Search

Previous studies have successfully applied empirical and equivalent circuit models (ECM) in battery management system (BMS) to perform highly accurate state estimation and other critical functions. However, these models struggle to meet the increasingly harsh requirements from modern electric vehicles, such as timely fault diagnosis that may require high-precision and detailed knowledge of battery cells. This paper starts with a comprehensive review to identify potential of applying electrochemical model, followed by proposed methods to solve problems induced from practical applications including data sampling precision and frequency. Considering the distribution of discharge currents in real driving scenarios, an improved Single Particle Model is introduced to simplify the complex electrochemistry model, and its practical applicability is verified. Moreover, the impact of sparse data collection frequencies under real vehicle conditions on parameter sensitivity is investigated, on top of which a novel optimization algorithm that combines Cuckoo Search with Particle Swarm Optimization has been proposed, facilitating grouped identification based on sensitivity analysis. The results of the virtual battery simulation verification indicate that the average absolute error of the algorithm in identifying parameters is 4.013%. The verification with actual vehicle data shows that the voltage fitting error of the algorithm is 13.05 mV. • Analyzing electrochemical parameter identifiability by varying data frame counts. • Hybrid cuckoo and particle swarm algorithm boosts search capability and efficiency. • Enhancing low-sensitivity parameter identification via sensitivity grouping. • Method shows lower error than experimental and surpasses existing algorithms. [ABSTRACT FROM AUTHOR]

The overheat abuse experiment of a 12S1P 37 Ah prismatic Lithium-ion battery module in a nominal energy of 1.65 kWh is conducted in this work. The cell behaviors and characterization in the process of thermal runaway propagation is investigated, including the gas eruption, the fire ejection, the flame combustion, the audio features, and the heat transfer, respectively. In the experiment, the central cell is heated on both sides until the pole temperature moves beyond 300 °C, the thermal runaway undergoes about 43 min and propagates from the central to both sides in the module, and all 12 cells burn. Results show that the first three runaway cells spout gas at first, and, then, emit sound with close amplitudes, frequencies, and energies, about 200 s earlier than the fire ejection. Then, the characteristic of the internal short circuit is the temperature rate zone of 1.0 K/s with time greater than 20 s. Moreover, the proposed thermal propagation coefficient is used to assess the thermal propagation capabilities of the runaway cells on their adjacent cells, and this explains the runaway sequence. It is anticipated that the experimental results can provide the deep understanding, thermal runaway warning, and evaluation method for the module safety design. [ABSTRACT FROM AUTHOR]

Thanks to the outstanding adaptability and relatively simple design process, machine learning-based energy management strategies (EMSs) show their superiority in reducing the energy consumption for multi-power powertrains. Given the popularity of distributed-drive in battery electric vehicles (BEVs), this study selects a dual-motor four-speed electric bus to investigate the proposed soft actor-critic (SAC)-based EMS. Two improvements are made to the traditional SAC algorithm to meet the special requirements of EMS in a distributed-drive electric bus in this study. Firstly, the introduced combination of Gumbel-SoftMax and actor-network allows the SAC agents to explore the hybrid action space (discrete operating modes and continuous torque distribution coefficients). Secondly, a heuristic rule-interposing action controller (HRIAC) is involved to reduce illogical exploration of SAC agents in searching for optimal power distributing ratio. Simulation results demonstrate that Gumbel-SoftMax facilitates agents' exploration and narrows the energy consumption gap between the proposed EMS and the global optimal EMS, meanwhile, HRIAC accelerates the convergence of agent training. The comparative results of proposed EMS performance in unknown driving cycles show that the proposed EMS outperforms other deep reinforcement learning (DRL)-based EMSs in adaptability, which is taken as a solid foundation for energy efficiency improvement of dual-motor electric bus in practice. • Gumbel-SoftMax and actor-network are involved in Soft Actor-Critic algorithm training. • Proposed Soft Actor-Critic algorithm enable output continuous-discrete actions. • Heuristic rule-interposing action controller is designed to accelerate SAC training. • Proposed EMS outperforms other typical EMSs in adaptability and energy saving. [ABSTRACT FROM AUTHOR]

