Premium
A novel dual time scale life prediction method for lithium‐ion batteries considering effects of temperature and state of charge
Author(s) -
Wang Xueyuan,
Li Rikang,
Dai Haifeng,
Zhang Nutao,
Chen Qijun,
Wei Xuezhe
Publication year - 2021
Publication title -
international journal of energy research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.808
H-Index - 95
eISSN - 1099-114X
pISSN - 0363-907X
DOI - 10.1002/er.6746
Subject(s) - state of charge , battery (electricity) , lithium (medication) , lithium ion battery , electrolyte , reliability engineering , computer science , chemistry , engineering , thermodynamics , electrode , physics , power (physics) , endocrinology , medicine
Summary Life prediction facilitates efficient management and timely maintenance of lithium‐ion batteries. Challenges are still faced in eliminating the effects of battery temperature or state of charge (SOC) on the life indicator to form a life prediction method for complex onboard working conditions. To fulfill the research gap, this paper focuses on three novelties about the life indicator, effect elimination, and life prediction method. First, impedance spectra at different temperatures, SOC, and aging cycles are comprehensively studied by experiments. By fitting the spectra with an equivalent circuit model, changes of ohmic resistance, solid electrolyte interphase resistance, and charge transfer resistance (CTR) are analyzed in detail. CTR is determined as a novel life indicator, and an empirical model describing the changing trend of CTR with aging cycles is established. Second, a multi‐factor coupled CTR model is applied to eliminate the strong effects of temperature and SOC during the prediction. Third, the tracking of the effects and the changing trend of the CTR with the aging cycles form a composite life prediction method with dual time scales. The results show that the battery life can be accurately predicted and the errors converge to within ±5% even though the indicator CTR is obtained at different temperatures and SOC. With this method, life prediction no longer depends on the indicator obtained in a specific state. It has great potential to broaden the implementation of life prediction for onboard conditions.