Abstract:With the aim of reducing the cost of developing internal combustion engines， while at the same time investigating different geometries， layouts and fuels， 3D-CFD-CHT simulations represent an indispensable part for the development of new technologies. These tools are increasingly used by manufacturers， as a screening process before building the first prototype. This paper presents an innovative methodology for virtual engine development. The 3D-CFD tool QuickSim， developed at FKFS， allows both a significant reduction in computation time and an extension of the simulated domain for complete engine systems. This is possible thanks to a combination of coarse meshes and self-developed internal combustion engine models， which simultaneously ensure high predictability. The present work demonstrates the capabilities of this innovative methodology for the design and optimization of different engines and fuels with the goal of achieving the highest possible combustion efficiencies and pollutant reductions. The analysis focuses on the influence of different fuels such as hydrogen， methanol， synthetic gasolines and methane on different engine geometries， in combination with suitable injection and ignition systems， including passive and active pre-chambers. Lean operations as well as knock reduction are discussed， particularly for methane and hydrogen injection. Finally， it is shown how depending on the chosen fuel， an appropriate ad-hoc engine layout can be designed to increase the indicated efficiency of the respective engines.

Abstract:Technological trends in the automotive industry toward a software-defined and autonomous vehicle require a reassessment of today’s vehicle development process. The validation process soaringly shapes after starting with hardware-in-the-loop testing of control units and reproducing real-world maneuvers and physical interaction chains. Here， the road-to-rig approach offers a vast potential to reduce validation time and costs significantly. The present research study investigates the maneuver reproduction of drivability phenomena at a powertrain test bed. Although drivability phenomena occur in the frequency range of most up to 30?Hz， the design and characteristics substantially impact the test setup’s validity. By utilization of modal analysis， the influence of the test bed on the mechanical characteristic is shown. Furthermore， the sensitivity of the natural modes of each component， from either specimen or test bed site， is determined. In contrast， the uncertainty of the deployed measurement equipment also affects the validity. Instead of an accuracy class indication， we apply the ISO/IEC Guide 98 to the measurement equipment and the test bed setup to increase the fidelity of the validation task. In conclusion， the present paper contributes to a traceable validity determination of the road-to-rig approach by providing objective metrics and methods.

Abstract:The traditional topology optimization method based on finite element method requires multiple finite element calculation and iterations， which consumes a lot of computational resources and time. In order to improve the efficiency of topology optimization， the paper takes topology optimization of cantilever beam as an example and proposes a generative convolutional neural network （CNN） model based on residual connections， which considers four optimization parameters： filter radius， volume fraction， loading point and loading direction. And the influence of different loss functions and number of samples on the accuracy of generative CNN model is discussed at length. The results show that the proposed model has high accuracy and generalization ability， and the mean structural similarity index between the model prediction and finite element method can reach 0.9720， the mean absolute error is 0.0143. And the prediction time of the model is only 0.0041 of finite element method， which significantly improves the efficiency of topology optimization.

Abstract:This study aims to achieve intelligent prediction of collision energy absorption characteristics of new structures in forward design of automotive parts. An energy-absorbing box is taken as the research object to generate training data sets by finite element crush deformation simulation. A graph-based encoder is adopted for geometric structure recognition. Long and short-term memory networks and graph convolutional neural networks were used to process adjacent temporal data. The novel neural network prediction system proposed can recognize geometric structures and memorize temporal data. The comparison between the model prediction results and simulation results shows that the predicted crush pattern of the energy-absorbing box is consistent with the finite element simulation results， and the prediction accuracy of the model for the crush deformation amount can reach up to 95.33%， while the prediction accuracy of the maximum energy absorption value can reach 99.98%. Compared with the finite element calculations， computational efficiency is 174.5 times and 210.5 times higher respectively， which manifested that the system can accurately and quickly predict the crash performance of the energy-absorbing box.

Abstract:In recent years， machine learning methods have been widely adopted for real-time vehicle energy consumption predictions. However， the accuracy and generalizability of these predictions are often hindered by challenges such as data imprecision， missing fields， multicollinearity， and substantial difference in driving conditions and driver behaviors among identical vehicle models. To address these issues， this study systematically considers factors such as feature redundancy， data balance， freight trip frequency， transport capacity， traffic congestion and driving duration. Subsequently， an energy consumption prediction model with high precision is developed using a combination of machine learning methods such as XGBoost， Random Forest （RF）， and Multilayer Perceptron （MLP）. The model utilizes key features selected through the Mutual Information （MI） method， along with a constructed driver profile that captures characteristic behaviors as an independent feature. The proposed method is validated using T-BOX data collected from 120 light trucks. Experimental results indicate that the prediction method significantly enhances the prediction accuracy of energy consumption under various driving behaviors and conditions. This research contributes to the development of models with high precision in estimating the fuel consumption of light trucks.

