Article
Chemistry, Physical
Tanvir R. Tanim, Sangwook Kim, Andrew M. Colclasure, Zhenzhen Yang, Kevin Gering, Peter J. Weddle, Michael Evans, Eric J. Dufek, Yulin Lin, Jianguo Wen, Francois Usseglio-Viretta, Alison R. Dunlop, Stephen E. Trask, Kandler Smith, Brian J. Ingram, Andrew N. Jansen
Summary: Achieving 10-minute extreme fast charging while maintaining good charge acceptance and cycle life is a challenging task in the design of lithium ion batteries. This study proposes combining multiple solutions, including materials-to-electrode design-to-charging protocols, to overcome limitations in lithium-ion transport and enable 10-minute extreme fast charging in batteries.
JOURNAL OF POWER SOURCES
(2023)
Article
Energy & Fuels
Junzhe Shi, Min Tian, Sangwoo Han, Tung-Yan Wu, Yifan Tang
Summary: Electric vehicles (EVs) are gaining popularity and it is crucial to accurately estimate the Remaining Charging Time (RCT) for EVs. However, finding an algorithm that achieves accurate RCT estimation is a challenge. This study proposes an algorithm that updates charging accuracy online and predicts charging current profiles to improve RCT estimation.
JOURNAL OF ENERGY STORAGE
(2022)
Article
Thermodynamics
Chen Zhang, Hongmin Wang, Lifeng Wu
Summary: This paper proposes a model for predicting the cycle life of lithium-ion batteries using charging data and discharge curve features. The model combines the features from both charging and discharging processes and uses random forest regression for prediction. Experimental results show that the proposed method outperforms existing algorithms in predicting battery cycle life.
Article
Energy & Fuels
Dinghong Chen, Weige Zhang, Caiping Zhang, Bingxiang Sun, Linjing Zhang, Xinwei Cong
Summary: This paper proposes a comprehensive data-driven rapid lifetime prediction method for lithium-ion batteries under different charging protocols. The method establishes the relationship between charging conditions and battery lifetime, and employs a deep neural network model with Bayesian optimization algorithm for accurate prediction. Experimental results show that the proposed approach outperforms traditional shallow machine learning algorithms in predicting battery degradation trajectory and end-of-life.
JOURNAL OF ENERGY STORAGE
(2023)
Article
Engineering, Electrical & Electronic
Yanchang Liang, Zhaohao Ding, Tianyang Zhao, Wei-Jen Lee
Summary: Battery swapping-charging systems (BSCSs) can provide better battery swapping services for electric vehicles (EVs) in large cities. This paper models the real-time optimization scheduling problem as a decentralized partially observable Markov decision process (Dec-POMDP) and solves it using multi-agent deep reinforcement learning (MADRL) algorithms. To address dynamic hard constraints, MADRL is combined with binary integer programming (BLP) to propose the VDN-BLP algorithm. Simulation results based on historical battery swapping data in Sanya City validate the effectiveness of the proposed method.
IEEE TRANSACTIONS ON SMART GRID
(2023)
Article
Automation & Control Systems
Mao Tan, Zhuocen Dai, Yongxin Su, Caixue Chen, Ling Wang, Jie Chen
Summary: With the increase in the number of electric vehicles, battery swapping is seen as promising due to its short waiting time. However, it is challenging to achieve efficient scheduling in a large scale battery swap station due to the uncertainty of the power grid and EV behavior. To address this, a new bi-level scheduling model is proposed, combining deep reinforcement learning for optimal power allocation and MILP subproblems for battery dispatching. Experimental results show excellent performance and cost reduction, benefiting both the battery swap station and the power grid in peak shaving and valley filling.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Engineering, Civil
Shuai Mao, Yan Wang, Quanxue Guan, Yunjian Xu
Summary: This study addresses the joint battery charging and replenishment scheduling problem in a battery swapping charging system, taking into account random electric vehicle arrivals, renewable generation, and electricity prices. The proposed approach integrates structural properties, such as threshold-charging and least demand first structures, to reduce the dimensionality of the action space. Experimental results demonstrate that the proposed approach outperforms various structural charging and replenishment policies as well as a vanilla soft actor-critic algorithm, achieving significant cost savings of 7.16%-78.61% and 6.53%-93.73%.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Thermodynamics
Chunsheng Hu, Liang Ma, Shanshan Guo, Gangsheng Guo, Zhiqiang Han
Summary: This paper proposes a method for estimating the state of charge (SoC) of LiFePO4 batteries during the charging process using a deep neural network (DNN). Battery data collected from different charging protocols are used to train the DNN model. The developed DNN can accurately estimate the battery's SoC during charging and can be used to calculate the SoC during discharging. Experimental results show that the maximum error and root mean square error of the SoC estimation using DNN are within an acceptable range.
