Subspace‐based modeling and parameter identification of lithium‐ion batteries

For reliable and safe operation of lithium‐ion batteries in electric vehicles, the monitoring of state‐of‐charge and state‐of‐health is necessary. However, these internal states cannot be measured directly, which are usually estimated through model‐based techniques. Therefore, an accurate application‐oriented battery model is of significant importance. The purpose of this paper is to present a novel method on battery modeling and parameter identification. In this work, a state‐space model with clear mathematical and electrochemical meanings is proposed on the basis of the electrochemical basics of lithium‐ion batteries. The frequency‐domain characteristics of the lithium‐ion batteries are also investigated. On the basis of the frequency analysis, an identification test profile that can excite the dynamic characteristics of the battery fully and persistently is proposed. A subspace‐based algorithm is then adopted to identify the parameters of the battery model. The performance and robustness of the estimated model are validated through some experiments and simulations. The validation results show that the proposed method can achieve an acceptable accuracy with the maximum error being less than 2%. Copyright © 2013 John Wiley & Sons, Ltd.     

Research progress on fre protection technology of LFP lithium-ion battery used in energy storage power station

随着电化学储能市场的蓬勃发展,电化学储能电池本身的安全性越来越受到关注,如何最大程度地降低储能电池组火灾风险是电化学储能大规模应用时亟需解决的问题。本文综述目前国内外针对锂离子电池热失控已有的研究成果,包括磷酸铁锂电池的燃烧特性、火灾危险等级以及在储能电站预警系统中应用的锂离子电池热失控及热扩散参数;梳理不同灭火剂对电池火灾的灭火效率;同时总结电化学储能电站的灭火系统选择,为电网储能工程应用提供参考,有效支持锂离子储能电池的大规模工程需求。     

Analysis of the effect of resistance increase on the capacity fade of lithium ion batteries

Lithium ion cells, when cycled, exhibit a two‐stage degradation behavior characterized by a first linear stage and a second nonlinear stage where degradation is rapid. The multitude of degradation phenomena occurring in lithium ion batteries complicates the understanding of this two‐stage degradation behavior. In this work, a simple and intuitive model is presented to analyze the coupled effect of resistance growth and the shape of the state of charge (SOC)‐open circuit voltage (OCV) relationship in representing the complete degradation behavior. The model simulations demonstrate that a single resistance that increases linearly on cycling can capture the transition from slow to fast degradation, primarily due to the shape of the SOC‐OCV curve. Further, the model simulations indicate that the shape of the degradation curve depends strongly on the magnitude of current at the end of discharge of the cycling protocol. To verify these observations, specific experiments are designed with minimal capacity loss but with shrinking operating voltage ranges that result in shrinking operating OCV range. The results of the experiments validate the observations of model simulations. Further, long‐term cycling experiment with a commercial lithium ion cell shows that the operating OCV range shrinks substantially with aging and is a major reason for the observed accelerated degradation. The analysis of the present work provides significant insights towards developing simple semiempirical models suitable for battery life management in microcontrollers.     

Electro-thermal battery model identification for automotive applications

This paper describes a model identification procedure for identifying an electro-thermal model of lithium ion batteries used in automotive applications. The dynamic model structure adopted is based on an equivalent circuit model whose parameters are scheduled on the state-of-charge, temperature, and current direction. Linear spline functions are used as the functional form for the parametric dependence. The model identified in this way is valid inside a large range of temperatures and state-of-charge, so that the resulting model can be used for automotive applications such as on-board estimation of the state-of-charge and state-of-health. The model coefficients are identified using a multiple step genetic algorithm based optimization procedure designed for large scale optimization problems. The validity of the procedure is demonstrated experimentally for an A123 lithium ion iron-phosphate battery.     

Development of a universal modeling tool for rechargeable lithium batteries

An interesting universal modeling tool for rechargeable lithium batteries is presented in this paper. The generic model is based on an equivalent circuit technique commonly used in electrochemical impedance characterization. Therefore, the parameters used in the model can be easily parameterized from the electrochemical impedance derivations, which provide a convenient integration with experimental cell characterizations. Such integration offers the universality in this modeling approach.     

