Power-Aware Smart Grid Integration for Sustainable Buildings Using Deep Reinforcement Learning
Keywords:
Building automation, Deep reinforcement learning, Energy management, Power-aware control, Smart buildings, Sustainable infrastructureAbstract
As urban environments shift towards sustainability, intelligent building systems play a pivotal role in optimizing energy use and reducing dependence on fossil fuels. This paper explores the use of deep reinforcement learning (DRL) to enable power-aware integration of smart buildings with the electrical grid. Specifically, we design and evaluate a DRL-based control framework that leverages deep Q-networks (DQN) to manage energy consumption in buildings equipped with renewable generation systems such as rooftop photovoltaic (PV) panels. The system interacts with a dynamic pricing environment and considers thermal comfort constraints while optimizing energy cost and usage. In our model, the smart building comprises heating, ventilation, and air conditioning (HVAC) systems, controllable electrical loads, and distributed PV sources. We define the building’s environment as a Markov decision process (MDP), where states represent time of day, indoor temperature, energy price, and solar output; actions correspond to discrete adjustments in load and HVAC settings; and rewards are shaped to minimize energy costs while preserving occupant comfort. The DQN agent is trained using a simulated environment based on real-world solar irradiance data and time-of-use (TOU) pricing models. Our experimental results demonstrate that the proposed DRL framework can achieve up to 24.6% reduction in energy costs compared to baseline rule-based systems. Moreover, peak load events are curtailed by 18%, and PV self-consumption increases by 32%, signifying improved grid responsiveness and energy autonomy. The intelligent agent learns to anticipate high-tariff periods and adapt its behavior by pre-cooling or rescheduling non-essential loads. These findings underscore the potential of DRL in enabling smarter, more sustainable buildings that not only respond to real-time grid signals but also enhance operational efficiency and occupant comfort. This interdisciplinary study bridges civil infrastructure design with electrical energy systems and machine learning, paving the way for future developments in demand-side flexibility and energy-aware building automation.
References
International Energy Agency, “Digitalization and Energy—Analysis,” IEA, Nov. 2017. Available: https://www.iea.org/reports/digitalisation-and-energy
F. Oldewurtel et al., “Use of model predictive control and weather forecasts for energy efficient building climate control,” Energy and Buildings, vol. 45, pp. 15–27, Feb. 2012, doi: https://doi.org/10.1016/j.enbuild.2011.09.022
M. Fayiah and M. K. Kallon, “Energy efficiency and smart grid technologies for environmental sustainability,” Environmental Science and Engineering, pp. 439–462, Jun. 2025, doi: https://doi.org/10.1007/978-3-031-85327-2_13
L. Chen et al., “Green building practices to integrate renewable energy in the construction sector: A review,” Environmental Chemistry Letters, vol. 22, pp. 751–784, Dec. 2023, doi: https://doi.org/10.1007/s10311-023-01675-2
G. Serale, M. Fiorentini, A. Capozzoli, D. Bernardini, and A. Bemporad, “Model predictive control (MPC) for enhancing building and HVAC system energy efficiency: Problem formulation, applications and opportunities,” Energies, vol. 11, no. 3, p. 631, Mar. 2018, doi: https://doi.org/10.3390/en11030631
S. Stavrev and Dimitar Ginchev, “Reinforcement learning techniques in optimizing energy systems,” Electronics, vol. 13, no. 8, pp. 1459–1459, Apr. 2024, doi: https://doi.org/10.3390/electronics13081459
Y. Boutahri and A. Tilioua, “Reinforcement learning for HVAC control and energy efficiency in residential buildings with BOPTEST simulations and real-case validation,” Discover Computing, vol. 28, no. 1, Apr. 2025, doi: https://doi.org/10.1007/s10791-025-09544-y
L. Jia and Y. Pei, “Recent advances in multi-agent reinforcement learning for intelligent automation and control of water environment systems,” Machines, vol. 13, no. 6, p. 503, Jun. 2025, doi: https://doi.org/10.3390/machines13060503
M. Mshragi and I. Petri, “Fast machine learning for building management systems,” Artificial Intelligence Review, vol. 58, no. 7, Apr. 2025, doi: https://doi.org/10.1007/s10462-025-11226-6
F. Behrooz, N. Mariun, M. Marhaban, M. Mohd Radzi, and A. Ramli, “Review of control techniques for HVAC systems—Nonlinearity approaches based on fuzzy cognitive maps,” Energies, vol. 11, no. 3, p. 495, Feb. 2018, doi: https://doi.org/10.3390/en11030495
N. Myo Lin, X. Tian, M. Rutten, E. Abraham, J. M. Maestre, and N. van de Giesen, “Multi-objective model predictive control for real-time operation of a Multi-Reservoir System,” Water, vol. 12, no. 7, p. 1898, Jul. 2020, doi: https://doi.org/10.3390/w12071898
A. Mugnini, G. Coccia, F. Polonara, and A. Arteconi, “Performance assessment of data-driven and physical-based models to predict building energy demand in model predictive controls,” Energies, vol. 13, no. 12, p. 3125, Jun. 2020, doi: https://doi.org/10.3390/en13123125
E. Ginzburg-Ganz et al., “Reinforcement learning model-based and model-free paradigms for optimal control problems in power systems: Comprehensive review and future directions,” Energies, vol. 17, no. 21, pp. 5307–5307, Oct. 2024, doi: https://doi.org/10.3390/en17215307
P. Michailidis, I. Michailidis, and E. Kosmatopoulos, “Reinforcement learning for optimizing renewable energy utilization in buildings: A review on applications and innovations,” Energies, vol. 18, no. 7, p. 1724, Mar. 2025, doi: https://doi.org/10.3390/en18071724
J. Terven, “Deep reinforcement learning: A chronological overview and methods,” AI, vol. 6, no. 3, p. 46, Feb. 2025, doi: https://doi.org/10.3390/ai6030046
S. Tabassum et al., “Solar energy in the United States: Development, challenges, and future prospects,” Energies, vol. 14, no. 23, p. 8142, Dec. 2021, doi: https://doi.org/10.3390/en14238142
P. Michailidis, I. Michailidis, and E. Kosmatopoulos, “Reinforcement learning for optimizing renewable energy utilization in buildings: A review on applications and innovations,” Energies, vol. 18, no. 7, p. 1724, Mar. 2025, doi: https://doi.org/10.3390/en18071724
D. Deltetto, D. Coraci, G. Pinto, M. S. Piscitelli, and A. Capozzoli, “Exploring the potentialities of deep reinforcement learning for incentive-based demand response in a cluster of small commercial buildings,” Energies, vol. 14, no. 10, p. 2933, May 2021, doi: https://doi.org/10.3390/en14102933