Article
Energy & Fuels
Arpitkumar J. Patel, Amit V. Sant
Summary: The widespread development of charging infrastructure is crucial for the projected growth of electric vehicles. This paper proposes a control method for shunt active power filters based on a moving window min-max algorithm to prevent current distortions from affecting the grid in charging stations. Experimental results demonstrate that the proposed method offers faster dynamic and accurate steady-state responses.
Article
Computer Science, Information Systems
Sang-Won Lee, Young-Kyun Cho
Summary: This paper proposes a single-stage wireless battery charging circuit that is capable of constant current and constant voltage charging modes without the need for complex control algorithms or additional components, showcasing excellent performance.
Article
Engineering, Industrial
Xiaodong Xu, Shengjin Tang, Xuebing Han, Languang Lu, Yu Wu, Chuanqiang Yu, Xiaoyan Sun, Jian Xie, Xuning Feng, Minggao Ouyang
Summary: This paper proposes a fast capacity prediction method, AM-Bi-LSTM neural network, which incorporates a novel deep aging mechanism and bidirectional long-short term memory layers. The method accurately predicts the battery capacity and reflects the entire charging curves using a physical informed aging mechanism layer and a deep learning framework. Case studies demonstrate the method's effectiveness and discussions are provided on the influence of voltage window length on capacity prediction. The results show that the proposed method achieves higher accuracy, faster prediction speed, and stronger robustness compared to other data-driven methods.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Chemistry, Multidisciplinary
Zafer Acar, Phu Nguyen, Kah Chun Lau
Summary: Ionic liquids have great potential for energy storage and conversion devices, but their practical application is limited due to unfavorable melting points. Accurate prediction of the melting points is important for fine tuning their properties. A deep-learning model was used to predict the melting points of various ionic liquids, achieving high accuracy and providing useful design rules for tuning their melting points.
APPLIED SCIENCES-BASEL
(2022)
Article
Thermodynamics
Zhang Chen, Wenjing Shen, Liqun Chen, Shuqiang Wang
Summary: In this paper, a long short-term memory network-based transfer learning model is proposed to predict battery capacity under fast charging. By introducing a novel voltage feature and using a cross-validation method to derive optimal hyperparameters, the adaptability and prediction accuracy of the model are improved. Experimental results demonstrate the effectiveness and applicability of the proposed method.
Article
Construction & Building Technology
Wei Jin, Qiming Fu, Jianping Chen, Yunzhe Wang, Lanhui Liu, You Lu, Hongjie Wu
Summary: This paper proposes a novel method for building energy consumption prediction using Deep Reinforcement Learning (DRL) with consideration of fluctuation points. The method improves prediction accuracy and convergence stability compared to other comparable methods and representative DRL methods. The experimental results show the effectiveness of the proposed method.
