Abstract:The sharp decrease in automotive powertrain noise due to electric vehicles and stringent emissions regulations has highlighted wind noise and its mitigation as a key area of automotive development. Consequently， the accurate identification of aeroacoustic noise sources is critical to the improvement of vehicle acoustic comfort in an efficient and cost-effective manner， especially since many now prevalent noise sources were previously masked by powertrain noise. Beamforming techniques， which employ one or more microphone arrays to identify noise sources on a virtual plane close to the vehicle， are the current industry standard for external aeroacoustic evaluation. However， not all noise sources lie on the virtual plane defined， which can result in the misinterpretation of the intensity and location of the out-of-plane noise sources. To improve the accuracy of its beamforming techniques， Pininfarina currently utilizes a multi-plane formulation for the overhead microphone array. This involves the definition of several planes that better describe the height variation of the vehicle. The natural evolution of the multi-plane approach is to directly map the noise sources on a 3-D scan of the car. The aim of this work is to develop a new algorithm that directly represents acoustic maps on the vehicle surface by combining data from all three arrays into a single， unified map. The goal is to create a tool that minimizes errors in source localization and strength caused by focus deviations.

Abstract:This research investigates methods to reduce the operating costs of Range-Extended Electric Vehicles （REEV） at low temperatures through efficient utilization of range extender residual heat. The study presents an operating cost optimization approach incorporating range extender active battery preheating and examines its impact on low-temperature operating costs for REEVs. The research methodology involves developing a vehicle thermal management simulation model based on a commercial REEV， incorporating data from component thermal load testing and road trials. The study establishes an integrated control strategy combining thermal and energy management， utilizing a comprehensive low-temperature operating cost evaluation function that accounts for fuel consumption， electricity usage， battery degradation， and range extender active battery preheating. The effectiveness of the active preheating method was evaluated across various initial battery charge levels. Simulation results demonstrate that the integrated control strategy achieves an 8% to 17% reduction in low-temperature operating costs over four WLTC cycles compared to conventional thermostat energy management strategies.

Abstract:The dual-motor controller system is one of the key components in the multi-power architecture of new energy vehicles and is crucial to the safe operation of the vehicle. From the perspective of functional safety， this article takes the torque safety goal of "avoiding unexpected acceleration of the vehicle due to unexpected acceleration torque" as an example to explore various software and hardware architecture implementation methods， and analysis the test scheme covering multiple stages such as HIL/PowerHIL testing， bench testing and vehicle testing to verify the effectiveness of the safety path. Through the verification of the safety strategy， it provides innovative ideas for the functional safety design and testing of the electric drive control system， which is expected to provide a reference for improving the overall safety of new energy vehicles.

Abstract:The roof-mounted camera serves as a critical component in intelligent vehicle perception systems， yet its performance degradation caused by blurred vision during rainy conditions remains a significant challenge. This study investigates the dynamic characteristics of raindrops with varying diameters impacting inclined camera surfaces under both quiescent and airflow conditions， with particular focus on the influence of surface inclination angle. Experimental results demonstrate that， under no airflow conditions， the maximum spreading diameter occurs at a 70° inclination angle， where inertial forces and surface tension reach equilibrium. When the airflow velocity in the low-turbulence model wind tunnel is 40 km/h， raindrop impact velocity increases by 90% with a corresponding 25% expansion in liquid film spreading diameter. However， at a 75° inclination angle， airflow shear effects cause intensified splashing of the liquid film. Reducing the droplet diameter from 3.0 mm to 2.7 mm decreases the Weber number by 19%， consequently reducing the spreading diameter.

Abstract:To enhance the reconstruction accuracy of in-vehicle acoustic signals in automotive NVH performance development， this study presents a method based on order extraction and overlap-add. The method extracts rotational speed curves corresponding to characteristic orders in time-frequency representations by combining path tracking with short-time Fourier transform， considering the order characteristics of rotating machinery sound sources. Signal reconstruction is accomplished through waveform similarity matching within a specific rotational speed range and an overlap-add algorithm. Furthermore， psychoacoustic evaluation metrics quantify perceptual similarity using objective parameters. The proposed method demonstrates superior performance in reconstruction accuracy and perceptual consistency， establishing a novel technical approach for in-car sound signal simulation.

