The uninterruptible power supply system plays a crucial role in ensuring the continuity of power supply for critical loads. However, during situations such as power outage, restoration, or bypass switching, the traditional mode switching mechanism of the uninterruptible power supply often leads to voltage and frequency fluctuations, and even temporary power interruptions, which may cause potential damage to sensitive electronic equipment such as servers, medical devices, and precision instruments, affecting their stable operation and data integrity. To address this technical challenge, this paper proposes a new smooth mode switching strategy for uninterruptible power supply systems. This research mainly focuses on online uninterruptible power supply systems and studies the key technologies for achieving seamless and smooth mode switching in the event of power supply fluctuations. A method for feedforward control of the uninterruptible power supply output voltage is proposed, enabling the output voltage of the uninterruptible power supply to quickly return to stability after power supply fluctuations. A phase-locked control method for the uninterruptible power supply output voltage is also proposed to ensure the stability of the uninterruptible power supply during power supply fluctuations. Experimental verification using actual hardware prototypes shows that the proposed smooth switching strategy can control the amplitude of voltage fluctuations during the switching process within 2

Large-capacity electrochemical energy storage power stations have experienced multiple power oscillation incidents caused by the failure of multi-controller coordination during grid-connected operation. Based on the analysis of two typical power fluctuation events in energy storage power stations, this paper establishes a formal model of the original control logic on site and extracts a four-link closed-loop oscillation mechanism: control authority switching, then power step, then electrical quantity disturbance, and finally triggering misjudgment. On this basis, a hierarchical decoupling control method based on a finite state machine is proposed: a priority interlocking state machine is designed at the discrete layer to eliminate command conflicts, and a frequency-adaptive smoothing algorithm is constructed at the continuous layer to block the power-frequency coupling. Simulation results show that the proposed method effectively suppresses oscillations in both types of scenarios, and cross-scenario validation demonstrates the universality of the method.

Silicon carbide (SiC) power modules, with its advantages of high-voltage tolerance, high-temperature operation capability, high-switching frequency, and low power loss, are progressively replacing traditional silicon-based devices in applications such as new energy vehicles, renewable energy generation, and rail transportation. However, as the voltage and current ratings of SiC power modules increase, their power density escalates significantly, exacerbating thermal nonuniformity within the modules. This article proposes a regionalized multiobjective optimization method that explicitly addresses thermal nonuniformity, alongside module thermal resistance and pumping power. The methodology integrates a genetic algorithm with finite element analysis. Using the Pin-Fin heatsink of an automotive power module as a case study, an optimized design was developed and validated through a back-to-back test platform. Experimental results show that, compared to the traditional Pin-Fin heatsink, the optimized design achieves a 43.7% improvement in thermal uniformity, a 6.2% reduction in thermal resistance, and a 9.8% decrease in pumping power-without added costs.

New energy delivery system based on half-bridge sub-module MMC has a broad development prospect. Existing half-bridge MMCs need to be additionally fitted with high-capacity DC circuit breakers and energy dissipation devices to cope with system overvoltage and overcurrent challenges caused by AC and DC faults, but the DC circuit breakers and energy dissipation devices contain hundreds of classes of expensive power electronic switches, which limits their large-scale applications. To address the above problems, this paper proposes a scheme to integrate DC circuit breakers and energy dissipation devices to achieve both DC on-off and resistance casting functions under faults by multiplexing the power electrical switches, which results in significant savings in device costs. The working principle and parameter design method of the proposed scheme are further analysed, and a system simulation model is built to verify the effectiveness of the proposed scheme. Finally, the techno-economic analysis of the proposed scheme is carried out, and the cost is reduced by more than 33

To overcome the limitations of conventional fault ride-through (FRT) schemes, the energy-dissipation MMC scheme has been proposed. By integrating energy dissipation units, this scheme possesses energy balancing capability and demonstrates promising application prospects. To address the challenge of topology selection arising from the diverse placement positions of energy dissipation units, this paper proposes a multi-dimensional evaluation model for the energy-dissipation MMC covering three dimensions: performance, cost, and structure. This model adopts a linear weighting method to construct a comprehensive score and sets weight coefficients under three typical scenarios based on the Pareto optimality principle, thereby providing a quantitative decision-making basis for topology selection of the energy-dissipation MMCs in different engineering scenarios.

