基于充电曲线特征的锂离子电池容量估计

准确的容量估计对锂离子电池管理具有重要意义。通过电池循环老化试验，归纳出老化过程中与电池容量衰减相关的充电曲线特征。通过计算充电曲线特征与衰减容量的相关系数进一步确定特征的电压区间。建立以径向基函数为核函数的相关向量机模型，以5个特征为输入量、电池容量为输出量进行数据训练，然后以筛选出的相关矢量进行在线容量估计。结果表明，该电池容量估计算法精度在2.2%以内，且算法收敛性较好。

Capacity Estimation of Lithium-ion Batteries Based on Charging Curve Features

Accurate capacity estimation plays an important role in lithium-ion battery management. The charging curve features related to the battery capacity attenuation during the aging process are summarized through the battery cycle aging experiments. By calculating the correlation coefficient between curve features and attenuation capacity, the voltage range of curve features is further determined. A relevance vector machine with radial basis function as the kernel function is established. Five features are adopted as input and battery capacity as output for data training, and then the trained relevant sparse vector is used for online capacity estimation. The estimation results show that the accuracy of the data-driven capacity estimation algorithm is less than 2.2% and the convergence of the algorithm is rapid.