【作者】 吴杰；

【导师】 徐惠钢；

【作者基本信息】 中国矿业大学 ， 电子信息（专业学位）， 2023， 硕士

【摘要】 锂离子电池本身的特性比较复杂,如果工作中的荷电及健康状态不能得到实时估计,将会导致锂离子电池发生过充、过放等问题,进而引发过温、热失控等安全事故。研究能够对锂电池荷电及健康状态进行准确估算的储能电池管理系统具有重要意义。针对当前储能电池管理系统存在的问题,开展如下研究工作:(1)分析储能锂离子电池的工作原理、技术参数,构建其二阶RC等效电路模型,并通过脉冲放电实验实现对SOC(电池荷电状态)与开路电压的关系标定,进一步对模型参数辨识。其次,利用多新息最小二乘算法(MILS)的多个新息对输出观测值与模型估计值之间的差值进行修正,解决递推最小二乘法辨识精度低、收敛速度慢的问题。相比于递推最小二乘法而言,所研究的MILS算法在电池参数辨识上具有算法精度高、收敛速度快的优点。(2)研究电池荷电与健康状态联合估计问题。研究了具有RTS平滑结构的迭代无迹卡尔曼粒子滤波算法(RTS-IUPF),通过迭代无迹卡尔曼优化粒子滤波算法的建议分布函数,解决了传统粒子滤波算法后期存在粒子贫乏的问题。相比于传统算法,所提出的算法在估计电池荷电状态(SOC)、电池健康状态(SOH)上,算法精度及收敛速度得到提高。(3)选择功能强大、运行速度较快且移植算法方便的TMS320F28035为主控制器,以采集精度高的BQ76940芯片组成的外围电路作为前端模拟信号采集单元,并通过CAN总线实现下位机与上位机之间数据的可靠传输。(4)分析所研究的算法在实际平台上估算的结果。结果表明,估算的电池荷电状态(SOC)的绝对误差不超过1.911%,估算的电池健康状态(SOH)的绝对误差低于4%,综合性能强。该论文有图49幅,表14个,参考文献63篇。更多还原

【Abstract】 The characteristics of lithium-ion batteries themselves are relatively complex.If the charging and health status during work cannot be estimated in real-time,it will lead to overcharging,discharging,and other issues of lithium-ion batteries,which can lead to safety accidents such as overheating and thermal runaway.Studying an energy storage battery management system that can accurately estimate the charge and health status of lithium batteries is of great significance.This article focuses on the problems existing in the current energy storage battery management system and conducts the following research work:(1)Analyze the working principle and technical parameters of energy storage lithium-ion batteries,construct their second-order RC equivalent circuit model,and calibrate the relationship between SOC(battery state of charge)and open circuit voltage through pulse discharge experiments,further identifying the model parameters.Secondly,multiple innovations of the Multi Innovation Least Squares(MILS)algorithm are used to correct the difference between the output observation values and the model estimation values,solving the problems of low identification accuracy and slow convergence speed of the recursive least squares method.Compared with the recursive least square method,the MILS algorithm has the advantages of high accuracy and fast Rate of convergence in battery parameter identification.(2)Study the joint estimation problem of battery charge and health status.We studied the iterative unscented Kalman particle filter algorithm(RTS-IUPF)with RTS smoothing structure.By optimizing the suggested distribution function of the particle filter algorithm through iterative unscented Kalman,we solved the problem of particle poverty in the later stage of traditional particle filter algorithms.Compared with traditional algorithms,the accuracy and Rate of convergence of the proposed algorithm are improved in the estimation of battery state of charge(SOC)and battery state of health(SOH).(3)Choose TMS320F28035 as the main controller with powerful functions,fast running speed,and convenient algorithm transplantation.The peripheral circuit composed of a high-precision BQ76940 chip is used as the front-end analog signal acquisition unit,and reliable data transmission between the lower computer and the upper computer is achieved through CAN bus.(4)Analyze the estimated results of the algorithm studied on the actual platform.The results show that the absolute error of the estimated battery state of charge(SOC)does not exceed 1.911%,and the absolute error of the estimated battery state of health(SOH)is less than 4%,indicating strong overall performance.This thesis has 49 figures,14 tables,and 63 references.更多还原