Autonomous driving with active obstacle avoidance in dynamic urban environment has attracted significant attention to improve road safety and traffic efficiency. In this paper, an online optimization-based trajectory planning method for autonomous overtaking is developed to prevent collision and realize safe efficient driving. Unlike traditional methods which solve the overtaking problem with segmented reference path in multi-stages with integer variables, this paper proposes a novel dual-variable trajectory planning framework to get the optimal trajectory for active collision avoidance. First, the nonlinear non-differentiable collision-free constraint between vehicles is reformulated using dual problem optimization. Then, the optimal trajectory for longitudinal and lateral movement is obtained jointly in a receding horizon optimization framework based on the optimized dual variable considering the safety and dynamic performance. The key novelty lies in the reformulation of the nonlinear optimal trajectory planning problem and the formulated whole optimization framework can be solved using efficient open-source solvers for online computation. Simulation studies in different dynamic cases are provided to show the efficiency and robustness for safe driving. The method reaches safe overtaking performance with much less calculation time. Real-world experiment with a static obstacle demonstrates its capability for the safe efficient trajectory planning and online implementation in autonomous driving. • A real-time optimal trajectory planning method for autunomous overtaking driving is proposed. • The optimal trajectories in longitudinal and lateral directions are calculated jointly, which are kino-dynamically feasible. • Nonlinear and non-differentiable collision-free constraint is reformulated using duality of convex optimization. • The proposed method is computationally efficient and is promising for real implementation. • Complex dynamic driving scenarios are studied to evaluate the robustness. [ABSTRACT FROM AUTHOR]

• Three-layer voltage fault detection method for lithium-ion batteries is proposed. • Box-Cox transformation approach is used to normalize the raw voltage data. • The computation process has been significantly simplified by the improved K-means. • Real accident vehicles and normal vehicles data are both applied to verify the method. It is vital to detect the safety state and identify faults of the battery pack for the safe operation of electric vehicles. The voltage faults such as over-voltage and under-voltage imply more serious battery faults including short-circuit and thermal runaway. The voltage abnormal fluctuation is a warning signal of short-circuit, over-voltage and under-voltage. This paper proposes a scheme of three-layer fault detection method for lithium-ion batteries based on statistical analysis. The first layer fault detection is based on the thresholds of over-charge and over-discharge of a battery pack. In the second layer, confidence interval estimation is applied to identify risky cells. In the third layer, correlation and variability of all cells in one battery pack are analyzed by using an improved K-means method to identify abnormal voltage fluctuation over a certain period. The validity and feasibility of the proposed method are verified by real vehicle data from the National Big Data Alliance of New Energy Vehicles. [ABSTRACT FROM AUTHOR]

Battery state foretasting and health management are significant tasks for ensuring safety and stability of battery systems. Accurate state estimation can not only provide valuable parameters for energy management but also may prolong battery usage lifespan. Comprehensive theoretical analysis and practical application, differential thermal voltammetry analysis method has great potentials in actual operation. This paper proposes a closed-loop battery capacity estimation framework, Gaussian process regression and multi-output Gaussian process regression for constructing battery dynamic state-space function, to improve the accuracy and robustness of battery SOH estimation. Firstly, a time-series model of battery capacity degradation is established as the state equation using Gaussian process regression. Secondly, two strong correlation indicators are treated as observed parameters to construct an observation equation through multi-output Gaussian process regression, where the health indicators are extracted from the partial smoothed curves by two filter methods. Thirdly, particle filter algorithm is employed to correct the prior estimated capacity and suppress noise perturbations for achieving closed-loop control. Additionally, the performances of particle filter algorithm with different particle sizes are discussed and analyzed from accuracy and computational time aspects. Verification of three types of batteries indicates that the proposed method has an excellent capability for battery capacity estimation. • A data-fusion and closed-loop battery capacity estimation framework is proposed. • The health indicators are extracted considering peak position, peak and valley values. • State-space equation of battery aging model is establish using GPR and MOGPR algorithms. • The performances of PF algorithm with different sizes of particle numbers are discussed and analyzed. [ABSTRACT FROM AUTHOR]

Electrochemical energy storage systems are fundamental to renewable energy integration and electrified vehicle penetration. Hybrid electrochemical energy storage systems (HEESSs) are an attractive option because they often exhibit superior performance over the independent use of each constituent energy storage. This article provides an HEESS overview focusing on battery-supercapacitor hybrids, covering different aspects in smart grid and electrified vehicle applications. The primary goal of this paper is to summarize recent research progress and stimulate innovative thoughts for HEESS development. To this end, system configuration, DC/DC converter design and energy management strategy development are covered in great details. The state-of-the-art methods to approach these issues are surveyed; the relationship and technological details in between are also expounded. A case study is presented to demonstrate a framework of integrated sizing formulation and energy management strategy synthesis. The results show that an HEESS with appropriate sizing and enabling energy management can markedly reduce the battery degradation rate by about 40% only at an extra expense of 1/8 of the system cost compared with battery-only energy storage. • Hybrid energy storage systems for electrified vehicle and smart grid are surveyed. • The operation principles and energy storage system requirements are provided. • System configuration, converter design and energy management are covered. • A case study is presented for integrated sizing and control strategy development. [ABSTRACT FROM AUTHOR]