Abstract:Sound source identification utilizing beamforming technology represents a critical testing methodology in aero-acoustic wind tunnels. The three-dimensional beamforming algorithm， developed through modifications to the steering vector of conventional two-dimensional beamforming， demonstrates enhanced capabilities for spatial sound source identification and proves particularly suitable for multi-array configurations. This research examines the influential parameters of three-dimensional beamforming algorithms， specifically investigating sound source frequency， array quantity， microphone count， and testing distance. Results indicate that increasing source frequency or array number enhances spatial resolution while reducing dynamic range. Conversely， increasing microphone quantity or decreasing testing distance improves both spatial resolution and dynamic range. This comprehensive analysis of influence parameters in three-dimensional beamforming algorithms provides valuable insights for optimizing parameter combinations to achieve superior sound source identification outcomes.

Abstract:In contrast to steady airflow conditions in wind tunnel tests， vehicles on actual roads experience airflow with varying velocity and direction due to natural wind and wake effects from upstream vehicles. This study employs a simplified vehicle model to investigate turbulence characteristics of unsteady wind and unsteady wind noise sources at the front side window area. The research utilizes DES model， implementing both SEM and preceding vehicle method to generate unsteady wind excitation. Analysis of airflow velocity data collected in front of the model and at the front side window area revealed turbulence intensity ranging from 10% to 25%. The probability density distribution of airflow velocity and yaw angle follows a Gaussian distribution， while the power spectral density of fluctuating wind velocity corresponds to the von Karman spectrum， aligning with real road test conditions. Through integration with APE， distinct convective and acoustic pressure fluctuations at the front side window area were obtained. Results demonstrate that SEM's convective pressure fluctuation energy approximates steady wind excitation across all frequencies， with notably elevated acoustic pressure fluctuation energy between 500-2500 Hz. The preceding vehicle method shows reduced convective and acoustic pressure fluctuation energies compared to steady wind excitation due to upstream vehicle influence. Under unsteady wind excitation， the kurtosis value at the external ear position of the side window increases， contrasting with near-zero values under steady conditions.

Abstract:This research investigates the flow field structure and aerodynamic noise characteristics in the pantograph region of high-speed trains operating within tunnels. The interaction between the pantograph and surrounding air intensifies during tunnel operation due to air compression ahead of the train. While the aerodynamic characteristics of pantographs during open-air operation have been extensively documented， limited research addresses the aerodynamic drag and noise generation when trains traverse tunnels. This study employs the IDDES method to analyze the flow field around a 1/6 scale high-speed train model equipped with a pantograph. The analysis reveals the transient flow field characteristics and determines velocity and pressure distributions in the pantograph region. Through detailed sound field analysis using WE， the research establishes the sound power levels， frequency distribution， and directivity patterns in relation to the incoming flow direction. The simulation results demonstrate a strong correlation between pantograph noise distribution and the surrounding flow structure. These findings provide valuable insights for noise control strategies in the pantograph region of high-speed trains operating within tunnels.

Abstract:In gear dynamics， time-varying mesh stiffness represents a primary internal excitation source， as recognized by numerous scholars. The accurate calculation of mesh stiffness holds significant importance for analyzing vibration in helical gear systems. During operational conditions， spalling defects emerge due to wear， affecting the tooth surface contact state and causing mesh stiffness fluctuations. Therefore， establishing an analytical model for calculating time-varying mesh excitation under defective tooth surface conditions becomes essential. This paper employs tooth surface contact projection analysis to obtain helical gear tooth surface contact lines by simulating actual three-dimensional rotation angles and contact relationships. Based on this analysis， the principle of potential energy superposition is utilized to propose a method for calculating helical gear pair mesh stiffness， incorporating the effects of gear base and transition curves. In modeling mesh stiffness with irregular spalling defects， regional methods correct changes in moment of inertia and cross-sectional area of variable-section cantilever beams caused by defects， effectively describing the impact of irregular defects on time-varying mesh stiffness. Finally， a finite element method validates the analytical calculations through a contact model of gear pairs with defects.

Abstract:Interior acoustics modulation through sound field reproduction represents a crucial component of vehicular intelligent cockpits. However， challenges regarding solution accuracy and control robustness in practical applications require further investigation. To address the impact of occupant postural changes on sound field reproduction performance， this paper presents a targeted modulation strategy based on mapping relationships between facial features and binaural positions. This approach aims to optimize reproductive performance by adjusting control parameters in response to occupant postural variations. Both simulation and experimental results demonstrate that， with the main-driving and co-driving positions designated as acoustic reproduction regions， the proposed control framework effectively enhances inter-regional acoustic contrast within the target frequency range through numerical computation of subject binaural positions.

