RECsim—Virtual Testbed for Control Strategies Implementation in Renewable Energy Communities

In recent years, the building sector has been pointed out as critical for the energy transition, and the Energy Communities have been introduced to allow the aggregation of multiple buildings to jointly manage their energy generation and consumption. At cluster level, buildings can provide flexibility services to the energy grid by coordinating the energy consumption in order to match local energy generation. The aim of this work is to present a virtual testbed for the implementation of control strategies in Energy Communities and the assessment of their performance under different climatic conditions, generation systems, penetration of renewable energy sources or availability of energy storage. The virtual environment embeds models of the thermal building dynamics and of various systems and it is wrapped into an OpenAI gym to easily allow the imple-mentation of advanced control algorithms such as reinforcement learning.