Reducing the precious Pt loading in catalyst layers (CLs) while maintaining fuel cell performance is critical for cost reduction, necessitating the development of an optimized CL microstructure, but conventional empirical expertise-dependent trial-and-error optimization faces severe efficiency limitations when dealing with multiple transfers coupled with electrochemical reactions inside the multi-component hierarchical porous architecture. Here, we present Generative Artificial Intelligence for Controllable Electrode Synthesis (GAI4CES), a framework that enables rapid CL microstructure synthesis, achieving nearly 500x speedup for constructing a representative CL microstructure of 128 nm3 (in 0.36 s compared to 168 s required by traditional numerical methods). It can tailor key property constraints, such as Pt loading and electrochemical surface area (ECSA), to generate statistically representative microstructures satisfying target specifications. This capability enables an exhaustive search for the optimal CL structural design through large-scale geometry-constrained synthesis coupled with fast performance prediction. The optimized CL microstructure achieves an improvement of 11.6% in ECSA and a reduction of 32.9% in average ionomer thickness surrounding the Pt/C at ultralow Pt loading (e.g., 0.05 mg cm- 2), and contributes to a maximum 70.2% voltage increase at 3 A cm- 2, indicating that the GAI4CES is powerful in promoting the development of advanced low-cost and high-performance fuel cells.

Efficient gas transport and abundant triple-phase boundaries (TPBs) are vital for thick cathode catalyst layers (CCLs) in proton exchange membrane fuel cells (PEMFCs), yet remain challenging to realize. In this work, we introduce a carbon-based nanotrap architecture functionalized with pyrrolic-N groups, which reorganizes the Pt-carbon-ionomer interface to enhance local oxygen supply. These nanotraps simultaneously confine Pt nanoparticles and ionomer, forming continuous pathways for oxygen, protons, and electrons, thereby significantly increasing active TPB density. The underlying enhancement mechanism is validated by x-ray tomography and 3D two-phase flow simulations. Using a 15.5 µm-thick CCL, the optimized electrode achieves peak power densities of 1940 mW cm-2 in H2/O2 and 1410 mW cm-2 in H2/Air-improvements of ∼30% and ∼80%, respectively. Moreover, it exhibits a good stability, with a voltage decay rate of only 43.8 µV h-1 at 1.5 A cm-2 over 1000 h. This nanotrap concept offers a versatile interfacial design strategy for advanced gas-diffusion electrodes in energy conversion technologies.

High-temperature proton exchange membrane fuel cells (HT-PEMFCs) are promising energy-conversion devices owing to their simplified water management and improved tolerance to carbon monoxide in reformate H2. However, their practical performance and durability are constrained by anode carbon monoxide (CO) poisoning, cathode catalyst utilization, and long-term degradation of phosphoric-acid-doped polybenzimidazole (PBI) membranes. In this study, a three-dimensional, non-isothermal numerical framework was established to investigate the coupled effects of CO contamination at the anode, cathode platinum loading, and phosphoric acid doping level degradation on HT-PEMFC performance. The proposed model explicitly describes the competitive CO/H₂ adsorption and electrochemical oxidation processes on anode platinum sites, employs an agglomerate-based formulation to represent oxygen reduction kinetics and oxygen transport in the cathode catalyst layer, and incorporates the conductivity decay of PBI membranes associated with phosphoric acid loss during prolonged operation. The model predictions were validated against experimental polarization curves obtained under CO-containing anode fuel conditions. The results indicate that CO poisoning causes pronounced performance deterioration, particularly at elevated CO concentrations, lower operating temperatures, and medium-to-high current densities, while increasing the temperature partially alleviates the anode poisoning effects. A higher cathode platinum loading improves the cell output by enhancing the electrochemical reaction kinetics, but concurrently strengthens the oxygen transport limitations within the catalyst layer. Moreover, progressive phosphoric acid depletion leads to a substantial reduction in the membrane ionic conductivity and a corresponding increase in ohmic losses, and after approximately 6,000 h of operation, severe doping loss renders the fuel cell nearly inoperable. These results demonstrate that coordinated optimization of the operating temperature, cathode platinum loading, and membrane durability is essential for achieving stable and efficient HT-PEMFC operation with CO-containing fuels.