Abstract:This paper proposes a multi-scale feature fusion model based on Swin Transformer for driver pose estimation. The model enhances semantic expression capabilities through layer-by-layer downsampling， achieving an effective balance between image details and semantic features. To address pose estimation challenges in low-light driving scenarios， a teacher-student model optimization strategy is introduced， significantly improving estimation accuracy under low-light conditions. Experimental results demonstrate that the proposed model achieves an AP of 0.719 on the COCO validation set， surpassing the previous state-of-the-art performance of 0.717 for similar models. On our D1-DDB driving dataset， the model improves AP from 0.702 to 0.710 under normal lighting conditions and from 0.386 to 0.513 under low-light conditions. These results validate the effectiveness and robustness of the proposed method in real-world driving scenarios.

Abstract:Current research on vehicle aerodynamics often neglects the effects of sidewind and turbulence on the road. In this work， the influence of both sidewind and turbulent flow conditions on the aerodynamics of a square-back Ahmed model is experimentally studied. The aerodynamic drag force， surface pressure and wake are measured at different yaw angles （β∈0°,5°） and free-stream turbulence （Ix∈0%~15.4%）. At β =0° with turbulence， Cd and Cb increase monotonically with turbulence intensity， when Ix< 8.7%， Cd and Cb are lower than uniform inflow condition. At β =5°， the sensitivity of the aerodynamic force to turbulence intensity decreases. When Ix< 8.7%， little effect has been noted for Cd， but at Ix= 15.4%， Cd increases significantly. The wake length decrease monotonically with increasing turbulence intensity， and it is insensitive at β =5° compared to β =0°. The Reynolds stress are also more insensitive to the increase of turbulence intensity at β =5° compared to β =0°. This study can provide some reference for the impact of real road turbulence on vehicle aerodynamics.

Abstract:This research addresses aerodynamic drag reduction in modular flying vehicles through a novel nested configuration design that optimizes aerodynamic matching between the cabin， driving module， and flight module. CFD simulation methods were employed to analyze and compare the aerodynamic characteristics of nested and suspended configurations during flight mode. The numerical simulation results indicate that， compared to the suspended configuration flying vehicle developed by Airbus， the overall drag coefficient decreased from 0.7399 to 0.3417， representing a 53.82% reduction and demonstrating substantial improvement in aerodynamic performance. These findings provide valuable insights for configuration optimization and low-drag design of modular flying vehicles， offering significant potential for practical applications.

Abstract:Given the not exact same working conditions， there are differences between aerodynamic forces measured in wind tunnel and those on real road， so in order to get a more accurate and reliable result， it’s quite necessary to study how to clarify the differences and correct the wind tunnel test result. First， a classical wind tunnel error correction theory is introduced. Then， the theory is used to correct the measured values of aerodynamic drag of various car models in two wind tunnels. Next， the influence of the length of central moving belt on the aerodynamic forces， especially drag， was analyzed through CFD simulation. The research proves that the structures of current full-scale wind tunnels can meet the measurement requirements， the differences from real road conditions are predictable， and the data will be more reliable after correction.

Abstract:Based on the model of aerodynamic-acoustic vehicle wind tunnel， a numerical wind tunnel model including accurate geometrical and physical boundary conditions is established； applying the international DrivAer standard model， and applying the wind tunnel test and numerical wind tunnel methodology， the correlation between the quality of the flow field in the empty wind tunnel， the correlation between the numerical simulation results of the three types of back-type of the DrivAer standard model and the physical test is respectively investigated to verify the quality of the empty wind tunnel flow field， the correlation between the numerical simulation results of the three types of back-type of the DrivAer standard model and the physical test， and attempts to analyze the mechanisms for the differences， and to explore the numerical wind tunnel to explore the engineering applicability of the numerical wind tunnel. The results show that the established numerical wind tunnel can better reproduce the wind tunnel test environment， and the error of the wind resistance coefficient of the DrivAer scalar model ranges from 2.6% to 5.9%， which is closer to the test data than that of the open road calculation. The consistency and accuracy of SAWTC's numerical wind tunnel based on the wind tunnel computational domain with the physical reality are verified， which provides a reference and basis for the optimization of automotive wind tunnel test， the improvement of numerical wind tunnel method and the aerodynamic prediction.

