20 October 2026, Volume 50 Issue 10

    

• Select all
  |

Review
• Select
  Study on multi-level safety standard system for electrochemical energy-storage batteries

  CHEN Wentao, WEI Zhen, LI Qian, LIU Shiqiang

  Chinese Journal of Power Sources. 2026, 50(10): 1893-1901. https://doi.org/10.3969/j.issn.1002-087X.2026.10.001

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  With the accelerated global transition toward a renewable-dominated energy system, electrochemical energy storage has become a critical infrastructure supporting the flexibility, stability, and high-quality operation of new-type power systems.Lithium-ion batteries, represented by LiFePO₄ and ternary chemistries, dominate large-scale energy storage applications due to their high energy density, mature industrial foundation, and rapid cost reduction. However, a series of large-scale energy storage fires and explosions have exposed significant weaknesses in thermal management, system integration, operational control, and safety standards. This review systematically analyzed the multi-level coupling mechanism of thermal runaway in energy-storage lithium-ion batteries-from materials, electrodes, cells, and modules to system-level and summarized key triggers such as internal short circuits, overcharge abuse, cooling failure, cell inconsistency, and insufficient environmental tolerance. Typical accidents were examined to reveal chain-reaction propagation pathways. Furthermore, a comparative assessment of the safety standards in China, the United States, and Europe identifies that China has established relatively comprehensive cell-level testing requirements, whereas system-level thermal propagation evaluation, scenario-based environmental adaptability assessment, and lifecycle-oriented safety governance remain to be strengthened. Based on these findings, this paper proposed several directions for improving storage-system safety, including system-level thermal propagation and full-scale fire tests, multi-climate adaptability standards, AI- and digital-twin-enabled proactive safety management, improved operation-and-maintenance and decommissioning protocols, and enhanced international standard coordination. These measures are expected to support the high-safety and high-quality development of China’s energy storage industry.
• Select
  Progress of sulfide solid-state battery technology

  LIU Jia

  Chinese Journal of Power Sources. 2026, 50(10): 1902-1914. https://doi.org/10.3969/j.issn.1002-087X.2026.10.002

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  Sulfide-based all-solid-state batteries (ASSBs) have emerged as a critical candidate for next-generation energy storage technologies with high energy density and superior safety, owing to their high ionic conductivity, favorable processability, and wide temperature adaptability. This paper systematically reviewed the research progress and existing challenges in material systems, cell structures, and fabrication processes of sulfide ASSBs. It emphatically analyzed key technological breakthroughs including the improvement of air and electrochemical stability of sulfide electrolytes, suppression of volume expansion effects of anodes, and regulation of interfacial kinetics of cathodes. Meanwhile, the contributions of bipolar stacking configuration and dry electrode technology to the enhancement of energy density were discussed, highlighting that the independent research and development of advanced manufacturing processes and dedicated equipment is the core prerequisite for large-scale industrial production. Finally, future prospects were proposed from the perspectives of material-interface synergistic regulation, process and equipment innovation, and industrialization pathways. This review aims to provide theoretical support and practical guidance for fundamental research, technological iteration, and engineering application in the field of sulfide all-solid-state batteries.
• Select
  Research progress of carbon materials in sodium-ion capacitors

  LUO Yuanrong, CHUAN Xing, NIU Zhenyong, PENG Yiming, ZHANG Yang

  Chinese Journal of Power Sources. 2026, 50(10): 1915-1921. https://doi.org/10.3969/j.issn.1002-087X.2026.10.003

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  Carbon-based sodium-ion capacitors have gained considerable attention due to their combination of the high energy density of sodium-ion batteries and the high power characteristics of capacitors. The design and fabrication of low-cost, high-performance carbon-based electrode materials are key to advancing the further application of sodium-ion capacitors. Based on a summary of the energy storage mechanisms of sodium-ion capacitors and the requirements for electrode materials, this paper reviewed carbon-based materials and their modification methods, and discussed prospects for future development.
Research and design
• Select
  Lightweight SOH prediction framework for lithium-ion batteries based on multi-scale convolution and channel attention