Fundamental scientific aspects of lithium ion batteries(ⅫⅠ) —Electrochemical measurement

电化学测量方法在锂离子电池研究中有着广泛的应用,常用于电极过程动力学基本信息的测量。本文首先简述了电化学测量的基本原理、电化学极化、测量方法特点等,并讨论了常见的测量方法在锂电池基础研究中的应用,包括循环伏安,电化学阻抗谱、恒电流间歇滴定、电位弛豫技术。     

全固态锂电池界面的研究进展

与传统锂离子电池相比,基于无机固体电解质的全固态锂电池,具有安全性能高、循环寿命长、能量密度高等优点,是目前锂电池研究领域的热点之一,未来有望在电动汽车和智能电网等领域得到广泛应用。全固态锂电池中,电极与固体电解质之间的固固接触相比固液接触具有更高的界面接触电阻,同时,界面相容性和稳定性也显著影响全固态锂电池的循环性能和倍率性能。而在固体电解质中,晶界电阻决定了电解质整体的离子电导率,因此,界面问题是决定电池电化学性能的关键所在。本文重点综述了全固态锂电池中各种界面问题的研究现状,主要包括界面调控机理、修饰方法,并指出全固态锂电池中界面调控面临的挑战。     

A comparative study of lumped equivalent circuit models of a lithium battery for state of charge prediction

Battery modeling plays an important role in remaining range prediction and battery management system development. An accurate and realistic battery model is essential to design an efficient electric storage system. The goal of this paper is to investigate the performance of different circuit topologies for diffusion process in the equivalent circuit models (ECMs). The theory of diffusion process approximation by using resistive‐capacitor (RC) networks is explained in frequency domain. The terminal voltage predictive capabilities of the ECMs are compared and validated with test data. The numerical simulation results show that model prediction accuracy and computation burdens increase along with the number of RC pairs. The ECM with three RC networks is the best choice in terms of the balance between accuracy and complexity for ternary lithium batteries. In addition, a novel method of combining unscented Kalman filter (UKF) algorithm with initial state of charge (SOC) acceleration convergence for SOC estimation is proposed. The results of urban dynamometer driving schedule (UDDS) show that ECM with three RC networks has the best comprehensive performance on calculation cost and SOC estimation accuracy.     

全固态锂离子电池关键材料研究进展

全固态锂离子电池采用固态电解质替代传统有机液态电解液,有望从根本上解决电池安全性问题,是电动汽车和规模化储能理想的化学电源。为了实现大容量化和长寿命,从而推进全固态锂离子电池的实用化,电池关键材料的开发和性能的优化刻不容缓,主要包括制备高室温电导率和电化学稳定性的固态电解质以及适用于全固态锂离子电池的高能量电极材料、改善电极/固态电解质界面相容性。本文以全固态锂离子电池关键材料为出发点,综述了不同类型的固态电解质和正负极材料性能特征以及电极/电解质界面性能的调控和优化方法等,阐述了未来全固态锂离子电池关键材料的发展方向以及界面问题的解决思路,为探索全固态锂离子电池产业化前景奠定基础。     

Second-order RC modeling and parameter identification of electric vehicle power battery

建立精确的动力电池模型是电池管理系统（battery management system,BMS）开发过程中的重要环节,电池系统具有较强的非线性特性,其模型参数随多种因素的变化而变化。在电池模型参数辨识过程中,考虑的可变因素越多,辨识结果越准确,但模型的运行速度将降低,影响其实际应用。在各种可变因素中,电池荷电状态（state of charge,SOC）对电池模型参数的影响最为显著,对不同SOC下电池模型参数进行辨识并应用于电池模型,将在提高模型精度的同时保持较好的实时性。本文以动力锂电池为对象,采用二阶RC等效电路模型,通过试验得到电池组在不同SOC下的回弹电压数据,采用最小二乘拟合法辨识不同SOC状态下的模型参数。在此基础上搭建模型参数随SOC变化的实时仿真模型,并对模型进行仿真和试验验证,结果表明模型具有较高的精度和实时性。     

Fundamental scientific aspects of lithium ion batteries (XV) ——Summary and outlook

自摇椅式可充放锂电池概念由Armand M等人在1972年提出,锂离子电池的基础研究历经43年,在材料体系、电化学反应机理、热力学、动力学、结构演化、表界面反应、安全性、力学行为等方面不断取得更为深入广泛的认识,并最终推动锂离子电池技术发展和成功实现了商业化。锂离子电池面临着电池性能需要全面提升、应用领域需进一步拓宽的强劲需求,因此要求基础研究能够提供创新的、更好的技术解决方案,对锂离子电池材料复杂的构效关系能精确认识,对于电池在制造和服役过程中的失效机制有全面的理解,对各种控制策略的效果能提供可靠的科学依据。同时,锂离子电池的发展也在促进着固态电化学、固态离子学、能源材料、能源物理、纳米科学等交叉基础学科的发展。作为“锂离子电池基础科学问题”讲座的最后一篇文章,本文对锂离子电池基础研究的科学问题,存在的难点、发展趋势进行了总结。     

