under a strong grid. Furthermore, there is a certain contradiction between the stability and performance of the GCIs. To improve the dynamic performance of the GCIs while ensuring stability, a neural network-based multiple-model adaptive control (MMAC) strategy is proposed in this paper. First, a small-signal stability analysis is conducted for different control modes, and a control-switching scheme is introduced to ensure stable GCI operation across a wide SCR range. The proposed method employs the virtual dynamic short-circuit ratio for GCI control design, enabling a unified stability margin while maximising dynamic performance. Then, the online application of multiple-model adaptive control is realised by the neural network, which can adapt to the fluctuations of grid impedance and power. The experimental results verify the effectiveness of the proposed method.

In environments with numerous converters, there is a need for highly efficient. The LLC resonant converter is a good candidate for microgrid applications, as it features low switching losses and a wide voltage regulation range. However, designing magnetic components with lower losses remains a challenge for designers. This paper presents a methodology for reducing losses in the LLC resonant converter by integrating the tank design with the construction design of the transformer and resonant inductor. An algorithm generates a list of feasible components, filtered by constraints on flux density, core temperature, inductance, and winding resistance. The filters select resonant tank element designs that operate at the ideal core temperature, with fewer turns, and that limit current circulation. The tank elements were constructed for validation through experimental results on a 1 kW prototype. Experimental results demonstrated a peak efficiency of 97.48% under ideal thermal conditions. Losses in the resonant tank were reduced by 6.6% compared to ambient operation. Further efficiency (up to 97.52%) was achieved by increasing the magnetizing inductance. The results confirm that the integration of the magnetic design with the design of the resonant tank allows the reduction of losses in LLC converters.