Lithium-ion (Li-ion) batteries have been used in a wide variety of applications, ranging from portable electronics to electric vehicles. During repetitive charging and discharging, a battery's capacity fades due to electrochemical reactions such as solid electrolyte interphase growth. Li-ion batteries reach an end-of-life (EOL) point, after which using them is not recommended. However, some unhealthy batteries reach their EOL sooner than expected. A qualification test is usually conducted to evaluate the reliability of Li-ion batteries and classify unhealthy batteries, but this test requires several months. This paper develops a data-driven method to reduce the qualification time by detecting anomalies before EOL. This method detects an anomaly in the capacity fade curve of unhealthy batteries based on their capacity fade trend. Since the developed method detects anomalies of unhealthy batteries before EOL, the method is effective in reducing the time for the qualification test of Li-ion batteries.

中文翻译：

---

基于预测和健康管理缩短锂离子电池鉴定时间


锂离子 (Li-ion) 电池已用于多种应用，从便携式电子产品到电动汽车。在重复充电和放电过程中，电池的容量会因固体电解质界面生长等电化学反应而衰减。锂离子电池达到使用寿命 (EOL) 点，之后不建议使用它们。然而，一些不健康的电池比预期更早达到停产状态。通常会进行资格测试来评估锂离子电池的可靠性并对不健康的电池进行分类，但该测试需要几个月的时间。本文开发了一种数据驱动的方法，通过在 EOL 之前检测异常来缩短鉴定时间。该方法根据不健康电池的容量衰减趋势来检测其容量衰减曲线的异常。由于所开发的方法可以在EOL之前检测出不健康电池的异常情况，因此该方法可以有效减少锂离子电池的鉴定测试时间。