Developing large-area flow fields is critical for high-power proton exchange membrane fuel cells (PEMFCs), but compression-induced gas diffusion layer (GDL) deformation poses severe challenges to mass transport. This work combines finite element analysis (FEA) and computational fluid dynamics (CFD) to study the coupled effects of GDL deformation and flow field design. A three-dimensional two-phase PEMFC model is employed with real GDL deformation under a multi-branch flow field. Results reveal that compression causes remarkable non-uniform GDL deformation, degrading mass transport and cell performance. At 2.1 A cm−2, power density is reduced by 8.38%, with concentration loss dominating performance degradation especially at high current densities. Increasing reactant stoichiometric ratio and operating pressure effectively alleviate performance loss. Moreover, flow field structure significantly influences performance: reducing branching divisions enhances flow uniformity, lowers pressure drop, and improves output. This study offers practical guidance for the design and operation optimization of large-area PEMFCs.

Reliable cold starts under subfreezing conditions are essential for the commercialization of proton exchange membrane fuel cell (PEMFC) systems. This study investigates the cold start performance of a 2 kW PEMFC stack at ambient temperatures ranging from -10 degrees C to -30 degrees C using a voltage-controlled approach. Comprehensive experimental analysis reveals that while the ice-melting threshold is reached within approximately 25 s at -10 degrees C, stable operation is achieved after further temperature rise. Attempts at -20 degrees C and lower fail due to end-cell freezing, voltage reversal or insufficient heat generation. Results show that lower voltage leads to faster warm-up but also increases water production and icing risks, highlighting a trade-off between startup speed and reliability. Cold start failures exhibit two distinct temperature-dependent mechanisms. At -20 degrees C, partial icing in end-cells still permits limited reactant transport, which promotes voltage reversal and triggers protective shutdown. In contrast, at -25 degrees C and -30 degrees C, rapid and severe ice blockage suppresses current generation before reversal develops, resulting in current stagnation and thermal failure. End-cell freezing disrupts voltage consistency among individual cells and plays a critical role in determining the cold start viability. Additionally, post-start analysis suggests that implementing controlled load strategies and protective measures can minimize performance degradation. This study provides valuable insights into failure mechanisms and identifies key design and control strategies essential for achieving reliable cold starts in extreme climates.

Proton exchange membrane fuel cells (PEMFCs) encounter significant challenges during cold start at subzero temperatures due to water freezing and ice blockage. This study experimentally investigates the impact of cathode air starvation on cold start behavior of a 12-cell PEMFC stack at different subzero ambient temperatures (-10, -20, and -25 degrees C). Cathode air starvation is induced by reducing the cathode stoichiometric ratio to either 1.2 or 0.9 during startup, monitoring the voltage, temperature, and internal resistance responses. Results demonstrated successful cold starts at -10 and -20 degrees C, with startup times between 30 and 47 s under both the starvation conditions, facilitated by additional heat generation from exothermic reactions. However, at -25 degrees C, rapid ice accumulation caused voltage reversals and hence stack failure under both the starvation conditions. The mechanistic observations indicate that the coupled effects of starvation-induced self-heating, localized ice formation, and subsequent melting govern the temperature-dependent voltage divergence and recovery behavior. Post-test analysis showed minimal degradation following successful cold starts, whereas severe cold start failures resulted in slight catalyst-layer damage and increased internal resistance. This study provides experimental evidence supporting controlled cathode air starvation as a promising strategy for improving PEMFC cold start capabilities, while defining its operational limits at extremely low temperatures.