Abstract:A circular and sustainable economy for the private transport sector requires a holistic view of the emitted CO₂ emissions. Looking at the energy supplied to the vehicle in terms of a circular economy leads to defossilisation. The remaining energy sources or forms are renewable electric energy， green hydrogen and renewable fuels. A holistic view of the CO₂ emissions of these energy sources and forms and the resulting powertrain technologies must take into account all cradle-to-grave emissions for both the vehicle and the energy supply. In order to compare the different forms of energy， the three most relevant forms of powertrain technology are considered and a configuration is chosen that allows for an appropriate comparison. For this purpose， data from the FVV project “Powertrain 2040” are used^［1］ and combined with research data on the energy supply chain for passenger cars. The three comparable powertrain configurations are a battery electric vehicle， a fuel cell electric vehicle and an internal combustion engine hybrid vehicle fueled with electric fuel. First， the three selected powertrain configurations are presented in terms of their performance， weight， technology and other characteristics. A comparative analysis is carried out for different CO₂ emissions of the electricity mix. The electricity mix is used for both the production of the vehicle and the energy. The results are presented in the form of cradle-to-wheel emissions， which consider the total CO₂ emissions of the vehicle over its life cycle. Finally， the results are analyzed and discussed to determine which powertrain technology fits best into which energy sector CO₂ emissions window.

Abstract:The thermal management systems for vehicular electronic devices often operate under vibrational conditions， and effective vibrational control can enhance convective heat transfer of the fluid. However， the current understanding of the influence of vibration on the heat transfer performance of single-phase spray cooling remains unclear. To investigate the thermal performance of single-phase spray cooling under various vibrational conditions， a closed-loop vibrating surface spray cooling system has been established in this work. Through experimental investigation， the study explores the influence mechanisms of Vibrational Reynolds Number （Re_v）， Dimensionless Acceleration Number （A_c）， amplitude， and frequency on single-phase spray cooling performance. The findings indicate that an increase in Re_v and amplitude results in a suppression of heat transfer coefficient and heat transfer enhancement factor； specifically， spray cooling is suppressed by 37% when the heating power is 200 W and Re_v is 4 947. Conversely， higher Ac and frequency can improve the thermal performance of spray cooling， enhancing it by 17% when the heating power is 200 W and the Ac number is 3.6. The results from this study can be utilised for optimising spray cooling thermal management systems under vibrational conditions.

Abstract:Diesel engines equipped with turbochargers is an effective way to alleviate energy shortage and reduce gas emissions， but their compressor aerodynamic noise emissions have become an important issue that needs to be addressed urgently. In the studies of diesel engine turbocharger compressors noise emissions， the pattern of compressor aerodynamic noise emissions in the near-surge condition， near-choke condition and its relationship with the internal flow characteristics are still unclear. Therefore， in order to study the aerodynamic noise emission characteristics and mechanism of a diesel engine turbocharger compressor in the near-surge and near-choke conditions， the experimental and numerical simulation methods were used to analyze the aerodynamic noise of a turbocharger compressor in this study. The analysis of experiment results showed that total sound pressure level （SPL） of the aerodynamic noise for the compressor increased with an increase in the speed under the near-surge and near-choke conditions. At low speed， the total SPL of aerodynamic noise was influenced by the mass flow rate of the compressor more obviously. In the compressor aerodynamic noise， the blade passing frequency （BPF） noise was dominated. With the increase of speed， the contribution of BPF noise to the total SPL of aerodynamic noise was greater， and its proportion was up to 75.35%. The analysis of simulation results showed that in the near-surge and near-choke conditions， there were obvious stall phenomena in the internal flow of the compressor. Among them， in the near-choke condition， the rotor and the diffuser regions of the compressor were dominated by multiple monophonic noise， and the dynamic-static interference between the impeller and the diffuser had a high contribution to both the axial frequency and its harmonic frequency noises. In the near-surge condition， the compressor inlet and outlet regions were dominated by low-frequency noises， and the interference between the impeller blades and the incoming air flow had a more obvious contribution to the induced low-frequency noises.

