Coupling effects between building and energy system design Variables: An analysis of different building types

Traditional building design and energy system sizing are often treated as decoupled processes, neglecting complex interactions that hinder overall performance improvement. This study proposes a comprehensive framework integrating building performance simulation, energy system optimization, and interpretable machine learning to systematically quantify the coupling effects between building and energy system design variables. Annual dynamic load simulations and multi-objective optimizations were conducted for 1000 design scenarios across six typical building types (commercial, school, residential, hotel, office, and hospital) in the Hot-Summer and Cold-Winter zone. A dual-validation approach using second-order Sobol indices and SHAP interaction values was then applied to identify key coupling mechanisms. The results reveal three distinct load patterns (continuous-high, intermittent-high, and stable-low) determined by building functions, which dictate differentiated energy system configurations. While Ground Source Heat Pumps (GSHP) and battery storage emerged as universal "fixed configurations," other technologies showed high priority variances. Crucially, the interaction between window thermal properties and absorption chiller capacity was identified as the dominant physical constraint. Finally, a hierarchical "Co-optimization Design Workflow" is proposed, guiding practitioners from typology-based baselining to targeted coupling optimization. This study facilitates a paradigm shift from "passive energy saving" to "source-load interaction," providing actionable guidance for low-carbon integrated design.