Deep-Learning and Blockchain-Empowered Secure Data Sharing for Smart Grid Infrastructure

Abstract

The Smart Grid represents a modernized electrical infrastructure integrating information and communication technology for bidirectional data exchange between power providers and consumers. This advancement enables seamless digital connectivity among intelligent devices such as Smart Meters, Demand Response Control Units, and Service Providers, facilitating remote data management for optimized energy distribution. However, the reliance on unsecured wireless communication channels poses significant security vulnerabilities, including replay, impersonation, man-in-the-middle, and physical capture attacks. To address these challenges, this study introduces a pioneering approach called Deep-Learning and Blockchain-enabled Secure Data Sharing. Specifically, Deep-Learning techniques are leveraged to develop an effective Intrusion Detection System. The proposed RENS (intRusion detEction aNd clasSification) combines Variational AutoEncoder with Attention-based Bidirectional Long Short-Term Memory for feature extraction and attack detection. Moreover, normal instances identified by RENS are utilized in a blockchain-based access control mechanism, ensuring secure and immutable data exchange among Smart Grid entities. In this framework, participating Service Providers form a peer-to-peer network responsible for generating blocks associated with individual SMs. These blocks undergo validation and are appended to a private blockchain ledger using a smart contract-based Proof-of-Authentication consensus mechanism. Experimental results and security analysis demonstrate the superiority of the DBSDS framework over conventional BiLSTM techniques, confirming its effectiveness in safeguarding Smart Grid operations.