Demand response in buildings: Unlocking energy flexibility through district-level electro-thermal simulation

Amin Amin (1) , Oudom Kem (2) , Pablo Gallegos (3) , Philipp Chervet (4) , Feirouz Ksontini (2) , Monjur Mourshed (1)

1 Cardiff School of Engineering
2 LI3A (CEA, LIST) - Intelligence Artificielle et Apprentissage Automatique (CEA, LIST)
3 ENGIE
4 CSEM - Centre Suisse d'Electronique et de Microtechnique SA [Neuchatel]

European Union households account for 26% of the final energy consumption, yet their participation in demand response (DR) schemes is virtually non-existent. Relatively small amounts of flexibility primarily from appliances and the lack of DR-enabled energy management solutions in homes have been identified as the main barriers. As part of a Horizon 2020 funded collaborative research and development project, we developed a cloud-based framework for intelligent optimisation of electricity consumption and generation in buildings to upscale and enhance their demand response effectiveness at a district level. Through electro-thermal simulations that unlock building flexibility while considering low-voltage grid constraints, our approach aims to extend technical and economic potential of the whole solution for wider replication at a low computational cost. The framework simultaneously schedules the operation of controllable appliances such as dishwashers, washing machines and dryers in buildings to further reduce peak loads and energy costs of the district. The framework was tested in a real UK district comprising 66 dwellings, which demonstrated its effectiveness in minimising electricity peak loads and costs by prioritising local consumption of the electricity produced from photovoltaics and exploiting the asymmetry on the time-varying electricity tariff. The results show the ability to achieve an average 30% reduction in energy costs for the district while adhering to low-voltage grid constraints.

Mots clés: Demand response; Energy prediction; Energy optimisation; District simulation; Smart grid; Artificial intelligence; machine learning; Distributed learning; multi-agent system; intelligent optimisation; electricity consumption; electricity generation; buildings; simulation; modelling; electro-thermal simulation; building flexibility; low-voltage grid constraints; controllable appliance; demonstration; local electricity consumption

Domaines: Système multi-agents [cs.MA]; Intelligence artificielle [cs.AI]; Energie électrique

Amin Amin, Oudom Kem, Pablo Gallegos, Philipp Chervet, Feirouz Ksontini, et al.. Demand response in buildings: Unlocking energy flexibility through district-level electro-thermal simulation. Applied Energy, 2022, 305, pp.117836. ⟨10.1016/j.apenergy.2021.117836⟩. ⟨hal-04573559⟩