Sodium-ion batteries (SIBs) are widely regarded as a promising alternative to lithium-ion batteries because of their low cost and the abundance of sodium resources. However, although SIBs share similar electrochemical operating principles with lithium-ion batteries, differences in battery materials and operating voltage windows hinder the direct transfer of charging strategies developed for lithium-ion systems. Therefore, to simultaneously achieve high charging rates and enhanced safety, a comprehensive experimental investigation combined with numerical simulations is conducted to analyze the fast-charging behavior of SIBs and to systematically develop an optimized charging strategy. Experimental results demonstrate that heat generation during the fast-charging process of SIBs can be divided into three distinct stages: an initial rapid temperature rise, a subsequent moderate heating stage, and a final sharp temperature increase. Based on these thermal characteristics, an optimal charging strategy incorporating multi-stage charging is developed using an adaptive particle swarm optimization (APSO) algorithm. Experimental validation indicates that, compared with the conventional 3C constant-current charging protocol, the proposed strategy reduces the maximum temperature rise by 2.32 degrees C. Furthermore, the effectiveness of the optimized strategy is corroborated by numerical modeling results. Overall, this study provides a feasible and effective approach for improving fast-charging safety and enhancing thermal management performance in sodium-ion batteries.

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.

All-soluble all-iron flow batteries are considered a promising technology for low-cost and large-scale energy storage. In the past few years, efforts have been taken to design various iron chelates to enhance the cycling stability of negative electrolytes, while ignoring the kinetic mismatch and the corresponding battery design strategies, which greatly limited the performance of all-soluble all-iron flow batteries. In this regard, combining experimental analysis and numerical simulation, this work analyzed the kinetic performance of iron chelates on the negative side and conducted further investigations on battery asymmetric structure design to balance mass transport and electrochemical reactions within the battery. Results show that the reaction rate constant of the Fe(II)(BIS - TRIS)2- /Fe(III)(BIS - TRIS)- redox couple is 1.48 & times; 10-5 cm s- 1, significantly lower than that of the positive electrolyte (6.0 & times; 10-5 cm s- 1), which limits the performance of the battery. Utilizing an asymmetric electrode design to increase active reaction sites and enhance convection is a critical strategy in achieving balanced mass transport and reaction activity between the positive and negative electrolytes. More notably, the battery with asymmetric geometric characteristics demonstrates a remarkable energy efficiency reaching up to 80.17 % at 80 mA cm- 2, which is 7.95 % higher than that of the symmetric structure. This research provides theoretical guidance for the structural design of key components of batteries and reduces the cost of trial and error.

Efforts to increase power density, primarily motivated by the need to reduce capital cost, have become a central focus in research on redox flow batteries (RFBs). However, the increase in power density intensifies the parasitic hydrogen evolution reaction (HER) at the negative electrode, posing a significant operational challenge for RFBs. Conventional engineering countermeasures frequently involve lowering the charging cut-off voltage to mitigate HER. However, this approach reduces electrolyte utilization, ultimately leading to increased electrolyte costs. To enable more rigorous control strategies and elucidate the mechanistic basis of hydrogen evolution, a three-dimensional model incorporating HER phenomena in vanadium RFBs is developed in this work. The numerical simulation and accompanying experiments indicate that HER displays pronounced spatial heterogeneity across the porous electrode, producing hotspots for gas formation and accumulation, where active-species concentrations are low and under-rib convection is weak. Enhancing mass transport and improving the uniformity of active-species distribution are shown to substantially mitigate HER. We find that increasing the electrolyte flow rate from 1 to 11 mL min(-1) cm(-2) reduced the hydrogen gas fraction within the electrode from similar to 1.2% to similar to 0.4% while raising state of charge (SOC) from 0.8 to 0.9, a strategy that suggests simultaneous improvement of SOC and suppression of the HER side reaction is achievable, but this method incurs higher pumping losses. Moreover, HER is sensitive to other operating conditions (e.g., current density and cut-off voltage), which similarly imply trade-offs between instantaneous power density, electrolyte utilization, and parasitic losses. Collectively, the model and supporting experiments provide mechanistic insight into HER behavior during RFB operation, offering guidance to mitigate these trade-offs and minimize parasitic reactions, thereby enhancing overall system efficiency and durability.

