Courtyard design impact on indoor thermal comfort and utility costs for residential households: Comparative analysis and deep-learning predictive model

A courtyard is an architectural design element which often known as microclimate modifiers and responsible to increase the indoor occupant comfort in traditional architecture. The aim of this study conduct a parametric evaluation variants residential building different climates with focus on thermal utility costs. brute-force approach applied generate wide range alternatives simulation workflow conducted by Grasshopper together environmental plugins Ladybug Honeybee. main objective air-conditioned building, energy load, cost analysis, derived from variables including geometry, window-to-wall ratio, envelope materials, heating, cooling set-point dead-bands, geographical location. Furthermore, Deep Learning model developed using inputs outputs analysis transform outcomes into algorithmic tangible environment feasible for predictive applications. results suggest that regarding loads, costs, index (PMV), there are high correlations between outdoor weather variation dead-band ranges, while extreme such Singapore, spaces might not be efficient enough expected. Finally, highly accurate deep learning also developed, delivering superior capabilities costs designs.