Advanced assess and test close to real field operation are of great significance to ensure the reliability of modular multilevel converter (MMC). Thus, mission profile emulator (MPE) is proposed as an economical and efficient testing scheme, which aims to create an emulated mission profile for submodules of MMCs with simplified testing circuits. A challenge of MPE is to create a testing current that has as small level of current ripple as the arm current in MMC. To limit the current ripple of the testing circuit, the existing methods try to increase the filter inductance of the testing circuit. However, when the filter inductance increases to a certain level, the influence from the increase of filter inductance will be offset by the increase of DC supply voltage, and the current ripple will not be reduced anymore. As a result, a new MPE with a resonant filter is proposed to further reduce the current ripple; meanwhile, a control strategy and a design method are proposed to cope with the resonant filter and achieve higher emulation accuracy. Simulation and experimental results are presented for the validation of the proposed method.

Modular multilevel converters (MMCs), with IGBTs as core switching elements, are crucial for grid-forming in new energy systems. However, their limited overcurrent-carrying capability constrains the regulating effect of grid-forming control on system stability. Despite IGBTs' theoretical high current turn-off capability, their sharply increased conduction losses under overcurrent conditions frequently lead to excessive temperature rise, thus limiting their current switching potential and becoming a critical bottleneck for MMC overcurrent capability. To address this, a novel thyristor-enhanced semiconductor switch is proposed. This solution effectively reduces the IGBT's conduction losses and extends its overcurrent operating boundary by utilizing a bypass thyristor to carry the current during overcurrent conditions. Furthermore, by introducing a passive auxiliary commutation (PAC) module, reliable current transfer from the thyristor back to the IGBT is achieved, successfully addressing the inherent challenge of the thyristor's inability to actively turn off current. Multiple 1.6 kA-10.7 kA double-pulse experiments demonstrate the proposed switch's ability to perform switching operations across a wide current range and achieve 10.7 kA current transfer and turn-off. Piecewise linear electrical circuit simulation (PLECS) simulations confirm the thyristor-enhanced switch's effectiveness in enhancing the transient overcurrent capability of MMCs: at 2 p.u. overcurrent, converter valve duration extends from 66.6 ms to 4.64 s. The cost curves for the thyristor-enhanced semiconductor switch (TES) switching scheme and the IGBT parallel-connection scheme as a function of overload duration under a 2 p.u. condition are provided. The results demonstrate that the proposed TES scheme possesses superior economic efficiency.

In recent years, with the rapid development of the new energy industry, the scale of new energy storage facilities centered around lithium-ion batteries has continued to climb. Additionally, there have been several significant safety incidents in energy storage power plants across the globe in recent years, raising concerns for the industry. It is worth noting that the widely used traditional Battery Management Systems (BMS) can only detect the structural integrity and physical parameter abnormalities of the battery, which has obvious monitoring limitations and cannot detect condensation inside of the battery pack. In this paper, we propose an Edge Aware Instance Segmentation Network (EAIS-Net) based on the visual features of condensation inside of battery packs. Specifically, the proposed EAIS-Net is used to enhance the perception ability of condensation phenomenon in battery images, and its core components is the Edge Perception Module (EPM). EPM is committed to enhancing the blurred edge structure of condensation on the surface of battery cell caused by factors such as light exposure and scattering, highlighting the edge characteristics of condensation. The proposed algorithm can provide early warning for energy storage power plants, and experimental results show that the proposed network is superior to other advanced algorithms.

