Dual-Discriminator cGAN-Based Defense Against False Data Injection Attacks in Power Systems

With the deepening integration of cyber and physical layers in power systems, the risk of cyberattacks has become increasingly severe. Among these, false data injection attacks (FDIA) can stealthily manipulate measurement data, disrupt state estimation, and pose significant risks to the safe and stable operation of the grid. To address this challenge, this paper proposes an FDIA defense method based on a dual-discriminator conditional generative adversarial network (DD-cGAN). The proposed approach first incorporates attack localization masks to construct a credible measurement dataset, enabling precise reconstruction of the compromised data. Furthermore, a dual-discriminator architecture is designed to establish a global-local collaborative discrimination mechanism, enhancing the accuracy and robustness of the repair process. Finally, simulation experiments on a modified IEEE 14-bus system are conducted, and the results demonstrate the superior performance of the proposed abnormal data recovery model and the effectiveness of the FDIA defense method.