Blockchain and machine learning-based decentralized energy trading platform

Luo, Xiaojun

Abstract

In most buildings, gas and electricity are supplied through centralised energy systems operated by major energy and utility companies. This results in a high burden on central management systems, while energy prices are determined by these companies and affected by the country’s economic status. On the contrary, decentralised multi-energy network among buildings can address UK’s Industrial Strategy Clean Growth Challenge. However, state-of-the-art blockchain-enabled energy trading market lacks predictive control and focuses on electricity only. Therefore, this paper develops an integrated machine learning and blockchain-enabled energy management platform for multiple forms of energy (i.e., heat and electricity) allocation and transmission among different types of buildings. Deep learning models, especially long-short term memory models, is used to predict day-ahead hourly energy generation and consumption patterns of each prosumer and consumer within the multi-energy network. Meanwhile, evolutionary optimisation algorithms, especially particle swarm optimisation, is used to determine optimal operating schedules of each energy devices within the buildings. Moreover, blockchain platform is established to facilitate automated energy allocation among energy users through peer-to-peer energy transactions. This energy-trading platform focuses on energy-matching from both the supply and demand sides, while encouraging direct energy trading between producers and consumers. The security and fairness of energy trading are also enhanced through using smart contracts to strictly execute the energy trading and bill payment rules. In the case study, the proposed platform is implemented on 4 real-life buildings. It is found that the weekly energy cost savings are xxx for the typical winter and xxx for a typical summer, respectively. Meanwhile, greenhouse gas emissions are also reduced by xxx for one year’s implementation. This could be a powerful tool to transform traditional buildings into net-zero ones. It is expected that the proposed platform can involve a larger number of prosumers and consumers within the community to decentralise multiple energy trading, reduce greenhouse gas emissions and enhance environmental sustainability.

• Presentation Conference Type: Presentation / Talk
• Conference Name: 7th Edition of the Green Urbanism (Gu)
• Start Date: Dec 11, 2023
• End Date: Dec 12, 2023
• Deposit Date: Jan 12, 2024
• Public URL: https://uwe-repository.worktribe.com/output/11605082