Lithium iron phosphate electrodes are widely used in lithium-ion batteries, but their low intrinsic electronic conductivity can produce non-uniform reaction and heat-generation distributions during high-rate operation. In this study, a deterministic three-dimensional microscale electrochemical model with one-way thermal coupling was developed to examine how electronic conduction, ionic transport, active-particle size, temperature, and conductive-binder-domain (CBD) distribution affect local electrode behavior. The model reproduces the rate-dependent voltage response of LFP half-cells and provides spatial information that is not directly available from macroscopic discharge curves. Under 3 C discharge, the solid-phase potential drop reached 198.75 mV at 0.6 depth of discharge, whereas the liquid-phase potential drop was only 4.13 mV, indicating that solid-phase electronic conduction strongly affected reaction localization. The CBD network accounted for 87.17–91.22% of the total Joule heat over the examined temperature range. Reducing particle size improved macroscopic capacity and increased the local current-density peak near the current collector, while decreasing the through-thickness variation of reaction current. The CBD gradient results further showed that abrupt low-conductivity regions can introduce local polarization losses under fixed total CBD content. These results provide a controlled microscale analysis of coupled transport and heat-generation behavior in LFP electrodes and support the design of conductive networks with continuous electron-transport pathways.

Under stringent environmental regulations, multi-stack fuel cell commercial vehicles are emerging as a key technology for zero-emission transportation. However, due to the “barrel effect,” disparities in the state of health (SOH) among fuel cell stacks can accelerate system degradation and shorten the overall service life. In this study, we firstly propose an index to quantify the SOH disparity among fuel cell stacks and incorporate it into the objective function to be minimized. Then, an Enhanced Soft Actor-Critic (ESAC) reinforcement learning framework is developed, which embeds a three-layer rule-based strategy. Hardware-in-the-loop test results demonstrate that introducing the SOH-disparity-aware term into the objective function effectively mitigates the SOH imbalance phenomenon. Meanwhile, ESAC promotes long-term operation of the multi-stack system at identical constant power levels, which alleviates fuel cell degradation and further reduces SOH disparity among stacks. These findings provide a critical pathway for intelligent energy management in next-generation fuel cell trucks.

The thermal runaway (TR) and its propagation in high-energy-density lithium-ion batteries (LIBs) remain critical safety bottlenecks that urgently need to be overcome. Here, a two-phase immersion cooling thermal protection strategy is proposed to suppress propagation of TR. Thereafter, a semi-open TR experimental system is constructed. Based on nail penetration-induced internal short circuit (ISC) experiments on a NCM pouch cell, a three-dimensional (3D) immersion-cooling TR model is established. Simultaneously, the influences of liquid filling ratio, initial temperature, and coolant thickness on multiphase flow characteristics and the spatiotemporal evolution of TR are systematically examined. A stepwise increase in the peak temperature of TR is observed with decreasing liquid filling ratio, while the peak time is significantly advanced from 13.41 s to 9.78 s. The onset of phase change is strongly affected by the initial temperature. Under the 40 °C condition, the peak temperature of 470.44 °C is reached at 8.79 s. A critical heat flux condition is approached as the coolant thickness increases. A saturation effect in the suppression of temperature rise is observed with increasing coolant thickness, with the peak temperature reduced only from 338.28 °C to 334.61 °C. This study offers important quantitative guidance for the optimization of next-generation battery thermal management systems (BTMSs), with the objective of achieving an optimal trade-off between extreme thermal safety and high system-level energy density.

