Blockchain-enabled secure communication for Vehicle-to-Grid networks using deep learning anomaly detection framework

The convergence of Electric Vehicles (EVs) and the smart grid through Vehicle-to-Grid (V2G) networks introduces critical cybersecurity vulnerabilities. Traditional security mechanisms present single points of failure and are inadequate for the dynamic and high-volume data environment of V2G ecosystems. This paper proposes an integrated security framework that synergizes blockchain technology with a deep learning-based anomaly detection system to strengthen V2G communication. A permissioned blockchain architecture establishes a decentralized and immutable ledger for secure messaging between EVs. A deep Long Short-Term Memory (LSTM) autoencoder model is deployed to perform real-time multivariate time-series analysis on grid command data and energy flow metrics. The proposed framework was evaluated using a synthetic V2G network dataset. EV Electrical Vehicles Dataset (3k records, 2025)” from Kaggle, which was augmented by the authors to generate V2G transaction fields, digital signatures, grid frequency deviations, and three attack types (False Data Injection, replay, and coordinated load spikes). The study relies entirely on synthetic data and simulated communications; real-world validation is a necessary next step (see Sections 6.8 and Appendix E). In this proof-of-concept simulation, the proposed framework achieved 91.48% resilience to attacks, 97.1% detection F1-score, 77.73 kbps communication overhead, 92.48% detection accuracy, and 134.26 ms transaction latency. These results are obtained from a proof-of-concept simulation and should be interpreted as promising indications; real-world validation on physical hardware and live grid data is required before any operational claims can be made.