锂离子电池容量预测方法研究

Study on capacity prediction method of Li-ion battery

以280 Ah磷酸铁锂电池为研究对象,在标准环境下测得电池在0.5 C的放电曲线。先将电池以0.5 C放电至30%荷电状态(SOC)并记录静置3 h后稳定的开路电压(OCV),然后继续在0.5 C下从30%SOC放电至10%SOC,过程中每步放电2%SOC后静置3 h记录相应OCV,最终获得电池的实际SOC-OCV曲线。该区间内SOC-OCV曲线非线性,特别在24%~26%SOC区间存在明显斜率变化。因此,对30%~10%SOC区间数据进行多项式拟合,得到的多项式拟合方程的拟合R值为0.9965;同时对26%~32%SOC区间数据进行线性拟合,得到的线性方程拟合R值为0.99949;两种拟合结果均有较好的拟合度。将电池充电至100%SOC后再放电至接近30%SOC以保证电池性能和出货带电态,并将流程结束电池进行静置,记录电池从100%SOC放电至接近30%SOC时的总放电容量Cy、实际放电总容量C以及静置过程中的OCV。根据电池的SOC、Cy、C和OCV之间的关系,推出基于上述两种拟合方程下的容量预测模型。将OCV和Cy代入容量预测模型进行容量预测。研究表明:静置12 h后OCV值接近稳定,利用该OCV值进行容量预测的预测误差在±1.5%以内;在静置时间较短的情况下,引入校正因子来修正预测容量值可保证同样的预测误差;在本研究的SOC区间内,该预测方法具有可行性且两种预测模型下的预测误差基本一致。

Taking 280 Ah lithium iron phosphate(LFP)battery as the research object,the discharge curve of the battery under the current rate of 0.5 C was measured in the standard environment.First,the battery was discharged at 0.5 C to the state of charge(SOC)of 30%and rested for 3 h to record the stable open circuit voltage(OCV).Then the battery was discharged from 30%SOC to 10%SOC at 0.5 C,meanwhile the test was interrupted after every single step of 2%SOC and rested for 3 h to record the corresponding stable OCV.Finally the SOC-OCV curve was obtained.The result shows that the SOC-OCV curve of the battery between 30%-10%is nonlinear,especially in the range of SOC at 24%-26%with a significant slope change.Therefore,polynomial fitting of SOC to OCV between 30%-10%theoretical SOC is performed,and the adjusted R-square of polynomial fitting is 0.9965.Meanwhile the linear fitting of SOC to OCV between 26%-32%SOC is performed,and the adjusted R-square of linear fitting is 0.99949.Both fitting equations are of desirable goodness-of-fit.The battery is charged to 100%SOC and then discharged to nearly 30%SOC in order to ensure the performance and the outgoing state of charge.The OCV of battery during standby period,the discharge capacity from 100%SOC to nearly 30%SOC and the total discharge capacity are recorded as OCV,Cy and C,respectively.On the basis of the relation between SOC,Cy,C and OCV,this study proposes two capacity prediction models based on the above mentioned fitting equations,then OCV and Cy are brought into the capacity prediction model to obtain the prediction capacity.The following conclusions are drawn from this study:the capacity prediction errors under the stable OCV after rest for 12 h were within±1.5%;in the case of a short standby time,a correction factor of predicted capacity should be introduced to ensure the same level of prediction error;the prediction methods are feasible,and the prediction errors under two models are basically the same in the range of SOC selected.