点击次数：

影响因子：9.8

DOI码：10.1016/j.est.2025.117325

发表刊物：Journal of Energy Storage

关键字：Lithium-ion battery; Capacity estimation; Discharge rate compensation model; Error compensation functions; Different discharge rates

摘要：Accurate estimation of LFP battery capacity is important for improving system safety and extending battery life. Most existing research focuses on capacity estimation at a single discharge rate, neglecting the impact of discharge rate on battery capacity. Moreover, the trained models are often only applicable to specific discharge rates. To overcome this challenge, this paper proposes an adaptive capacity estimation method based on a discharge rate compensation model. Initially, a comparative analysis was conducted to examine the correlation between selected features and battery capacity at diverse discharge rates, revealing highly correlated feature ranges across all rates. Subsequently, optimal data-driven models were obtained by leveraging these features and employing Bayesian optimization. Finally, an error compensation function was incorporated into the optimal data-driven model to construct a discharge rate compensation model (DRCM), enabling capacity estimation across multiple discharge rate scenarios. The comparison results with deep learning and traditional machine learning show that the proposed DRCM has the best estimation performance, achieving 0.74% Mean Absolute Percentage Error (MAPE) and 0.99% Root Mean Square Percentage Error (RMSPE) for LFP batteries, and 1.56% MAPE and 2.12% RMSPE for NMC batteries, with training time only 1/6–1/20 of machine learning.

备注：中科院2区

合写作者：Changpeng Tan,Ji Wu,Duo Yang

第一作者：Li Wang

论文类型：期刊论文

通讯作者：Muyao Wu

论文编号：117325

学科门类：工学

文献类型：J

卷号：131

ISSN号：2352-152X

是否译文：否

发表时间：2025-06-23

收录刊物：SCI、EI

发布期刊链接：https://www.sciencedirect.com/science/article/pii/S2352152X25020389