Optimizing classroom modularity and combinations to enhance daylighting performance and outdoor platform through ANN...

The global COVID-19 pandemic has increased attention to the relationship between the built environment and health, particularly in educational settings where students spend a significant amount of their time. Traditional side daylighting used in schools, while cost-effective and easy to construct, can result in uneven indoor daylighting. To address this issue, this paper proposes a terraced teaching building design model for primary and secondary schools in Guangzhou based on the design experience of an "open-air school movement" during a historical respiratory epidemic in the early 20th century. The proposed design relies on skylight for lighting, and each classroom has an outdoor platform. An optimization algorithm based on Spatial Daylight Autonomy (sDA), Uniformity of Daylighting (UOD), Annual Sunlight Exposure (ASE), Outdoor Platform Area (OPA), Gable Wall Length (GWL), and Space Utilization (SU) is used to obtain the optimal concrete form of the building. To speed up the simulation process, a set of Artificial Neural Network (ANN) based rapid prediction network models for complex forms is proposed. This group prediction method improves the simulation speed by 357 times and grossly speed up the optimization process based on six indexes in the early design stage, resulting in four terraced teaching buildings that meet the above criteria. Overall, the proposed design provides a novel architectural form that ensures overall visual comfort while promoting students' learning and physical health.