  JIANG Ming, WANG Meng, LIU Mingyang, GAO Fei, SU Yuefeng

  Chinese Journal of Power Sources. 2026, 50(10): 1922-1932. https://doi.org/10.3969/j.issn.1002-087X.2026.10.004

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  Existing methods for lithium-ion battery state of health (SOH) prediction often struggle to capture multi-scale temporal degradation and lack refined channel feature discrimination. To bridge this gap, we proposed a lightweight framework integrating multi-scale convolution (MSConv) and efficient channel attention (ECA). First, health features such as constant current charging statistics, incremental capacity (IC) peaks, and charging time were extracted as model inputs. The MSConv-ECA architecture then leveraged parallel convolutions to extract multi-scale temporal features while adaptively capturing inter-channel correlations via ECA. This approach enhanced the representation of critical degradation information with minimal parameters. The experimental results on NASA and CALCE datasets demonstrate high precision, with root mean square errors (RMSE) of 0.716 1% and 1.259 7%, respectively. Compared with MSConv-Transformer models, the proposed framework reduces parameter count by 44.9% and training time by 38.3%. This study offers a high-accuracy, high-efficiency solution ideal for edge deployment in battery management systems.
• Select
  Research on online monitoring method of internal temperature of lithium battery pack cells based on PWM switching oscillations

  XIE Qiang, LI Hao, ZHAO Huaijian, XIANG Dawei

  Chinese Journal of Power Sources. 2026, 50(10): 1933-1940. https://doi.org/10.3969/j.issn.1002-087X.2026.10.005

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  Internal temperature is a crucial state variable reflecting the safe, efficient, and reliable operation of lithium batteries. To enable online, safe, and timely acquisition of internal temperature information for individual cells within a battery pack, this paper proposed a novel online monitoring method based on PWM switching oscillations. Initially, an RLC resonant network was artificially constructed for each battery cell. Subsequently, the self-excited PWM switching oscillation current was extracted online from the battery system’s power electronic converter in a contactless manner to estimate the high-frequency impedance of the lithium battery cells. Ultimately, based on the temperature-sensitive characteristic of the high-frequency impedance, the internal temperature of individual cells was continuously monitored online. Experiments on lithium battery pack cell temperature imbalance demonstrate that this method can reliably detect abnormal temperature deviations among lithium battery cells in real-time without affecting the normal operation of the system, providing support for lithium battery fault warning and protection.
• Select
  Effect of cycle aging on overcharge thermal runaway behavior of NCM622/graphite pouch lithium batteries

  WANG Haonan, YANG Ye, LU Suyang, LI Genpeng

  Chinese Journal of Power Sources. 2026, 50(10): 1941-1951. https://doi.org/10.3969/j.issn.1002-087X.2026.10.006

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  To investigate the effect of cycle aging on the overcharge safety of lithium-ion batteries, fresh 53 Ah nickel cobalt manganese oxide (NCM622)/graphite pouch cells were compared with those cycled 250 times. Their thermal runaway behavior and temperature evolution under 1 C overcharge were analyzed via charge-discharge tests, thermocouple temperature measurement, and COMSOL electro-thermal coupling simulation, with the overcharge process divided into four stages. The cycled cells trigger thermal runaway earlier, showing an average heating rate of 33 ℃/s during runaway (5 ℃/s for fresh cells) and a peak temperature approximately 450 ℃ higher. Cycle aging increases cell internal resistance by 35%, reduces solid electrolyte interphase (SEI) film stability, and causes cathode structure damage, which synergistically lower the thermal runaway trigger barrier and intensify heat generation. This work clarifies the influence of cycle aging on the heat generation mechanism at each overcharge stage, and the proposed safety optimization strategies provide a theoretical basis for the safety design of battery systems.
• Select
  Research on thermal runaway and gas production characteristics of 314 Ah lithium iron phosphate battery