The accelerated degradation of the cathode catalyst layer (CCL) under dynamic operating conditions is a primary factor limiting the durability of proton exchange membrane fuel cells (PEMFCs). However, the coupled feedback mechanisms between microscopic structural evolution and macroscopic performance remain insufficiently understood. In this study, a dual-scale coupled model was developed, integrating a 1D CCL degradation model with a 3D PEMFC multiphysics performance model to quantitatively describe the structural evolution and its subsequent impact on cell performance under voltage cycling. At the microscopic scale, the 1D model accounts for critical degradation processes, including platinum oxidation, dissolution, Ostwald ripening, and carbon corrosion. At the macroscopic scale, the model couples multi-component gas transport, electrochemical reactions, and electronic/protonic conduction. Based on this model, the effects of operating conditions and CCL structural parameters on performance degradation were systematically analyzed. The results indicate that 80°C is the optimal temperature to balance reaction kinetics with microstructural stability. The upper potential limit (UPL) exerts a decisive influence on cell lifetime; elevating the UPL to 1.1 V exacerbates platinum dissolution and redeposition, thereby accelerating performance decay. Furthermore, low platinum loading intensifies mass transport polarization, particularly in the high-current-density region, where non-uniform degradation of the local microstructure leads to more severe performance losses. This study elucidates the degradation mechanisms under multiphysics fields, providing a theoretical foundation for optimizing CCL design and extending the operational lifetime of PEMFCs.

The performance of proton-exchange membrane fuel cells (PEMFCs) is constrained by coupled transport and reaction processes within the catalyst layer (CL). In this study, oxygen mass transfer in reconstructed porous CL structures was analyzed using the lattice Boltzmann method (LBM) and the quartet structure generation set (QSGS). A three-dimensional microstructure was reconstructed by prescribing phase fractions and growth probabilities. Pt agglomeration, carbon-support corrosion, and ionomer degradation were represented to analyze morphological evolution and its influence on transport. The results reveal that integrated degradation triggers a complex, non-monotonic evolution of mass transfer resistance (Rcl). In the early stages, the thinning of the ionomer film and initial structural reconfiguration shorten the gas diffusion pathways, leading to a transient reduction in Rcl. However, as degradation intensifies, the massive loss of electrochemical active surface area (ECSA), the structural collapse of the carbon skeleton, and the diminished proton conductivity of the ionomer phase collectively lead to a non-linear spike in Rcl. The resulting structure-transport mapping provides a pore-scale basis for durability-oriented catalyst layer design.

Large-scale groove-type flow fields are critical for high-power proton exchange membrane fuel cells (PEMFCs), yet their design must account for both cooling water management and assembly-induced compression. In this study, a three-dimensional, two-phase PEMFC performance model is developed, integrating finite element method (FEM)-based gas diffusion layer (GDL) deformation due to clamping force and coolant boundary conditions. The effects of thermal boundary conditions, assembly compression, and cathode flow field configurations-including channel layout, flow direction, inlet temperature, and velocity-on cell performance and voltage losses are systematically investigated. This study quantifies the concentration, activation, and ohmic losses and identifies the dominant mechanisms leading to performance variation. Results show that compression primarily aggravates concentration loss by reducing GDL porosity and permeability, whereas realistic coolant boundaries elevate ohmic loss via membrane dehydration, though the elevated cell temperature induced by such boundaries slightly reduces activation loss. Increasing coolant inlet temperature reduces output voltage, mainly through enhanced concentration and ohmic losses, while higher coolant velocity improves performance by mitigating membrane dehydration. Among flow field layouts, increasing the number of channel branches induces severe reactant maldistribution, leading to up to 42.2% performance reduction. These findings provide critical insights into the coupling among mechanical compression, thermal management, and cathode flow field design, and offer quantitative guidance for optimizing large-area PEMFCs operated at high current densities.

This study aims to improve cathode gas distribution uniformity in a U-type 140-cell PEMFC stacks by developing an improved experimentally-assisted flow network method (IFNM) that combines computational efficiency with accuracy. The proposed improved flow network method (IFNM) replaces the traditional empirical correction for straight channels with a porous-medium pressure-drop model, providing a feasible framework for describing complex flow field resistance. An experimental-assisted explicit vapor-generation model is proposed and incorporated to account for the influence of electrochemical water production on gas distribution. The porousmedium parameters are experimentally identified from measured flowrate-pressure drop relationships and further validated under multiple operating conditions to ensure the reliability. Comparative results show that IFNM achieves less than 5 % deviation from three-dimensional CFD predictions of flow distribution while offering a two-order-of-magnitude reduction in computational time. Moreover, the IFNM introduces a geometrybased domain partition to distinguish bridge and reaction regions within the flow field, enabling a full-factor analysis of manifold cross-sectional and bridge geometric effects on distribution uniformity. The results demonstrate that manifold geometry is the dominant factor on cathode maldistribution, with the cross-sectional length and width showing strong positive correlations - simultaneous 20 % increases in both dimensions reduce the maldistribution indicator from 7.69 % to 3.64 %.

