Boosting battery state of health estimation based on self-supervised learning

State of health(SoH) estimation plays a key role in smart battery health prognostic and management.However,poor generalization,lack of labeled data,and unused measurements during aging are still major challenges to accurate SoH estimation.Toward this end,this paper proposes a self-supervised learning framework to boost the performance of battery SoH estimation.Different from traditional data-driven methods which rely on a considerable training dataset obtained from numerous battery cells,the proposed method achieves accurate and robust estimations using limited labeled data.A filter-based data preprocessing technique,which enables the extraction of partial capacity-voltage curves under dynamic charging profiles,is applied at first.Unsupervised learning is then used to learn the aging characteristics from the unlabeled data through an auto-encoder-decoder.The learned network parameters are transferred to the downstream SoH estimation task and are fine-tuned with very few sparsely labeled data,which boosts the performance of the estimation framework.The proposed method has been validated under different battery chemistries,formats,operating conditions,and ambient.The estimation accuracy can be guaranteed by using only three labeled data from the initial 20% life cycles,with overall errors less than 1.14% and error distribution of all testing scenarios maintaining less than 4%,and robustness increases with aging.Comparisons with other pure supervised machine learning methods demonstrate the superiority of the proposed method.This simple and data-efficient estimation framework is promising in real-world applications under a variety of scenarios.

Deep learning-based battery state of charge estimation:Enhancing estimation performance with unlabelled training samples

The estimation of state of charge(SOC)using deep neural networks(DNN)generally requires a considerable number of labelled samples for training,which refer to the current and voltage pieces with knowing their corresponding SOCs.However,the collection of labelled samples is costly and time-consuming.In contrast,the unlabelled training samples,which consist of the current and voltage data with unknown SOCs,are easy to obtain.In view of this,this paper proposes an improved DNN for SOC estimation by effectively using both a pool of unlabelled samples and a limited number of labelled samples.Besides the traditional supervised network,the proposed method uses an input reconstruction network to reformulate the time dependency features of the voltage and current.In this way,the developed network can extract useful information from the unlabelled samples.The proposed method is validated under different drive cycles and temperature conditions.The results reveal that the SOC estimation accuracy of the DNN trained with both labelled and unlabelled samples outperforms that of only using a limited number of labelled samples.In addition,when the dataset with reduced number of labelled samples to some extent is used to test the developed network,it is found that the proposed method performs well and is robust in producing the model outputs with the required accuracy when the unlabelled samples are involved in the model training.Furthermore,the proposed method is evaluated with different recurrent neural networks(RNNs)applied to the input reconstruction module.The results indicate that the proposed method is feasible for various RNN algorithms,and it could be flexibly applied to other conditions as required.

基于自适应H_(2)/H_(∞)滤波的锂电池SOC和SOH联合估计

准确、实时地估计电池的荷电状态(state of charge,SOC)和健康状态(state of health,SOH)是现代电池管理系统的关键任务。通过自适应H_(2)/H_(∞)滤波器可对锂电池的SOC和SOH进行联合估计。该方法基于锂电池的二阶RC等效电路模型,采用AFFRLS法在线辨识锂电池的模型参数,并利用H_(2)/H_(∞)滤波器估计锂电池的SOC,AFFRLS辨识与H_(2)/H_(∞)滤波交替进行,得到一种自适应H_(2)/H_(∞)滤波器。SOH依据AFFRLS辨识的电池内阻进行估计,实现了锂电池SOC与SOH的联合估计。实验结果表明:自适应H_(2)/H_(∞)滤波算法的估计精度高且鲁棒性强,电池的SOC和SOH的平均估计误差始终保持在±0.19%以内,相比于EKF和H_(∞)滤波算法有更高的估计精度与稳定性。

基于MCS D2P的动力锂电池管理系统主控软件

为克服传统的电池管理系统(battery management system,BMS)开发方法不仅不利于先进算法的软件实现,且存在开发周期长和费用高的缺点,基于MCS D2P快速控制原型开发平台搭建BMS硬件,针对动力锂电池由多节电芯串联而成的特点,分析BMS对电池组总电压、电流、电芯电压和温度的实时测量,对电池剩余电量实时评价的功能,以及由主控模块、检测模块和均衡管理模块构成的主要硬件结构.设计主控软件数据流、主程序流程图、数据采集程序流程图和荷电状态(state of charge,SOC)估算程序流程图.以由15节电芯构成的48 V、50 Ah磷酸铁锂电池组为试验对象,设计周期性充放电测试试验,得到电池总电压和电流的测量误差均在±0. 5%内,温度的测量误差在±0. 5℃内,SOC实测值与理论值差值小于8%,验证了主控软件设计的可行性.

