基于ICA和Box-Cox变换的锂离子电池SOH估计方法

SOH Estimation Method for Li-ion Battery Based on ICA and Box-Cox Transform

对锂离子电池的健康状态SOH(state of health)进行准确估计是锂离子电池安全稳定运行的重要保障,提出了一种基于容量增量分析ICA(incremental capacity analysis)和Box-Cox变换的锂离子电池SOH估计方法。首先,将电池恒流充电过程的IC曲线峰值高度ICP(peak of incremental capacity curve)作为健康特征HF(health factor),数学推导出ICP与健康状态的强相关性。结合卡尔曼滤波算法提取光滑的容量增量曲线。将电池容量衰退过程的前部分周期作为训练周期,通过Box-Cox变换将训练周期的ICP和SOH序列变换成线性关系,然后通过线性拟合来实现剩余周期的SOH估计。在Oxford和NASA数据集上进行实验验证,并与机器学习算法进行对比,结果表明所提方法具有较高的估计精度、较短的计算时间和较强的鲁棒性。

The accurate estimation of the state of health(SOH)of Li-ion batteries is an important guarantee for their safe and stable operation.In this paper,an SOH estimation method for Li-ion batteries based on incremental capacity analysis(ICA)and Box-Cox transform is proposed.First,the peak of incremental capacity curve(ICP)for a Li-ion bat⁃tery during its constant current charging process is taken as a health factor(HF),and the strong correlation between ICP and SOH is deduced mathematically.Then,the smoothed incremental capacity curve is extracted by combing the Kalman filter algorithm.The cycles before the starting point(SP)during the battery capacity degradation are regarded as the training cycles,whose ICP and SOH series are transformed into a linear relationship by the Box-Cox transform.Afterwards,the SOH estimation of the remaining cycles is realized by linear fitting.The experimental results on the Ox⁃ford and NASA dataset show that compared with machine learning algorithms,the proposed method has a higher estima⁃tion efficiency,a shorter computational time,as well as a stronger robustness.