  SONG Haoyu, YANG Kai, ZHANG Mingjie, CHEN Hao, WEI Bin

  Chinese Journal of Power Sources. 2026, 50(10): 1952-1960. https://doi.org/10.3969/j.issn.1002-087X.2026.10.007

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  To investigate the thermal runaway and gas production characteristics of large-format lithium iron phosphate (LFP) batteries, thermal runaway was induced in a 314 Ah LFP cell by overheating in both open and confined space. The thermal runaway behavior, thermal characteristics, and gas production were analyzed. The results show a distinct transition period between safety valve opening and violent gas ejection; without an external ignition source, no spontaneous combustion occurs. Upon ignition, a jet flame forms and is sustained for 585 s, with a maximum flame temperature of 993.4 ℃. Thermal runaway propagates from the local trigger point to the entire cell. The heated surface fails first, and heat transfers internally, leading to a higher temperature rise rate on the non-heated surface. The peak heat release rate (HRR) reaches 119.1 kW, and the total heat release (THR) is 12.483 MJ. The total gas production is approximately 182.88 L, primarily consisting of H₂, CO₂, CO and alkanes. The volume fractions of H₂ and CO₂ are 55.97% and 23.26%, respectively. The explosion limit range of the mixed gas is 5.02%~52.86%, indicating a high explosion risk upon encountering an ignition source in a confined space. These findings provide data to support thermal runaway prevention and fire safety design for large-format LFP batteries.
• Select
  Thermal analysis and optimization of serpentine flow channels for pure electric heavy-duty truck power batteries

  RAN Wenjiang, WEI Shengli, YANG Zengchang, XIE Weibo

  Chinese Journal of Power Sources. 2026, 50(10): 1961-1970. https://doi.org/10.3969/j.issn.1002-087X.2026.10.008

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  In order to enhance the thermal dissipation capability of a commercial vehicle power battery pack and reduce the energy consumption of the battery module cooling system, this study compared the heat dissipation effects of different liquid cold plate structures and proposed a novel dual-inlet dual-outlet configuration. Contour analysis of temperature and pressure drop was conducted. Taking a 230 Ah prismatic lithium iron phosphate (LFP) battery as the research subject, a simulation model of the liquid cold plate battery module cooling system was established. The effects of coolant inlet flow velocity, aspect ratio, number of flow channels, and coolant inlet temperature on the heat dissipation of the battery module were investigated. An orthogonal experiment was employed to analyze and optimize the results. The results show that the proposed liquid cold plate can effectively satisfy the heat dissipation requirements of the battery pack while maintaining good control over the pressure drop. Compared with the traditional serpentine flow channel, the new structure significantly reduces the temperature, with the maximum temperature decreasing by 2.18 ℃ and the maximum temperature difference decreasing by 2.72 ℃. At the same time, a substantial reduction in pressure drop was achieved. After optimization, the maximum temperature can be further lowered by 2.51 ℃, and the maximum temperature difference is reduced by 2.77 ℃.
• Select
  Research on mechanism of paraffin-based flame-retardant composite phase change materials in battery thermal management

  LI Xuxiong, HE Ruifeng, LIU Hongli, CHEN Peihui, LIANG Beiwen, LIU Xiaozhou, LI Xinxi

  Chinese Journal of Power Sources. 2026, 50(10): 1971-1981. https://doi.org/10.3969/j.issn.1002-087X.2026.10.009

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  This paper developed a flame-retardant composite phase change material possessing high thermal conductivity, excellent flame retardancy, and good structural stability for thermal management in lithium-ion battery modules. This material employed paraffin wax as the phase change matrix, styrene-ethylene-butylene-styrene copolymer as the flexible supporting framework, and expanded graphite acted as the thermal conductivity enhancer. Methyl MQ silicone resin and cyanuric acid served as the synergistic flame-retardant system. Experimental results indicate that when the addition levels of expanded graphite, methyl MQ silicone resin, and cyanuric acid are 2%, 14%, and 14% respectively, the composite material (PSEF2) exhibits optimal performance with a phase change latent heat of 141.15 J/g and a thermal conductivity of 1.112 W/(m·K). At discharge rates of 1 C, 1.5 C, and 2 C, the maximum temperatures of battery modules using this material were 41.9, 48.2, and 56.0 ℃, respectively, with maximum temperature differences of only 5.0, 7.2, and 8.8 ℃, demonstrating excellent temperature control and uniformity performance. This research provides an efficient and safe material solution for thermal management in power batteries, offering broad application prospects in new energy vehicles and other fields.
• Select
  Remaining useful life prediction of electric vehicle battery based on VMD-CNN-Transformer

