UTSTF锂离子电池时变参数估计与故障诊断

针对锂离子电池的参数偏差型故障诊断问题，提出基于无迹变换强跟踪滤波器（UTSTF）的电池时变参数估计与故障诊断方法.建立电池的开路电压（OCV）-荷电状态（SOC）特性曲线与一阶等效电路模型；将电池参数加入状态变量，建立状态与参数的联合状态空间方程，通过UTSTF算法得到电池参数的实时估计结果，并根据估计值设计故障诊断算法流程；以电池内部的接触型故障与扩散型故障为例，在变温环境下模拟故障发生并进行电池充放测试，得到电池参数在UTSTF与无迹卡尔曼滤波（UKF）下估计值与真实值的对比.实验结果表明，所提方法对于电池故障参数具有良好的跟踪效果、较高的估计精度与诊断可靠性.

Time-varying parameters estimation and fault diagnosis of li-ion battery using UTSTF

A method of estimation of time-varying parameters and fault diagnosis was proposed based on unscented transformation of strong tracking filter (UTSTF) according to parameter-biased fault of li-ion batteries. The open circuit voltage (OCV)-state of charge (SOC) characteristic mapping curve and One order equivalent circuit model were established. Then the joint state-space equation of battery was founded by introducing parameters into state variable. UTSTF was applied to battery parameters estimation in real-time. The fault diagnosis process was proposed according to parameters estimation results. Taking battery's internal contact-fault and diffusion-fault as a case, battery charging and discharging tests were conducted under the temperature-varying conditions for simulating fault occurrence. The estimation results of UTSTF and unscented Kalman filter (UKF) were compared with true value. The experimental results show that the proposed method can achieve better tracking performance, estimated accuracy and diagnosis reliability for battery fault parameters.