基于FPGA控制1tc6804-2的电池管理系统的研究

为了提高电池管理系统中多块单体电池电压的采集速度与效率,提出了一种基于FPGA控制ltc6804—2的电池管理系统;单个Ltc6804—2里的ADC采集功能仅在20μs内就可完成对12块单体电池的电压或电流的采集,并且多块ltc6804-2可进行级联从而可监测系统设备所需的单体电池电压,FPGA通过SPI总线与ltc6804—2实现通信,可快速获取多块单体电池的电压或电流从而可以在其内部实现系统的充放电均衡、荷电状态(SOC)估计等功能,并且精确度高,使整套设备更加简化与可靠,效率更高;该系统可广泛应用于电动汽车或通用航空机上。

社区自我管理小组干预对2型糖尿病患者综合控制情况的长期和短期效果研究

背景糖尿病自我管理是降低疾病影响及改善预后的重要手段。目前,糖尿病自我管理研究多以干预后短期效果评价为主,对长期效果的追踪报告较少。目的评价社区自我管理小组活动对2型糖尿病患者综合控制情况的长期和短期效果。方法于2014年3月,在北京市房山区招募500例成年2型糖尿病患者,随机分入对照组(n=241)和干预组(n=259)。两组患者均接受常规糖尿病随访服务,干预组患者在此基础上开展为期3个月的糖尿病自我管理小组活动。在干预前、3个月干预结束时、干预结束后2年和5年分别进行调查,收集两组患者的人口学信息、疾病信息、糖尿病综合控制指标[体质指数(BMI)、血压、空腹血糖(FPG)、糖化血红蛋白(HbA1c)、高密度脂蛋白胆固醇(HDL-C)、三酰甘油(TG)、低密度脂蛋白胆固醇(LDL-C)]。采用广义估计方程分析自我管理小组活动主效应,以及自我管理小组活动与干预后时间的交互效应。结果调整协变量后,自我管理小组活动在BMI、收缩压、舒张压、FPG、HbA1c、HDL-C、TG、LDL-C上的主效应不显著(P>0.05)。时间在各项指标上的主效应显著(P<0.05),相较于基线,患者BMI、收缩压、舒张压、FPG、HbA1c、HDL-C、LDL-C升高,TG降低。自我管理小组活动与干预后时间对BMI[β(95%CI)=-0.33(-0.62,-0.05)]、FPG[β(95%CI)=-1.03(-1.71,-0.35)]、TG[β(95%CI)=-0.54(-0.93,-0.14)]的影响存在交互效应:干预前干预组BMI比对照组低0.31 kg/m2,干预3个月后比对照组低0.64 kg/m2;干预前干预组FPG比对照组高0.19 mmol/L,干预后2年比对照组低0.84 mmol/L;干预前干预组TG比对照组高0.03 mmol/L,干预后5年比对照组低0.51 mmol/L。结论自我管理小组活动对2型糖尿病患者BMI控制有短期效果,对FPG、TG控制可能有长期效果。

基于STM32的全钒液流电池管理系统的设计与实现

为了解决PLC信息采样端口有限,扩展成本高以及功能多样化受限的问题,设计了基于ARM的全钒液流电池(VFB)管理系统;系统利用霍尔传感器实现电压电流的采集,采用修正卡尔曼滤波法实现剩余电荷量(SOC)估计,利用MCGS组态软件实现数据的实时监测以及故障诊断和显示,并实时控制钒电解液的活化和混合,自主切换工作模式;本电池管理系统可实现电池侧电压、电流和温度等信号的实时数据监测、系统的故障诊断以及保护,并能存储和远程上传数据;现场联机测试结果表明,SOC估计误差可控制在±2%,电池充电时间缩短25%,实时响应速度快;该系统能够提高电池的利用率,防止电池出现过充和过放,运行可靠,已投入批量研制和生产。

基于LTC6804-2煤矿后备电源电池管理系统的设计与实现

基于锂离子电池在煤矿后备电源应用领域的普遍性和实用性,给出了应用LTC6804-2芯片的电池管理系统硬件及软件设计、实现方案,并且给出了该系统在实际应用中的实验数据,数据表明本系统在精确性、稳定性方面表现突出。

基于S32K的电池管理系统主控单元设计

为提高电池组的使用寿命,设计了一种基于S32K的电池管理系统主控单元。该设计对绝缘检测系统进行了改进,提出了绝缘监控系统与绝缘检测系统相配合的绝缘监测系统,对系统的绝缘状态进行实时监测,有助于避免故障的发生;同时,对SOC估算方法进行了改进,提高了估算精度。实验表明:所设计的主控单元运行稳定,能够实时、准确地完成电流以及总电压的采集、SOC估算、实时通讯、绝缘检测等功能,保证了电池组的安全稳定运行,提高了电池组的使用寿命。