Medium voltage silicon carbide (SiC) power semiconductor modules with excellent electrothermal properties offer novel opportunities for revolutionizing high-power electronic converters and systems. Based on the requirement for MV high-capacity power modules from offshore wind power flexible high voltage direct current (HVDC) transmission systems, this article successfully developed a high-density 3.3 kV/2000 A SiC mosfet power module by paralleling 36 chips, which is the world's largest current capacity and highest power rating SiC module reported to date. First, the package design of this 3.3 kV/2000 A SiC power module is elaborated, including the parallel chip number selection, package architecture optimization, and material system design. Second, the critical electromagnetic and thermal performance of the SiC power module is evaluated in detail through simulation. Third, the key fabrication processes for the SiC power module are demonstrated, including the initial bare chip prescreening stage. Finally, comprehensive experimental evaluations were conducted on the fabricated SiC power modules. Compared to the widely deployed 3.3 kV/1500 A Si IGBT power module in flexible HVDC converter valves, the developed 3.3 kV/2000 A SiC power module exhibits significant advancements in switching frequency, power loss, and power density, holding promising potential in driving the transition of flexible HVDC converter valves toward compactness, lightweight, and high efficiency.

The renewable energy voltage sourced converter based high-voltage direct current (VSC-HVDC) system has a broad application prospect, and the ac chopper (ACC) is the core equipment to realise the system fault ride-through (FRT). The traditional ACC consists of multiple three-phase energy-dissipation valves, which require a large number of thyristors, cover a large area, and are difficult to accurately match dissipated power, which restricts its further promotion and application. In this article, the flexible thyristor switch modules based ac chopper (FTSM-ACC) is proposed, which adopts diode bridges for rectification and combines with thyristor energy-dissipation switches to reduce the cost and volume while realizing the flexible casting and cutting of the FTSM-ACC, which provides a new way of thinking for the improvement of the technological and economic performance of ACC. Further, the FTSM-ACC control strategy and key parameter coordination method are proposed. And the system simulation is carried out to analyze the system FRT characteristics of FTSM-ACC. On this basis, a 4-kV/3.2 MW FTSM-ACC prototype is built, and the experimental results verify the effectiveness of FTSM-ACC and demonstrate its advantages over traditional ACC. Finally, the performance and economic efficiency of different ACCs are compared. Compared with the traditional ACC, the FTSM-ACC improves the performance by more than 50%, and the total cost and volume are reduced by 33% and 46%, which has a great performance and economic advantage.

Abstract With the rapid growth of lithium-ion battery applications in new energy storage systems, safety incidents caused by internal faults have become a critical concern. Traditional battery management systems monitor electrical parameters, making it difficult to detect early mechanical and environmental anomalies. To address this gap, this paper proposes an machine vision-based abnormal visual identification method inside battery packs. First, a physical detection platform is built, consisting of a robotic arm and visual sensor. The robotic arm can be controlled to extend into the semi-enclosed battery pack through inspection ports, enabling non-destructive automated visual anomaly detection. Second, a Soft-Gated Cross-scale Fusion Module (SGCFM) is designed to optimize feature fusion in YOLOv8. SGCFM uses local and non-local branches to generate adaptive weights, dynamically re-weighting neck network outputs to enhance small target and large target representation. Experiments on a self-constructed dataset show the proposed method achieves 96.9% mAP50 and 66.7% mAP50-90 for screw loosening, 84.3% mAP50 mask and 55.1% mAP50-95 mask for condensation.

In recent years, with the rapid development of renewable energy in China, the installed capacity of new energy storage systems—particularly lithium-ion batteries—has grown significantly. However, several major safety incidents at energy storage stations worldwide have raised serious concerns about the reliability and safety of these systems. Traditional Battery Management Systems (BMS) are primarily designed to monitor structural and physical defects within the battery, but they are incapable of detecting externally invisible faults such as loose screws—an issue that often serves as an early indicator of battery failure. To address this limitation, this paper proposes a visual detection algorithm for identifying loose screw faults inside lithium-ion energy storage battery packs. The algorithm leverages deep learning and image processing techniques, employing the YOLOv8 object detection framework. Enhancements are made to both the backbone and neck of the network, and attention mechanisms are introduced to improve the recognition of small-sized objects such as screws. Experimental results demonstrate that the algorithm achieves a detection accuracy of up to 95.6%, enabling automatic identification of screw loosening and providing early fault warnings. This method offers a promising solution for intelligent operation and maintenance of energy storage stations.

