---

Energy Management Model for HVAC Control Supported by Reinforcement Learning

Heating, Ventilating, and Air Conditioning (HVAC) units are essential for providing a comfortable and healthy indoor environment for office workers. However, they also have a high consumption profile and contribute to greenhouse gas emissions. In this paper, it is proposed a smart and cost-effective solution for managing HVAC units in office buildings using data-driven methods. the proposed solution consists of three steps: (i) collecting data from an open-source building energy management system, (ii) developing a learning and predictive model that can forecast if users will be present in a certain location, and (iii) designing a decision model that can control the HVAC units based on the user prediction, the current weather conditions, and the current energy prices. Our results show that our predictive model achieved a high accuracy of 93.8% and that our decision model maintained the comfort level of users.

Publication of the scientific paper "Energy Management Model for HVAC Control Supported by Reinforcement Learning" in Energies journal (IF: 3.252). P. Macieira, L. Gomes, Z. Vale, “Energy Management Model for HVAC Control Supported by Reinforcement Learning”, Energies, 2021, 14, 8210. DOI: 10.3390/en14248210

Open publication