Early-Stage Optimization of Building Envelope Components

This research develops optimization models for early stage building envelope design that account for urban and environmental factors. It examines cross-scale impacts of the envelope by linking indoor and outdoor environmental qualities that are influenced by envelope design choices. The method uses machine learning to provide rapid predictions during preliminary design stages, when full envelope assemblies and material definitions are still unknown. By combining simplified envelope descriptions with urban form parameters, and climate and microclimate inputs, the models support rapid testing of alternatives and the identification of robust design directions. The research aims to help designers make better-informed decisions early, improving energy balance, daylight, indoor and outdoor thermal comfort, and carbon-related outcomes. The project aims to lay the foundations for a practical decision-support workflow that connects urban form and envelope design and bridges the gap between early design needs and time-intensive simulation workflows.