Abstract:Under extreme and severe weather conditions， vehicles frequently experience stability loss incidents such as skidding and rollovers. In intelligent transportation systems， vehicles enhance driving stability and safety through advanced environmental perception and efficient distributed drive systems. In crosswind and rain conditions， the complex coupling effects of non-uniform transient aerodynamic loads on the vehicle body， tire-road low adhesion， and hydroplaning effects remain critical factors affecting the precise automatic/assisted driving of electric vehicles and are essential for improving driving safety. This paper examines the aerodynamic characteristics and torque allocation of electric vehicles under extreme wind and rain conditions， aiming to enhance driving stability in complex meteorological conditions.The study establishes a numerical calculation model of vehicle body extreme aerodynamics under wind and rain flow fields to investigate the quantitative effects of varying side wind angles and rainfall on aerodynamic loads and driving stability of electric vehicles. Additionally， addressing the torque allocation requirements of the four-wheel independent drive system， this research proposes a torque allocation strategy for equivalent wind and rain aerodynamic loads under distributed drive， accurately reflecting their impact on torque distribution. Finally， incorporating distributed wind and rain aerodynamic loads on four wheels， a CarSim-Simulink co-simulation platform is developed to conduct extreme aerodynamic stability simulation experiments， supporting the advancement of extreme aerodynamic characteristics and driving stability design for electric vehicles.

Abstract:Based on a turbocharged direct-injection gasoline engine， the condensation phenomenon occurring after the intercooler of a low-pressure EGR （LP-EGR） system at high exhaust gas recirculation （EGR） rates was modelled， and the effects of EGR rates and condensate generated at the intake intercooler on engine combustion and emissions were further investigated by engine bench experiments. The results show that condensate has a large impact on engine operating stability as well as on combustion， the COV increased significantly to about 12% for some cylinders at 25% EGR rate. The amount of condensate generated after the intercooler can be calculated using a multiphase flow model that includes both heat exchange and phase change components. With the combined effect of EGR and post-intercooling condensate as combustion decelerators， the peek pressure is reduced. At the same time， the introduction of EGR and the condensate entering the cylinder will reduce the in-cylinder combustion temperature， which helps reducing the emission of particulate matter and NOx. As the EGR rate increases， the CO emission first decreases and then increases， while the THC emission keeps increasing. Besides， the fuel consumption first decreases and then slightly increases as the EGR rate increases. Finally， through the design of targeted tests， it was found that the condensate generated after intercooling can have a large impact on engine stability and combustion. In particular， it was observed that the physical shape of the intake manifold resulted in a large difference in the effect of condensate on combustion between cylinders.

Abstract:The implementation of TSN technology has substantially enhanced the communication performance of automotive Ethernet， enabling it to fulfill the performance requirements of connected automated vehicles for real-time network communication with minimal jitter and packet loss rate. However， in terms of cybersecurity， the integration of TSN inevitably introduces new security vulnerabilities to the automotive network， necessitating the adoption of appropriate security measures to mitigate these risks. This research initially conducts a comprehensive analysis of current mainstream automotive intrusion detection systems and traffic analysis methodologies， proposing an intrusion detection system architecture based on traffic analysis. Subsequently， within this framework and leveraging the transmission characteristics of the IEEE 802.1AS protocol， two distinct intrusion detection methods were developed： one based on parameter monitoring and another on information entropy. Finally， following AUTOSAR deployment specifications for intrusion detection systems， both methods were implemented in the vehicle domain controller to validate their feasibility and effectiveness. The findings demonstrate that both intrusion detection methods achieve high detection success rates while maintaining minimal computational and memory resource consumption in the vehicle domain controller.

Abstract:Wheel load constitutes a critical information source for autonomous driving technology， reflecting the status of load distribution and vehicle motion. Conventional wheel force transducers （WFTs） present significant implementation challenges—including system complexity and spatial incompatibility—restricting their use to specially instrumented test vehicles. This paper introduces an innovative load-sensing unit integrated within an idler-gear speed reducer of an EDW， compatible with standard passenger car wheels. The EDW thereby achieves integration of the sensing and executing units. The research examines extreme vehicle conditions with lateral， longitudinal， and vertical wheel loads， establishing these as the working boundaries of the load-sensing unit. Within these parameters， structural parameters of the load-sensing unit serve as design variables， utilizing optimization algorithms to balance measurement sensitivity and decoupling performance. A virtual prototype facilitates CAE simulation across each signal channel of the dimensional component. Results demonstrate clear linear relationships between main signals while maintaining coupling signals within 15% of main signal amplitude， confirming successful optimization balancing sensitivity and decoupling performance.

