Integrating machine learning and parametric design for energy-efficient building cladding systems in arid climates: Sport hall in Kerman

Climate change, driven by fossil fuel dependence, presents a significant challenge for the construction industry, particularly in energy-intensive regions like arid climates. This research investigates the potential of integrating machine learning and parametric optimization to enhance the energy efficiency of spatial structure domes in such environments. Focusing on a sports pavilion in Kerman, Iran, the study examines the crucial role of cladding systems in building energy performance. Employing a rigorous four-phase methodology, the research optimizes dome cladding materials for hot, dry climates using a dual objective function: energy cost and material cost. The process involves a comprehensive literature review, data-driven material selection, advanced energy simulations, and optimization analysis. Parametric modeling tools (Rhino, Grasshopper, Honeybee) facilitate the comparative analysis of various cladding systems. Multivariate Polynomial Regression (MPR) enables predictive modeling of energy consumption and material costs, streamlining the design process for architects. The optimized solution is a hybrid cladding model composed of 10 % polycarbonate and 90 % aluminum. Analysis reveals that the hybrid system offers superior energy optimization compared to pure aluminum (4.58 %) and polycarbonate (5.70 %). While polycarbonate has a lower initial material cost, the hybrid system achieves a balance between material expenditure and long-term energy efficiency. This highlights the importance of considering life-cycle costs when evaluating building envelope materials. This research advances a framework that leverages machine learning and parametric design for building envelope optimization. This framework empowers architects and engineers to create energy-efficient structures within arid environments, promoting a more sustainable built environment.