Battery Aging Dataset for HEV Drive Cycles applied to 2.5Ah A123 LFP Cells

Three 2.5Ah LFP battery cells are cycled until they reached a very low SOH (70% for the 300k km case, 40% for the 206k km case, and 15% for the 95k km case). The tests begin with an HPPC test and two repeated 1C discharge tests, which are followed by 300 repetitions of a power profile calculated for an HEV following the WLTP drive cycle. Each aging case has a different power profile, which is designed based off of an empirical aging model to acheive approximately 95,000 km, 206,000 km, or 300,000 km of driving before the cell ages to 80% SOH.

One possible application of the dataset could be parameterizing/training and testing SOC or SOH estimation models for hybrid electric vehicles. Since there are three cases which age at a different rate, two cases could be used to train a machine learning algorithm and the third case could be used to test it, for example. For full details regarding the design of the tests see the Applied Energy paper "Battery state-of-health sensitive energy management of hybrid electric vehicles: Lifetime prediction and ageing experimental validation".

本数据集选取3节2.5Ah磷酸铁锂电池（LFP）进行循环老化测试，直至其健康状态（SOH）降至极低水平：其中30万公里工况对应70%健康状态，20.6万公里工况对应40%健康状态，9.5万公里工况对应15%健康状态。测试流程首先开展混合脉冲功率特性测试（HPPC）与两次重复的1C倍率放电测试，随后进行300次循环的功率曲线测试，该功率曲线基于遵循世界轻型汽车测试循环（WLTP）的混合动力电动汽车（HEV）工况计算得到。每个老化工况对应专属的功率曲线，该曲线基于经验老化模型设计，旨在使电池在衰减至80%健康状态前，累计行驶里程分别达到约9.5万公里、20.6万公里或30万公里。

该数据集的潜在应用场景之一是为混合动力电动汽车的荷电状态（SOC）或健康状态（SOH）估算模型提供参数化、训练与测试支持。例如，由于三个老化工况的衰减速率各不相同，可选取其中两组数据用于训练机器学习算法，剩余一组数据则用于测试模型性能。若需了解测试设计的完整细节，请参阅《应用能源》期刊刊载的论文"Battery state-of-health sensitive energy management of hybrid electric vehicles: Lifetime prediction and ageing experimental validation"。