JOURNAL OF BUILDING ENGINEERING
(2023)
Article
Energy & Fuels
Yongjun Pan, Xiaoxi Zhang, Yue Liu, Huacui Wang, Yangzheng Cao, Xin Liu, Binghe Liu
Summary: In this paper, a rapid stress prediction method based on a deep neural network algorithm is proposed to select the thicknesses and materials of electric vehicle battery-pack system (BPS) under crush scenarios. By establishing a nonlinear FE model and training historical data, the method accurately predicts the stresses of the modules, demonstrating high efficiency in designing safe and durable BPS.
Article
Public Administration
Jiahuan Lu, Wan-Ju Hung
Summary: This study conducts a meta-analysis on the antecedents of contracting back-in and finds that it is driven by market management strategies and political moves. The study also highlights that government decisions to contract out and contract back-in may be based on different considerations.
INTERNATIONAL REVIEW OF ADMINISTRATIVE SCIENCES
(2023)
Article
Political Science
Jiahuan Lu, Shanshan Guan, Qiang Dong
Summary: The level of commercialization in Chinese nonprofits is modest overall, but varies significantly among organizations. Resource insufficiency, government connections, and environmental munificence drive nonprofit commercialization in China.
PUBLIC ADMINISTRATION
(2023)
Article
Chemistry, Applied
Rui Xiong, Jinpeng Tian, Weixiang Shen, Jiahuan Lu, Fengchun Sun
Summary: This study proposes a convolutional neural network (CNN) based method for accurate battery capacity estimation using only raw impedance spectra as input. An input reconstruction module is designed to effectively utilize impedance spectra without corresponding capacities, reducing the cost of collecting training data. The proposed method outperforms supervised benchmarks, even with fewer samples with measured capacities, and can reduce the root mean square error by up to 50.66%.
JOURNAL OF ENERGY CHEMISTRY
(2023)
Article
Green & Sustainable Science & Technology
Jackson Hannagan, Rhys Woszczeiko, Thomas Langstaff, Weixiang Shen, John Rodwell
Summary: In recent years, there has been a noticeable change in the flow of reactive power in power network systems worldwide. This change is influenced by the increasing use of LED lights and battery-powered devices with switch-mode power supplies in residential households. The study examined the power characteristics of 56 modern appliances and devices, revealing significant changes in their electrical behavior. LED technology and switch mode power supplies are identified as key contributors to increased household reactive power injection, particularly in line with government programs promoting their adoption. Addressing this issue may involve government requirements for appliance manufacturers to display power factor information or the development of products with built-in power factor correction. Overall, this research emphasizes the need to consider unintended consequences on overall electrical system sustainability in energy efficiency efforts.
Article
Engineering, Electrical & Electronic
Tao Rui, Zheng Yin, Cungang Hu, Geye Lu, Pinjia Zhang, Weixiang Shen, Wenping Cao, Xinghuo Yu
Summary: This paper proposes a novel modulated MFPCC method to improve the current performance of voltage source inverters. Real-time current gradient updating and sampling disturbance suppression are achieved. Simulation and experimental results demonstrate the advantage of the proposed method in reducing the total harmonic distortion of output current.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2023)
Article
Engineering, Electrical & Electronic
Yiming Xu, Xiaohua Ge, Ruohan Guo, Weixiang Shen
Summary: This article proposes a novel soft short-circuit fault diagnosis algorithm that can accurately estimate the SC resistance with limited charging data, providing simultaneous fault detection and estimation for EV batteries.
IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION
(2023)
Article
Energy & Fuels
Rui Xiong, Kui Zhang, Siyu Qu, Jinpeng Tian, Weixiang Shen
Summary: A new method using an improved wireless charging system to obtain onboard AC power is developed, which heats the lithium-ion battery by high-frequency AC. Experimental results show that the battery can be heated from -20 degrees C to 0 degrees C in 142s with an average heating rate of about 8.5 degrees C/min using 85 kHz AC. This proposed heating system provides a fast heating solution for lithium-ion batteries in electric vehicles.
