Experimental and machine learning-based identification of a solar thermal system for domestic hot water and direct solar floor heating

Abstract

This study investigates the energy performance of a combined solar underfloor heating and domestic hot water (DHW) system using an innovative approach that combines experimental data and mathematical modeling. The PieceWise Affine Auto-Regressive eXogenous (PWARX) model was employed to identify discrete operational states and optimize the system’s performance. Three configurations were analyzed under winter conditions: (1) the solar underfloor heating system achieved 130 % energy coverage, maintaining stable temperatures between 17 °C and 19.5 °C; (2) the DHW system with a 300 L storage tank recorded a 71 % coverage, optimizing circulator operation and thermal energy storage; and (3) the combined system demonstrated synergy between the components, balancing energy production with a minimum coverage of 45 %.

The PWARX model identified four distinct operational states, correlating solar radiation with the system’s thermal response, providing insights for energy management and system optimization. The findings underline the potential of the PWARX model to enhance the design and efficiency of solar thermal systems. This study contributes to the energy transition by proposing effective and adaptable solutions for maximizing solar energy utilization in the residential sector.