A coulombic efficiency-based model for prognostics and health estimation of lithium-ion batteries. (15th March 2019)

• Record Type: Journal Article
• Authors: Yang, Fangfang; Song, Xiangbao; Dong, Guangzhong; Tsui, Kwok-Leung
• Is Part Of: Energy. Volume 171(2019)
• Page Start: 1173
• Page End: 1182
• Publication Date: 2019-03-15
• Subjects: Degradation modeling -- Coulombic efficiency -- Lithium iron phosphate batteries -- Loss of lithium inventory -- Remaining useful life prediction -- Health estimation; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2019.01.083
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: Coulombic efficiency, as an important battery parameter, is highly related to the loss of lithium inventory, which is the dominant aging factor for lithium-ion batteries. In this paper, a semi-empirical model is derived from this relationship to capture the capacity degradation of lithium-ion batteries. The coulombic efficiency-based model effectively captures the convex degradation trend of lithium iron phosphate batteries and presents better fitting performance than the existing square-root-of-time model. To evaluate the proposed model, a battery cycle life experiment was designed, in which the subjects were continuously cycled under a federal urban driving schedule to simulate real-life battery usage. To perform online battery health estimation and prognostics, a particle filtering framework incorporating the proposed model was constructed to update the model parameters regularly with periodically measured data. Remaining useful life of the battery was then predicted by extrapolating the models with renewed parameters. The experimental results indicated that the proposed prognostic method can provide higher prediction accuracy than the existing square-root-of-time model. Highlights: Proposal of a coulombic efficiency-based model for LFP battery degradation. Investigation of the new model w.r.t. feasibility, goodness-of-fit, and robustness. Integration with particle filter for on-board battery health estimation. Dynamic loading profiles for on-board performance verification.