Bubble retention caused by the high viscosity of gelled propellants poses a significant challenge to combustion stability and performance in propulsion systems. To address this, the present study investigates the dynamic behavior of a single bubble in shear-thinning gelled propellants flowing through corrugated channels. Numerical simulations are conducted employing the Volume of Fluid (VOF) approach, with a modified Carreau-Yasuda model applied to represent the non-Newtonian viscosity characteristics. The effects of channel geometry, temperature, and inlet velocity on bubble dynamics and apparent viscosity are analyzed. The results indicate that bubble velocity is highest in trapezoidal channels, followed by sinusoidal and smooth channels. Increasing the corrugation amplitude enhances bubble speed, while higher temperatures reduce it. At low inlet velocities, the bubble maintains its shape; at moderate velocities, it deforms and recovers; and at high velocities, it splits. These findings provide valuable insights into bubble behavior in gelled propellants and contribute to the optimization of propulsion system design.

Two-phase loop thermosyphon is widely used in heat dissipation field because of its simple structure and excellent performance. However, the issue of local drying in loop thermosyphon under high heat flux conditions with low filling ratios imposes a limitation on its critical heat flux. At low heat flux, the thermal resistance of loop thermosyphon with high filling ratios is significantly greater, and as the heat flux increases, geyser boiling gradually emerges. Consequently, enhancing the heat transfer performance of a two-phase loop thermosyphon across a broad range of liquid filling ratios is crucial for expanding its application scenarios. In this paper, the heat transfer characteristics of hierarchical gradient mesh surfaces evaporator loop thermosyphon (HGMSE-TPLT) with different filling ratios are studied experimentally, and compared with smooth surface evaporator loop thermosyphon (SSE-TPLT). The results show that the hierarchical gradient mesh surfaces evaporator can improve the local drying phenomenon at low filling ratios (30 %) and medium filling ratios (60 %), and reduce the temperature fluctuation by 6.34 degrees C and 2.57 degrees C, respectively, but it is close to the smooth surface at high heat flux. The performance of high-filling ratio HGMSE-TPLT improved significantly, the critical heat flux increased to 450 W/cm(2), evaporator thermal resistance decreased by 48.8 %, temperature fluctuation decreased by 3.98 degrees C. The basic mechanism of enhanced heat transfer and stability in hierarchical gradient mesh surfaces evaporator is analyzed based on visual images. Sufficient bubble nucleation sites and stable bubble overflow channels are critical factors for enhancing its performance.

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.

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.

Gelled fuels have gained attention for their enhanced safety and higher energy density in aerospace propulsion, yet their complex rheological behavior poses challenges for atomization modeling and performance prediction. This study develops a constitutive model for thixotropic organic kerosene gel and integrates it with a VOF-LESbased numerical framework to investigate the primary atomization characteristics under varying inlet conditions and Reynolds numbers. The simulations explore breakup mechanisms, jet penetration, fuel-air mixing, and atomization efficiency, with conventional kerosene as a baseline for comparison. Under uniform inlet conditions, atomization is predominantly driven by the frontal impact between the jet tip and the quiescent ambient gas, leading to lateral dispersion and front-end ligament breakup with relatively stable liquid core structures. In contrast, turbulent inflows introduce strong perturbations and turbulent kinetic energy, intensifying surface instabilities, promoting earlier and more complete breakup, and producing more numerous and smaller droplets. Increasing Reynolds number enhances inertial forces and weakens the suppressive effects of gel rheology, though efficiency gains diminish at high values due to energy dissipation and droplet coalescence. Compared to conventional kerosene, gel fuels demonstrate lower atomization efficiency and mixing due to higher viscosity and internal structural resistance. These findings offer valuable insights into the atomization dynamics of gelled propellants, supporting improved injector design and combustion performance optimization.

Water in Proton Exchange Membrane Fuel Cells (PEMFCs) holds significance and complexity. The study of water is crucial for enhancing the efficiency and extending the lifespan of the batteries. This study used micro-CT technology to obtain tomographic images of the gas diffusion layer (GDL) in PEMFCs. Subsequently, the samples were reconstructed in three dimensions using Avizo, and the internal fluid flow in the GDL was simulated using the Volume of Fluid (VOF) method. The local and average porosities of all sample types were calculated, providing insight into the distribution of the internal pore structure of the GDL. Analyzed the impact of the pressure difference at the inlet and outlet(Delta P), contact angle, and the thickness of the model on the flow of liquid. The research results indicate that for the TGP-H-60 model of GDL, the Delta P must be at least 6 kPa to allow liquid water to flow from one end to the other. The contact angle within the GDL significantly impacts the removal of liquid water. In practical applications, the selection of GDL thickness must ensure mechanical strength while also considering fluid transport efficiency to enhance battery performance. An overly thick can make the flow of water more difficult, resulting in flooding phenomena.