Abstract:To enhance the acceptability of ACC while maintaining safety， ACC systems must accommodate drivers' requirements for personalized vehicle following control. This paper presents an adaptive vehicle-following control strategy that incorporates driving style recognition. Initially， a forward simulation model of the powertrain system for a PHEV is developed. Subsequently， based on comprehensive driving style evaluation， a linear mapping relationship between overall driving style scores and driving style adjustment coefficients is established. Furthermore， a desired vehicle-following distance control strategy is designed that considers the adjustment coefficient， preceding vehicle velocity， and acceleration. The effectiveness and stability of the proposed integrated driving style following control strategy are validated through offline simulations and hardware testing across three typical traffic scenarios： normal following， lateral vehicle insertion， and emergency braking of the preceding vehicle. Hardware testing results demonstrate that the proposed control strategy successfully maintains stable following distances appropriate for conservative， regular， and aggressive driving styles， while confirming the strategy's reliability and real-time performance capabilities.

Abstract:In autonomous navigation and robotics， particularly within intelligent transportation systems， efficient and precise path planning is essential for navigation through complex environments. While traditional path planning algorithms such as ACO show potential， they frequently encounter limitations in directionality and local optima challenges. This paper introduces an enhanced algorithm—ACO-ESD. Through the implementation of a Step Direction Judgement mechanism that considers pheromone concentrations， heuristic functions， and supplementary indices， the ACO-ESD algorithm significantly improves path search directionality， expedites convergence， and effectively circumvents local optima. Simulation results indicate that the ACO-ESD algorithm surpasses traditional ACO algorithms in path efficiency， accuracy， and convergence rate， offering an effective solution for path planning in complex weighted lattice maps.

Abstract:Extrinsic calibration plays a crucial role in autonomous driving， intelligent surveillance， and robotic perception. However， existing research primarily focuses on the calibration between LiDAR and pinhole cameras， while studies on LiDAR and fisheye cameras remain relatively limited. To fill this gap， this paper proposes a deep learning-based extrinsic calibration method for a LiDAR and a fisheye camera， incorporating a geometric optimization strategy as post-processing to significantly improve calibration accuracy. The proposed method first utilizes a neural network to predict the extrinsic transformation matrix， followed by geometric optimization to refine the predicted results. Experimental evaluations on 18 335 image–point cloud pairs demonstrate that the deep learning model reduces the point cloud alignment error by 91.6%. With the additional geometric post-processing， the error is further reduced by 52.4%， verifying the high accuracy， reliability， and robustness of the proposed approach in practical applications.

Abstract:Given the highly flammable and explosive characteristics of hydrogen， the safety of hydrogen fuel cell vehicles （HFCVs） remains a primary concern， with hydrogen leakage prevention representing a fundamental challenge for operational safety. This study systematically investigates hydrogen leakage diffusion mechanisms and control strategies in fuel cell vehicles. Using a domestic fuel cell passenger vehicle as the research subject， a three-dimensional physical model of the onboard hydrogen supply system was developed using computational fluid dynamics （CFD） principles through Fluent software to analyze hydrogen leakage diffusion processes. The research examines flow control effects of diversion devices， presenting two innovative deflector plate designs and analyzing their gas control mechanisms through comparative assessment. Simulation results demonstrate that Z-axis hydrogen leakage presents significantly higher safety risks compared to X and Y axes， with increased leakage volumes leading to rapid expansion of diffusion range， elevated concentrations， and heightened safety risks. The deflectors effectively channel hydrogen accumulation， facilitating transition from multi-point to single-point monitoring， thereby reducing costs. However， substantial optimization challenges remain. This research establishes theoretical foundations for hydrogen leakage prevention in HFCVs， contributing significantly to hydrogen energy commercialization advancement. Future research will incorporate environmental factors to enhance deflector designs， improving both safety and economic efficiency of hydrogen leakage control.

Abstract:The performance and lifespan of fuel cell vehicles are significantly impacted by airborne pollutants during operation. This study examines NH₃ as a specific pollutant and investigates various recovery strategies for NH₃-poisoned proton exchange membrane fuel cells （PEMFC）. Research demonstrates that fresh air purging effectively restores poisoned PEMFC functionality， albeit with gradual recovery， achieving 72.5% performance restoration within one hour. Following one hour of exposure to 5 ppm NH₃， a single polarization curve test results in 79.2% voltage recovery， indicating a strong correlation between dynamic operational conditions and performance restoration. Electrochemical impedance spectroscopy analysis reveals that NH₃ poisoning substantially affects cathode activation resistance and mass transfer resistance， while minimally impacting high frequency resistance. Furthermore， the cathode activation impedance demonstrates greater recovery potential compared to mass transfer impedance in poisoned PEMFC. These findings provide valuable insights for stable PEMFC operation in complex environments， significantly advancing the commercialization of fuel cell vehicle technology.

