Abstract: Lithium-ion battery health estimation is critical for ensuring their safe use and optimizing the energy management of electric vehicle battery packs. To overcome the shortcomings of existing methods in prediction accuracy, a hybrid model combining improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), osprey-cauchy-sparrow search algorithm (OCSSA), and deep extreme learning machine (DELM) is proposed. ICEEMDAN algorithm is used to decompose the battery capacity decay data and separate the high-frequency and low-frequency components. The sparrow search algorithm (SSA) is improved by chaotic mapping, osprey optimization and cauchy variation to improve the parameter optimization accuracy. OCSSA optimizes DELM model parameters to enhance prediction accuracy. In order to verify the effectiveness of the proposed method, this paper tests the data set of the national aeronautics and space administration (NASA) and the center for advanced life cycle engineering (CALCE). The experimental results show that the model has high prediction accuracy and stability, and the average absolute error and root mean square error of the model prediction are within 1%. Moreover, high prediction accuracy is still maintained when predicting multiple steps in the future.

Key words: lithium-ion battery; state of health prediction; hybrid model; parameter optimization

摘要： 锂离子电池健康状态估计对于确保其安全使用和优化电动汽车电池组的能量管理至关 重要。针对现有方法在预测精度方面的不足，文章提出了一种结合改进自适应噪声互补集合 经验模态分解（ICEEMDAN）、融合鱼鹰优化和柯西变异的麻雀搜索算法（OCSSA）以及深 度极限学习机（DELM）的混合模型。利用 ICEEMDAN 算法分解电池容量衰退数据，分离高 频和低频分量。通过混沌映射、鱼鹰优化和柯西变异改进麻雀搜索算法（SSA）提高参数优 化精度。利用 OCSSA 优化 DELM 模型参数，增强预测准确性。为验证所提出方法的有效性， 文章基于美国国家航空和宇宙航行局（NASA）和先进寿命周期工程中心（CALCE）的数据 集进行测试，实验结果证明该模型具有高预测精度和稳定性，模型预测的平均绝对误差及均 方根误差均在 1%以内，并且在预测未来多步时，仍然保持高的预测精度。

关键词: 锂离子电池；健康状态预测；混合模型；参数优化