Thermal runaway (TR) is a severe challenge to the widespread commercial adoption of high energy-density lithium-ion batteries (LIBs). Nonetheless, the current strategies lack responsiveness for both extreme heat dissipation and explosion suppression. Here, a thermal safety protection strategy based on liquid immersion cooling (LIC) is proposed. The peak temperature of overcharge-induced TR is decreased below 300 degrees C through boiling heat exchange of FS49, rapidly (3 min) stabilizing the LIB temperature around 49 degrees C. Simultaneously, critical radicals are captured by FS49 in the combustion chain reaction, reducing emissions of combustible toxic gases by approximately 62.65%. This effectively prevents LIB explosions and secondary re-ignition disasters. Surprisingly, when applied as 1 mm interlayers between cells for a pack with four LIBs, the FS49 not only suppresses the TR propagation but also maintains the adjacent LIB temperature at 53.89 degrees C. Additionally, it is further demonstrated that the thermal safety of a large-scale 36-cell LIB pack through finite volume method simulations. This strategy can represent a critical step forward in enhancing the safety performance of electric vehicles and grid-scale energy storage systems.

To meet the increasing demands for high energy density and safety in aerospace propulsion, gel propellants have drawn significant attention due to their dual solid-liquid characteristics. However, their high viscosity and complex rheology suppress interfacial instabilities and hinder jet breakup, posing critical challenges to efficient atomization. This study develops a high-fidelity numerical framework coupling the VOF method, large eddy simulation (LES), and adaptive mesh refinement (AMR), incorporating a thixotropic shear-thinning model based on experimental data for 5 % organic kerosene gel. Considering the strong mechanical vibrations in rocket engines, the primary atomization behavior under periodic perturbations with varying amplitude and frequency is systematically investigated, with a focus on jet evolution, breakup mechanisms, droplet characteristics, and mixing efficiency. Under high-speed injection, periodic forcing intensifies upstream core instability and promotes downstream fragmentation, generating numerous fine droplets. Increased perturbation amplitude enhances radial spreading and multiscale breakup, while frequency primarily adjusts droplet size uniformity but contributes little to penetration, with a saturation effect observed. Despite large number of small, slow droplets are generated near the jet core and upstream, the jet remains largely dominated by an unbroken liquid core and ligaments. Compared to kerosene, the gel exhibits significantly poorer atomization performance, producing larger, slower, and more localized droplets due to its rheological resistance to instability growth and spatial dispersion. This work provides quantitative insights into the atomization dynamics of gel fuels and establishes a theoretical foundation for rheological control strategies in propulsion applications.

Mass transport polarization induced by water blockage within the gas diffusion layer (GDL) of proton exchange membrane fuel cells (PEMFCs) constitutes a critical bottleneck limiting high-current-density performance. This study establishes a model of water invasion under compression and thickness changes, facilitated by in-situ X-ray computed tomography (X-CT) imaging and finite element modeling based on realistic geometric structures. Subsequently, it investigates the gas transport under liquid saturation. Extraction of the pore network model (PNM) reveals that both compression and water flooding significantly reduce the mean pore diameter, while exerting minimal impact on the coordination number. The results indicate that liquid water transport pathways exhibit scale-dependent characteristics. Furthermore, the effective diffusion coefficient (EDC) and permeability (K) vary linearly with thickness. Notably, 40 % compression causes an 80 % reduction in permeability, while significantly mitigating the pressure drop phenomenon. This work provides multiscale insights into mass transport limitations across various porous media.