Abstract:The increasing adoption of hydrogen fuel cell vehicles （HFCV） has raised significant concerns regarding their safety in indoor parking facilities. Assessment of hydrogen leakage behavior under various parking scenarios is essential for the successful integration of fuel cell vehicles. Semi-enclosed garages， a subset of indoor parking facilities， have become increasingly prevalent in urban areas due to their optimal ventilation characteristics and space utilization， making them critical subjects for safety evaluation. This study examines hydrogen safety considerations for HFCV in semi-enclosed parking environments. CFD simulations were employed to analyze the impact of leakage conditions and ventilation on hydrogen dispersion patterns. Results demonstrate that larger TPRD orifices generate more extensive flammable gas clouds， though these dissipate more quickly. In scenarios with multiple parked vehicles， hydrogen accumulates in intervehicular spaces， and increased vehicle spacing enhances hydrogen dispersion. Moreover， horizontal wind patterns prove more effective than longitudinal wind in reducing hydrogen concentration. These insights provide practical guidance for the safe implementation of hydrogen fuel cell vehicles in indoor parking environments， particularly in semi-enclosed configurations.

Abstract:The significant range degradation of new energy vehicles （NEVs） in low-temperature environments necessitates accurate remaining mileage prediction to alleviate range anxiety， optimize energy management strategies， and enhance battery performance. This research analyzes real-world driving data from 30 NEVs collected between January and March in 2023 and 2024， implementing systematic data cleaning， feature engineering， and model development. The data processing phase addressed temporal inconsistencies， missing values， and outliers through time-series reorganization， interpolation techniques， and segment filtering. A state-of-charge （SOC） threshold-based segmentation strategy substantially increased the effective sample size. In feature extraction， 22 characteristics were derived from driving behavior patterns， battery conditions， and static vehicle parameters， with 14 core features selected through correlation analysis. The incorporation of nominal battery energy and driving range enhanced model representation capabilities. The model development utilized a multi-vehicle data fusion approach based on the Boosting Tree algorithm as the foundation. A hierarchical modeling strategy improved predictions for short-range vehicles， while incremental learning enabled dynamic model updates to accommodate time-varying factors such as battery degradation. Experimental validation demonstrated root mean square errors （RMSE） of 24.483 km for long-range vehicles and 6.425 km for short-range vehicles， markedly superior to conventional single-vehicle models. This methodology not only enhances prediction accuracy under low-temperature conditions but also presents novel technical approaches for NEV battery management and energy optimization.

Abstract:The increasing emphasis on environmental protection and NO_x emissions reduction has heightened the significance of SCR in automotive applications due to its high conversion efficiency. This research focuses on developing a model to characterize the efficiency of a Cu-SCR system in catalytically reducing NO_x for a specific hydrogen engine. Given the complexity of the chemical reactions within the Cu-SCR， a GPR is implemented to predict the yield based on process parameters including temperature， inlet species of NO_x， exhaust mass flow， and the fraction of NO₂ in NO_x［1］. To address the substantial data requirements of the regression process， simulation data from a GT-model serves as training data， eliminating the need for extensive experimental measurements. The model undergoes validation with limited experimental data through an affine transformation， and the transformation variables are optimized using MATLAB's fmincon function with the Interior-Point algorithm to address this high-dimensional problem.

Abstract:Electric vehicles experience significant range limitations due to decreased lithium-ion battery performance at low temperatures and substantial energy consumption by air conditioning systems. This study presents an integrated thermal management system topology for electric vehicles that implements a triple-hybrid heating strategy， combining a heat pump system， auxiliary positive temperature coefficient heater， and motor waste heat recovery. This approach enhances battery preheating efficiency while maintaining cabin thermal comfort and reducing overall energy consumption. To address challenges in secondary loop flow distribution and heater power control， a vehicle-level model predictive controller was developed based on heat transfer principles. Additionally， a heating optimization strategy specifically designed for the proposed thermal management topology was introduced. The viability of the proposed topology was validated through co-simulation using Simulink and AMEsim. Results demonstrate that the proposed integrated thermal management system solution achieves energy savings of 6.99%， 4.75%， and 2.36% in vehicle SOC consumption across three WLTC driving cycles， compared to alternative solutions. Moreover， the MPC control strategy yields an additional 0.53% reduction in overall vehicle SOC consumption upon completion.