Rapid and accurate prediction of multi-physics fields is crucial for the performance optimization and health management of proton exchange membrane fuel cells (PEMFCs). While digital twin (DT) technology has emerged as a promising solution, existing models predominantly address simplified typical unit geometries. This study develops a high-fidelity, 3D multi-field DT surrogate model for a serpentine-channel PEMFC. The approach integrates a validated two-phase, non-isothermal computational fluid dynamics (CFD) model with a machine-learning-assisted reduced-order model. A snapshot database comprising 40 operating conditions is constructed using orthogonal experimental design and random sampling. Proper orthogonal decomposition (POD) is applied to extract dominant modes from the relative temperature, membrane water content, and liquid water saturation fields. Nine validation cases are projected onto POD modes to obtain reference coefficients. A fully connected feedforward artificial neural network, trained with 40 snapshots and monitored by the 9 validation cases, is established for each retained mode coefficient to map operating conditions (operating temperature, cathode relative humidity, cathode stoichiometric ratio) to the reduced-order representation. In-situ 3D field construction is then achieved by linear combination. The DT achieves global relative deviations of 0.52%, 4.38%, and 4.82% for the relative temperature, membrane water content, and liquid water saturation fields, respectively, and predicts cell voltage with an R2 of 0.999. Compared with a single CFD simulation requiring three days, the DT model delivers field prediction within one second. The proposed framework extends high-fidelity 3D DT modeling to serpentine flow fields and provides a validated foundation for in-situ monitoring and control of PEMFCs.

Proton exchange membrane fuel cells (PEMFCs) are promising power sources for clean transportation and distributed generation due to their high efficiency and zero emissions. However, reliable self-start at subzero temperatures remains a major barrier to large-scale deployment in cold regions. During cold start, heat generation facilitates stack warming, while ice formation within the membrane electrode assembly (MEA) blocks reactant pathways, leading to voltage collapse and irreversible damage. In this study, a one-dimensional, stack-scale transient model is developed and experimentally validated at-10 degrees C. The model considers the electrochemistry, multiphase water transport, phase change, and heat transfer, enabling quantitative analysis of cold start dynamics under realistic conditions. Parametric studies are conducted to assess the effects of initial membrane hydration, maximum current density, current ramp rate, and current loading strategies. The results show that maximum current density acts as a tipping point for the trade-off between heat generation and ice blockage. Ramp rate governs the transient voltage response and the time duration of ice accumulation. Initial hydration serves as a bifurcation parameter, where both insufficient and excessive water content shrink the feasible startup window. By integrating these factors, a cold start strategy map is established to visualize the transition boundaries between success, partial failure, and complete failure at the stack level. The map provides direct design guidelines for robust and energy-efficient cold start control, bridging mechanistic insight with practical application. The proposed framework establishes mechanistic guidelines for selecting energy-efficient and durable startup protocols, supporting fuel-cell-powered vehicles and stationary systems operating in cold climates.

Proton exchange membrane fuel cells (PEMFCs) provide great efficiency and zero-pollution, making them intriguing alternatives to traditional internal combustion engines. Optimizing the global and local performance of PEMFCs under varied operational conditions remains an important issue. However, limited by the computing efficiency of the high accuracy 3D computational fluid dynamics (CFD) model, present researches focus on either the detailed multi-physics fields or wide operational condition ranges. In this paper, to address both aspects, an optimization framework, combining the CFD method, digital twin (DT) technology, and optimization model, is proposed. Operational parameters, including temperatures, relative humidity values, pressures, and stoichiometric ratios, are selected as variables to optimize PEMFC performance. The 3D CFD model is utilized to generate snapshots as the training set for DT part, in which proper orthogonal decomposition - machine learning/interpolation models are used to construct data-driven surrogate models to predict multi-physics fields. Through three-step optimization, the optimal operational conditions for the studied PEMFC are obtained: T = 68.7 degrees C; RHA/C = 0/12.6 %; pA/C = 1.7/1.5 atm; StA/C = 1.34/3.00, resulting in a score of 98.12. This methodology could be further expanded to encompass more intricate fuel cell configurations or other energy-related applications.