SOC estimation of lithium battery based online parameter identification and AEKF

SOC的准确估计对提高电池的动态性能和能量利用效率至关重要,估计过程中,模型参数不准确以及系统噪声的不确定性都会对结果产生较大影响。为减小模型参数辨识和系统噪声对SOC估计精度的影响,本文采用二阶RC等效电路模型,结合自适应扩展卡尔曼滤波算法（AEKF）进行锂电池的SOC估计。用带有遗忘因子的最小二乘法对模型参数进行在线辨识,以减小由参数辨识引起的估计误差,AEKF可以对系统和过程噪声进行修正,从而减小噪声对SOC估计的影响。最后分别用EKF和AEKF进行SOC估计并比较其误差,结果表明,AEKF联合最小二乘法参数在线辨识具有更高的精度和更好的适应性。     

Low‐temperature reversible capacity loss and aging mechanism in lithium‐ion batteries for different discharge profiles

In this paper, reversible capacity loss of lithium‐ion batteries that cycled with different discharge profiles (0.5, 1, and 2 C) is investigated at low temperature (?10°C). The results show that the capacity and power degradation is more severe under the condition of low discharge rate, not the widely accepted high discharge rate. To shed some light on the aging phenomena, noninvasive electrochemical methods, ie, incremental capacity and differential voltage analysis, are applied to identify and quantify the effects of different degradation modes (DMs). Apart from the resistance increase, the DMs include the loss of lithium inventory (LLI) and the loss of active material (LAM). Both LLI and LAM decay to a greater extent for the cell cycled with lower discharge rate, and the growth of LAM is higher than that of LLI. Further, the analysis of state of charge (SOC) window shows that the earlier cutoff of the high discharge rate can lead to less mechanical and thermal stress on cathode materials, thus a lower degradation rate. Another cause is that the lithium plating on the anode materials can be mitigated by increasing the charging temperature which results from preceding high rate discharging.     

In situ detection of lithium dendrite in the failure of lithium-ion batteries

锂离子电池在应用过程中常出现一些失效现象如循环寿命缩短、自放电率变大、功率特性劣化等,甚至发生安全问题。负极析锂是导致这些失效甚至安全问题的主要因素之一,因此了解析锂的原因与过程就显得格外重要。探明析锂原因的关键是表征锂的存在和锂枝晶的生长过程。本文综述了锂离子电池负极析锂现象常用的原位检测技术,包括物理检测法和电化学法。物理检测法主要介绍：光学原位技术、原位X射线技术、原位核磁技术以及原位中子技术,电化学方法包含：充放电电压曲线法、Arrhenius法、内阻容量曲线分析法和容量衰减率法。本文针对物理检测法的原理、优缺点以及对应的特殊检测装置实例,电化学检测法的原理及分析方法等进行简要概述,并对目前析锂原位检测存在的问题进行总结及其研究方向进行展望。     

Dynamic electric behavior and open-circuit-voltage modeling of LiFePO4-based lithium ion secondary batteries

Accurate battery modeling is one of the key factors in battery system design process and operation as well. Therefore, the knowledge of the distinct electric characteristics of the battery cells is mandatory. This work gives insight to the electric characteristics of lithium ion batteries (Li-ion) comprising LiFePO₄-based cathode active materials with emphasis on their specific open-circuit-voltage (OCV) characteristics including hysteresis and special OCV recovery effects, which last for several minutes or even hours after a current load is interrupted. These effects are elucidated incorporating OCV measurement data of high power cells. Simple empiric models are derived and used in a model-based state estimation algorithm. The complete battery model includes an impedance model, a hysteresis model and an OCV recovery model part. The introduced model enables the assessment of the cells’ state-of-charge (SOC) precisely using model-based state estimation approaches.     

