Decarbonization of the built environment by electrifying energy systems and decarbonizing the electrical grid coupled with the digitization of these systems is a central strategy implemented by the European Commission (EC) to meet carbon reduction policies. The proliferation of technologies such as renewable energy sources (RES) and demand-side management (DSM) systems can be improved by using digital twins to predict and optimize their integration with existing systems. Digital twins in the built environment have been used for multiple purposes, such as predicting the performance of a system before its inception or optimizing its operation during use. To this end, a novel application of a combination of these technologies towards optimized DSM is peer-to-peer (P2P) energy trading, which can improve the local use of RES in the built environment. This paper investigates the potential of P2P energy trading in optimizing local RES of a remote island, Inishmore, Republic of Ireland, using a combination of data-driven and predictive digital twins towards the island’s journey to net zero. Data-driven digital twins are used to evaluate the current energy use at the pilot site. Predictive digital twins are applied to estimate the impact of applying P2P in the future and its influence on RES consumption at the pilot site. The findings show that in scenarios with limited RES coverage, P2P can significantly increase the local consumption of excess RES energy, reducing the risk of transmission or curtailment losses. However, P2P is limited in scenarios with widespread RES installation without storage or behavioral change to shift energy loads.

This paper presents an evaluation of the use of deep learning architectures for forecasting electrical energy consumption in residential environments. The main contribution of this study lies in the development and assessment of a hybrid forecasting framework that integrates multiscale temporal analysis and weather data, enabling evaluation of predictive performance across different temporal granularities, forecast horizons, and aggregation levels. Single and hybrid models were compared, trained with high-resolution data from a single residence, both considering only endogenous variables and including exogenous variables (weather data). The results showed that, among all models tested in this study, the hybrid LSTM + GRU model achieved the highest predictive performance, with R² values of 94.62% using energy data and 95.25% when weather variables were included. Intermediary granularities, particularly the 6 steps, offered the best balance between temporal detail and predictive robustness for the tests performed. Furthermore, short-time windows aggregation (1 to 5 min) showed better accuracy, while the inclusion of weather data in scenarios with larger aggregation windows and longer horizons provided additional gains. The results reinforce the potential of hybrid deep learning models as effective tools for forecasting residential electricity consumption, with possible practical applications in energy management, automation, and integration of distributed energy resources.