Hydrogen - electricity coupled DC microgrids (HE-DCMGs) represent a promising and sustainable solution for off-grid power supply. However, achieving high economic performance while ensuring DC bus voltage stability is a critically challenging task. This study proposes a hierarchical energy management and control strategy for HE-DCMGs that integrates an adaptive mutation Harris hawks optimization (AMH-HO) algorithm at the system level with a fractional-order sliding mode controller (FOSMC) at the device level. A multi-objective optimization problem is formulated to minimize hydrogen consumption and reduce degradation of proton exchange membrane fuel cells and lithium-ion batteries. The AMHHO algorithm, augmented with differential evolution and L & eacute;vy flight mechanism, determines the optimal power allocation among distributed sources, while the FOSMC provides robust DC bus voltage regulation. The proposed strategy is validated on a 750 V HE-DCMG experimental platform capable of 168 hours of off-grid operation. Experimental results show that the proposed strategy reduces long-term operating costs and improves energy-utilization efficiency, achieving an overall system efficiency of 80.49%-97.37%. The DC bus voltage is maintained with a response time of 0.02 s, a low overshoot of 3.7%, and a voltage fluctuation rate of 3.08%, all of which comply with the requirements of IEEE Std 1547-2018.

Accurate state of health (SOH) estimation of lithium-ion batteries is essential for reliable battery management but is particularly challenging under mixed non-Gaussian measurement noise and data scarcity. This article proposes a robust SOH estimation framework that combines noise-resistant feature extraction with data-efficient, interpretable continuous-time dynamic modeling. A mixture correntropy loss-based principal component analysis (MCL-PCA) method is first developed, which employs a hybrid Gaussian-Laplacian correntropy criterion to adaptively suppress mixed non-Gaussian noise and extract stable low-dimensional health features (HFs). These features serve as inputs to a newly designed Kolmogorov-Arnold-enhanced liquid neural network (KLNN), which augments continuous-time liquid dynamics with structured nonlinear mappings to improve stability and nonlinear generalization under limited small-sample conditions. These two components are integrated into a unified framework that yields coherent and physically consistent SOH degradation trajectories. Experimental results on laboratory and public datasets demonstrate that MCL-PCA significantly improves feature robustness, while KLNN achieves superior SOH estimation accuracy in noisy and small-sample scenarios, resulting in notably lower prediction errors than conventional PCA-based methods and consistent advantages over baseline models.

Current transfer learning-based methods for predicting the degradation of proton exchange membrane fuel cells (PEMFCs) commonly face insufficient elimination of inter-domain distribution shifts and low knowledge transfer efficiency. Traditional approaches typically extract domain-invariant features by filtering multi-dimensional operational parameters. This process is not only complex but may also introduce interference weakly correlated with voltage degradation, thereby compromising the stability and reliability of cross-domain predictions. To address these challenges, this paper proposes a time-frequency fusion transfer learning (TF-TL) architecture. Unlike existing methods that filter domain-invariant features from multi-dimensional operational parameters, the proposed approach directly aligns domains by leveraging the intrinsic frequency-domain characteristics of voltage signals. The integrated Frequency Domain Adaptation (FDA) technique employs Fast Fourier Transform (FFT) to extract intrinsic low-frequency components from voltage signals. It then aligns features between source and target domains using an enhanced Maximum Mean Discrepancy (MMD) metric. This voltage-centric, frequency-domain alignment avoids dependence on complex auxiliary parameters, thus improving the interpretability and stability of transfer learning. Experimental results show that TF-TL significantly improves generalization across devices and operating conditions, with FDA effectively reducing inter-domain distribution divergence. In multiple transfer tasks, the framework achieves highly accurate and reliable remaining useful life estimation with a relative error below 1.9%. The proposed architecture provides a theoretical foundation for deploying rapid and reliable online prognostic systems for PEMFCs under complex operating conditions.