储能锂电池系统综合管理研究进展

锂电池系统是新型储能系统的重要组成部分,电池管理技术对确保电池系统的安全、高效运行和延长使用寿命具有重要意义。从电池系统的模型、状态、故障、一致性、热管理等层面出发,综述了锂电池管理的研究进展。在电池模型方面,分析总结了电池在电、热、力等特性的建模及高效计算方法。对于电池的状态估计,总结了基于模型、数据驱动以及两者相结合的状态估计方法以及各自的优缺点,分析了构建数据驱动与电池领域知识融合框架的未来发展前景。在电池故障方面,分析了电池系统不同故障特征、类型、触发机制和诊断方法,对电池系统在早期故障预警、故障检测灵敏度提升等方面进行了展望。对于电池系统的一致性,总结分析了在均衡拓扑和均衡策略两个方面的研究现状,并对电池重构在均衡、快充、能量利用提升、故障隔离等方面的应用进行了分析。在电池热管理方面,分析了高温冷却以及低温加热的方法,并对热管理系统的优化设计和控制策略进行了总结,对开发高效换热、轻量化以及低能耗的先进热管理系统进行了探讨。此外,对电池管理系统的新技术和应用场景进行了概述,分析了数字孪生技术以及储能综合管理在电池系统上的应用,为未来电池管理的发展提供了参考。

电动汽车动力锂离子电池可靠性关键技术综述

随着动力锂离子电池在电动汽车行业发挥关键性作用,保证其工作可靠性已成为当下研究热点。文中综述锂离子电池材料和制造工艺;详细归纳总结电池状态估算方法以及剩余寿命预测方法,并讨论各种方法的优缺点;从电池管理系统层面,先后梳理均衡管理系统和热管理系统相关知识;从电动汽车混合储能系统层面阐述实际工况下性能退化机理及相关技术。最后总结电动汽车动力锂离子电池与可靠性相关的四个方面关键技术的现状,并展望未来发展可能。

储能电池建模与状态估计研究进展

储能电池具有能够平滑可再生能源输出,提高电力系统灵活性和应对电力需求峰谷等优势,有助于推动可再生能源发展,从而应对环境污染和能源紧缺的双重压力。目前市场主流的储能电池为锂离子电池,具有高比能特性,同时新型储能电池也在蓬勃发展,其中全钒液流电池具有高安全性的优势,液态金属电池具有超长循环寿命,在电力储能领域具有重要应用前景。储能电池的建模和状态估计对提高储能电池系统性能,确保其安全性以及优化维护效率至关重要,因此文中对锂离子电池、全钒液流电池和液态金属电池的建模和状态估计进行综述。首先,介绍了储能电池状态估计的整体框架,对基于实验的方法、基于模型的方法和基于数据驱动的方法进行整体介绍,并对荷电状态(state of charge,SOC)、健康状态(state of health,SOH)和剩余使用寿命(remaining useful life,RUL)进行概括;然后,从原理出发,分别总结了不同储能电池体系的内部工作过程、模型构建、状态估计与电池管理过程;最后,对不同储能电池体系的主要工作特性进行横向对比和总结,旨在为储能电池选择和发展提供启示。

基于模型的动力电池SOC估计方法研究进展

随着电动汽车应用的快速发展,负责行驶安全的电池管理系统(BMS)的重要性日益凸显。高精度的电池荷电状态(SOC)估计方法是BMS有效运行的基础。对近年来有关电池模型、模型参数辨识和SOC估计方法三个领域的研究进行了较为系统的综述,总结了主要电池模型的特点及发展趋势,介绍了模型参数辨识的一般过程以及常见的辨识方法,对四类常见SOC估计方法做了介绍,分析了它们的特点和局限性,并探讨了SOC估算功能在BMS行业中的应用研究。针对当前研究中存在的不足,提出了未来的研究方向。

A brief review on key technologies in the battery management system of electric vehicles

Batteries have been widely applied in many high-power applications,such as electric vehicles(EVs)and hybrid electric vehicles,where a suitable battery management system(BMS)is vital in ensuring safe and reliable operation of batteries.This paper aims to give a brief review on several key technologies of BMS,including battery modelling,state estimation and battery charging.First,popular battery types used in EVs are surveyed,followed by the introduction of key technologies used in BMS.Various battery models,including the electric model,thermal model and coupled electro-thermal model are reviewed.Then,battery state estimations for the state of charge,state of health and internal temperature are comprehensively surveyed.Finally,several key and traditional battery charging approaches with associated optimization methods are discussed.

