Large Language Model–Based Fault Diagnosis for Lithium-ion Batteries in Cloud-Edge Systems

This dataset was used in the paper Large Language Model–Based Fault Diagnosis for Lithium-ion Batteries in Cloud-Edge Systems. It comprises laboratory experimental data and real-world energy storage system (ESS)data.he laboratory data record battery current, voltage, and (in some cases) temperature during fault experiments, as well as the fault severity and fault type. The dataset includes both raw measurements and data obtained through augmentation. For fault cases, the samples retained are those closest to the time of fault occurrence, whereas samples farther from the fault time are labeled as normal.he real-world ESS data consist of current, voltage, and temperature measurements collected during the operation of an actual ESS power station. For commercial reasons, we do not disclose the exact fault locations or labels; however, the data are confirmed to include faults such as self-discharge and inconsistency. Note that, in the real-world ESS data, a positive current indicates battery discharge.