The integrated energy-consuming MMC (IEC-MMC) plays a critical role in dissipating surplus power generated during the low-voltage ride-through process in offshore wind systems to achieve low-voltage fault ride-through. As the core equipment of the converter station, the converter valve must undergo a complete type test to ensure normal operation and successful fault crossing. However, existing research has primarily focused on theoretical analysis and simulation of the energy dissipation process in the IEC-MMC, with limited investigation into equivalent testing methods for evaluating its fault ride-through performance at the module level. To address this research gap, this paper proposes an equivalent assessment index covering the entire fault ride-through process, taking the integrated energy-consuming MMC valve as the research object, and resolves the problem of the lack of assessment index for the fault ride-through equivalent test. Furthermore, an equivalent test circuit with integrated energy dissipation and a corresponding test methodology are proposed, offering a practical approach for replicating the stress variations experienced by the valve during fault ride-through. At the same time, it also provides an important reference value for the type test of the converter valve.

This study addresses the issue of inadequate safety-monitoring methods for lithium-ion battery energy storage systems. An image recognition approach based on a single-shot multibox detector (SSD) algorithm is proposed for detecting gas-liquid emissions within electrochemical energy storage compartments. An experimental platform is developed to simulate the actual operating conditions of lithium-ion battery storage units, and a dataset is constructed from the image data capturing gas-liquid emissions during the overcharging stage. To overcome the limitation of the original SSD algorithm, which features an excessively large model scale that restricts real-time detection, several modifications were implemented: the Visual Geometry Group (VGG) backbone is replaced with MobileNet-V3 to enhance computational efficiency; the squeeze-and-excitation (SE) attention module is substituted with the Coordinate Attention (CA) module to enhance feature extraction capabilities; and mean clustering optimization is applied to refine the default box scale sizes based on the dataset. The experimental results show that the improved SSD model achieves a 92.2% reduction in model size (from 91.9 MB to 7.2 MB), with a 1.54% increase in average accuracy (from 90.38% to 91.92%). The prediction speed increased from 15 to 58 frames per second (FPS), meeting the real-time detection requirements for lithium-ion battery energy storage compartments.

Even contact pressure distribution among submodules in a press-pack insulated gate bipolar transistor (PP-IGBT) is an important factor in reliability screening tests before engineering application. However, the current stress-sensitive method for contact pressure measurement is an invasive method where each measurement would change the contact pressure distribution of submodules. It is necessary to study a noninvasive measurement method for contact pressure distribution in PP-IGBT. In this article, a noninvasive measurement method based on the ultrasonic reflection coefficient is proposed to measure the contact pressure distribution within PP-IGBTs. First, the characteristic of ultrasonic wave propagation at the contact interface of two different materials is analyzed. The ultrasonic measuring platform for the contact pressure distribution in PP-IGBTs is designed, and the new assembly measured part has little effects on the pressure distribution within the device by different clamping force experiment. Second, an efficient contact pressure measurement system for PP-IGBTs is designed, and a calibration method of ultrasonic reflection coefficient for contact pressure measurement is proposed. Finally, the accuracy of the proposed ultrasonic measurement method for contact pressure distribution is verified by comparing the measurement results with those of the stress-sensitive film method, where the relative error of contact pressure within PP-IGBTs in three different conditions is less than 10%.