Accurate state of health (SOH) estimation of batteries in vehicle-to-grid (V2G) applications is critical for the safety of vehicles, chargers, and grids. In this paper, a novel SOH estimation model based on inverse ampere-time integration and natural gradient boosting (NGBoost) is proposed and verified by real operating data from V2G electric vehicles (EVs). Raw data is segmented, where segments of charging and grid-feeding with relatively smooth currents are selected to constitute the dataset for research. On top of this, five model inputs have been selected through Pearson correlation and physical meaning analysis. Hyperparameters of the model are optimized and model performance is compared with the other nine commonly used machine learning methods. The results show that the NGBoost model has the highest estimation accuracy with the mean absolute percentage error and root mean squared error of 1.484% and 2.302 A h, respectively. To increase the transparency of the model, the Shapley additive explanation (SHAP) method is utilized to provide a full explanation of its predictions. Furthermore, the model is validated to be robust to noise and shows great potential for integration into embedded battery management systems for fast and accurate SOH estimation in V2G applications. • A novel SOH estimation model based on NGBoost is described in detail. • Real-world operation data from V2G electric vehicles is derived. • Effectiveness and robustness are verified by real-world data and data with noise. • A credible comparison with nine other machine learning methods is proposed. • A full explanation of the model's predictions is offered by using the SHAP method. [ABSTRACT FROM AUTHOR]

Detecting battery safety issues is essential to ensure safe and reliable operation of electric vehicles (EVs). This paper proposes an enabling battery safety issue detection method for real-world EVs through integrated battery modeling and voltage abnormality detection. Firstly, a battery voltage abnormality degree that is adaptive to different battery types and working conditions is defined. Then an integrated battery model is developed by combining an electrochemical model, an equivalent circuit model (ECM), and a data-driven model to evaluate the normal voltage. To ensure normality of input current, a current processing model is presented. The performance of the proposed scheme is examined under random loading profiles using operating data collected from real-world EVs. The results show that the integrated battery model can precisely predict normal battery terminal voltage, with mean-squared-errors of 1.034e−4 V 2 , 7.221e−5 V 2 , and 4.612e−5 V 2 for driving, quick charging, and slow charging, respectively. The accuracy in classifying faulty and normal batteries is verified based on the operating data collected from 20 EVs. • A robust parameter estimation method is proposed for equivalent circuit model. • An integrated battery model is proposed for normal voltage evaluation. • A voltage abnormality degree is defined for battery safety issue detection. [ABSTRACT FROM AUTHOR]

• A scheme is firstly proposed to select battery safety-representing feature in real-world scenarios. • A method is proposed to construct the actual feature distribution. • A model is constructed to analyze the real-world influencing factors. • Statistical analysis of battery safety from four dimensions is deployed. • All the schemes/models are based on numerous real-world data to reflect real-world conditions. Battery safety issues have been raising anxiety of electric vehicle (EV) consumers and threatening to the occupant's life. Many battery safety management/monitoring methods and features have been reported, but which features can better represent battery safety in real-world EVs has been rarely discussed and compared in the unified level. In this study, the multi-feature and multi-dimension statistical analysis for battery pack safety in numerous real-world electric EVs is deployed. Firstly, an EV state distinction scheme is proposed to cope with the different characteristics of EV driving and EV charging, and twenty statistical features are constructed to extract the characteristics of battery voltages. Then, a scheme is proposed to select the feature which can effectively represent battery safety in real-world EVs. A multiple iteration Gaussian mixture model is proposed in the scheme to construct the actual feature distribution and cope with the random sensor error and data outlier. Finally, a model is constructed to analyze the statistical dimensions of selected battery safety-representing feature, and statistical analysis from different seasons, mileages, state-of-charges, and EV states is deployed. Results showcase that the selected feature is competent to distinguish between normal and unsafe batteries in real-world EV packs of different conditions. Various interesting conclusions based on the selected feature are also summarized to explore some common laws of battery safety in real-world EV application and support for the establishment of battery safety management method. [ABSTRACT FROM AUTHOR]

