Labeled Cyber\u2013Physical Microgrid Dataset for Secondary-Control Cyberattack Detection

This dataset contains detailed and labeled data created from a MATLAB\/Simulink simulation of an inverter-based microgrid. The microgrid operates under both normal conditions and cyberattacks on the secondary control system. The system includes several droop-controlled distributed generators that are coordinated using a consensus-based secondary control method.Six types of cyberattacks are added to the secondary control layer: Additive Bias, Ramp, Slow Ramp, Sinusoidal, Coordinated Stealth, and Denial-of-Service, along with normal operation. The dataset records time-aligned electrical measurements such as voltage, current, active power, reactive power, and frequency deviation for each generator.Each data sample is labeled for both binary classification (normal or attack) and multiclass classification (specific attack type). The dataset is designed to support research on microgrid cybersecurity, cyber-physical system resilience, and machine-learning-based attack detection under realistic noise and communication disturbances.