Detection and Prevention Strategies for False Data in Distributed Load Frequency Control

In modern power systems, distributed load frequency control systems in microgrids are widely used in various energy management applications due to their flexibility and efficiency. However, False Data Injection Attacks (FDIAs) pose a serious threat to the stability and security of microgrids. Therefore, a detection and defense strategy based on Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) and Enhanced Dwarf Mongoose Optimization Algorithm (EDMOA) is proposed to counter FDIAs. First, the CNN-LSTM model is utilized for accurate detection of false data by effectively identifying attack patterns through feature extraction and temporal analysis. Second, the EDMOA algorithm is employed to optimize Proportional-Integral-Derivative (PID) controller parameters, enhancing the system's adaptability to frequency fluctuations and resistance to attacks. Experimental results show that the proposed method significantly outperforms traditional methods in terms of detection accuracy and defense effectiveness, providing strong protection for the secure operation of microgrids.