量化光伏双层幕墙的参数交互：基于二阶Morris方法的敏感性分析

Abstract The numerous parametric analyses of photovoltaic double skin façades (PV-DSF) have provided valuable insights for its practical design, whereas inconsistent optimal values have been reported, typically attributed to variations in climate conditions. Other studies have also identified the coupling effect between parameters, which introduces potential errors to optimization results. To address such discrepancies in parametric impacts, this paper applies the Morris method to perform both first-order and second-order sensitivity analyses against three categories with ten parameters. A trajectory group with maximum Euclidean distance is adopted to achieve efficient analysis. The first-order sensitivity analysis reveals that solar energy richness, indoor illuminance setting value ( L set ), and the etching ratio of semi-transparent photovoltaic modules ( ϑ STPV ) exert the strongest individual impacts, with the complexity of ϑ STPV and cavity depth ( D cav ) remarkably outperforms others. The second-order sensitivity analysis further shows that the most significant coupling effect occurs between L set and other daylighting-related parameters, followed by interactions between key PV-DSF design parameters (such as ϑ STPV and D cav ) and outdoor air temperature . Consequently, it is recommended that researchers avoid relying on single-parameter optimization results without verification, and instead apply intelligent optimization methods and machine learning technology for multi-parameter analysis, considering local climate conditions and illuminance requirements.