Abstract:Aiming at requirements of intelligent modulation in vehicle acoustic environment for drivers and passengers，a targeted sound field reproduction method is proposed based on stabilized biconjugate gradient （BiCGSTAB） to collectively improve the acoustic energy contrast and sound field planarity. Considering parametric results of the contribution weight on the in-vehicle sound acoustic arrays to the acoustic environment， and combining the automotive scenario to divide the target acoustic regions， the cabin acoustics model for the drivers and passengers are first constructed. By defining the acoustics problems related to sound field reproduction and combining with the parametric results that the sound pressure response of target regions with respect to sound source distribution positions， the interior acoustics model for drivers and passengers is constructed. Furthermore， the targeted reproduction framework of interior sound field is constructed embracing the given expected sound field， and the driving signals of sound sources are iteratively obtained based on BiCGSTAB. Focusing on acoustics energy contrast， sound field planarity and other indicators， the interior sound field reproduction performance is analyzed in detail. The results show that the proposed method can effectively avoid the abnormal convergence in solving the acoustics complex linear equations， and can achieve the high-quality reproduction about the acoustics amplitude and phase information in vehicle target regions

Abstract:Unsteady numerical simulations were carried out using different RANS/LES hybrid methods for the outflow field of the MIRA generic model. The applicability of RANS/LES hybrid methods （DDES， IDDES， SBES， SDES） and DDES based on different RANS models （RKE， SA， SST k-ω， GEKO k-ω） in calculating the automotive outflow field was explored through detailed comparative analyses with the results of the wind tunnel experimental measurements of aerodynamic forces and pressures. The study shows that the results of different hybrid methods are all on the high side in predicting aerodynamic coefficients， with the DDES-GEKO model having the smallest relative error. For surface pressure coefficients， the predictive results of different hybrid methods for the vertical center line are in good agreement with the experimental values， among which the SBES-GEKO model performs better. The embedded RANS models in the DDES show significant differences in predicting pressure on the rear windshield， with the SA model being better. The pressure predictions for the underbody are deviated by different hybrid methods， with the pressure predictions for the front part being smaller than experimental values， while the SBES-GEKO model gives better results for the rear part. Additionally， the SBES-GEKO model is able to identify the unsteady flow structures in the wake region well.

Abstract:Based on the MIRA model group of full scale （squareback S， fastback F and notchback N）， numerical simulations are done to investigate the effect of longitudinal spacing （0.1-1L） on the uniform （vehicles in the platoon are the same） and ununiform （vehicles in the platoon are varied） three car platoon aerodynamic drag. Analyses of drag variation are conducted by pressure drag and flow field. The results indicate that the drag reduction of the three uniform platoons diminishes with increased spacing. The drag reduction of the head car dwindles basically with increased spacing， which is owing to the drop of backpressure recovery on the verticalbase and rearwindow. The drag reduction of squareback middle car declines gradually with increased spacing， that of other two configurations plunge rapidly in the spacing of 0.1-0.2L， then keep a low drag reduction. The drag reduction of squareback tail car also declines gradually with increased spacing， and experiences drag increase at the spacing 0.5L， other two configurations basically keep a low drag reduction or increase. the drag reduction of the three ununiform platoons also diminishes with increased spacing. The drag variations above are mainly resulted from the comprehensive effects of backpressure recovery， the impairing of positive pressure on the frontface and the crippling of negative pressure on the A-pillar and so on.

Abstract:The statistical analysis was performed to the vehicle velocity during coastdown test. It was found that the velocity deviation is approximately linear to the mean value. Based on the coasting velocity， the road load and air drag coefficient were fitted by v-t method and v-F method. At the same time， the convexity of corresponding optimization problem and the uncertainty of road load and air drag coefficient were analyzed. It was shown that fitting problem of v-F method is convex and that of v-t method is non-convex. During the fitting process， we need to choose the algorithm of high generalization capability and specify the initial values and bounds of parameters to obtain the reasonable fitting parameters and road load. Besides， the uncertainty of air drag coefficient is lower for v-t method and its confidence interval is narrower. At last， based on the statistical law of coasting velocity， the Monte Carlo method was applied to simulate the standard coast down test. It was found that the distribution of road load and air drag coefficient is approximately normal. And influence of randomness is highlighted for v-F method. The air drag coefficient and road load fitted by v-t method is more stable.

