锂离子电池（LIB）已经在电动汽车和电网能量储存市场占据了主导地位。准确监测电池健康状况一直是电池行业最关键的挑战之一。机器学习（ML）已广泛应用于电池健康评估和预测。香港中文大学卢怡君团队在EcoMat发表综述性文章，通过研究特定的特征和目标，全面讨论了在电池健康领域各种场景下ML辅助任务的实施。本综述探讨了基于多级单元衰减的ML辅助任务。作者强调了在ML模型训练期间考虑潜在的特征-目标对的机会和意义，以识别更多关于LIB的健康信息，并为新的应用场景设计任务提供指导。
Figure 1 The hierarchical relationships between the root cause (Level 4), the relevant degradation mechanism (Level 3; the darker the blue, the higher number of coupling cause factors), the corresponding degradation mode (Level 2), and resulting effect (Level 1)
Figure 2 Flowchart of training process for supervised ML model starting from the selection of feature and target
Figure 3 Overview of ML tasks implemented in multiple levels of degradation in battery health.
Zijie Huang, Lawnardo Sugiarto, Yi-Chun Lu,* Feature–target pairing in machine learning for battery health diagnosis and prognosis: A critical review, EcoMat
. 2023; e12345
原文链接：https://doi.org/10.1002/eom2.12345
