Smart Contract-Based Decentralized Energy Trading Platform with AI-Enhanced Decision Support

Abstract

The rapid digitalization of the energy sector and the growth of distributed energy resources have exposed the limitations of traditional centralized energy management and trading models. This shift has created a need for more flexible, transparent, and user-focused solutions. Blockchain technology addresses these needs by enabling secure, traceable, and direct transactions through a decentralized and immutable record system. Peer-to-peer energy trading platforms on the public Ethereum network, for example, allow producers and consumers to exchange energy securely without intermediaries. This study presents a blockchain-based system architecture for peer-to-peer energy distribution and trading, known as the Decentralized Energy Management System (DEMS). The system is built on a permissioned Ethereum blockchain (PEDNET) using the Istanbul Byzantine Fault Tolerance (IBFT 2.0) consensus mechanism, and automates energy exchanges and payments using smart contracts, which enable secure, auditable, and traceable transactions through the use of energy tokens. An artificial intelligence-powered decision support module comprising three specialized neural network models has also been integrated to optimize users' energy purchasing preferences, achieving approximately 90% recommendation quality. The system has been validated through comprehensive testing with 500 simulated users over a 3-month period, demonstrating a 32% reduction in average transaction time and an 18% increase in user satisfaction compared to non-AI baselines. Performance benchmarking shows sub-2-second transaction finality and throughput exceeding 500 TPS on the PEDNET network. The study also addresses security considerations, regulatory compliance requirements, and provides a detailed cost analysis of smart contract operations. The study demonstrates the practical impact of combining blockchain and artificial intelligence technologies in P2P energy systems.