文档名：基于数据预处理和VMDLSTMGPR的锂离子电池剩余寿命预测
摘要：锂离子电池的剩余使用寿命(RUL)是健康管理中重要参数,其准确评估对于保证电池设备的安全稳定运行非常重要.该文提出一种数据预处理联合变分模态分解(VMD)、长短期记忆网络(LSTM)和高斯回归过程(GPR)的预测框架.首先选取充放电循环过程中的信息作为间接健康因子(HI),并通过核主元分析方法(KPCA)实现间接HI的特征提取,完成数据预处理;其次通过VMD-LSTM方法实现健康因子的分解、预测和重构,并将重构得到的数据应用于RUL预测的GPR模型,完成预测模型搭建;最后以NASA锂电池数据集作为算法测试数据,结果表明,所提取的健康因子能够准确跟踪锂电池的退化过程;所提预测方法能够准确地估计电池的剩余寿命,同时具有较高的可靠性和稳定性.
Abstract：Theperformanceoflithium-ionbatteriescandeterioratewiththedecreaseofcapacityandtheincreaseofimpedanceduringcontinuouscharginganddischargingprocess,whichposesariskofequipmentandsystemfailures,includingcatastrophiclosses.Accurateandreliablepredictionoftheremainingusefullife(RUL)oflithium-ionbatteriesiscrucial.However,previouspredictionmethodsmainlysupportedpointpredictionswithoutofferingaclearmathematicalrepresentationoftheconfidencelevelofthepredictionresults.Thenoiseandcapacityreboundphenomenaintheoriginaldataareignored.Therefore,thispaperproposesapredictionframeworkbasedondatapre-processingcombinedwithvariationalmodedecomposition(VMD),longandshort-termmemorynetwork(LSTM),andGaussianregressionprocess(GPR).Firstly,anindirecthealthindicator(HI)reflectingthelifedegradationtrendoflithium-ionbatteriesisextractedfromthecharge/dischargecurve.PearsonandSpearmancorrelationcoefficientsareusedtoverifythecorrelationbetweentheextractedindirectHIandcapacity.Thekernelprincipalelementanalysis(KPCA)methodreducescomputationalcomplexitybyremovingredundantcomponentsofindirectHI,transformthemintofusionHI.Then,theVMDdecompositionmethoddecomposesthefusionHIintomultiplemodalcomponents.Basedonthecentralfrequencyofmodalcomponentsandcapacity-relatedcoefficients,multiplemodalcomponentsaredividedintothreeparts:globalattenuation,localregeneration,andothernoise.TimeseriespredictionoftheglobalattenuationandlocalregenerationcomponentsareperformedseparatelyusingLSTMneuralnetworkstoobtainthepredictedvaluesafterthepredictionstartingpoint.ThevaluesofeachcomponentaresummeduptoobtainthereconstructedHI.Finally,thereconstructedHIandcapacityserveastheinputandoutputoftheGPRmodelRULprediction,respectively.NASAlithium-ionbatterypublicdatasetisusedforexperimentalvalidation,anddifferentpredictionstartingpointsaresetforeachbattery.Thestartingpointsetisatabout40%ofthetotalcycle.Theexperimentalresultsshowthatthepredictedcapacityvaluescloselytrackthebatteryagingtrendandeffectivelyestimatethelocalregenerationphenomenonduringtheagingprocess.Thetruecapacityvaluesgenerallyfallwithinthe95%confidenceintervalofthepredictedvalue,andthepredictioneffectimprovesasthetrainingdataincrease.Regardingevaluationindexesforcapacitypredictionresults,themaximumrootmeansquareerror,meanabsoluteerror,andmeanabsolutepercentageerrorare2.98%,2.34%,and1.81%,respectively.Errorsintheremainingusefullifepredictionareallwithin2cycles.Thefollowingconclusionscanbedrawnfromtheexperimentalanalysis:(1)TheKPCAalgorithmremovesredundantinformationbetweenindirecthealthindicatorstoreducedatacomplexityandachievedatapre-processing.(2)TheVMDdecompositionmethodminesintrinsicinformation,capturingthelong-termdownwardtrend,localregeneration,andnoisecomponenttoremoveimpliednoiseinthedata.(3)TheGPRmodelsupportsprobabilisticpredictionandprovidesconfidenceintervalsforcapacitypredictionresults.
作者：李英顺 阚宏达 郭占男 王德彪 王铖 Author：LiYingshun KanHongda GuoZhannan WangDebiao WangCheng
作者单位：大连理工大学控制科学与工程学院大连116000沈阳顺义科技股份有限公司沈阳110000
刊名：电工技术学报 ISTICEIPKU
Journal：TransactionsofChinaElectrotechnicalSociety
年，卷(期)：2024, 39(10)
分类号：TM912
关键词：锂离子电池 剩余寿命 健康因子 变分模态分解 高斯回归过程 长短期记忆
Keywords：Lithium-ionbattery remainingusefullife healthindicator variationalmodedecomposition Gaussianregressionprocess longandshort-termmemory
机标分类号：TH17TP277TM912.9
在线出版日期：2024年5月31日
基金项目：辽宁省兴辽英才计划资助项目基于数据预处理和VMD-LSTM-GPR的锂离子电池剩余寿命预测[
期刊论文] 电工技术学报--2024, 39(10)李英顺 阚宏达 郭占男 王德彪 王铖锂离子电池的剩余使用寿命(RUL)是健康管理中重要参数,其准确评估对于保证电池设备的安全稳定运行非常重要.该文提出一种数据预处理联合变分模态分解(VMD)、长短期记忆网络(LSTM)和高斯回归过程(GPR)的预测框架.首先选取...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
基于数据预处理和VMD-LSTM-GPR的锂离子电池剩余寿命预测 Prediction of Remaining Useful Life of Lithium-Ion Battery Based on Data Preprocessing and VMD-LSTM-GPR
基于数据预处理和VMD-LSTM-GPR的锂离子电池剩余寿命预测.pdf
- 文件大小:
- 3.35 MB
- 下载次数:
- 60
-
高速下载
|
|