Abstract:The flow control of the fan wake in the wind tunnel can improve the efficiency of the fan and optimize the quality of the flow field. For the closed-circuit wind tunnel， the global model with the entire flow channel must be used to verify the simulated flow structure and velocity distribution against the measurement data. Results show that for different truncated length of the fan tail， there are two mechanisms of wake control that are closely related to the near-wall large eddy structure. One is the interaction between the large eddies and the small ones induced by the separation flow from the vertical base， and the other one is changing the large eddy structure， which formed from the flow separation and can be strongly influenced by the separation positions due to the surface curving. Both mechanisms can affect the distance between the large vortex structure and the tail cone， and change the velocity distribution of the near-wall shear layer， causing the shear layer to deflect inwardly， resulting in a decrease of the low-velocity area of the wake， increasing the total pressure at the outlet of the main diffusion section， and improving the uniformity of the flow field. After the vortex generators are introduced， the small vortex structure generated by the flow passing through can change the large vortex structure of the wake， then the total pressure and the flow uniformity of the main diffusion section are further improved.

Abstract:When conducting aerodynamic test of vehicles in a wind tunnel， the discrepancy between the vehicle's ride height in the tunnel and its actual ride height on the road results in differences between the measured aerodynamic data and the actual values. A method of simulating the actual ride height of vehicles on the road is explored in this study using the "floating mode" of the wind tunnel balance system to test the ride height and aerodynamic performance data of 9 different vehicles under initial fixed ride height， wheel rotation ride height， and actual road ride height. The impact of wheel rotation and aerodynamic lift/torque on vehicle ride height and aerodynamic performance is analyzed， and recommendations for minimizing the disparity between wind tunnel test data and actual road test data are provided.

Abstract:Low aerodynamic drag are of great significance for energy conservation and emission reduction of fuel vehicles， as well as increasing the range of electric vehicles. In recent years， the active jet technology is mostly applied to high aerodynamic vehicle models such as square back model， but its drag reduction and mechanism on low drag vehicles are not clear. In this paper， the effects of steady jet position， jet momentum coefficient and jet angle on drag reduction and net saving are studied， and the flow field analysis is given. The results show that the vertical rear jet （J3， J4， J5） of the fast back vehicle is an effective energy-saving measure， the low momentum coefficient has a good net saving rate， and the aerodynamic drag reduction and net saving rate have a maximum value with the change of jet angle. The best working condition to get the back jet in the end is： J3， J4 and J5 jet grooves arranged at the tail of the model are used for jet，the momentum coefficient is 1% and the jet angle is 45 °， the aerodynamic drag was reduced 2% and the net saved energy can reach 9.5% and 129.7W， respectively.

Abstract:When a car is driven at high speed， there exists large pressure difference with strong pressure fluctuation between the car exterior and interior near the side window area， so the side window sealing system is likely to lose contact and thus generates leak noise， which can influence the interior sound quality seriously. The glassrun seal of the front side window was taken as the research object， with steady pressure penetration method and transient dynamic analysis， the condition of sealing failure occurrence was discussed. The main conclusions are as follows： with steady pressure penetration method， the critical static pressure difference is in the order of 10⁴ Pa. The influence factors of the critical pressure difference include glassrun shape， glass positioning， the rubber material parameters and the fixing manner etc. While with the transient dynamic analysis， as the window glass was offset with 0.2mm inward， the outer lip lost contact under the transient pressure difference in the order of 10³ Pa. The findings provide a meaningful reference for prevention of leak noise generation through improvement of sealing system design.

Abstract:Wind tunnel testing is a pivotal methodology for investigating automotive aerodynamics. In order to simulate test scenarios encompassing a broad spectrum of real-world driving conditions， wind tunnel test sections typically demand a relatively low level of turbulence intensity. However， as road traffic research becomes increasingly intricate， the notion of simplifying vehicle inflow conditions to uniformly low turbulence is no longer adequate. On one hand， the evolution of autonomous driving technologies has led to reduced inter-vehicle spacing at high speeds， even giving rise to platooning scenarios. Under such conditions， the trailing vehicle contends with inflow characterized by heightened turbulence intensity. On the other hand， vehicles navigating through non-open-road environments， like those flanked by trees or buildings， confront inflow conditions that are far from being characterized by low， uniform turbulence. The aerodynamic flow structures and drag-increasing mechanisms experienced by vehicles in these high-turbulence real-world settings undergo transformation. To furnish a flow environment more closely aligned with these road conditions， this study has devised a passive turbulence generator， aimed at emulating elevated turbulence intensity scenarios prevalent in specific driving conditions. Initially， numerical simulation techniques were employed to assess the turbulence-generating efficacy of the generator， and its resemblance to actual road flow spectra was examined. Subsequently， an analysis of the wind tunnel test section's flow field quality and jet shear layer characteristics revealed the amplification of low-frequency flutter phenomena attributed to the turbulence generator. To address this issue， a vibration-damping variant of the harbor seal whisker nozzle structure was incorporated， further investigating its impact on turbulence and flow field characteristics. Finally， model-scale wind tunnel experiments were conducted to validate both the simulation outcomes and the generated turbulence effects. The results underscore the capability of the passive turbulence generator to aptly emulate the higher turbulence intensity natural inflow conditions while still maintaining low-frequency flutter at manageable levels within the wind tunnel environment.

