Modeling the energy community members’ willingness to change their behaviour with multi-agent systems: A stochastic approach

• Published in: Renewable energy 2022-07, Vol.194, p.1233-1246
• Main Authors: Simoiu, Mircea Stefan, Fagarasan, Ioana, Ploix, Stéphane, Calofir, Vasile
• Format: Article
• Language: English
• Subjects: Behaviour model; Energy community; Multi-agent simulation; Net energy exchanged with the grid; Recommendation system; Systems with humans
• ISSN: 0960-1481; 1879-0682
• DOI: 10.1016/j.renene.2022.06.004

Collective actions in the context provided by energy communities and more sobriety from energy users could both represent a potential solution for significantly reducing carbon emissions in residential areas. However, research are needed to properly understand, model and simulate the collective behaviour of communities owning a PV plant with energy sharing mechanisms suggested by an European directive. In this context, we propose a multi-agent modeling framework for simulating energy communities that is built upon a stochastic interpretation of the willingness of energy users to modify their consumption. The proposed concept includes an intelligent decision support system that assists community members during their daily activities and provides optimal recommendations to minimise the collective net-energy-exchanged-with-the-grid. The paper includes a case study where we present the impact of different community configurations. We emphasize that a community with just 25% of enthusiastic members provides a significant decrease in net-energy-exchanged-with-the-grid comparing to the reference scenario. •A multi-agent simulation framework is proposed to model an energy community.•Consumer willingness to change consumption is modeled using a stochastic approach.•Personalised recommendations are computed by optimisation to minimise the NEEG.•Community members are distributed into clusters.•Recommendations are developed to coach members to develop a more flexible behavior.