  LI Lianbing, GAO Xin

  Chinese Journal of Power Sources. 2026, 50(10): 1982-1990. https://doi.org/10.3969/j.issn.1002-087X.2026.10.010

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  The remaining useful life (RUL) of electric vehicle power battery is an indispensable and important indicator in battery health management. In order to improve the prediction accuracy of battery RUL, a prediction method based on rolling variational mode decomposition (VMD) and CNN-Transformer was proposed. Firstly, the VMD algorithm was used to perform rolling feature decomposition on the preprocessed data to obtain multiple modal components with different trends and frequencies, which effectively reduced the influence of data nonlinearity and noise and improved the smoothness of data. Then the correlation between different components was calculated by Spearman correlation coefficient method. After that, the CNN-Transformer model was established to predict the residual life of lithium battery. The CNN model was used to extract the local information of different modal components, and the Transformer model was used to model the long-term dependence between each modal component. In addition, the size of the data set directly affected the training effect of the model. In order to solve the problem of insufficient battery capacity attenuation data, the WGAN-GP algorithm was used to generate data for pre-training of the model, and the real capacity data was used for fine-tuning to improve the model prediction effect. Finally, the CALCE lithium battery data set provided by the University of Maryland and the generated data set were simulated and tested, and the prediction results of different algorithms were compared. The results of experiments show that this method can accurately predict the change trend of lithium battery RUL.
• Select
  Cross-condition SOC estimation of lithium iron phosphate batteries with integrated stress

  LIU Ziyang, QIAO Xinlei, DONG Qianli, WANG Lujun

  Chinese Journal of Power Sources. 2026, 50(10): 1991-2002. https://doi.org/10.3969/j.issn.1002-087X.2026.10.011

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  Accurate estimation of the state of charge (SOC) of lithium-ion batteries under complex operating conditions directly affects the safety and reliability of battery management systems. To address the insufficient robustness of conventional SOC estimation methods that rely solely on voltage and current features in cross-condition scenarios, this paper proposed a cross-condition SOC prediction method based on an ALA-BiTCN-BiGRU framework with integrated stress features. Voltage, current, and stress signals were jointly used as model inputs to fully exploit the electrochemical-mechanical coupling information, while the artificial lemming algorithm (ALA) was employed to perform global optimization of key hyperparameters in the BiTCN-BiGRU network, thereby enhancing temporal feature extraction capability and generalization performance. Cross-condition test sets were constructed using four dynamic operating profiles, namely FUDS, UDDS, DST, and NEDC, and a leave-one-condition-out strategy was adopted for validation. The experimental results demonstrate that incorporating stress features significantly reduces prediction errors, and the proposed method maintains high accuracy and stability under unseen operating conditions, confirming its effectiveness and engineering applicability.
• Select
  Effect of external pressure on volume expansion and cycle life of LiNi_0.8Co_0.1Mn_0.1O₂||Si/C pouch cells

  CAI Shujun, DING Zhengping, SI Yufeng, LIU Jiachen, HUANG Yi, REN Yurong

  Chinese Journal of Power Sources. 2026, 50(10): 2003-2013. https://doi.org/10.3969/j.issn.1002-087X.2026.10.012