Enhancing the cold start performance of short stacks at low temperatures, which is largely attributed to their significant end-plate effect, proves essential for advancing commercialization. The constant current control strategy remains the most prevalent approach during startup operations. In this study, a 12-cell short stack was employed to investigate experimentally the effects of ambient temperatures and current load rates on cold start behavior. Results demonstrate that decreasing ambient temperatures intensifies the end-plate effect, while implementing lower current load rates effectively prolongs cold-start duration. At elevated ambient temperatures, the generated heat predominates over potential freezing risks for successful rapid startups. The currentcontrolled cold start process can be divided into three stages: initial performance recovery, moderate membrane water adsorption, and icing onset, which are primarily associated with the hydration state of the membrane electrode assembly. Variations in high frequency resistance evolution and reverse polarity characteristics emerge across these stages. Membrane electrode assembly damage induced by reverse polarity and icing leads to substantial increases in both high frequency resistance and membrane resistance, accompanied by a marginal reduction in the roughness factor of catalyst layer. The protective voltage strategy sustains stable stack performance, demonstrating high consistency at -10 degrees C. However, significant performance degradation manifests below -20 degrees C, particularly near end-plates, emphasizing the critical requirement for maintaining internal thermal uniformity under extreme low-temperature conditions.

Nowadays, with the accelerating development of the hydrogen industry, the analysis of proton exchange membrane fuel cell (PEMFC) attracts wide attention. Based on the research process of the authors’ team, the numerical models for PEMFC multi-physics analyses can be classified into three generations according to different thermal and liquid water assumptions. In this paper, the three generation numerical models are compared and analyzed under identical conditions. For the polarization curve, results suggest that under the studied condition, the simulating polarization curve using the third generation model fits best compared with the experimental one. Within the activation polarization control region, the three models result in nearly identical voltage. Within their respective ohmic polarization control regions, the three models result in similar voltage with a maximum relative deviation of 2.2

Air-cooled proton exchange membrane fuel cell stack's performance can be greatly affected by the arrangement and operational characteristics of the fans that generate the air flow. However, the influence mechanisms are complex and poorly understood. In this study, a one-dimensional analytical model specifically for air-cooled stacks integrating important fan features is developed to quickly (e.g. within 5 min for a 100-cell stack) and accurately predict the water and thermal states inside stacks and stack performance under various fan arrangements and operational characteristics. The model accuracy was validated against the experimental results of a 95-cell air-cooled stack, including the stack polarization curves, temperature distribution and air flow rates at different fan's duty cycles. The model results showed that for high air flow resistance stacks, the air flow rate and maximum net power for fans arrangement in series are higher than those in parallel by about 10% and 0.5%, respectively. Moreover, integrating a single large fan can be better than two small fans as the former induces a higher air flow rate and static pressure. On the other hand, altering the air flow channel length or depth changes the air flow resistance in stacks, leading to a shift of the working point for a fixed fan at a certain duty cycle and greatly influencing stack performance. This one-dimensional air-cooled stack model can be a powerful tool to facilitate stack design and optimization.

