Prediction of the thermodynamic performance evolution of a novel aerogel composite insulation material under service conditions based on data-driven methods

The thermodynamic properties of building insulation materials are influenced by the temperature and humidity of the service environment. Selecting appropriate design parameters of building insulation materials based on actual meteorological changes can effectively reduce energy waste caused by errors in building heating and cooling load calculations. This study proposes a high-precision, high-generalizability prediction model for a novel aerogel composite insulation material(ACIM). Experiments were conducted under different temperature and humidity conditions that reflect the service environment, aiming to investigate the changes in the thermodynamic properties and to construct a comprehensive dataset of ACIM. A data-driven prediction model with two inputs (temperature and humidity) and three outputs (temperature difference, heat flux density, and thermal conductivity) was constructed to determine the relationship between the ACIM’s thermodynamic properties and the environmental temperature and humidity. Finally, the proposed model was validated using experimental data and compared with the performance of interpolation and linear regression methods. The results indicate that the thermal conductivity of ACIM increases under higher environmental temperature and humidity conditions. The proposed prediction model performs well and the deep learning model performs excellently on all three target variables. The proposed method demonstrates good generalization and scalability, and provides insight in the relationship between the thermodynamic properties of ACIM with the service environment conditions.