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  This study investigated the effect of external mechanical pressure on the cycle life of NCM811||Si/C lithium-ion pouch cells. To mitigate the detrimental effects of significant volume changes in the Si/C anode, cells were cycled under different applied pressures. The results demonstrate that higher external pressure effectively suppresses cell expansion, thereby significantly improving capacity retention. In-situ pressure monitoring and post-mortem analyses reveal that this performance enhancement is attributed to improved electrode interface contact and reduced irreversible loss of active lithium. High pressure maintains electrode interface integrity during cycling, whereas low pressure leads to severe interfacial expansion and contact failure. This work elucidates the mechanism by which appropriate mechanical constraint enhances cycle stability through volume expansion inhibition, offering an effective strategy for the practical application of high-energy-density batteries.
• Select
  Research on overcharge testing of commercial energy storage cells

  GU Zebo, HUANG Kun, ZHANG Siyao, TANG Xinya, LIU Guorong, FENG Hao, CHEN Hao, WANG Wentao, CHEN Jun

  Chinese Journal of Power Sources. 2026, 50(10): 2014-2019. https://doi.org/10.3969/j.issn.1002-087X.2026.10.013

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  This work investigated the impact of test fixtures and pre-tension on overcharge safety of commercial prismatic energy storage cells based on GB/T 36276-2023. Comparing overcharge tests with no fixture, fixture without pre-tension, and fixture with pre-tension, it reveals their crucial role in suppressing thermal runaway. Fixtures significantly reduce temperature rise and thermal runaway severity, while pre-tension further delays and mitigates thermal runaway. Scanning electron microscope analysis shows that fixtures and pre-tension maintain electrode integrity, reducing active material shedding and structural damage. This research provides scientific basis for safety assessment and theoretical support for refining overcharge test standards of energy storage cells, enhancing system safety and reliability.
• Select
  Study on interface mechanism of graphene modified solid-state battery anode copper foil

  GENG Zhenkun, LIU Kexin, LIU Hongli, ZENG Yueyu, ZHANG Guoqing, LI Xinxi

  Chinese Journal of Power Sources. 2026, 50(10): 2020-2028. https://doi.org/10.3969/j.issn.1002-087X.2026.10.014

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  A series of graphene modified copper foils (G-Cu) with mass fractions of 70%, 80% and 90% were prepared by solution coating method to enhance the interface stability of current collector in sulfide solid-state batteries, and the effects of coating microstructure on interface protection mechanism and electrochemical performance were systematically analyzed. The results show that the 80% G-Cu sample achieves the best trade-off between the coating densification degree and the porosity. 80% G-Cu shows excellent interfacial compatibility, low interfacial polarization and long cycle life in both static interface contact and dynamic lithium deposition/stripping cycles. This is mainly due to its uniform pore structure, which effectively blocks the direct contact between the copper current collector and the sulfide electrolyte, constructs an efficient lithium ion transport channel, and significantly inhibits the interfacial side reactions and dendrite penetration. This study shows that precise regulation of the mass ratio of graphene to binder is a key way to improve the electrochemical compatibility of current collectors in sulfide all-solid-state batteries, and 80% is determined as the optimal mass ratio of graphene coatings.
• Select
  Bituminous coal-derived hard carbon anode for rechargeable sodium battery

  PAN Guanghong, DONG Chunwei, HE Guofeng, KONG Junli, SU Zhijiang, DONG Yang

  Chinese Journal of Power Sources. 2026, 50(10): 2029-2035. https://doi.org/10.3969/j.issn.1002-087X.2026.10.015

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  The utilization of coal-based hard carbon as an anode material for high-performance sodium-ion batteries holds significant importance in advancing the commercialization of sodium-ion batteries. However, owing to the large ionic radius of sodium, sluggish diffusion kinetics, and an unclear storage mechanism, the specific capacity, rate performance, and cycling stability of such anode materials still require further enhancement. This study utilized a series of low-rank bituminous coals as precursors to fabricate hard carbon anodes for SIBs through purification and carbonization processes. The experimental results demonstrate that as the fixed carbon content on a dry basis in the raw material increases, the specific surface area and porous structure of the resulting hard carbon material decrease. When used as an anode material in sodium-ion batteries, it exhibits a specific capacity of 284 mAh/g, an initial coulombic efficiency (ICE) of 85.8%, a 65% capacity retention at 2 C/0.1 C, and 91.2% capacity retention after 100 cycles. Analysis of the sodium storage mechanism reveals that the capacity of coal-based hard carbon is primarily contributed by a sloping region dominated by surface adsorption and a plateau region resulting from a composite mechanism involving intercalation and pore filling.
• Select
  Preparation and property study of composite graphite bipolar plates for redox flow battery