Abstract:Accurate pose estimation is of paramount importance for intelligent vehicles， serving as the foundation for decision-making， planning， and control. To enhance the accuracy of heading estimation in vehicle pose， a novel fusion localization algorithm based on observability is proposed in this paper， utilizing GNSS （Global Navigation Satellite System）， IMU （Inertial Measurement Unit）， and vision sensors. Firstly， to assess the observability of the error state in the GNSS/IMU system， a novel method for relative observability analysis is introduced， revealing the existence of four weakly observable states within the traditional GNSS/IMU system. Subsequently， a fusion localization algorithm grounded in relative observability is proposed， utilizing the relative heading angle estimated by Visual Odometry. The experimental results indicate that the proposed localization algorithm achieves a maximum heading error of 2.76° and an RMS heading error of 1°， highlighting the effective enhancement of vehicle heading accuracy in weakly observable states by the proposed algorithm.

Abstract:This paper presents a Nonlinear Model Predictive Controller （NMPC） for the path following of autonomous vehicles and an algorithm to adaptively adjust the preview distance. The prediction model includes vehicle dynamics， path following dynamics， and system input dynamics. The single-track vehicle model considers the vehicle’s coupled lateral and longitudinal dynamics， as well as nonlinear tire forces. The tracking error dynamics are derived based on the curvilinear coordinates. The cost function is designed to minimize path tracking errors and control effort while considering constraints such as actuator bounds and tire grip limits. An algorithm that utilizes the optimal preview distance vector to query the corresponding reference curvature and reference speed. The length of the preview path is adaptively adjusted based on the vehicle speed， heading error， and path curvature. We validate the controller performance in a simulation environment with the autonomous racing scenario. The simulation results show that the vehicle accurately follows the highly dynamic path with small tracking errors. The maximum preview distance can be prior estimated and guidance the selection of the prediction horizon for NMPC.

Abstract:Developing a fast and accurate trajectory tracking method is a core control task in the development of autonomous driving technology. This paper establishes a dynamic residual model based on the vehicle preview distance and constructs a trajectory tracking controller using a model predictive control （MPC） algorithm with the front wheel steering angle as the control input. By designing an optimized preview distance method that considers the vehicle state and the reference trajectory， the tracking accuracy is improved. The objective function and constraint conditions are designed to comprehensively consider the accuracy of vehicle tracking control and driving stability. A jiont simulation platform combing Carsim and Simulink is built to verify the effectiveness of the algorithm. Conduct simulation experiments at different speeds under double lane change conditions to explore the impact of preview distance on control results and design optimization methods for preview distance. The results of the simulation experiments demonstrate that compared to the case without optimized preview distance， the trajectory tracking controller designed in this paper can achieve high-precision control. When the vehicle is traveling at high speed， the lateral displacement error is less than 10 cm， the heading angle error is less than 1°， and the vehicle stability is significantly improved.

Abstract:With the development of Internet technology， attacks against vehicles occur frequently. Hackers can flash malware into the automotive MCU to maliciously control the vehicle or steal important information. For the above attack scenarios， it is necessary to add secure boot to the automotive MCU. Functions are protected to ensure that only legitimate software can be run. This paper first conducts an in-depth study of the current mainstream secure boot schemes， summarizes their performance and shortcomings in terms of safety， stability， and low time-consuming， and then proposes an optimization scheme for low-time-consuming performance ， and finally verified the feasibility and effectiveness of the optimization scheme through the built physical system. The results show that sampling verification and compilation optimization can protect as much important code as possible while satisfying the time-consuming verification.

Abstract:To investigate the multi-objective control problem of trajectory tracking and vehicle stability， a hierarchical coordinated control strategy of trajectory tracking and yaw stability was proposed for four-wheel-independent-drive autonomous electric vehicles. In the upper controller， the linear-time-varying model predictive control （LTV MPC） was employed to generate the desired front road wheel steering angle and yaw moment， and the PID speed control embedded in the model predictive optimization solutions was introduced to generate the desired total driving/braking torque. In the lower controller， the generalized forces from the upper layer were allocated to the four wheels based on quadratic programming. An 8-degree-of-freedom （DOF） vehicle model was used as the prediction model and a high-fidelity 14-DOF vehicle model with longitudinal and lateral combined brush tire model was used as the plant. Numerical simulation results under different speeds， road adhesion coefficients and conditions of whether to consider yaw stability control， demonstrate that the proposed controller possesses good trajectory tracking performance and robustness， which improves the tracking accuracy while ensuring the yaw stability under the limit condition.

