bsebench-org/lg-hg2-stroebl-2024-raw

LG INR18650HG2电池循环数据集（Stroebl等人，2024）的Tier 1原始镜像子集，专注于具有硅-石墨复合阳极（C-SiO混合石墨，含3-6% SiO）的NMC电池，在WLTC（全球统一轻型车辆测试循环）动态负载下的循环老化数据。该数据集是BSEBench V1.0的一部分，包含28个CSV文件（约7.9 GB），覆盖12个参数集（P065-P076）和3个重复实验，数据包括高分辨率日志（LOG）、循环结束数据（EOC）、老化日志（LOG_AGE）、电化学阻抗谱（EIS）和脉冲测试（PULSE）。电池在0°C、10°C、25°C和40°C温度下进行测试，老化持续时间长达449天。数据集特别强调硅-碳阳极的路径依赖滞后效应（约10%），这对传统等效电路模型（2-RC ECM）构成挑战，适用于高级滤波方法（如BMA集成、神经状态空间、粒子滤波混合模型）的基准测试。数据集用于电池状态估计（如SOC、SOH）研究，支持BSEBench V1.1评估。Tier 1 raw mirror subset of the LG INR18650HG2 battery cycling dataset by Stroebl, Mayer, Lienkamp (KIT, 2024), focusing on NMC cells with silicon-graphite composite anode (C-SiO blended graphite, 3-6% SiO) under WLTC (Worldwide Harmonized Light-duty Vehicles Test Cycle) dynamic load for cyclic aging. This dataset is part of BSEBench V1.0 and includes 28 CSV files (~7.9 GB) covering 12 parameter sets (P065-P076) with 3 replicates each, containing data types such as high-resolution logs (LOG), end-of-cycle data (EOC), aging logs (LOG_AGE), electrochemical impedance spectroscopy (EIS), and pulse tests (PULSE). Cells were tested at temperatures of 0°C, 10°C, 25°C, and 40°C, with aging durations up to 449 days. The dataset highlights the path-dependent hysteresis (~10%) of silicon-carbon anodes, which is uncorrectable by classical 2-RC equivalent circuit models, making it a stress-test for advanced filtering approaches (e.g., BMA ensembles, neural state-space, particle-filter mixture models). It is designed for battery state estimation (e.g., SOC, SOH) benchmarking and supports BSEBench V1.1 evaluation.