In 2020, China launched the Dual Circulation Strategy to tackle difficult international trade situations and COVID-19. To examine the environmental pressure induced by China's domestic and international trade from the perspective of dual circulation, this study quantifies the flows of embodied energy and virtual water in China's domestic and international trade and analyzes the driving factors to the net outflow, using the multi-regional input-output model and spatial structural decomposition method. The main findings of this study are: (1) Domestic trade of China alleviated the pressures of both energy and water shortage through the flow of embodied energy from Northern to Southern and Eastern regions and the flow of virtual water from Southern to Northern regions. In contrast, international trade yet increased China's environmental burden and tends to be unsustainable. (2) Production structure effect was the key factor to the net outflow of embodied energy (contribution degree: 58%) and virtual water (contribution degree: 46%) to major economies. For China, the export scale effect and export structure effect played an important role in increasing the net outflow of China's embodied energy, with contribution degrees of 17% and 23%, respectively, while the export structure effect and water efficiency effect reduced the net outflow of virtual water, with contribution degrees of −23% and −19%, respectively. Multidimensional Policies, including industrial structure transition, renewable energy deployment, resource efficiencies improvement, trade pattern upgrade, and regional free trade promotion, are suggested for implementing the Dual Circulation Strategy, considering environmental sustainability and industrial competition advantages. • Domestic trade alleviated the pressures of both energy and water shortage. • International trade increased China's environmental burden. • Nearly half of energy and water net outflow by the production structure effect. • 40% increase of energy net outflow by export scale and structure effects. • 42% decrease in water net outflow by export structure and water efficiency effects. [ABSTRACT FROM AUTHOR]

The state of resistance growth (SOR) for batteries on hybrid electric vehicles (HEV) is difficult to be estimated precisely online, due to the current and especially the temperature are uncontrollable, which make the battery health state always is hard to be known. Accurate estimation of resistance online and gaining the resistance of battery at the begin of life (BOL) are the two keys to address the problem. As direct current resistance (R dc) is a world-wide used evaluation parameter for battery through off-line test, algorithm for online R dc calculating is proposed in this paper, adopting a fractional order model and recursive least square identification with the virtual current design. Comparing with the common resistance estimation, R dc estimated here is comparable with the offline tested, with an error less than 5%. With the R dc provided in the specification for batteries at BOL, SOR with small error (1.3%) and high universality is obtained. It can accurately diagnose batteries health for HEV, and give early alarm for cell with abnormal or large resistance growth. It addresses the urgent evaluation problem especially for the old vehicles operated for years since this algorithm doesn't rely on the historical estimated resistance data. [Display omitted] • Direct current resistance is estimated online with an error less than 5%. • Computational burden is relieved by skipping the parameter solve process. • Resistance growth state is obtained online with an average error of 1.3%. • It provides a novel alarm function against the resistance growth especially for HEV. • SOR owns high universality as it doesn't rely on the historical estimated data. [ABSTRACT FROM AUTHOR]

• The proposed method is based on unsupervised learning, avoiding the problems of lack of samples; • A knowledge-based guidance matrix is established to accelerate training speed and to ensure convergence; • An interpretable Autoencoder is used to ensure extensibility among different vehicles; • Effectiveness has been verified using data from real-world EVs. The abnormality detection of lithium-ion battery pack is crucial to ensure the safety of electric vehicles (EVs). However, the dynamic and complex operating conditions of EVs making it challenging for algorithms designed under laboratory conditions to perform properly. In this study, a novel data-driven framework for abnormality detection is developed through establishment of a neural network with interpretable modules on top of an Autoencoder using data from real EVs to recognize abnormality while charging. The encoding guide matrix proposed in this method greatly accelerates the training speed, which also helps retains the learning ability of the neural network with consideration of the influence from each feature to provide supplementary information. The proposed algorithm is validated with data from real EVs. The results show that, compared with most existing algorithms, evidently higher accuracy can be achieved with shorter training time and lower computational cost, where the accuracy remains above 94% for all tested sample and the average root mean square error (RMSE) is as small as 0.03913. The proposed method can be utilized for both cloud-based and vehicle-based battery fault diagnoses. [ABSTRACT FROM AUTHOR]

