中文摘要：

论文以三元锂离子电池为实验对象,在利用线性粒子群优化算法对4种典型的等效电路模型参数进行参数辨识的基础上,比较了这4种模型的复杂度和精度。4种模型中,开路电压均取自标准脉冲实验的主迟滞回线的充电开路电压值和放电开路电压值的平均值。对4种典型模型预测结果误差进行分析,证明了电池极化过程引起的迟滞效应是造成预测结果误差的主要原因。基于此,提出一种改进方法,即构造分段线性迟滞电压函数,修正平均开路电压值和荷电状态之间的关系,以减小迟滞效应引起的预测误差。实验结果表明,提出的方法对多种等效电路模型具有普遍适用性,能够有效提高各等效电路模型的预测精度。

英文摘要：

The paper uses NMC batteries （with LiCoxNiyMnzO2 as positive pole material） as test objects. It adopts linear particle swarm optimization （PSO） algorithm to identify parameters of the most typical battery equivalent circuit models （ECM） and compares complexity and accuracy of four ECMs. For each model, open circuit voltage （OCV） is extracted from average of charging and discharging OCVs of major hysteresis loop under standard pulse test. However, battery polarization processing causes hysteresis effect, and it is verified by analysis that the hysteresis effect is the major contribution to ECM prediction error. To mitigate this problem, a novel method＂ constructing a piecewise linear function for hysteresis voltage is proposed to modify relationship between OCV and SOC. Experiment results show that the proposed algorithm is feasible and universally applicable for lithium-ion battery model, and improves ECM prediction accuracy.