bsebench-org/severson-2019

Severson 2019 LFP 快充数据集是BSEBench规范（Tier 2）版本，源自Severson等人2019年发表的商业LFP（磷酸铁锂）电池快充循环数据。该数据集已转换为BSEBench标准时间序列模式（TimeSeriesSchema），存储为Parquet格式，并使用Zstandard压缩，按cell_id分区以便并行读取。数据集包含124个电池单元（从原始140个通道中排除不合格单元），每个单元包含时间（秒）、电压（伏特）、电流（安培，遵循BPX-1.1符号约定：充电为正，放电为负）、温度（摄氏度）、循环次数（1起始索引）、步骤标识（如充电、放电等）和容量（安时）等列。数据集用于电池状态估计（如SOC和SOH）的基准测试，支持通过bsebench-datasets Python包或直接使用Pandas/Arrow加载。数据集遵循CC-BY-4.0许可，是原始.mat文件的衍生作品。The Severson 2019 LFP Fastcharge dataset is a BSEBench-canonical (Tier 2) harmonization of the commercial LFP (lithium-iron-phosphate) battery fast-charging cycling data from Severson et al. 2019. It is converted to the BSEBench standard TimeSeriesSchema, stored in Parquet format with Zstandard compression and partitioned by cell_id for parallel reading. The dataset includes 124 battery cells (after excluding unfit cells from the original 140 channels), each with columns such as time (seconds), voltage (volts), current (amperes, following the BPX-1.1 sign convention: positive on charge, negative on discharge), temperature (Celsius), cycle number (1-indexed), step identifier (e.g., charge, discharge), and capacity (ampere-hours). It is designed for battery state estimation (e.g., SOC and SOH) benchmarking and can be loaded via the bsebench-datasets Python package or directly with Pandas/Arrow. The dataset is licensed under CC-BY-4.0 and is a derivative work of the original .mat files.