Deep Reinforcement Learning based Smart Inverters in Microgrids with Electric Vehicle Charging Systems Considering False Data Injection Attacks Detection
Chou-Mo Yang, Pei-Min Huang, Chun‐Lien Su, Mahmoud Elsisi · 2025
The integration of electric vehicle (EV) charging systems into microgrids introduces complexities in energy management and vulnerability to cyber-attacks, such as false data injection attacks (FDIAs). Smart inverters play a critical role in regulating power flow and voltage stability but are susceptible to FDIAs that compromise grid reliability. This paper proposes a deep reinforcement learning (DRL)-based framework to enhance smart inverter operations in microgrids with EV charging while detecting FDIAs. The DRL agent optimizes real-time power dispatch and voltage regulation, simultaneously learning to identify anomalous data patterns indicative of attacks. By integrating detection mechanisms within the control strategy, the framework ensures resilient operation against malicious data manipulation. Results demonstrate the approach's effectiveness in maintaining operational efficiency and accurately detecting FDIAs. The proposed solution offers a robust, adaptive strategy for securing microgrids with high EV penetration against evolving cyber threats, ensuring both stability and security in decentralized energy systems.