PatchTrans-Diffusion-Based Defense Approach for Distributed Generation Dispatch and Control Systems Against False Data Injection Attacks

Distributed generation (DG) plays a crucial role in the global energy transition, with its dispatch and control (D&C) systems enabling intelligent management of distributed energy resources. However, these systems are vulnerable to cybersecurity threats such as false data injection attacks (FDIAs), which compromise state estimation and control reliability. To address this challenge, this paper proposes a defense model based on PatchTrans-Diffusion. The model integrates a conditional diffusion process with a patch-based transformer denoising network, unifying the generative capacity of diffusion models with the sequence modeling capability of transformers for accurate power data reconstruction. By exploiting historical measurements as conditional information and partitioning time-series data into patches, it captures the temporal dynamics and spatial correlations of electrical data, enabling the effective generation of substitute data that remain statistically consistent and physically coherent with system dynamics. This design allows the defense mechanism to dynamically replace corrupted measurements after FDIA detection, guaranteeing the the continuity and reliability of state estimation and control. Simulation experiments conducted on two CIGRE-Networks with different DG configurations demonstrate that the proposed method achieves highly consistent data reconstruction and outperforms comparative methods. The results demonstrate that PatchTrans-Diffusion enhances the resilience of DG D&C systems against FDIAs, contributing to more robust and secure DG networks.