Reconstructing flow fields from sparse observations constitutes a fundamental challenge in aerodynamic analysis, since existing methods often fail to reconcile accuracy with computational efficiency under severe data constraints. We introduce a Graph Transformer Reconstruction Network (GTREN), a novel model designed to infer latent data patterns from extremely sparse measurements and to reconstruct high-fidelity, full-field flow solutions. The approach represents discrete grid points as nodes within a graph data structure, enabling a principled encoding of spatial relationships. We integrate a Transformer-style attention mechanism into the graph network's message-passing operations, allowing the model to selectively emphasize salient neighbor interactions while attenuating irrelevant signals; this facilitates efficient and stable information propagation. After training, GTREN uncovers implicit correlations among sparse sensor measurements and successfully generalizes these relationships to reconstruct the entire flow field. Results indicate that, with only ten measurement points, the model accurately reconstructs both pressure and velocity distributions across the field. When compared to Computational Fluid Dynamics simulations, the mean squared error (MSE) of the reconstruction is as low as 0.05%. By contrast, a conventional graph neural network yields an MSE of 18.31%. We also provide a systematic analysis of how both the number and spatial arrangement of measurement points affect reconstruction accuracy. Our analysis reveals that, rather than merely increasing sensor count, strategically locating sensors within the core vortex-shedding region of the wake substantially improves reconstruction accuracy-offering practical guidance for flexible sensor deployment. Overall, GTREN achieves high-precision full-field reconstruction with very few pressure sensors, enabling real-time sparse-sensing flow-field monitoring.

With the rapid miniaturization and performance enhancement of electronic devices, high heat flux dissipation has emerged as a critical bottleneck restricting their reliable and stable operation. The Two-Phase Loop Thermosyphon (TPLT), as a representative passive phase-change heat transfer technology, boasts inherent advantages including a simple structure, no reliance on external power, and superior heat transfer efficiency, thus being widely recognized as a promising solution for high-heat-flux cooling scenarios. However, the complex internal gas-liquid two-phase flow mechanism leads to insufficient accuracy of existing models in predicting the coupling relationship between heat transfer performance and flow characteristics, limiting the optimized design of TPLT for target cooling scenarios. To address this gap, a one-dimensional steady-state momentum cycle model was established. Void fraction acts as the core intermediate variable linking flow and heat transfer: as heat flux density increases, void fraction gradually rises, regulating the balance between driving force and frictional resistance to drive the circulation flow rate to first increase and then decrease, while simultaneously affecting the dominant mode of heat transfer and thus the loop thermal resistance. Combined with experimental tests and visualization technology, the coupling characteristics of void fraction, circulation flow rate, flow regime, loop thermal resistance, and pressure drop were systematically analyzed under different filling ratios (30%, 60%, 80%) and heat flux densities (30 similar to 390 W/cm(2)). The results demonstrate that the proposed model can effectively predict both the flow performance and heat transfer performance of TPLT, with the predicted trends of key parameters being highly consistent with experimental data. Further visualization observations and performance analyses under different filling ratios complement the mechanism clarification. Visualization results confirm that a low filling ratio (30%) causes early fragmentation of annular flow into droplet flow, leading to premature peak flow, increased thermal resistance, and poor heat transfer stability at high heat flux. In contrast, a high filling ratio (80%) maintains a continuous liquid phase in the flow regime across the entire heat flux range, achieving the latest peak flow, the highest flow efficiency, and the lowest thermal resistance at high heat flux, proving its suitability for high heat flux dissipation. This study clarifies the intrinsic coupling mechanism between the flow and heat transfer performance of TPLT, and the established model provides a reliable tool for predicting its key performance indicators under complex working conditions, offering valuable theoretical support for the optimized design of passive cooling systems in high-heat-flux electronic devices.

Rapid thermal estimation for Multichip modules (MCMs) is crucial for structural design and optimization of electronic equipment. Although existing thermal estimation methods based on convolutional neural networks (CNN) can rapidly predict temperature distributions for various chip arrangements and power dissipations, they typically require large datasets and extensive training time, resulting in high computational and resource costs. To overcome these limitations, this paper proposes a sparse data-driven model integrating meta-learning with CNNs, specifically tailored for constructing geometry-adaptive heat transfer prediction models for MCMs, thus significantly reducing dependence on extensive training datasets. When adapting to new tasks, the proposed model requires only approximately 2 s of fine-tuning using merely 20 new data samples to achieve geometric adaptability, attaining an impressive prediction accuracy of approximately 99 %. This accuracy is comparable to conventional CNN-based surrogate models but reduces data requirements by approximately 80 %. Furthermore, the proposed model can estimate the temperature field within 10 ms, which is three to four orders of magnitude faster than traditional numerical simulations. The results demonstrate the model's significant potential for efficient few-shot multitask learning in thermal estimation scenarios, substantially improving the utilization efficiency of historical MCM heat transfer datasets and effectively supporting real-time thermal estimation and rapid optimization of chip configurations.

