Beyond energy trading: AI-enhanced Peer-to-peer energy bartering

Abstract

Peer-to-peer energy trading supports efficient utilization of renewable energy but fails to address energy poverty and financial transaction barriers, particularly during economic or infrastructural crises. Consumers may lack funds or means to purchase energy or face difficulties in making transactions for various reasons, while producers may struggle to sell surplus energy through conventional markets. To overcome these challenges, energy bartering offers a promising non-monetary alternative by facilitating direct energy exchange among peers within smart grid environments. This paper proposes AI-P2P-EB, a novel artificial intelligence-enhanced peer-to-peer energy bartering algorithm that improves producer profitability and drives broader participation in decentralized energy systems. Three bartering algorithms and their influencing factors across two scenarios were analyzed. Using two real-world datasets, it is demonstrated that the studied bartering algorithms reduced reliance on the utility grid, saving up to 1,310 megawatt-hours, and bartered between 3.8 megawatt-hours to 1.9 gigawatt-hours. Energy bartering maximizes monetary benefits for peers when the grid does not buy all surplus energy. In contrast, when the grid buys surplus, consumer benefits outweigh producer gains, except in the centralized approach. Our findings categorize energy transactions based on driving incentives and demonstrate the superior performance of the AI-enhanced algorithm in increasing producer profits while maintaining a high amount of bartered energy.