Operating temperature is an important factor that affects the efficiency, durability, and safety of proton exchange membrane fuel cells (PEMFC). Thus, a thermal management system is necessary for controlling the appropriate temperature. In this paper, a novel thermal management system based on two-stage utilization of cooling air is first established, whose core characteristic is utilizing the temperature difference between the cooling air leaving the main radiator and the auxiliary radiator. The novel thermal management system can reduce the parasitic power of the fan by 59.27% and improve the temperature control effect to a certain extent. The traditional feedforward decoupling control based on system identification is first adopted to control the temperature and surpasses dual-PID on all the 5 indexes, which are Integral Absolute Error Criterion (IAE) of temperature difference, IAE of inlet coolant temperature, parasitic power of fan, average overshoot of temperature difference and average overshoot of inlet coolant temperature. The multi-objective decoupling control based on multi-objective optimization is then proposed to further improve the temperature control effect on the basis of traditional feedforward decoupling control. The above 5 indexes are chosen as the optimization objectives, the decoupling coefficients are chosen as the decision variables, and the Pareto set is obtained by NSGAⅡ and NSGAⅢ. The results show that the proposed multi-objective decoupling control has the main advantages as follows: (1) It can provide comprehensive optimization options for different design preferences; (2) It can significantly optimize a certain objective while other objectives are not too extreme; (3) It has the ability to surpass traditional feedforward decoupling control on all the 5 indexes; (4) It does not rely on the system identification.

AbstractImproving the performance of proton exchange membrane fuel cells (PEMFCs) requires deep understanding of the reactive transport processes inside the catalyst layers (CLs). In this study, a particle-overlapping model is developed for accurately describing the hierarchical structures and oxygen reactive transport processes in CLs. The analytical solutions derived from this model indicate that carbon particle overlap increases ionomer thickness, reduces specific surface areas of ionomer and carbon, and further intensifies the local oxygen transport resistance (Rother). The relationship between Rother and roughness factor predicted by the model in the range of 800-1600 s m-1 agrees well with the experiments. Then, a multiscale model is developed by coupling the particle-overlapping model with cell-scale models, which is validated by comparing with the polarization curves and local current density distribution obtained in experiments. The relative error of local current density distribution is below 15% in the ohmic polarization region. Finally, the multiscale model is employed to explore effects of CL structural parameters including Pt loading, I/C, ionomer coverage and carbon particle radius on the cell performance as well as the phase-change-induced (PCI) flow and capillary-driven (CD) flow in CL. The result demonstrates that the CL structural parameters have significant effects on the cell performance as well as the PCI and CD flows. Optimizing the CL structure can increase the current density and further enhance the heat-pipe effect within the CL, leading to overall higher PCI and CD rates. The maximum increase of PCI and CD rates can exceed 145%. Besides, the enhanced heat-pipe effect causes the reverse flow regions of PCI and CD near the CL/PEM interface, which can occupy about 30% of the CL. The multiscale model significantly contributes to a deep understanding of reactive transport and multiphase heat transfer processes inside PEMFCs.

The temperature control of the air-cooled proton exchange membrane fuel cell (PEMFC) is important for effective and safe operation. To develop a practical and precise controller, this study combines the Radial Basis Function (RBF) neural network with Back Propagation neural network adaptive Proportion Integration Differentiation (BP-PID), and then a metaheuristic algorithm is used to optimize the parameters of RBF-BP-PID for further improvement in temperature control. First, an air-cooled PEMFC system model is established. To match the simulation data with the experimental data, Teaching Learning Based Optimization–Differential Evolution (TLBO-DE) is proposed to identify the unknown parameters, and the maximum relative error is <3.5 %. Second, RBF neural network is introduced to identify the stack temperature and provide the accurate ∂y(k)∂u(k) for BP-PID, which solves the problem of using sign function sgn(∂y(k)∂u(k)) to approximate the ∂y(k)∂u(k) in BP-PID. Regarding the temperature control of air-cooled PEMFC, several controllers are compared, including PID, Fuzzy-PID, BP-PID and RBF-BP-PID. The proposed RBF-BP-PID achieves the best control effect, which reduces the integrated time and absolute error (ITAE) by 3.4 % and 15.8 % based on BP-PID in the startup and steady phases, respectively. Since the ∂y(k)∂u(k) provided by RBF changes softly and continuously during the control process, the parameters self-tuning ability of RBF-BP-PID is better than BP-PID. Third, to improve the control effect of RBF-BP-PID further, TLBO-DE is adopted to optimize the parameters of RBF neural network and BP neural network.