A Control Optimization Model for a Double-Skin Facade Based on the Random Forest Algorithm

This study addresses the current difficulties in accurately controlling the indoor temperature of double-skin facades (DSFs), and its optimization, with a focus on the window opening angles of double-skin facades. The Spearman correlation coefficient method was used to select the main meteorological factors, including outdoor temperature, dew point temperature, scattered radiation, direct radiation, and window opening angle. A modified random forest algorithm was used to construct the optimization model and 80% of the data were used for model training. In the experiments, the average accuracy of the optimization model was as high as 93.5% for all window opening angles. This study provides a data-driven method for application to double-skin facades, which can effectively determine and control the window opening angles of double-skin facades to achieve energy saving and emission reduction, reduce indoor temperature, improve comfort, and provide a practical basis for decision-making. Future research will further explore the applicability and accuracy of the model under different climatic conditions.