OICC Press Signal Processing and Renewable Energy (SPRE) 2588-7335 9 3 2025 09 30 A Hybrid Defense Framework for Multi-Area AGC Systems Integrating RBPF-UKF, Noise-Adaptive DRL, and Lightweight Blockchain with Post-Quantum Cryptography 10.57647/j.spre.2025.0903.17 EN Navid Rashtian Department of Electrical Engineering, Islamic Azad University, CT.C. (Central Tehran Branch), Tehran, Iran https://orcid.org/0000-0001-8609-0649 Alireza Yari ICT Research Institute, Tehran, Iran Nahid Ardalani https://orcid.org/0009-0008-6884-7888 Payam Rabbanifar Department of Electrical Engineering, Islamic Azad University, CT.C. (Central Tehran Branch), Tehran, Iran https://orcid.org/0000-0003-0820-2320 Seyed Javad Mirabedini Journal Article 2025 09 30  Modern smart grids are facing increasing cyber threats, including False Data Injection (FDI), delays, and combined attacks, which can destabilize multi-area Automatic Generation Control (AGC) systems and lead to cascading blackouts. This study proposes a hybrid defense framework that integrates three advanced components: a Rao-Blackwellized Particle Filter with Unscented Kalman Filter (RBPF-UKF) for precise nonlinear state estimation under non-Gaussian noise, reducing frequency deviation to below 15 MHz within 2.5–4 seconds; a noise-adaptive Deep Reinforcement Learning (DRL) module utilizing a lightweight two-layer MLP, achieving 99% accuracy in detecting FDI, 97% for delays, and 98.5% for combined attacks; and a lightweight Practical Byzantine Fault Tolerance (PBFT) blockchain secured with CRYSTALS-Kyber/Dilithium post-quantum cryptography, ensuring data integrity with a 0.3-second consensus latency. Tested on a five-area IEEE 39-bus system, the framework surpasses traditional methods like LSTMs, DDPG, and Hyperledger Fabric, reducing estimation errors by 60% and meeting standards such as IEC 62443-4-2 and NIST SP 800-208. Its modular design allows seamless integration with legacy SCADA systems. Additionally, FPGA-based acceleration and energy-efficient variants of McEliece cryptography enhance scalability and sustainability in IoT and distributed grid environments. While effective under extreme conditions (0.5–4 seconds latency, 20% packet loss), scaling to ultra-large networks with over 100 nodes remains challenging, which encourages future research into alternative consensus mechanisms, such as DPoS and ABFT.