电流信号采样偏差下动力电池荷电状态估计研究  

Estimation of Power Battery State of Charge under Current Signal Sampling Deviation

作　　者：刘逸群 李猛猛 刘涛[1] 杨娜[1] 李卫华 王剑锋[1] LIU Yiqun;LI Mengmeng;LIU Tao;YANG Na;LI Weihua;WANG Jianfeng(School of Automotive Engineering,Harbin Institute of Technology at Weihai,Weihai 264209)

机构地区：[1]哈尔滨工业大学(威海)汽车工程学院,威海264209

出　　处：《机械工程学报》2023年第16期288-299,共12页Journal of Mechanical Engineering

基　　金：国家自然科学基金(51905121);长三角哈特机器人产业技术研究院(HIT-CXY-CMP2-PPTLC-21-01)资助项目。

摘　　要：电动汽车电池管理系统实际运行的过程中,电流信号易受到有色噪声的干扰而产生信号采样偏差,这会造成荷电状态(State of charge,SOC)估计精度的急剧下降。针对该问题,分析电流信号分别在两种不同类型的有色噪声干扰下的SOC估计问题,提出一种三层组合估计结构用于同时实现电流采样信号的校正、电池模型参数的在线更新以及SOC的高精度估计。该组合结构首先基于状态扩维后二阶RC等效电路模型,并利用自适应重组遗传算法(Adaptive recombination genetic algorithm,ARGA)辨识出模型参数,由自适应平方根容积卡尔曼滤波(Adaptive square root cubature Kalman filter,ASRCKF)算法在线校正产生偏差的电流信号;然后基于校正后的电流信号和二阶RC模型,通过偏差补偿遗忘因子递推最小二乘(Bias compensation recursive forgetting-factor least squares,BCFRLS)算法与ASRCKF算法相结合进行协同估计,实现模型参数和SOC值的在线更新。最后在DST工况下进行验证,试验和仿真结果表明,在电流信号中掺杂有色噪声信号而产生采样偏差时,所提出的组合估计结构仍能保证SOC估计的高精度性,其平均相对误差可维持在1%以下。In the actual operation of the electric vehicle battery management system,the current signal is susceptible to the interference of colored noise and the sampling signal deviation occurs,which causes the problem of a sharp drop in the accuracy of the SOC estimation.The SOC estimation problem of the current signal under the interference of two different types of colored noise is analyzed,and a three-layer combined estimation structure is proposed to realize the correction of the current sampling signal,the online update of the battery model parameters and the high-precision estimation of the SOC at the same time.Based on the second-order RC equivalent circuit model after state expansion,the model parameters are identified through adaptive recombination genetic algorithm,and the deviated current signal is corrected online by the adaptive square root cubature Kalman filter algorithm(ASRCKF).Based on this,bias compensation forgetting factor recursive least squares algorithm(BCFRLS)and ASRCKF algorithm for collaborative estimation,to achieve online update of model parameters and SOC values.It is verified under DST conditions.The experimental and simulation results show that the proposed combined structure can still guarantee the high-precision estimation of the SOC value under the current sampling deviation,and its average relative error can be maintained below 1%.

关 键 词：锂离子电池 荷电状态估计 有色噪声 卡尔曼滤波 

分 类 号：TG156[金属学及工艺—热处理]