Periodic Reporting for period 1 - OPENAI (OPTIMISATION OF ENERGY MARKET BY ARTIFICIAL INTELLIGENCE)
Reporting period: 2020-11-01 to 2022-01-31
The innovation idea of the project was to create an AI-based solution for energy efficiency and energy management services from the producer and consumer’s point of view that could contribute towards the improvement of decision-making processes in smart grids and ultimately result in the reduction of environmental impacts. We wanted to extend the current prediction module of our smart home solution system by enabling the use of multi-source information fusion-based methodologies for energy consumption prediction, as well as to examine the possibilities of construction of distributed and digitalized electricity markets. The specific focus was given on the further integration of the developed prediction module in the energy marketplaces empowered by blockchain technologies aiming to optimize energy trading activities. The innovation ideas intended to add value to the society by proposing a concept of a decentralized energy trading marketplace that could add more transparency to the energy markets, improve overall operating costs and energy prices, and reduce the ecological impact. For the project implementation, three chief objectives have been examined: 1) Development of a study concerning several AI-based techniques for the analysis and forecasting purposes of the energy-related data; 2) Designing a marketplace capable of joining and smart managing energy supply and demand based on blockchain technologies; 3) Preparation of a detailed market study for Spanish electricity markets and investigation of the business potential for the newly examined research line.
The OPENAI was a one-year project focusing on the improvements of smart grid functionalities and energy efficiency services, as well as contributing to the development of new generation energy markets. The project has started in February 2021 fully conforming with the plan specified in Annex 1 of the Grant Agreement. By the end of the project timeline, all major deliverables and milestones are completed in line with the Description of Action (DoA), confirming that the project has achieved its aims for the reporting period. With respect to the exploitable results, the work done in 12 months were divided into two areas: 1) The first and the main area of the research for the OPENAI project was the artificial intelligence (AI) applications in the energy sector. In this regard, the project aimed to develop an innovative methodology applicable for the prediction of energy consumption and demand. The core idea was to use AI and other statistical models for the analysis and optimization of energy production and distribution systems and the improvement of decision-making processes in smart grids. 2) The second area of the research included the research and designing of a fully digitalized energy marketplace. The key idea was to investigate the capabilities of newly developed blockchain technologies together with AI techniques in the development of distributed energy marketplaces enabling to conduct of energy exchange and trading activities without third-party intermediaries. By the end of the one-year project, the research study has proposed a novel multi-stage methodology for building a prediction system for electricity consumption combining an XGBoost feature selection algorithm, discrete wavelet transform-based (DWT) denoising scheme, and a statistical forecasting tool called SARIMAX. In addition, a conceptual design of a fully decentralized energy trading platform integrating blockchain technologies and P2P trading mechanisms has been presented.
The main purpose of the OPENAI project was the optimization of legacy energy systems and the research on the development of energy trading marketplaces. The developed approach which was successfully tested through the experiments using real-world data is based on the idea of analyzing the correlation between energy consumption with different external factors such as weather data, socio-economic indicators, sensors data for generating forecasts for future energy consumption. The proposed prediction module has shown higher accuracy results compared to the other methodologies from the state-of-the-art based on the experiments conducted on real-world consumption data. The obtained results allow to further adopt the proposed prediction module into the energy efficiency services for the optimization of energy demand and ultimately reduce the environmental impact related to unnecessary energy waste. The part of the research aimed to develop a study concerning blockchain technologies and their possible use in the development of decentralized energy markets. The research outcomes regarding to the development of blockchain-based energy marketplaces can be further implemented as a working prototype of a new generation energy marketplace which could promote the adoption and sustainable use of all forms of renewable energy sources and significantly minimize the electricity costs.