Traditional time–frequency domain methods face critical limitations in predicting voltage degradation of proton exchange membrane fuel cells (PEMFCs). Time-domain models struggle to robustly separate long-term degradation-related low-frequency trends from contaminated voltage signals under highly dynamic and non-stationary conditions, while conventional frequency-domain analysis loses essential time-localized information during feature extraction. Both approaches exhibit significantly degraded prediction performance under limited data conditions. To overcome these challenges, this paper proposes a time–frequency fusion algorithm that integrates TimesNet with long short-term memory (LSTM), effectively combining 2D frequency-domain representations with 1D temporal memory to enhance voltage degradation prediction under dynamic conditions. Based on the capability of TimesNet-LSTM to extract low-frequency voltage features, a transfer learning technique grounded in low-frequency consensus knowledge (LCK-TL) is further developed. By selectively transferring low-frequency voltage features that robustly reflect aging patterns, LCK-TL considerably reduces distribution discrepancy between source and target domains, achieving joint optimization of predictive modeling and transfer mechanisms. Leveraging the inherently low computational cost of transfer learning, LCK-TL enables rapid multi-step predictions while maintaining accuracy, providing effective guidance for cross-device and cross-condition fuel cell health management.

The gas diffusion layer (GDL) of proton exchange membrane fuel cells (PEMFCs) is a critical component for the transport of reactants. The efficiency of reactant gas transport remains a major technical challenge in the field today. The anisotropic structure of the GDL gives rise to substantial variations of gas diffusion as well as permeability in different directions. The study employs X-CT technology to obtain the actual GDL's geometry, aiming to investigate a spatial structure at the microscale and its gas transport characteristics. The computational fluid dynamics (CFD) method is used to simulate and study the gas diffusivity and gas permeability of GDL with four different thicknesses. The numerical simulation results show that the diffusivity and permeability in the through-plane (TP) direction are lower than those in the in-plane (IP) direction. Moreover, the effective diffusion coefficient (EDC) decreases with increasing thickness, but is also dependent on the solid fibre structure of GDL. Horizontal alignment of the carbon fibers and the disc-shaped adhesive contributes to the anisotropy between the TP and IP directions, resulting in anisotropic gas transport. The purpose of the study is to supply critical references for manufacturing techniques and optimization of gas transport in GDLs.

The performance of solid oxide electrolysis cells (SOECs) is closely related to its flow channel structure, stoichiometric ratio, and operating temperature. In this study, multiphysics numerical models were developed, incorporating coupled heat and mass transport alongside electrochemical reaction processes. The model validity was confirmed by comparing simulated results with experimentally measured I-V curves. The impacts of the fuel-to-air stoichiometric ratio (F:A from 2:1 to 2:4), channel aspect ratio (L:W from 1:1 to 4:1), temperature (from 873.15 K to 1073.15 K), and flow arrangements (co-current vs. counter-current) on the performance of a single-channel electrolyzer were systematically investigated, complemented by an analysis of multi-channel behavior under cross-flow conditions. The findings reveal that among the investigated parameters, temperature exerts the most significant influence on cell performance. Increasing the temperature facilitates the substitution of electrical energy with thermal energy, reducing the cell voltage from 1.5364 V to 1.1142 V at 1.2 A/cm2. Furthermore, varying the stoichiometric ratio effectively improves the oxygen partial pressure in the catalyst layer, thereby reducing concentration polarization. At the same current density, the required cell voltage decreased from 1.5364 V to 1.5153 V. Increasing the channel aspect ratio improves mass transport, reducing the required cell voltage from 1.5444 V to 1.5053 V. In contrast, flow arrangements were found to have a negligible impact on overall performance, though the counter-flow arrangement demonstrated marginal superiority over the co-flow arrangement.

This paper presents a novel algorithmic framework for efficiently solving the pseudo-two-dimensional (P2D) model of lithium-ion batteries. The proposed approach reformulates the original P2D model, typically expressed as a system of coupled nonlinear partial differential-algebraic equations, into a system of quasilinear partial integro-differential equations (PIDEs). Through this reformulation, intermittent algebraic states, such as local potential and current terms, are effectively eliminated, thereby reducing the model complexity. This enables the identification of a generic fixed-point iterated function for solving the P2D model's nonlinear algebraic equations. To implement this iterated function, the finite volume method is employed to spatially discretize the PIDE system into a system of ordinary differential equations. An implicit-explicit (IMEX) time integration scheme is adopted, and the resulting quasilinear structure facilitates the development of a single-step numerical integration scheme that admits a closed-form update, providing stable, accurate, and computationally efficient solutions. Unlike traditional gradient-based approaches, the proposed framework does not require the Jacobian matrix and is insensitive to the initial guess error of the solution, making it easier to implement and more robust in practice. Due to its significantly reduced computational cost, the proposed framework is particularly well-suited for simulating large-scale battery systems operated under advanced closed-loop control strategies.

