基于容量增量分析的动车组蓄电池分选方法
Battery sorting method for EMU based on increment capacity analysis
摘要
动车组镉镍电池的不一致性问题严重影响列车性能,不仅影响电池组荷电状态的准确评估及故障诊断,还会加剧电池性能的不一致性.为实现蓄电池的精准分选,文章提出一种基于容量增量分析的动车组蓄电池分选方法.该方法首先对电池进行容量增量分析,表征电池在充电过程中的性能变化特性;然后利用t-分布随机近邻嵌入(t-SNE)算法提取容量增量曲线中的不一致性参数,在保留关键信息的同时降低特征维度,以优化模型性能;最后采用迭代自组织聚类算法(ISODATA)对不一致性参数进行聚类分析,实现对电池动态性能的分选.试验结果表明,相较于传统基于静态参数的方法,该方法能更精准地识别电池间的性能差异,显著提高蓄电池分选的准确性和可靠性,为动车组蓄电池的管理与优化提供了理论依据与技术支持.
The inconsistency of nickel-cadmium batteries in electric multiple units(EMUs)has a significant impact on train perfor-mance.It not only hinders the accurate evaluation of the state of charge and fault diagnosis of the battery packs but also exacerbates the uneven performance of the batteries.To achieve precise sorting of batteries,an EMU battery sorting method based on capacity increment analysis is proposed.Firstly,the capacity increment analysis was performed to characterize the performance variation of the batteries dur-ing charging processes.Secondly,the t-distributed stochastic neighbor embedding(t-SNE)algorithm was used to extract inconsistency parameters from the capacity increment curves,aiming to retain critical information while reducing the number of features to optimize model performance.Finally,the iterative self-organizing data analysis technique(ISODATA)algorithm was employed to cluster the in-consistency parameters,enabling the sorting of battery dynamic performance.According to the test results,compared to traditional meth-ods based on static parameters,this method more accurately identifies performance differences between batteries,significantly improving the accuracy and reliability of battery sorting.This provides theoretical support and technical assistance for the management and optimi-zation of EMU batteries.
年，卷(期)：
2026, (1)
分类号：
TM912
U266
机标分类号：
TP301.6
TM912
P315.7
在线出版日期：
2026年6月8日
基金项目：
2022YFB4301201-05:国家重点研发计划，