• The overcharging behaviors of lithium-ion batteries are investigated. • The influence of overcharge depth and number of cycles are examined. • The EIS and IC-DV analysis are used to diagnose cells degradation. • Degradation modes during overcharging are identified and quantified. • Non-destructive fault diagnosis method is put forward. This paper explores the major degradation characteristics of commercial lithium-ion battery cells with nickel–cobalt-aluminum-oxide (NCA) electrode during cyclic overcharging, and proposes non-destructive methods for detecting overcharging degradation failure. The experimental results show that battery capacity drops significantly with increasing overcharge depth and number of cycles especially during the first three cycles and when the charging termination voltage is set to 5 V. At the same time, the cell overcharge tolerance decreases with the cyclic overcharging. The combination of the electrochemical impedance spectroscopy and the incremental capacity and differential voltage analysis is used to diagnose cell degradation during cyclic overcharging. Three main degradation modes are identified and quantified by extracting characteristic parameters such as internal resistance and peak, valley, and curve position changes of incremental capacity curves. It is concluded that loss of lithium inventory and loss of active materials are the most dominant degradation modes during cyclic overcharging. Besides, the sharp increase of the third peak on incremental capacity curves has been identified as a unique feature of overcharging degradation, which can be used for diagnosing cyclic overcharging-induced degradation for batteries with NCA cathode. [ABSTRACT FROM AUTHOR]

• A synchronous multi-time-step and multi-parameter prediction model is constructed. • A method is firstly proposed to achieve local-optimized hyperparameters for data-driven model. • A scheme is proposed to consider complex factors for real-world operation. • A scheme is designed for real-world application of fault prognosis and energy consumption prediction. • All the schemes/models are trained/verified by real-world data to reflect real-world conditions. Proton exchange membrane fuel cell vehicle (PEMFCV) is considered to be a promising way to cope with the environment pollution and energy exhaustion. But PEMFCV is also suffering from some difficulties for the real-world application, including different kinds of faults and hydrogen refueling. This raises requirement for timely and accurate fault prognosis and energy consumption prediction. However, limited and sparse parameters obtained by onboard sensors and random influencing factors during the real-world vehicular operation make the PEMFCV hardly be modelled. To cope with the issue, this study firstly puts forward a vehicle state-driving behavior factor (VDF) construction of PEMFCV to consider as comprehensive factors as possible and uses maximal information coefficient (MIC) to extract related factors for model training. Then a data-driven model is constructed to achieve synchronous multi-time-step and multi-parameter prediction for the fuel cell and hydrogen system in real-world PEMFCVs. The model considers noise denoising, spatial feature processing and temporal feature processing by combining convolutional neural network (CNN) and gated recurrent unit neural networks (GRU). To optimize numerous hyperparameters of constructed data-driven model, a "discrete gradient-based optimization" method (DGO) is first proposed to achieve local-optimized hyperparameters as well as reduce the time complexity of grid searching. Based on the predicted parameters of PEMFCVs, a scheme is designed for fault prognosis and energy consumption prediction. All the procedure is trained and verified by real-world operation data to reflect the real-world applicable conditions for different seasons and vehicles. Results show that the proposed model can achieve accurate 30-time-step synchronous fuel cell temperature, hydrogen temperature, and hydrogen pressure prediction with mean-relative-errors (MREs) of 0.54%, 0.85% and 0.71%; The fault can be accurately prognosed five minutes ahead with MRE of 0.71% to provide driver sufficient time to deal with the fault; The small MRE of 6.3% for energy consumption prediction also indicates the effectiveness of proposed model. The proposed whole method can be applied in the real-world PEMFCVs for long-time-step parameter prediction, fault prognosis, and energy consumption prediction. The proposed DGO can be used to optimize other data-driven methods. [ABSTRACT FROM AUTHOR]

Electrifying transportation in the form of the large-scale development of electric vehicles (EVs) plays a pivotal role in reducing urban atmospheric pollution and alleviating fossil fuel dependence. However, the rising scale of EV deployment is exposing problems that were previously hidden in small-scale EV applications, and the lack of large-scale EV operating data deters relevant explorations. Here, we report several issues related to the battery utilization and energy consumption of urban-scale EVs by connecting three unique datasets of real-world operating states of over 3 million Chinese EVs, operational data, and vehicle feature data. Meanwhile, by incorporating climatic data and EV data outside China, we extend our models to several metropolitan areas worldwide. We find that blindly increasing the battery energy of urban EVs could be detrimental to sustainable development. The impact of changes in the energy consumption of EVs would be exacerbated in large-scale EV utilization, especially during seasonal shifts. For instance, even with a constant monthly driving demand, the average energy consumption of Beijing light-duty EVs would change by up to 21% during winter-spring shifts. Our results may also prove useful for research on battery resources, urban power supply, environmental impacts, and policymaking., Competing Interests: The authors declare no competing interest.