Abstract:The large-scale commercial application of electric Robotaxi fleets requires well-developed charging infrastructure as a prerequisite. However， there are still issues with insufficient quantity， low utilization rate， and inappropriate distribution of charging facilities. In addressing the site selection and capacity planning of charging stations for electric Robotaxi fleets， firstly， a minimum fleet size calculation method based on the trip network articulation and Hopcroft-Karp algorithm was proposed to meet passenger temp-spatial travel demands. Based on this， the spatio-temporal distribution of charging demands of Robotaxi fleets was quantified through Monte Carlo simulation. Then， considering the construction and operation costs of charging stations， grid loss costs， no-load driving， queuing for charging and loss of order opportunity costs of Robotaxi fleets， an optimization model for the site selection and capacity planning of charging stations for electric Robotaxi fleets was constructed with the objective of minimizing overall costs， and an improved particle swarm optimization algorithm based on genetic operators and adaptive inertia weight was proposed. Finally， the effectiveness of the proposed model and algorithm was validated using real user travel orders and geographic data from the city of Chengdu， China.

Abstract:The defossilisation of transport is of central importance for meeting global climate targets. In addition to introducing battery-electric vehicles that today are mainly intended for individual and local passenger transport， fuels based on renewable resources offer the possibility of fulfilling energy-intensive transport tasks with low global greenhouse gas emissions. Hydrogen produced from renewable energy is gaining particular importance here， as its use in mobile applications ensures not only overall low global warming potential but also the lowest tailpipe greenhouse gas emissions.In this paper， the various aspects of the use of hydrogen in vehicles are briefly presented and the work at FKFS and IFS is illustrated using the development of an H₂-DI combustion process for ICE applications as an example.The presented FVV research project investigates a high-pressure direct injection concept on a single-cylinder passenger car engine test bench. It aims to operate a spark-ignited engine near stoichiometric conditions to produce a significant power output with low boost pressure demand. However， for a hydrogen engine， a stoichiometric mixture leads to an increased knocking tendency towards higher loads. To avoid pre-reactions in the end gas， the injection starts shortly before TDCF and the hydrogen jet is ignited by the spark plug. The injection duration and therefore the maximum hydrogen mass flow through the injector nozzle influence the combustion duration. Challenges of the investigated hydrogen combustion process are， among others， the higher NO_X emission level compared to the lean operation and the hydrogen slip into the exhaust system.

Abstract:The degradation process of lithium-ion batteries is intricately linked to their entire lifecycle as power sources and energy storage devices， encompassing aspects such as performance delivery and cycling utilization. Consequently， the accurate and expedient estimation or prediction of the aging state of lithium-ion batteries has garnered extensive attention. Nonetheless， prevailing research predominantly concentrates on either aging estimation or prediction， neglecting the dynamic fusion of both facets. This paper proposes a hybrid model for capacity aging estimation and prediction based on deep learning， wherein salient features highly pertinent to aging are extracted from charge and discharge relaxation processes. By amalgamating historical capacity decay data， the model dynamically furnishes estimations of the present capacity and forecasts of future capacity for lithium-ion batteries. Our approach is validated against a novel dataset involving charge and discharge cycles at varying rates. Specifically， under a charging condition of 0.25 C， a mean absolute percentage error （MAPE） of 0.29% is achieved. This outcome underscores the model's adeptness in harnessing relaxation processes commonly encountered in the real world and synergizing with historical capacity records within battery management systems （BMS）， thereby affording estimations and prognostications of capacity decline with heightened precision.

Abstract:To address the issues of state identification and diagnosis for cells in lithium-ion battery modules， this paper proposes using electrochemical impedance spectroscopy and distribution of relaxation time curves with the affinity propagation （AP） clustering algorithm for abnormal identification of battery modules. The AP algorithm is compared with the density-based spatial clustering of applications with noise （DBSCAN） algorithm using 10 normal samples and multiple abnormal samples. The results show that AP performs better than DBSCAN in terms of accuracy， robustness， and parameter sensitivity （overlapping data， uneven density， etc.）. In addition， the extreme gradient boosting （XGBoost） classifier is introduced， and after storing a certain amount of data corresponding to the battery， the same battery can be directly diagnosed for abnormalities through the XGBoost classifier. The anomaly detection rate is 100%， and the accuracy of identifying anomaly types exceeds 92%. Finally， a battery module abnormal identification and diagnosis system is proposed， which includes key steps such as data collection， feature extraction， identification， and diagnosis.