The continuous increase in chip power density has posed severe challenges for thermal management in strongly nonlinear conjugate heat-transfer systems. Reinforcement learning (RL), as a model-free and adaptive optimization approach, shows great potential in this field. However, in high-dimensional, time-delayed, and nonlinear coupled thermal environments, its training process is often hindered by reward sparsity and unstable convergence, making it difficult to achieve robust training and control performance. This study first establishes the feasibility of DRL for such problems by using microchannel cooling as a testbed and implementing an open-loop Proximal Policy Optimization (PPO) scheme. To address the core difficulty arising from the delayed and nonlinear thermal responses of microchannel heat transfer, a PPO variant augmented with multi-step temporaldifference learning, n-step return, is introduced from a physical perspective. By incorporating n-step return into advantage estimation, the method explicitly accounts for the temporal coupling between fluid flow and heat transport, thereby alleviating perturbations in value evaluation caused by pronounced hysteresis and nonlinearity of the thermo-fluid state, yielding notably improved training stability and sample efficiency. The optimized agent performs favorably on benchmark tests, achieving average temperature reductions of about 6 K in twodimensional and 4 K in three-dimensional microchannels compared with an empirical manual control baseline, demonstrating strong control capability for complex heat-transfer systems. Robust closed-loop temperature regulation across a range of unseen initial conditions further indicates the agent's stability and generalization. After reward restructuring, actuation power was reduced by more than 80%, while the control quality was maintained at a comparable level. Overall, the enhanced DRL framework demonstrates stable and efficient thermal control potential for complex conjugate heat transfer, and illuminates deeper connections between physical fields and reinforcement learning, opening avenues for subsequent applications.

The two-phase flow transport mechanisms in the 3D flow field and gas diffusion layer (GDL) of proton exchange membrane fuel cells are not fully understood. In this study, the GDL structure was reconstructed using micro-CT technology, and the two-phase flow behavior in the 3D fine-mesh flow field (FMFF) and GDL was simulated and analyzed using the phase-field lattice Boltzmann method. The results show that the FMFF significantly enhances gas convection in both the flow field and GDL, reduces liquid water saturation, increases the pore area for gas transport at the interface between the two, and makes the distribution of gas and liquid in the GDL more uniform. Liquid water tends to accumulate above the air stagnation zone and below the concave baffle. Additionally, increasing the gas flow rate reduces the liquid water volume transported to the upper side of the flow field plate. Hydrophilic flow field plates facilitate liquid water transport above the plate but tend to accumulate liquid water in the flow field; conversely, hydrophobic flow field plates exhibit the opposite behavior. Comprehensive analysis reveals that flow field plates with a contact angle between 90 degrees and 110 degrees offer the most balanced performance.

The two-phase loop thermosyphon (TPLT), known for its excellent heat transfer performance, simple and compact structure, and lack of need for a pump, effectively addresses heat transfer challenges in confined spaces under high heat loads. Compared to TPLT with low filling ratio, high-filling-ratio TPLT not only exhibit a higher maximum heat transfer capacity but also has a more complex heat and mass transfer process, leading to increased sensitivity to the working fluid's properties. Therefore, studying the impact of the working fluid on the operational state of high-filling-ratio TPLTs is crucial for understanding their heat transfer mechanisms. In this paper, comprehensive experiments were conducted on TPLT filled with H2O and R134a as working fluids in a wide filling ratio range (30 % -90 %), and their heat transfer performance and flow characteristics were compared. Heat transfer diagram, two-phase flow pattern diagram, and the distribution of gas-liquid two-phase of the TPLT was established with different filling ratio and heat input. Due to differences in latent heat of vaporization, the maximum heat transfer capacity of the H2O-TPLT (390 W/cm2) is greater than that of the R134a-TPLT (270 W/cm2). In the H2OTPLT, the predominant large-volume slug flow leads to significant flow resistance. Whereas in the R134a-TPLT, the flow pattern is primarily dominated by small-volume bubbly flow and churn flow, resulting in low resistance. High viscosity and flow pattern in the H2O-TPLT cause oscillation phenomena, leading to significant temperature and pressure fluctuations. Under high filling ratio, both types of TPLT experiences geyser boiling phenomena causing periodic temperature and pressure fluctuations and flow pattern changes. In summary, R134a is the preferred working fluid when heat transfer requirements are met, as it effectively reduces temperature fluctuations while dissipating heat. When exceeding the R134a-TPLT's maximum heat transfer capacity and less stringent temperature control is acceptable, H2O may serve as the working fluid.