Abstract:The internal dynamics of fuel cells are highly complex， involving coupled mechanisms such as electrochemical reactions， reactant transport， and membrane water transfer， which significantly influence their performance. Traditional equivalent circuit models struggle to independently quantify these dynamics processes. To address this， a two-dimensional impedance model of agglomerates is developed， based on mass， momentum， and component conservation equations， and validated under various operating conditions. By independently configuring system dynamics and analyzing the impedance spectroscopy and the distribution of relaxation times， this study decouples and examines electrochemical reactions and transport dynamics within the catalytic layer， diffusion layer， and flow channels. Results show that medium- and high-frequency impedance is linked not only to proton transport and charge transfer in the cathode catalytic layer but also to oxygen diffusion in both the catalytic and gas diffusion layers. Furthermore， the quantitative effects of agglomerate parameters， particularly platinum loading， on internal dynamics are analyzed under different conditions. This work provides critical tools and theoretical insights for advancing the understanding of fuel cell internal mechanisms.

Abstract:In this paper， the effects of air temperature， flow velocity and motion state of the lower plate on the heat transfer of the upper plate are studied by CFD simulation and response surface analysis. The results show that the air temperature significantly affects the total heating of the upper plate， but has little effect on the radiant heating. The motion of the lower plate thins the boundary layer and enhances convective heat transfer. The response surface analysis further confirmed that the temperature of the lower plate was the main factor affecting the radiative heat flux density， while the air temperature and flow velocity had a more significant effect on the total heat flux density.

Abstract:The bipolar plate serves a crucial function in the PEMFC regarding fuel supply quality and electrochemical performance. Achieving uniform gas distribution in the flow field presents significant challenges， particularly in transition areas where channel configurations directly affect reactant gas transport. This study presents a novel cathode flow field design for bipolar plates featuring multiple parallel sub-channels， analyzed through computational fluid dynamics modeling and simulation， with particular attention to the Coanda effect in the transition area. The investigation evaluates and analyzes mass flow distribution and pressure loss characteristics. The flow channels undergo further refinement through optimization of the transition region structure to achieve uniform mass flow distribution and reduced pressure loss. The results demonstrate that channel width allocation based on uniform distribution principles enhances mass flow uniformity. Additionally， incorporating fillets in the transition region with a jet width to fillet radius ratio below 0.5 amplifies the Coanda effect， thereby reducing adverse pressure gradients and flow separation phenomena， resulting in decreased pressure drop. Furthermore， when modifying gas outlet pressure to minimize pressure loss， an adjustment strategy for gas reactant concentration maintains consistent mass flow rates. The findings provide valuable reference points and optimization guidelines for flow channel design in bipolar plates， considering both fluid distribution uniformity and pressure loss factors.

Abstract:Methanol-reforming hydrogen production fuel cell （MRFC） systems demonstrate significant advantages in storage safety， energy density， and adaptability for distributed power generation and propulsion applications. However， existing systems below 30 kW face considerable limitations in dynamic response and power output capacity， while current research lacks comprehensive dynamic simulation methodologies and parameter optimization strategies for 100-kW-class implementations. This study establishes a 200-kW MRFC system model with several key innovations： An Integrated Thermodynamic-Kinetic Framework developed in Aspen Plus? couples essential modules （reformer， membrane separation with 65%－95% H? purification efficiency， and proton exchange membrane fuel cells （PEMFC） stack） to address technical challenges in large-scale system design. Global Sensitivity Analysis quantifies the impact of critical operating parameters—steam-to-methanol ratio （S/C=1.0－1.5）， catalytic conversion efficiency （80%－95%）， and purification efficiency （65%－95%）—revealing purification efficiency contributes 59.1% to overall system efficiency. The Dynamic Optimization Strategy， based on energy hub theory， resolves the coupling mechanism between efficiency and stability during load transitions， achieving 50%－100% load changes within 188 s （validated against 5 kW experimental data with <3% deviation）. The system's rated efficiency improved from 41.0% to 45.4% post-optimization， exceeding typical small-scale systems （<40%）. This research provides a scalable 1D dynamic simulation method that reduces extrapolation errors from >15% to <3% compared to conventional small-scale models， while implementing a hierarchical parameter sensitivity strategy that reduces system tuning cycles by 40%， offering valuable theoretical guidance for industrial-scale system design.

