A 14-day Ground-Truth Dataset for FDI Attack Detection in a Smart Microgrid Cyber-Physical Testbed

A physically consistent, 14-day, 1-second-resolution ground-truth dataset of a 0.4 kV radial smart microgrid, generated with pandapower. Each record contains the true and PMU-based WLS-estimated system state, measurement residuals, and a continuous chi-squared bad-data-detection statistic. The dataset is the foundation for synthesizing and detecting stealthy False Data Injection (FDI) attacks. Generation and validation code: https://github.com/EhsanEtesami/microgrid-fdi-testbed