  LIU Zi, WANG Yuwei, HAO Qiangzhou, ZHANG Qi, HE Yingnan, LIU Zhen, FENG Wei, WU Jingbo, NI Shenglan

  Chinese Journal of Power Sources. 2026, 50(10): 2036-2040. https://doi.org/10.3969/j.issn.1002-087X.2026.10.016

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  Using expanded graphite and graphene as conductive materials and epoxy-acrylate resin as binder, composite graphite bipolar plates for all-vanadium flow batteries were successfully fabricated through a combined rolling-impregnation process. Comparative investigations were conducted on the electrical conductivity, mechanical properties, swelling behavior, and electrochemical performance of composite bipolar plates with different thicknesses. Subsequently, single-cells were assembled for charge-discharge performance evaluation. The results demonstrate that through constructing continuous conductive network structures and utilizing the synergistic effects of resin filling and structural reinforcement, the prepared bipolar plates achieve a flexural strength exceeding 30 MPa, electrical conductivity surpassing 300 S/cm, and remarkably low dynamic potential corrosion current density (on the order of 10^-6 A/cm²), while exhibiting negligible swelling in electrolyte. Benefiting from these superior characteristics, the assembled all-vanadium single-cell battery maintains 82.5% voltage efficiency after 800 charge-discharge cycles under 150 mA/cm² current density.
• Select
  Influence of initial activation temperature of heating composition on thermal battery safety

  CHEN Xu, AN Jianmin, LIU Yufei

  Chinese Journal of Power Sources. 2026, 50(10): 2041-2045. https://doi.org/10.3969/j.issn.1002-087X.2026.10.017

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  In order to explore the correlation between the initial activation temperature of heat pellets and the safety performance of thermal batteries, a series of experimental studies were carried out. The results show that the activation temperature of thermal batteries should be controlled below the theoretical critical temperature. Under the no-load operating condition, there is no effective heat absorption mechanism inside the battery. The continuously accumulated self-generated heat will trigger the thermal decomposition of cathode materials and produce a large amount of gas, which further leads to safety hazards such as battery swelling and leakage.
• Select
  Application of thermally crosslinked ePTFE-reinforced SPES membrane in intermediate-temperature PEM water electrolysis

  WANG Baoan, HAO Jinkai, ZHU Yingran, ZHANG Hongjie, SHAO Zhigang

  Chinese Journal of Power Sources. 2026, 50(10): 2046-2053. https://doi.org/10.3969/j.issn.1002-087X.2026.10.018

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  In intermediate-temperature proton exchange membrane (PEM) water electrolysis, proton exchange membranes often suffer from excessive swelling and mechanical degradation under high-temperature and high-pressure liquid water, which limits electrolysis performance and long-term stability. To address this issue, a composite proton exchange membrane based on highly sulfonated polyethersulfone (SPES) was prepared by combining thermal crosslinking with an expanded polytetrafluoroethylene (ePTFE) porous reinforcement. The crosslinking conditions were optimized and the physical confinement of the ePTFE porous scaffold effectively suppressed membrane swelling, improving dimensional stability and mechanical strength. Scanning electron microscopy and energy-dispersive spectroscopy reveal that the SPES matrix fully infiltrates and fills the porous ePTFE structure. The composite membrane exhibits a proton conductivity of 168.3 mS/cm at 110 ℃ under pressurized liquid water. Electrolysis tests show that the membrane electrode reaches a current density of 7.52 A/cm² at 110 ℃ and 1.9 V, significantly higher than that of a commercial Nafion membrane electrode (3.34 A/cm²).
• Select
  Study on pulsed-load strategy to improve high-potential stability of PEMFCs