State-of-health estimation of lithium-ion batteries based on electrochemical impedance spectroscopy: a review

Lithium-ion batteries(LIBs)are crucial for the large-scale utilization of clean energy.However,because of the com-plexity and real-time nature of internal reactions,the mechanism of capacity decline in LIBs is still unclear.This has become a bottleneck restricting their promotion and application.Electrochemical impedance spectroscopy(EIS)contains rich electrochemical connotations and significant application prospects,and has attracted widespread atten-tion and research on efficient energy storage systems.Compared to traditional voltage and current data,the state-of-health(SOH)estimation model based on EIS has higher accuracy.This paper categorizes EIS measurement methods based on different principles,introduces the relationship between LIBs aging mechanism and SOH,and compares the advantages of different SOH estimation methods.After a detailed analysis of the latest technologies,a review is given.The insights of this review can deepen the understanding of the relationship between EIS and the aging effect mechanism of LIBs,and promote the development of new energy storage devices and evaluation methods.

基于EKF的SOC估算的Simscape实现

储能行业是当今热门产业之一,其中储能系统发挥着至关重要的作用,然而,该系统中的电池管理系统(Battery Management System,简称BMS)是储能系统的核心,其精确度和稳定性是储能系统正常工作的基础,为了提高其可靠性,SOC的精确估计至关重要。本文使用扩展卡尔曼滤波(Extended Kalman Filtering,简称EKF)算法来进行单电池的SOC估计,使用Simulink/Simscape搭建二阶RC电池模型来等效18650锂离子电池,经放电实验所得SOC和实际SOC进行对比,其误差不超过2%,验证了该算法具有较高精度。

Implementation for a cloud battery management system based on the CHAIN framework

An intelligent battery management system is a crucial enabler for energy storage systems with high power output,increased safety and long lifetimes.With recent developments in cloud computing and the proliferation of big data,machine learning approaches have begun to deliver invaluable insights,which drives adaptive control of battery management systems(BMS)with improved performance.In this paper,a general framework utilizing an end-edge-cloud architecture for a cloud-based BMS is proposed,with the composition and function of each link described.Cloud-based BMS leverages from the Cyber Hierarchy and Interactional Network(CHAIN)framework to provide multi-scale insights,more advanced and efficient algorithms can be used to realize the state-of-X es-timation,thermal management,cell balancing,fault diagnosis and other functions of traditional BMS system.The battery intelligent monitoring and management platform can visually present battery performance,store working-data to help in-depth understanding of the microscopic evolutionary law,and provide support for the development of control strategies.Currently,the cloud-based BMS requires more effects on the multi-scale inte-grated modeling methods and remote upgrading capability of the controller,these two aspects are very important for the precise management and online upgrade of the system.The utility of this approach is highlighted not only for automotive applications,but for any battery energy storage system,providing a holistic framework for future intelligent and connected battery management.

An improved neural network model for battery smarter state-of-charge estimation of energy-transportation system

The safety and reliability of battery storage systems are critical to the mass roll-out of electrified transportation and new energy generation.To achieve safe management and optimal control of batteries,the state of charge(SOC)is one of the important parameters.The machine-learning based SOC estimation methods of lithium-ion batteries have attracted substantial interests in recent years.However,a common problem with these models is that their estimation performances are not always stable,which makes them difficult to use in practical applications.To address this problem,an optimized radial basis function neural network(RBF-NN)that combines the concepts of Golden Section Method(GSM)and Sparrow Search Algorithm(SSA)is proposed in this paper.Specifically,GSM is used to determine the optimum number of neurons in hidden layer of the RBF-NN model,and its parameters such as radial base center,connection weights and so on are optimized by SSA,which greatly improve the performance of RBF-NN in SOC estimation.In the experiments,data collected from different working conditions are used to demonstrate the accuracy and generalization ability of the proposed model,and the results of the experiment indicate that the maximum error of the proposed model is less than 2%.

新能源电池化及换电标准化电池管理系统研究

基于提出的新能源电池化概念,制定电池化标准目标,实现换电标准化管理系统。采用二阶电路等效电池模型及相应的荷电状态实时估计算法,结合模拟前端芯片ml5238和ml610q486单片机,实现标准化换电电池管理系统的核心,实现一种可以投入商业应用的新型电池管理系统,其处理时间小于100μs,准确率高达99.2%,成本比同类系统降低30%。

基于多新息理论的电池模型参数辨识及SOC估计

准确的锂离子电池荷电状态(SOC)估计对于保障电池管理系统(BMS)安全稳定运行具有重要意义。以三元锂离子电池为对象进行电气特性测试,基于此数据建立一阶RC等效电路模型,再采用引入遗忘因子的多新息最小二乘法(FFMILS)进行模型参数在线辨识,联合加权多新息扩展卡尔曼滤波算法(MIEKF)进行锂离子电池的SOC估计。通过美国联邦城市行驶工况(FUDS)对算法进行验证,结果表明FFMILS-MIEKF算法的SOC估计准确性和稳定性相较于传统EKF与UKF算法均有不同程度提高。并且该算法在不同初始SOC值下均能快速收敛,具有较好的鲁棒性,能够适用于实际电池管理系统之中。