The integrated hydrogen energy utilization system (IHEUS) exhibits great potential for microgrid applications. However, its practical deployment faces significant challenges, primarily due to the low energy conversion efficiency and rapid aging of electrolyzers and fuel cells, especially when handling highly fluctuating power flows. To address these issues, this study proposes a multi-objective optimal dispatch scheme for off-grid IHEUS operations, incorporating waste heat recovery and life cycle cost considerations. First, a mechanistic model is established to characterize the electric-hydrogen-heat output characteristics of the system, with a specific focus on waste heat recovery and utilization subsystems. By correlating the aging behavior and lifetime degradation to voltage decay, a life-cycle operational cost function is formulated for the multi-objective optimization (MOO) model. Within this framework, comprehensive energy efficiency and energy supply loss probability are adopted as performance metrics to enhance energy utilization and stability. The resulting MOO problem is solved and prioritized using a proposed NSGA-III combined entropy-weighted TOPSIS strategy. Comparative studies demonstrate that this strategy effectively identifies the optimal dispatch scheme, achieving operational cost reductions of at least 17.53%, comprehensive energy efficiency improvements ranging from a 0.13% decrease to a 0.61% increase, and a limited increase in energy supply loss probability (4.14%).

To address the challenges of poor noise immunity and limited generalization performance in Li-ion battery modeling and state estimation (SE), a novel robust framework for parameter identification (PI) and joint estimation of state of charge (SOC) and surface temperature is proposed in this study by leveraging physical information and nonlinear extension techniques. Initially, a robust forgetting factor recursive maximum total correntropy algorithm is developed for PI, providing a solid foundation for SE under noisy conditions. Subsequently, a robust SOC estimation method is formulated by embedding the maximum correntropy criterion (MCC) with an adaptive kernel width into the square-root cubature Kalman filter, effectively replacing the conventional mean square error with MCC to enhance noise resilience. Next, a multidimensional feature input set is constructed using the PI results, including total internal resistance as auxiliary physical information, along with SOC estimates and raw measurements. A subinput structure is further designed using partial correlation analysis, and then the extreme learning machines are utilized to project the subinputs into a high-dimensional (HD) feature space to extract latent correlation features. Finally, by integrating nonlinear extended features with raw data in parallel, the input to the bidirectional gated recurrent unit model is generated, enabling simultaneous extraction of global representations from both HD and low-dimensional spaces. Experimental results demonstrate that the proposed method outperforms existing advanced approaches in SE under strong noise interference and complex operational conditions.

Accurate estimation of lithium-ion battery state of health (SOH) in electric vehicles (EVs) under real-world conditions is much more challenging than using well-designed laboratory cycling data due to unreliable SOH labeling and segmented charging behavior, and the results often lack interpretability. To address these issues, this paper proposes a physics-informed deep gated recurrent unit (PIDGRU) architecture for robust and interpretable SOH estimation without requiring explicit physical modeling. First, a modified inverse ampere-hour integral method is combined with the Bayesian estimator of abrupt, seasonality, and trend (BEAST) algorithm to estimate battery capacity and characterize SOH uncertainty. A universal feature extraction and selection framework is then developed to handle segmented EV charging data, utilizing a hybrid linear-nonlinear redundancy analysis to ensure an optimal input feature set. The PIDGRU integrates empirical degradation modeling and nonlinear dynamic degradation learning through a deep gated recurrent unit network, utilizing a deep hidden physics model (DeepHPM). A Bayesian inference uncertainty-constrained (BIUC) strategy is introduced to enhance training reliability and uncertainty quantification. Extensive evaluations on both in-vehicle and cross-vehicle datasets demonstrate that the proposed method achieves high accuracy, robustness, and generalizability, with the mean absolute error and root mean squared error consistently below 1.10% and 1.31%, respectively.

