Sudden Fault Diagnosis of Lithium Battery Pack Based on Double Fault Interval Location and Random Forest Algorithm
Sudden faults of lithium batteries often cause great harm in a short period of time,restricting their further development.Therefore,this paper investigates sudden faults and proposes a complete framework for diagnosing sudden faults in lithium battery packs.First,the voltage data sequence is decomposed using empirical modal decomposition,and the subsequence is constructed into a new data sequence and imported into the binary fault interval localization model,and the correlation coefficient of dichotomous ideas and the correlation coefficient based on the time window are used to diagnose the fault interval of the sudden fault.Then,the fault features are extracted using kernel principal component analysis.Finally,the faults are trained and classified using the random forest algorithm and compared with the remaining methods.In addition,a fault injection platform is established to physically trigger external short circuit faults and contact faults on a series-connected four-cell battery pack.The experimental results show that the method is accurate and reliable for the diagnosis of sudden faults.
国家科技期刊平台
NSTL
万方数据