Composite electrodes of disordered carbon and graphite for improved battery state estimation with minimal performance penalty

Voltage based state of charge (SOC) estimation is challenging for lithium ion batteries that exhibit little open circuit voltage (OCV) change over a large SOC range. We demonstrate that by using a composite negative electrode composed of disordered carbon and graphite, we were able to introduce additional features to the OCV-SOC relationship that facilitate voltage-based SOC estimation. In contrast to graphite, the potential of disordered carbon is sensitive to the state of charge; this behavior, when manifested in a lithium ion battery, gives rise to additional beneficial features of the cell OCV-SOC relationship in terms of state estimation. We have demonstrated the effectiveness of the approach by comparing model simulations and corresponding experimental data of a cell composed of LiFePO₄ positives and graphite + disordered carbon composite negative electrodes. Last, we find that although the graphite material has a higher coulombic capacity, very little (dynamic) performance loss is manifest with the mixed graphite + disordered carbon composite is employed.     

A novel multiobjective charging optimization method of power lithium‐ion batteries based on charging time and temperature rise

Power lithium‐ion batteries have been widely utilized in energy storage system and electric vehicles, because these batteries are characterized by high energy density and power density, long cycle life, and low self‐discharge rate. However, battery charging always takes a long time, and the high current rate inevitably causes great temperature rises, which is the bottleneck for practical applications. This paper presents a multiobjective charging optimization strategy for power lithium‐ion battery multistage charging. The Pareto front is obtained using multiobjective particle swarm optimization (MOPSO) method, and the optimal solution is selected using technique for order of preference by similarity to ideal solution (TOPSIS) method. This strategy aims to achieve fast charging with a relatively low temperature rise. The MOPSO algorithm searches the potential feasible solutions that satisfy two objectives, and the TOPSIS method determines the optimal solution. The one‐order resistor‐capacitor (RC) equivalent circuit model is utilized to describe the model parameter variation with different current rates and state of charges (SOCs) as well as temperature rises during charging. And battery temperature variations are estimated using thermal model. Then a PSO‐based multiobjective optimization method for power lithium‐ion battery multistage charging is proposed to balance charging speed and temperature rise, and the best charging stage currents are obtained using the TOPSIS method. Finally, the optimal results are experimentally verified with a power lithium‐ion battery, and fast charging is achieved within 1534 s with a 4.1°C temperature rise.     

Estimation of power battery SOC based on firefly BP neural network

针对BP神经网络算法对电动汽车电池荷电状态（state of charge,SOC）估算的缺陷,提出一种基于萤火虫（firefly algorithm,FA）神经网络的SOC估算方法。以磷酸铁锂电池为测试对象,在ARBIN公司生产的EVTS电动车动力电池测试系统装置上进行测试,收集锂电池的各项性能参数。采用端电压和放电电流作为输入参数,SOC作为输出参数,建立FA-BP神经网络模型,用于估算锂离子电池充放电过程中的任一状态下的SOC。仿真实验结果表明,与现有的BP神经网络估算方法相比,基于FA-BP神经网络的锂电池SOC估算方法准确度高,具备很好的实用性。     

5kW锂电池模组BMS的研究与开发

设计了以MSP430为控制核心的用于5 k W锂电池管理系统(BMS).建立了关于电池荷电状态的模型,在实际估计中,采用开路电压和按时积分相结合的方法且有较高的精度;采取电池均衡充电的方案,补偿了电池容量的差异性,进而使得电池组的使用寿命延长.电池荷电状态估算的改进方案解决了按时计量法无法确定初始荷电状态、难以精准测得库仑效率等问题,确保了电池管理系统处于稳定工作状态.该系统具有抗干扰能力极强、硬件电路可靠、且十分经济的特点.经过实验验证,利用该系统进行SOC剩余容量估计的结果较为精确.     

Capacity and power fading mechanism identification from a commercial cell evaluation

An 18650 lithium ion cell was evaluated using dynamic stress test (DST) protocol for cycle life study. Reference performance tests were applied every 60 DST cycles to quantify capacity fade and peak power capability (PPC) degradation with cycles to the end of life. A quantitative incremental capacity analysis (ICA) was applied to identify extrinsic and intrinsic contributions to capacity fading, whereas the open circuit voltage (OCV) measurements were utilized to determine the correct state of charge (SOC) in order to accurately correlate the capacity fade with SOC. The analysis also helps us identify that cell polarization resistance change in the cycles is the primary culprit that bifurcates to both extrinsic and intrinsic origins in capacity fade and PPC degradation. This analysis allows us to develop better understanding in predicting battery performance and life in the rechargeable lithium batteries.