Zonal Control of Radiant Cooling and Heating Ceiling to Enhance Comfort and Energy Efficiency: A Case Study in Palermo

Building operations still generate a relevant share of global energy-related emissions. Contemporaneously, the digitalization wave is providing new solutions for increasing efficiency in space heating and cooling. One of the most promising approaches for reducing energy use in climate control within buildings involves the development of personal and zonal environmental control systems. Additionally, the integration of artificial intelligence, particularly machine learning algorithms, is becoming increasingly essential in optimizing energy management and environmental comfort parameters. This study introduces an innovative methodology for optimizing HVAC control in partially occupied rooms, integrating thermal comfort estimation, machine learning algorithms, simulated data, and real-time measurements. The results of the dynamic simulation presented are used to assess the effectiveness of implementing a real HVAC system in Palermo, Italy. The model includes a radiant ceiling used for both heating and cooling, which can be controlled by sub-zones. By monitoring indoor conditions, machine learning algorithms will be employed to improve local thermal comfort estimations through limited measurements and investigate their relationship with energy consumption. The direct impact of the radiant ceiling on the mean radiant temperature of the room will be explored through the novel presented framework that combines physics-based approaches with data-driven methods. The scope is to develop a zonal control system based on conventional real-time measurements coupled with learning techniques capable of considering user preferences and local parameters. The results of the simulations demonstrated how it is possible to reduce heating consumption by up to 19% and ensure comfort for users by controlling zones.