A Dataset of Real-World Oscillograms from Electric Power Grids

This dataset comprises 50,765 oscillograms recorded from electrical power grids under various operational and fault conditions. Detailed preprocessing—including standardization, anonymization, and normalization—ensures that the data is well-suited for research in fault diagnosis and the development of intelligent relay protection systems. Additionally, the dataset supports machine learning applications by providing labeled examples (both manually annotated and automatically sorted via neural network models) that capture events ranging from normal operation to severe faults