The lithium plating reaction in graphite electrodes acts as a root cause for the accelerated degradation and the internal short circuits in lithium-ion batteries. Here, an electrochemical model based on multi-scale microstructural images was established to identify lithium plating-stripping processes, thereby supporting the predictive outcomes of electrochemical monitoring techniques. Experiments revealed that the open-circuit voltage differential curve (dOCV/dt) led to ambiguous delineation of the safe state-of-charge (SOC) operating range. The established lithium plating-stripping model was used to compare with experimental results, revealing the dynamic evolution of electrode-scale kinetics and quantified the impact of lithium metal residue on electrode performance. Ex situ X-ray computed tomography (XCT) captured micrometer-resolution microstructural details of graphite electrodes and plated lithium, enabling further correlation of spatially heterogeneous lithium plating-stripping reactions with electrode microstructure. The sensitivity of lithium plating to electrode microstructure was examined at the particle scale, attributed to competition between electrode kinetic rates and active reaction areas. Theoretical mechanism analysis and experimental results from high-energy-density electrodes demonstrated that positioning small particles on the current collector side effectively mitigates solid-state diffusion polarization while confining side reactions to a limited area. The integration of experiments and multiscale modeling elucidates the relationship between lithium plating-stripping reactions and electrode structure, providing mechanistic insights for similar structural optimization designs.

This paper introduces a modified relative voltage loss rate as an enhanced degradation indicator to characterize the primary operating points of eight onboard fuel cell city buses. The modified relative voltage loss rate is derived using a sliding window strategy combined with locally estimated scatterplot smoothing, effectively incorporating voltage decay acceleration under primary operating current conditions. Based on this refined degradation metric, a deep sequence model incorporating a self-attention gated recurrent unit is established to predict the remaining useful life of vehicle-level fuel cells. The self-attention mechanism enhances feature extraction, while the gated recurrent unit model captures temporal dependencies, forming a comprehensive time-series prediction framework. Comparative analysis demonstrates that the deep sequence model outperforms various conventional neural network architectures, including the baseline gated recurrent unit model and the self-attention long short-term memory model. The proposed deep sequence model achieves a mean absolute percentage error below 3.79% for all eight vehicle-level fuel cell systems, highlighting its robustness and generalizability.

Carbon corrosion induced by anode localized flooding severely compromises the durability of proton exchange membrane fuel cell (PEMFC). Limited by the computational stability and efficiency, existing simulations are always in 2D or single-channel scales, which overlooks the influence of the actual flow field structure in commercial PEMFC on carbon corrosion behavior. In this study, a performance-coupled 3D carbon corrosion model is established to investigate the carbon corrosion behavior and performance degradation in a 306 cm2 commercial-scale PEMFC under anode localized flooding conditions. The research demonstrates that the carbon corrosion zone exhibits a quasi-trapezoidal distribution influenced by hydrogen transport and in-plane proton conduction. Carbon loading undergoes rapid loss during the initial flooding phase, followed by a gradual leveling off. After 120 min of local flooding, the PEMFC exhibits an electrochemically active surface area (ECSA) loss of 23.97 % and an output power loss of 16.83 %. This model provides deeper insights into carbon corrosion behavior under localized flooding in large-scale PEMFC and offers a valuable reference for formulating carbon corrosion mitigation strategies.