Article
Energy & Fuels
Kaixuan Zhang, Rui Xiong, Qiang Li, Cheng Chen, Jinpeng Tian, Weixiang Shen
Summary: This study proposes a pseudo-open circuit voltage (OCV) modeling method to improve the performance of a closed-loop feedback correction. First, the relationship between the derivative of OCV and SOC is analyzed to select the appropriate SOC interval for OCV curve construction. Then, a pseudo-OCV curve is established and compared with the real OCV to determine the optimal construction interval. Finally, SOC estimation is carried out using the constructed full pseudo-OCV and compared with the reference SOC, with results showing an SOC estimation error of less than 3% over the full SOC range.
Article
Electrochemistry
Chen Zhu, Liqing Sun, Cheng Chen, Jinpeng Tian, Weixiang Shen, Rui Xiong
Summary: In this paper, a method is proposed to achieve reliable degradation diagnosis and accurate State-of-Health (SOH) estimation of lithium-ion batteries (LiBs). The method includes reconstructing open-circuit voltage, developing a degradation diagnosis model, and selecting an appropriate voltage range for improved accuracy. Experimental results show a maximum error of 1.44%.
ELECTROCHIMICA ACTA
(2023)
Article
Engineering, Civil
Zeyu Chen, Rui Xiong, Xue Cai, Zhen Wang, Ruixin Yang
Summary: The regenerative braking control strategy of DDEVs under varying road slope is studied in this paper. The vehicle dynamic characteristics during downhill driving are analyzed, and the specific impacts of road slope on braking control are disclosed. An online co-estimation method for road slope and vehicle mass is proposed, and the power allocation is optimized to achieve optimal braking torque split. The proposed strategy provides better braking performance and higher energy recovery compared to traditional methods, with up to 9.62% improvement in energy recovery under certain driving conditions.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Engineering, Civil
Ruohan Guo, Cungang Hu, Weixiang Shen
Summary: This paper proposes an adaptive approach for online co-estimation of SOC and SOP of lithium-ion batteries considering temperatures. It constructs a fractional-order multi-model system (FO-MMS) by integrating three sub-models at different temperatures and adapts the contribution coefficient of each sub-model through a temperature-embedded algorithm. An FO-MM-PIO is designed for SOC estimation and a iterative approaching algorithm is derived to estimate SOP. Experimental validations show that the proposed method achieves high accuracy in SOC and SOP estimation, even at sub-zero temperatures.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Engineering, Mechanical
Rui Xiong, Baoqiang Zhu, Kui Zhang, Yanzhou Duan, Fengchun Sun
Summary: This study developed a battery big data platform to manage battery operation data in real time. The platform includes an electric vehicle, charging and heating systems, a data transmission network, and a battery management system. Experimental validation demonstrates the efficiency and reliability of the platform, with battery state of charge estimation as an example.
CHINESE JOURNAL OF MECHANICAL ENGINEERING
(2023)
Article
Automation & Control Systems
Ruohan Guo, Yiming Xu, Cungang Hu, Weixiang Shen
Summary: This study incorporates the Butler-Volmer equation and the fractional-order model representation into a model-based physics-informed neural network to simulate the dynamics of battery charge transfer in electric vehicles under different operating conditions. By using an adaptive neural network-based fractional-order observer, accurate online state of charge estimation is achieved.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2023)
Article
Automation & Control Systems
Ruohan Guo, Weixiang Shen
Summary: This article proposes an IA-MAFF-RLS method to identify model parameters of lithium-ion batteries in electric vehicles. The method utilizes information analysis and adaptive strategies to accurately identify the parameters and avoid accuracy loss in model transformation.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Chemistry, Physical
Yu Tian, Cheng Lin, Hailong Li, Jiuyu Du, Rui Xiong
Summary: A deep learning method is proposed for the detection and quantification of lithium plating based on charge curves. The method achieves high accuracy and adaptability, and can accurately predict lithium plating under uncertain conditions.