Federated Learning and LLMs for Secure Peer-to-Peer Energy Trading

Document Type

Article

Source of Publication

IEEE Network

Publication Date

1-1-2026

Abstract

The global transition towards renewable energy demands intelligent, robust, and trustworthy management systems. Blockchain-based Peer-to-Peer (P2P) energy trading has emerged as a promising solution enabling producers to exchange surplus energy with consumers in a distributed manner. Such a solution relies on various smart devices and distributed algorithms to provide various functionalities, in essence, introducing significant cybersecurity challenges. This paper presents an intelligent security framework that integrates Large Language Models (LLMs), Graph Neural Networks (GNNs), and Federated Learning (FL) for blockchain-based P2P energy trading. The framework is able to detect and correlate sophisticated multistage attacks in real-time while maintaining scalability and decentralization. Nodes employ locally fine-tuned LLMs for log parsing, anomaly detection, and semantic event correlation. A global model is then derived through federated aggregation, hence preserving data privacy. GNN then uses the aggregated global model to capture inter-node relationships and identify complex cross-peer attack patterns. This layered design approach strengthens the energy trading system’s ability to proactively detect and mitigate sophisticated cyber threats. The proposed approach lays the foundation for a new generation of intelligent, adaptive, and privacy-preserving security frameworks for distributed energy ecosystems.

ISSN

0890-8044

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Disciplines

Computer Sciences

Keywords

Computer science (0.56) | Computer security (0.46) | Business (0.44) | Energy (signal processing) (0.39) | Industrial organization (0.36) | Environmental economics (0.3) | Commerce (0.29) | Efficient energy use (0.29) | Electronic mail (0.28) | Energy consumption (0.28) | Computer network (0.28) | International trade (0.26) | Telecommunications (0.26) | The Internet (0.26) | Government (linguistics) (0.26) | Transaction processing (0.26) | Internet privacy (0.25) | Energy conservation (0.25)

Scopus ID

105043575582

Indexed in Scopus

yes

Open Access

no

Link to Full Text

Share

COinS