Abstract:Open circuit voltage （OCV） is an important variable for accurately estimating the State of Charge （SoC） of lithium-ion batteries in electric vehicles （EV）. Since the OCV-SoC mapping relationship changes continuously as battery ages， the OCV-SoC function determined at a specific aging stage cannot be applied for SoC estimation throughout the battery's entire lifecycle， thus necessitating regular OCV testing and calibration. However， traditional OCV testing can typically take several days to obtain one or more complete charge-discharge-cycle data due to the hysteresis phenomenon of the OCV-SoC curve， making real-time OCV test and calibration impractical in real EV operation scenarios. Here， we proposed a fast and flexible OCV-SoC extraction method based on the smoothness hypothesis of the OCV-t curve during the battery discharge process. The non-dominated sorting genetic algorithm （NSGA-II） was utilized to extract the OCV-SoC relationship based on arbitrary current-voltage measurement data. Meanwhile， the present method was validated using the extended Kalman filter （EKF） based on WLTC and UDDS driving cycles. The results show that the OCV-SoC curve can be effectively constructed based on the smoothness hypothesis， where the maximum estimation error of SoC is 2% and will not be affected by inaccurate initial values.

Abstract:The fabricating process of polymer electrolyte fuel cell membrane electrode assembly significantly affects its power generation performance， particularly the hot-pressing process. However， the internal structure evolution mechanism of the catalyst layer （CL） during the hot-pressing period is not fully understood. Therefore， the effect of the hot-pressing temperature and the treatment time on the CL structure variations is investigated in this study， utilizing orthogonal experiments， characterization， and performance analysis. Based on the scanning electron microscope data， the compressed behavior of the pore structure in the CL can been observed after the hot-pressing treatment and increasing the hot-pressing temperature and the treatment time can deteriorate this compressed effect， which is consistent with the pore diameter analysis. In addition， the secondary pores are mainly compressed. These phenomena illustrate the soften behavior of the ionomer in the CL above the glass transition temperature. Furthermore， the electrical performance analysis demonstrates the mass transfer resistance significantly increments due to the hot-pressing treatment and the mass transfer resistance increment of the CL hot-pressed at 160 ℃ for 9min is the most significant， which is 2.77 folds more than that of the CL prepared by direct spraying method and it is the result of the compressed pore structure in the CL. This study provides more sights into comprehending the effect of the ionomer thermal response behavior in the CL on the structure evolution process of the CL， and offers theoretical support for the critical parameter optimization of the hot-pressing procedure.

Abstract:Air Supply Subsystem of multi-stack fuel cell system（MFCS） consists of air compressor， buffer tank， inlet flow throttle valve and backpressure valve， because of strong coupling between air mass flow and back pressure control in the air supply system，proposed decoupling control needed to be adopt in MFCS. Based on Air Supply Subsystem of MFCS， the transfer function matrix of the air supply subsystem for MFCS is identified near steady state point and then air flow and back pressure is decoupled via feed-forward compensation decoupling control technique．Experimental results show that the designed controller can really achieve the decoupling control of air flow and back pressure.

Abstract:With the rapid development of hydrogen fuel cell vehicles （HFCVs）， it is possible for hydrogen fuel cell vehicles to be tested in wind tunnels. In this paper， the flow diffusion law and distribution of hydrogen in wind tunnel under the condition of HFCV leak were studied by numerical simulation. When the wind tunnel simulates the driving condition of the HFCV at 80 km/h， the leaking hydrogen moves against the surface of the car and into the wake area. The hydrogen concentration decreases with increasing distance from the car. The hydrogen entering the flow channel from the collecting port is concentrated in the middle and bottom， with an obvious concentration gradient， and becomes uniform after passing through the fan section. After 10s， the hydrogen leaking from the bottom of the car returned to the test section and merged with the hydrogen leaking from the car， resulting in a continuous increase in the concentration of hydrogen in the entire area. The concentration of hydrogen inboard path increases faster than outboard path due to the deflectors at the corner of the channel. When the wind tunnel simulates the idling condition， the leaking hydrogen mostly diffuses in the test section and almost no hydrogen is detected in the flow channel. Affected by airflow disturbance in the test section， the hydrogen on two sides of the vehicle presents an unsymmetrical distribution， the hydrogen at the top is concentrated on the left side of the HFCV.Finally， according to the numerical simulation results， the installation position of the hydrogen sensors is suggested.

