Optimization and prediction of office building shading devices for energy, daylight, and view consideration using genetic and BO-LGBM algorithms

In modern office buildings, the pursuit of a transparent and aesthetically pleasing facade often results in a high window-to-wall ratio, leading to excessive solar radiation entering the interior. This increases air conditioning energy consumption to ensure indoor comfort. Architectural shading design can block solar radiation, but inappropriate shading design can reduce daylight levels or block view out. Therefore, effectively balancing daylight and view out while reducing building energy consumption remains a pressing issue. In this study, an office building in Jinan, Shandong Province as a baseline model. Using Grasshopper and Python, a data-driven method integrating analysis, optimization, and prediction was constructed to optimize various design parameters of shading devices, such as spacing, width, and rotation angle of shading panels, using a genetic algorithm. The BO-LGBM algorithm was used to establish a classification prediction model for office building performance, verifying the model’s accuracy and exploring the contribution of different shading design parameters to performance prediction. The research results indicate that the optimized shading scheme not only improves visual comfort and ensures sufficient natural light and good view out, but also reduces building energy consumption by 0.63%–2.17%. Key factors for energy prediction include the spacing, rotation angle, and width of south facade panels, and the rotation angle of east facade panels. This method improves the interactive feedback efficiency between shading design decisions and building performance assessment, providing a valuable theoretical reference for designers in the early stages of architectural shading design.