基于机械振动信号的锂离子电池组连接故障诊断  

Connection Fault Diagnosis of Lithium-ion Battery Pack Based on Mechanical Vibration Signals

作　　者：申东旭 吕超[1] 葛亚明 张刚[1] 杨大智 王立欣 SHEN Dongxu;LÜChao;GE Yaming;ZHANG Gang;YANG Dazhi;WANG Lixin(School of Electrical Engineering and Automation,Harbin Institute of Technology,Harbin 150001;Education Center of Experiments and Innovations,Harbin Institute of Technology,Shenzhen 518071;School of Mechanical Engineering and Automation,Harbin Institute of Technology,Shenzhen 518071)

机构地区：[1]哈尔滨工业大学电气工程及自动化学院,哈尔滨150001 [2]哈尔滨工业大学(深圳)实验与创新教育中心,深圳518071 [3]哈尔滨工业大学(深圳)机电工程与自动化学院,深圳518071

出　　处：《机械工程学报》2022年第22期56-68,共13页Journal of Mechanical Engineering

基　　金：广东省重点研发资助项目-乘用车三元动力电池系统主动安全防控技术资助项目(2020B090919004)。

摘　　要：为满足高电压大容量的实际应用场景和需求,锂离子电池组往往由成百上千的电池单体通过螺栓、焊接等方式串并联组成。电池组单体间的连接故障会导致接触电阻升高和连接处异常发热,严重影响电池组的性能和安全。提出一种基于机械振动信号的锂离子电池组连接故障诊断方法。利用压电陶瓷传感器实现电压信号和振动信号的相互转换,在每种故障模式下采集振动信号;基于稀疏测度指标和熵测度方法在频域和时域提取故障特征以描述锂离子电池组在不同连接故障模式下的故障特性;利用最大相关最小冗余算法降低高维特征空间的冗余度,选择出最重要的特征;在此基础上,建立基于差分进化算法优化的支持向量机诊断模型。结果表明,该方法诊断准确度为0.963,可以准确检测到锂离子电池组的连接故障并明确故障发生的位置。In order to meet the practical application scenarios and needs of high voltage and large capacity, lithium-ion battery packs are often composed of hundreds or thousands of battery cells connected in series and parallel through bolts and welding. The connection failure between the cells of the battery pack can lead to increased contact resistance and abnormal heating at the connection, which can seriously affect the performance and safety of the battery pack. A connection fault diagnosis method for lithium-ion battery packs based on mechanical vibration signals is proposed. The piezoelectric ceramic sensor is used to realize the mutual conversion of the voltage signal and the vibration signal, and the vibration signal is collected in each fault mode. Based on sparse measure metrics and entropy measure method, fault features are extracted in frequency and time domains to describe the fault characteristics of lithium-ion battery packs under different connection fault modes. The maximum relevance minimum redundancy algorithm is used to reduce the redundancy of high-dimensional feature space and select the most important features;On this basis, the diagnosis model of support vector machine optimized by differential evolution algorithm is established. The results show that the diagnostic accuracy of this method is 0.963, which means that this method can accurately detect the connection fault of lithium-ion battery pack and determine the location of the fault.

关 键 词：锂离子电池组 连接故障诊断 压电传感技术 特征提取 

分 类 号：TM912[电气工程—电力电子与电力传动]