Abstract:Ammonia is used as a green hydrogen storage fuel with great potential for application， with zero carbon emissions compared to traditional hydrocarbon fuels. Liquid ammonia has a low boiling point of only 239.7 K at 1 atm， which makes it highly susceptible to flash boiling. In this paper， numerical simulations of liquid ammonia fuel were carried out in the Eulerian-Lagrangian framework to investigate the morphology， penetration distance， particle size and ammonia vapor mass fraction of the flash-boiling sprays at different superheat degrees， injection pressures， fuel temperatures and ambient temperatures. The results show that the spray penetration increases with the larger superheat of liquid ammonia， and the collapse of the spray front becomes more pronounced. The higher the spray pressure， the liquid ammonia collapses are enhanced and then weakened， and the particle size is smaller. The spray characteristics of liquid ammonia are almost independent of the ambient temperature， and are less affected by superheat degree at low fuel temperatures， and the opposite is true at high fuel temperatures. The ammonia vapor mass fraction increases with increasing fuel temperature.

Abstract:This paper summarizes and analyzes the hydrogen safety accidents of hydrogen compressors based on the description of skid-mounted compressor systems； then， a fault-tree-based hazard identification of skid-mounted hydrogen compressor systems is carried out for combustion and explosion accidents caused by hydrogen leakage； after that， an event tree analysis is applied to study the consequences and probability of hydrogen safety accidents considering system protection measures； finally， the consequences of jet flame， flash fire and explosion accidents caused by the hydrogen leakage considering protection measures is carried out based on FLACS software， combined with personal and building injury standards； besides， the optimization suggestions of safety equipment are proposed. The study shows that hydrogen leakage is mainly triggered by design faults and operational errors. Besides， failure of protection measures has a greater impact on fire or explosion events caused by hydrogen leakage； the probabilities of hydrogen safety accidents are all less than 3.3×10^-3/year， and jet fire happens the most frequently. In addition， the results of the accidents consequences show that， for the jet fire， the max temperature is 2329.8℃， and the maximum heat flux density is 399.0kW/m²； for the flash fire accident， the maximum damage radius is 7.7m， the highest temperature is 3069.5 ℃； for the explosion accident， the maximum damage radius is 5.5m， the maximum overpressure of 10 bar. These accidents will cause serious personal injury and container structure damage. In addition， the safe distance should be larger than 7.7m. Moreover， the safety of the system can be effectively enhanced by improving the sensitivity of the hydrogen concentration sensor， optimizing the layout position of the hydrogen concentration sensor and enhancing the exhaust air volume of the exhaust system.

Abstract:In this paper， a torque reserve control strategy for dual-rotor motor for vehicles is proposed to address the issue that the rate of change of motor stator current is limited by the inductance of the stator windings， which affects the rapid change of motor torque. Firstly， the mathematical model of the dual-rotor motor is derived， and the basic expression for the spatial rate of torque change is analyzed along with its relationship to the torque-angle displacement characteristic. Next， a current control strategy considering the torque reserve is formulated based on the torque-angle displacement characteristic of the PMSM. Specifically， the stator current is preloaded before the rapid torque changes， while maintaining the stator current constant during the rapid torque change stage， aiming to achieve the rapid torque output by the change of the torque angle driven by the rotation of the rotor. Finally， a simulation model and experimental platform for the torque reserve control strategy are created， and the correctness and feasibility of the proposed control strategy are verified through simulations and experiments.

Abstract:Electromagnetic vibration and noise prediction method of 16 poles-24 slots In-Wheel-Drive （IWD） permanent magnetic synchronous electrical motor were analyzed， which was then validated by bench test in anechoic chamber. Firstly， anisotropic material stator core and stator system with concentrated winding is established and modal results of stator system were validated by modal impact test. Secondly， for the assembly of IWD permanent magnetic synchronous electrical motor， electromagnetic， structural and acoustic multi-physical coupled model was built. Electromagnetic noise is predicted on WOT condition. Finally， electromagnetic， structural and acoustic multi-physical coupled model of permanent magnetic electrical motor was validated by bench test results in anechoic chamber. The research achievements will contribute to other permanent magnetic synchronous electrical motor design and mechanism analysis of electromagnetic vibration and noise.