Consumers' sensitivities and preferences modelling and integration in a decentralised two levels energy supervisor

摘要

To address the new challenges arising from the higher penetration of renewable energy in electrical grid, Demand Response (DR) aims to involve the residential consumers in the grid equilibrium. Ensuring benefits for both utility and users requires the consumers sensitivities to be understood and then included in the Energy Management System (EMS). For this purpose, the cost is the predominant and most often only factor taken into account in the literature, although in the residential sector other concerns influencing electricity consumption behaviour have been observed. This paper presents a two levels EMS applied to a neighbourhood of consumers mathematically modelled at the level of their appliances and incorporating 5 consumers profiles along three sensitivities: cost, environment and appliances shifting comfort. The first level is a day ahead supervision based on a multi-agent optimisation lead by a central aggregator but performed locally by the household using Dynamic Programming (DP), thus ensuring privacy protection for the stakeholders. The second level is a real time supervision using the same decentralised structure and based on fuzzy logic. Both levels are evaluated in this paper, with a focus on the balance between grid and consumers objectives.