Despite over a decade of research into the electrochemical reduction of carbon dioxide, achieving high selectivity for single, value-added products remains an unresolved challenge. While significant progress has been made in reactor-level performance and activity, our mechanistic understanding of why specific products emerge under different catalytic and interfacial conditions is often fragmented or contradictory. Most existing reviews categorize results by broad product classes or specific catalyst levers, often failing to resolve the critical mechanistic branching points between similar multicarbon products. Consequently, the field lacks a coherent, predictive structure that connects elementary reaction steps and microenvironmental effects to single-product selectivity. This manuscript addresses this gap by introducing a descriptor-based, multidimensional framework designed to unify the mechanistic origins of both C1 and C2+ pathways. Rather than merely cataloging performance, this approach identifies selectivity-determining intermediates to establish rules that transcend individual material systems. By reconciling long-standing debates and bridging molecular-scale mechanisms with practical system design, we herein provide the necessary foundation for the rational development of next-generation electrochemical carbon dioxide reduction technologies.

Efficient and reliable feature extraction engineering plays a crucial role in improving the accuracy and real-time operational capability for data-driven state-of-health (SOH) estimation of lithium-ion batteries. However, conventional feature extraction methods are time-intensive and prone to manual bias, particularly under the multi-step constant current fast-charging protocols prevalent in practical scenarios. To address this problem, this study introduces a generative pre-trained transformer (GPT)-powered automated feature extraction (AFE) method that identifies a series of health features from voltage and capacity aging data in both the time domain and frequency domain. Observation and analysis reveal that with increasing temperature, capacity-related features exhibit a more stable correlation with the SOH compared to voltage-related features. Based on this insight, we propose a two-stage kernel extreme learning machine-autoencoder (TS-KELM-AE) model, which integrates deep feature extraction and nonlinear mapping capabilities to estimate SOH under varying temperatures. Compared to manual feature extraction (MFE) and mainstream deep learning as a feature extractor, the proposed AFE method is feasible and efficient, and the TS-KELM-AE model achieves higher accuracy. Furthermore, analyses of feature importance, model multicollinearity, and cross-battery generalization under diverse operating conditions demonstrate that integrating AFE with the TS-KELM-AE framework establishes a robust and scalable solution for practical applications.

The rapid growth of renewable energy sources (RESs) has introduced significant uncertainty and instability to the operation of smart microgrids. To address these challenges, energy storage systems (ESSs) have become essential, as they can effectively mitigate the impact of the fluctuations of RESs. However, ESSs consisting of single-type energy storage devices face limitations in meeting the various requirements of microgrid applications. A hybrid energy storage system (HESS) that integrates multiple types of ESSs offers a popular solution by exploiting their complementary characteristics. This article provides a comprehensive review of the state-of-the-art technologies of HESSs, covering their types, modeling methods, control strategies, and particularly, the safety considerations that have often been overlooked in the existing review papers. The pros and cons of the existing works across these four dimensions are also critically analyzed. Finally, future trends and potential directions are highlighted to carry out further research and the practical implementation of HESS technologies.

To address the reliable and economical operation of virtual power plants containing a large number of adjustable resources in the day-ahead and intraday markets, a two-stage model predictive control (MPC) based optimization scheduling strategies for virtual power plants (VPPs) in the electricity spot market is proposed and discussed in this paper. Firstly, a TCN-GRU-attention hybrid prediction algorithm is developed for forecasting output of inelastic loads, wind power and photovoltaic (PV) in VPPs. Secondly, a hierarchical bi-level mixed-integer linear programming model integrating multiple resources with EV participation is established to enable that EV charging load, GT generation and ESS can dynamically adjust their behaviour under different operation cost. Furthermore, an improved MPC-based intraday dispatch strategy embedded with a bidirectional dynamic penalty mechanism is proposed to rapidly respond to real-time fluctuations of uncontrollable resources. The proposed two-stage MPC-based scheduling strategies can thereby enhance the flexibility and operational efficiency of the VPP system. Simulation results demonstrate that the two-stage collaborative optimization improves the total revenue of VPPs by 5.06%, fully verifying the robust economic advantages of the proposed solution in complex market environments.