  ZHANG Hanwei, ZHANG Hui, MEI Jian, ZHOU Fen, PAN Mu

  Chinese Journal of Power Sources. 2026, 50(10): 2054-2062. https://doi.org/10.3969/j.issn.1002-087X.2026.10.019

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  To address proton exchange membrane fuel cell (PEMFC) performance degradation caused by Pt oxidation at high potentials, a non-short-circuit pulsed load strategy was proposed. This strategy effectively suppressed voltage decay during high-potential operation. It is found that the mitigation effect depends on pulse frequency and pulse voltage trough. Pulse frequency must balance oxide removal with Pt catalyst stability. At an extremely low frequency (0.01 Hz), long oxidation time leads to incomplete oxide removal and limited improvement. At a lower frequency (0.1 Hz), Pt catalyst stability deteriorates. In contrast, a higher frequency of 2 Hz achieves the best mitigation effect. Compared with the non-pulsed sample, the 24 h total voltage decay drops from 51.12 mV to 23.90 mV. Additionally, significant improvement occurs only when the pulse voltage trough is below 0.6 V. This reduces the voltage decay by at least 45%.
• Select
  Parameter identification of static model for proton exchange membrane fuel cells based on starfish optimization algorithm

  JIN Chang, LIU Miao, FANG Yesheng, XING Yanfeng, LI Chuanchang

  Chinese Journal of Power Sources. 2026, 50(10): 2063-2072. https://doi.org/10.3969/j.issn.1002-087X.2026.10.020

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  To address the challenges of difficult parameter identification and low efficiency in static models of proton exchange membrane fuel cells (PEMFCs) due to numerous unknown parameters and strong coupling, this paper proposed a parameter identification method based on the starfish optimization algorithm (SFOA). First, a low-dimensional static model with only three unknown parameters was constructed based on electrochemical theories such as the Nernst equation, Tafel relations, Ohm’s law, and diffusion limits, reducing the identification dimensionality and correlation. Then, SFOA was used to accurately define the unknown parameters. Finally, steady-state polarization data at 50, 60, 70, and 80 ℃ were collected on a 30 kW fuel cell stack platform to verify the effectiveness of the method. Experimental results show that the sum of squared errors for the four different temperature conditions are 0.126 4, 0.558 9, 0.105 6, and 0.553 3, respectively. These results demonstrate higher accuracy and stability compared with the porcupine optimization algorithm, the velvet worm optimization algorithm, and the gray wolf optimization algorithm. The I-V and I-P curves output by the model closely match the experimental data points and exhibit excellent generalization ability, providing a reliable foundation for performance prediction and power optimization control of high-power PEMFCs.
• Select
  Fault diagnosis of PEMFC based on CNN-FEDformer considering overcurrent condition

  XU Lousong, YUAN Zhi, ZOU Zifan, LI Ji

  Chinese Journal of Power Sources. 2026, 50(10): 2073-2083. https://doi.org/10.3969/j.issn.1002-087X.2026.10.021

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  This paper proposed a dual-channel feature fusion model based on convolutional neural network (CNN) and frequency-enhanced decomposition transformer (FEDformer) to address issues in the fault diagnosis of proton exchange membrane fuel cells (PEMFC). These issues included confusion between over-current and fault states, the use of single evaluation metrics, and model complexity. The proposed model used a parallel structure that combined the local spatial feature extraction capability of the CNN with the FEDformer’s global temporal modelling advantage. Furthermore, a gated recurrent unit (GRU) has been introduced to optimize temporal features further, enabling high-precision diagnosis of PEMFC operating states. The experiments were conducted on a dataset provided by Loughborough University, which was pre-processed and divided into training and testing sets. The results demonstrate that the proposed method achieves an accuracy, precision, recall and F1-score of 99.96% on the test set, with a false positive rate and rejection rate of 0.000 1 and 0.000 4 respectively. This significantly outperforms comparative models. By incorporating the over-current state into the diagnostic scope and employing six metrics for a comprehensive performance evaluation, this study provides an efficient, robust solution for the intelligent diagnosis of PEMFC faults.
• Select
  Modeling on multi physical field distribution based on a 3D two-phase flow model and performance optimization of proton exchange membrane electrolytic cell

