Enhancing Building Energy Efficiency and Comfort with AI-Driven Predictive Control for HVAC Systems
摘要
This paper investigates the application of advanced process control (APC) techniques to improve HVAC system management. The implemented control strategies include optimizing thermo-ventilation units by modulating supply and return airflows based on detected CO₂ levels, ensuring both energy efficiency and indoor air quality. Additionally, a dynamic control algorithm regulates the operation of polyvalent heat pumps, switching them on and off based on real-time heating and cooling demands. Lastly, the system optimizes circulation pumps within the heat pump and secondary circuit, activating them only when necessary to maintain optimal indoor conditions while minimizing energy consumption. These control strategies are not only reactive but also predictive considering not only factors like monitored indoor temperatures, CO2 levels and predefined comfort boundaries, but also forecasted occupancy to optimize the system performance further. The proposed APC system was deployed in a retrofitted office building, serving as a case study for its effectiveness. The results demonstrate a monthly reduction in electrical energy consumption ranging from 10.3% and 60.2% calculated from March ‘24 to December ‘24 compared to the same months in 2023. These savings highlight the potential of advanced control strategies to improve energy efficiency and reduce operational costs in HVAC systems.