Software Architecture for Residential Prosumer Agents in a Transactive Energy System

Transactive Energy (TE) systems decentralize the grid management by allowing different kinds of agents to agree on power demand and selling prices. Accordingly, to succeed in the management and guarantee the dynamic balance between consumption and supply, it is essential to count on reliable agents that can fulfill their agreements. Currently, generation companies and large consumers have experience as decision-makers in wholesale electricity markets, and this condition facilitates their participation in TE systems. In contrast, the residential sector has mostly had a passive role in electricity markets. As a consequence, the expansion of TE systems to distribution grids requires not only the implementation of communication channels and market rules but also the development of tools that engage small consumers and producers. Facilitating and automating the participation in transactions will build trust in the TE systems and accelerate their deployment. In this regard, this paper presents an agent architecture based on a beliefs-desires-intention approach to automate residential prosumers’ decision process. The architecture is composed of six behaviors that consider the interaction with internal and external household environments, appliances modeling, and consumption planning. This architecture was illustrated in simulated transactions employing a heating system as a controllable load. The results suggest how data-driven models of aggregated demand signals could be useful to participate appropriately in a forward transactive energy market.