Abstract:As hydrogen-powered vehicles emerge as a viable alternative to fossil fuel-driven vehicles due to their low emissions and sustainability potential， increasing investments have made hydrogen safety a critical concern requiring comprehensive investigation. Given that hydrogen has not yet become mainstream in daily transportation， its unique safety considerations warrant further examination. This review synthesizes potential risks and safety challenges associated with hydrogen implementation and provides a detailed comparative analysis of hydrogen-related challenges in road vehicles across four distinct scenarios： open areas， tunnels， parking spaces， and refueling stations. This analysis aims to inform future research directions in hydrogen technology advancement and risk mitigation strategies.

Abstract:This study utilized the FLACS software to evaluate the dispersion and explosion characteristics of hydrogen released through thermally activated pressure relief devices （TPRDs） when parking a hydrogen fuel cell vehicle （HFCV） in an enclosed parking facility. The research investigated the effects of crossbeams on hydrogen dispersion patterns， concentration distribution， and post-explosion overpressure. Results indicate that crossbeams impede lateral hydrogen diffusion along the ceiling， resulting in concentration increases of up to 73% near the leak point and decreases of up to 25% in more distant areas. While crossbeams reduce the total combustible volume within the parking facility， they lead to higher overall concentrations of combustible gases. Moreover， the presence of crossbeams results in a 23% increase in maximum post-explosion pressure within the parking facility， thereby heightening the potential for casualties and injuries during an explosion event.

Abstract:With the rapid development of the new energy vehicle industry， the safety issues of power battery systems have become increasingly prominent. Therefore， researching safety risk assessment and early warning of power batteries is of great significance for ensuring the sustainable development of the industry. However， traditional assessment methods suffer from limitations such as high cost and long cycle times. Existing studies mainly focus on the analysis of accident evolution laws， but lack in-depth exploration of the main triggers of safety accidents - power batteries， especially their safety performance under extreme environments. To address this， this study proposes a safety risk assessment method for power batteries of pure electric vehicles that integrates EWM （Entropy Weight Method） and TOPSIS （Technique for Order of Preference by Similarity to Ideal Solution） based on machine learning technology， and introduces RF （Random Forest Algorithm） to construct a prediction model. The research obtained three-level alarm data from 4 faulty vehicles monitored by an automobile enterprise's platform. After data cleaning and feature extraction， monomer and pack-level characteristic parameters with correlation coefficients > 0.7 were screened as model inputs. Then， the EWM+TOPSIS assessment model and RF prediction model were trained based on the feature dataset and validated using full-scale combustion experiment data of 2 pure electric vehicles. The results show that under extreme working conditions， the model established in this study can provide early warnings of thermal runaway 1-3 minutes earlier than the automobile enterprise's monitoring platform. The output quantitative scores and alarm level predictions can intuitively reflect the changes in battery status， and the assessment results are highly consistent with the actual alarm levels. This study provides a new technical solution for power battery safety early warning， which has reference value for improving the safety of new energy vehicles.

Abstract:As an important path to achieve the goal of “dual-carbon”， fuel cell vehicles have attracted a lot of attention in the process of technological development and industrialization. In this paper， we firstly make a comprehensive systematic combing and in-depth analysis on the development situation and technology status of fuel cell vehicle industry in China and abroad. Then we mainly focus on the performance of electric stack， hydrogen storage technology， etc.， analyze the latest situation of global fuel cell automobile industry development and technical status， and summarize the bottlenecks faced by the fuel cell automobile industry at the present stage of development. Combined with the development trend of fuel cell automobile industry， put forward measures and suggestions for the high-quality development of fuel cell automobile industry and provide theoretical support for the breakthrough of the industry.

Abstract:This study focuses on the complex computation and difficult convergence of whole fuel cell stack based on the finite element simulation models for automotive applications. It represents a method of equivalent modelling to simplify the flow channel and sealing part of a unit cell into uniform anisotropic materials for the flow channel of bipolar plates which is achieved by calculating the equivalent stiffness. Furthermore， the equivalent beam model of the fuel cell stack is established through the difference in the mechanical properties of the beam structure. The downward bowing mechanism of the fuel cell stacks for vehicles subjected to impact is analyzed， and an equivalent model of the fuel cell stacks containing multiple unit cells is established. Based on this model， impact simulation is carried out， and the different numbers of unit cells in the stack， the impact acceleration， assembly force of the fuel cell stack， and the thickness of the end plate are systematically analyzed in order to evaluate their influence on the downward bowing of the fuel cell stack subjected to impact of the vehicle on road. Research shows that more unit cells and greater impact acceleration lead to more pronounced downward bowing. Appropriately increasing the encapsulation force and improving the end plate shape can reduce the degree of downward bowing in the stack.