  QIN Ruyi, WANG Yi, AN lei, SHEN Keju, LU Xiaohong, LIU Min, YE Xiaming, YANG Yang, HE Zhengjie, PENG Jiaying

  Chinese Journal of Power Sources. 2026, 50(10): 2084-2093. https://doi.org/10.3969/j.issn.1002-087X.2026.10.022

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  An in-depth understanding of the heat and mass transfer, electrochemical processes and multi-field distribution characteristics of the proton exchange membrane electrolytic cell(proton exchange membrane electrolysis cell, PEMEC) can provide theoretical support for the optimized design and operation management aimed at improving performance and durability. In this paper, a three-dimensional PEMEC model was developed, coupled with the heat and mass transport, two-phase flow and electrochemical processes, and the distribution characteristics of multiple fields were simulated and studied. Subsequently, the influence of operating conditions was evaluated. The main conclusions are as follow. The gas saturation is relatively high downstream of the flow channel and at the rib plate, while the liquid phase saturation is relatively low. The temperature of the catalyst layer on the cathode side of the membrane electrode is the highest. When the voltage is 2 V, the temperature difference between the inside and outside of the membrane electrode reaches 7 K. Even when contact resistance is not taken into account and the membrane is sufficiently hydrated, the ohmic loss is still considerable, especially when operating at high voltage. When operating at 2 V, the activation loss is 56%, the ohm loss accounts for 30.3%, and the mass transfer loss increases to 13.7%. Appropriately increasing the operating temperature can reduce the activation loss and ohmic loss. Increasing the flow rate of liquid water at the anode can reduce mass transfer losses. Within the considered operating conditions, the electrolytic cell reaches its optimal working state when the velocity liquid water in the channel is 2.5 m/s and the temperature is 80 ℃. The conclusions of this paper can be a reference for the design and operation management of PEMEC.
• Select
  Parameter sensitivity analysis of operating conditions for high-power automotive proton exchange membrane fuel cells

  LIU Zhen, ZHOU Yafen, XIAO Kuan

  Chinese Journal of Power Sources. 2026, 50(10): 2094-2100. https://doi.org/10.3969/j.issn.1002-087X.2026.10.023

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  To investigate the output response of high-power automotive proton exchange membrane fuel cells (PEMFCs) under multi-parameter fluctuations and identify key sensitive parameters, an experimental study on operating-parameter sensitivity was carried out. Five typical current density points (0.2, 0.5, 1.0, 1.6, and 1.9 A/cm²) were selected. Nine operating parameters, including hydrogen and air stoichiometric ratios, gas pressures, humidification dew-point temperatures, as well as coolant flow rate, temperature and pressure, were adjusted by ±10% to examine their effects on output voltage. The results show that the fuel cells operate stably at all current densities. Under most ±10% parameter perturbations, the performance variation rates are less than 2%, demonstrating favorable adaptability to operating-condition disturbances.Coolant temperature, air stoichiometry, and air pressure were identified as highly sensitive parameters. Among them, coolant-temperature deviation produces the most pronounced influence: a performance degradation of 5.82% is observed at 1.9 A/cm² for a -10% coolant-temperature perturbation. Hydrogen pressure and coolant pressure belong to low-sensitivity parameters, whose impacts on cell performance are close to zero under most test conditions. The magnitude of parameter deviation impact increases with current density; the performance fluctuation amplitude at high current densities (1.6-1.9 A/cm²) is 5-9 times larger than that at low current density (0.2 A/cm²). Furthermore, the inhibitory effect of low-temperature deviation on performance is stronger than the promoting effect of high-temperature deviation. The results provide an experimental basis for operating condition optimization, control strategy formulation, and error tolerance design of high-power automotive fuel cells.