基于自愈特性的电容剩余寿命预测  

Prediction of capacitor remaining useful life based on self-healing characteristics

作　　者：成庶[1] 邹阳 CHENG Shu;ZOU Yang(School of Traffic and Transportation Engineering,Central South University,Changsha 410075,China)

机构地区：[1]中南大学交通运输工程学院,湖南长沙410075

出　　处：《铁道科学与工程学报》2025年第2期900-908,共9页Journal of Railway Science and Engineering

基　　金：国家自然科学基金资助项目(52072414)。

摘　　要：薄膜电容的寿命状态与动车组的安全运行息息相关,而电容的自愈现象是影响电容寿命的关键。目前,对于电容的寿命研究多为使用算法手段预测退化数据,少有对具体失效形式及其影响进行分析和研究。为揭示薄膜电容自愈现象与寿命的关系,提出一套基于自愈性能评估的电容退化分析方法,该方法能够很好地描述电容发生自愈时的累积损伤。首先通过分析自愈导致电容损伤的机理,确定了自愈参数中对电容影响显著的2个参数:自愈能量与自愈频次。在此基础上开展电容的加速老化实验,并结合电荷补偿的方式对电容自愈能量和次数进行检测,对自愈能量特征进行统计分析,并结合改进的粒子群算法和霍尔特指数平滑法对电容的计数过程进行建模。然后基于预测的计数过程和电容自愈特征分布情况,使用非齐次复合poisson过程计算电容的退化特征,得到老化过程中的累积能量预测及90%置信区间。同时根据电容容值损伤和电容累积自愈能量的关系,制定自愈的老化预测方案的失效阈值,得到自愈损伤的寿命分布。最终,根据容值得到的电容退化寿命与基于自愈特征的电容寿命进行比较。研究结果表明,在直流电压作用下,通过自愈能量特征所得的电容寿命与通过容值所得电容寿命相对误差约为2.55%,验证了根据自愈性能的退化分析和寿命预测方法的有效性和可靠性。The lifespan of film capacitors is directly related to the safe operation of electric multiple units,and the self-healing phenomenon of the capacitor plays a critical role in determining their lifespan.Presently,most research on capacitor lifespan utilizes algorithms to predict degradation data,yet in-depth analysis of specific failure modes and their impact is relatively scarce.To elucidate the relationship between the self-healing phenomenon of film capacitors and their lifespan,this paper proposed a degradation analysis method based on self-healing performance evaluation,which can accurately describe the cumulative damage during the self-healing process.First,an analysis was carried out on the mechanisms through which self-healing causes capacitor damage,identifying two significant parameters from the self-healing process:self-healing energy and self-healing frequency.Following this,accelerated aging experiments were performed on the capacitors,and the self-healing energy and frequency were detected using the charge compensation method.Statistical analysis of self-healing energy characteristics was performed,and the counting process of capacitors was modeled using an improved particle swarm optimization algorithm in combination with the Holt exponential smoothing method.By using the predicted counting process and the distribution characteristics of self-healing,a non-homogeneous compound Poisson process was employed to calculate the degradation characteristics of the capacitor,resulting in the prediction of cumulative energy during aging and its 90%confidence interval.Furthermore,a failure threshold for the self-healing aging prediction scheme was established based on the relationship between capacitor capacitance damage and cumulative self-healing energy,allowing for the lifespan distribution of self-healing damage to be obtained.Finally,the degradation lifespan of the capacitors,based on both capacitance and self-healing characteristics,was compared.The results showed that under direct current voltage,the re

关 键 词：非齐次复合poisson过程 金属化薄膜电容器 寿命预测 自愈 霍尔特指数平滑法 

分 类 号：U266[机械工程—车辆工程] TM533.3[交通运输工程—载运工具运用工程]