Occupant-centred thermal comfort space heating control via occupant position detection and multiphysics simulation

• Proposed occupant-centred heating control for spatial temperature variability • Validated 3D Multiphysics model predicting occupant-surrounding temperature • Achieved robust real-time occupant localization using Density Peak Clustering • Reduced energy use by >20% and improved thermal comfort by up to 38% Conventional space heating control methods often rely on single-point or averaged temperature measurements, without consideration of non-uniform temperature distributions within the heated space. This may lead to extra energy consumption with no benefits to thermal comfort. This study proposes an occupant-centred space heating control method that explicitly accounts for spatial temperature variability. The key success of the method is due to the development of a new adaptive multi-target localization method to identify the real-time occupant position, which is based on Density Peak Clustering (DPC) algorithm. Then, a three-dimensional Multiphysics thermal model is developed to predict the spatial indoor temperature distribution providing the temperature information surrounding the occupant. The experimental tests show that the average relative error of the occupant surrounding temperature across multiple test cases is below 1%. The control strategy adopts a Hierarchical State Machine (HSM) structure, which determines the optimal control actions based on the feedback of localized temperature information. The core innovation of this approach lies in utilizing the temperature variation at the occupant’s real-time location as the control input. The proposed control scheme achieves more than 20% reduction in energy consumption while maintaining desired thermal comfort levels. Additionally, the thermal comfort experienced by occupants has improved up to 38%.