To address the issues of low capacity and unstable output voltage in existing Uninterruptible Power Supply (UPS) systems, a phase control method for UPS output voltage with a bypass mode is proposed. A grid-following-like control strategy (input voltage feedforward control) is adopted when the UPS operates in bypass or normal mode, while a grid-forming-like control strategy (output voltage phase-locked control) is used during input voltage fluctuations or interruptions, thereby stabilizing the output voltage of various operating conditions of the UPS.

Accurate and prompt diagnosis of internal short circuits at an early stage is critical for preventing severe safety incidents and ensuring the reliability and safety of lithium-ion batteries. However, existing early-stage internal short-circuit diagnosis methods often rely heavily on high-precision battery models and large volumes of high-quality labeled training data, limiting their practicality and robustness in real-world applications. To address these limitations, this paper proposes a novel method for early detection and quantitative assessment of internal short circuits in lithium-ion battery packs, based on differential voltage (DV) analysis and Mahalanobis distance. The proposed approach extracts a median DV curve from the sorted terminal voltages of individual cells within a battery pack, which serves as a reference to characterize the normal cell behavior. The Mahalanobis distance between each cell's DV curve and the reference curve is then calculated and compared against a threshold to distinguish short-circuited cells from healthy ones. For the identified faulty cells, the short-circuit current and resistance are estimated by analyzing the differences between charging voltage curves across adjacent cycles, enabling precise quantification of fault severity. Experimental validation is conducted using simulated internal short circuits with varying severities. Results show that the proposed method can accurately detect shortcircuited cells when the short-circuit resistance is less than or equal to 300 Omega. The maximum and minimum relative errors of short-circuit resistance estimation are 5.21 % and 1.20 %, respectively, demonstrating the effectiveness and accuracy of the proposed method.

The safety of energy storage systems relies heavily on thermal runaway early warning protection and cooling intervention of fire extinguishing agents for large-capacity lithium iron phosphate batteries (LiFePO4). An integrated platform was established to trigger thermal runaway fires through heating abuse and coordinate with extinguishing agents at the cell/module level. Key indicators of the battery thermal runaway evolution process were constructed, identifying 16 multidimensional signal characteristics representing thermal runaway failure, with the hydrogen gas production rate and contribution ratio being highlighted as critical early warning signals. Based on the maximum voltage drop rate and duration of soft short circuits, the cooling effects of perfluorohexanone-based extinguishing agents were explored to reduce the risk index of thermal runaway explosion to the lowest level. Through gas concentration tests under various conditions, a three-level early warning mechanism for thermal runaway was established, setting thresholds for hydrogen, volatile organic compound, carbon dioxide, smoke, and gas temperature. The establishment of warning indicators and mechanisms is crucial for improving the early warning and protection strategies of energy storage systems against thermal runaway.

A three-step etched ultrahigh-voltage 4H-SiC drift step recovery diode (DSRD) with the record high length utilization efficiency is fabricated and measured. The depth-optimized junction termination extension (JTE) obtains a 10 kV-class breakdown voltage without ion implantation. The process of the termination based on a thick p-base region was investigated. A reverse blocking voltage of 9.95 kV at a leakage current density of 2.1 µA/cm2 is achieved. The JTE length efficiency achieves 83 V/µm, which is approximately 66% higher than that of other terminations with the same voltage level of 10 kV (~50 V/µm).

In silicon carbide (SiC) multichip power modules, chip parameter dispersion is the critical factor causing parallel current sharing imbalance. This paper presents for the first time the characterization and analysis of the parameter dispersion in medium voltage (MV) SiC MOSFETs. First, the characterization platform at the bare die level and probing test considerations are elaborated. Second, the key static characteristics of 30 state-of-the-art 3.3 kV SiC MOSFET chips are measured across a wide junction temperature range (25 degrees C to 125 degrees C). Then, three novel indicators are proposed to indirectly reflect the high voltage and high current region IV characteristics and SiO2/SiC interface properties of SiC MOSFETs. Finally, the dispersion of key static parameters of 3.3 kV SiC MOSFETs is statistically evaluated and compared, providing valuable reference for chip pre-screening of MV SiC multichip power modules.