基于混合模型的储能用锂离子电池剩余寿命预测方法研究  

RESEARCH ON REMAINING LIFE PREDICTION METHOD OF LITHIUM-ION BATTERY FOR ENERGY STORAGE BASED ON HYBRID MODEL

作　　者：夏向阳[1] 吕崇耿 吴小忠 曾小勇[1] 刘代飞[1] Xia Xiangyang;Lyu Chonggeng;Wu Xiaozhong;Zeng Xiaoyong;Liu Daifei(School of Electrical and Information Engineering,Changsha University of Science and Technology,Changsha 410114,China;State Grid Hunan Electric Power Co.,Ltd.,Changsha 410118,China)

机构地区：[1]长沙理工大学电气与信息工程学院,长沙410114 [2]国网湖南省电力有限公司,长沙410118

出　　处：《太阳能学报》2024年第10期726-735,共10页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金(51977014);湖南省教育厅优秀青年项目(19B007)。

摘　　要：针对储能系统在调峰、调频等工况下存在电池老化等安全问题,提出一种储能用锂离子电池剩余使用寿命(RUL)预测方法。该方法通过研究锂离子电池充电曲线特点,从电池恒流、恒压充电过程中提取短时电压电流片段数据作为表征电池老化的特征,提出一种B样条曲线的方法对得到的片段数据进行降噪处理,得到与容量高度相关的新健康因子(HF),然后运用蝴蝶优化策略改进的麻雀搜索算法对卷积神经网络-长短期记忆网络混合模型参数进行优化,最后以NASA数据集和实验平台实际运行数据对锂离子电池RUL进行预测并与其他方法对比验证,结果表明,NASA数据集预测结果均方根误差RMSE、平均绝对误差MAE范围控制在3.45%、2.42%之内,实验平台电池老化数据集预测结果RMSE、MAE范围控制在0.96%、0.83%之内,所提方法对锂离子电池RUL预测具有较高的准确性,能有效解决储能电池安全预警问题,可为电网稳定控制提供保障。Aiming at the safety problems of battery aging in energy storage systems under peak and frequency regulation conditions,a method for predicting the remaining useful life(RUL)of lithium-ion batteries for energy storage is proposed.The method extracts short-time voltage and current fragment data from the constant current and constant voltage charging process of lithium-ion batteries as a characteristic to characterize the battery aging by studying the characteristics of the charging curve of lithium-ion batteries,proposes a B-spline curve method for noise reduction of the obtained fragment data,and obtains the new health factor(HF)which is highly correlated with the capacity,and then applies the sparrow search algorithm improved by the butterfly optimization strategy to optimize the parameters of the convolutional neural network-Long and short-term memory network hybrid model parameters are optimized,and finally the RUL of lithium-ion battery is predicted by NASA dataset and experimental platform actual operation data and verified by comparing with other methods.The results show that the predicted results of the NASA dataset RMSE and MAE are controlled within 3.45%and 2.42%,and the predicted results of the battery aging dataset of the experimental platform are controlled within 0.96%and 2.42%,and the predicted results of the experimental platform are controlled within 0.96%and 0.96%.The range of RMSE and MAE is controlled within 0.96%and 0.83%,and the method proposed in this paper has high accuracy in predicting the RUL of lithium-ion batteries,which can effectively solve the problem of safety warning of energy storage batteries,and provide a guarantee for the stable control of power grids.

关 键 词：锂离子电池 储能 剩余使用寿命 B样条曲线 新健康因子 蝴蝶优化策略 

分 类 号：TM911[电气工程—电力电子与电力传动]