Edge-Fog Computing-Based Blockchain for Networked Microgrid Frequency Support

Microgrids have gained increasing adoption in recent years due to the growing demand for renewable energy integration. Frequency support has become a critical issue in networked microgrid operation, primarily due to the intermittent nature of renewable energy sources and the inherently low system inertia of microgrids. Sudden system events, such as faults that lead to generator disconnections, can result in significant frequency drops. This paper proposes a novel method for providing frequency support during generator disconnection events in networked microgrids. The proposed approach integrates an edge-fog hierarchical blockchain architecture with a Long Short-Term Memory Model-Free Predictive Controller (LSTM-MFPC). The parameters and hyperparameters of the LSTM-MFPC are optimized using the Bayesian Adaptive Direct Search (BADS) algorithm. The root mean square error (RMSE) of the current obtained using the traditional model predictive control (MPC) and the proposed LSTM-MFPC applied to the inverter are 0.1970 and 0.1432, respectively. These results were obtained through experiments conducted on MATLAB/Simulink-based power hardware-in-the-loop (PHIL) platforms. This demonstrates that the proposed method significantly improves the frequency response of networked microgrids following fault events and exhibits strong potential for real-world implementation.