A simple and efficient machine-learning based approach for optimal heating control of radiant floor heating systems: Proposal and validation

Radiant Floor Heating (RFH) systems are increasingly favored for their comfort, efficiency, and energy-saving features. However, their large thermal inertia and dynamic delay pose challenges in maintaining thermal comfort and energy efficiency, particularly in intermittently used buildings. This study proposes an optimized start-stop control strategy for RFH systems utilizing the Decision Tree (DT) algorithm based on indoor and outdoor temperatures. The proposed method is highly adaptable, computationally efficient, and straightforward to implement, making it particularly well-suited for application in existing buildings. Data were collected over two months from a building in Ansan, Korea, and analyzed using correlation coefficients to develop a time prediction model for system start-stop control. Validation was performed using a combined TRNSYS and Python simulation platform. The results demonstrate the proposed model's adaptability and efficacy, providing a simple and feasible alternative to conventional control strategies. The proposed model effectively reduces operational energy consumption by 29.67 % and achieves a 26.37 % thermal comfort increase compared to manually regulated systems.