Fig. 4: AI surrogate mode optimization.
From: Reciprocal frame design for large-scale timber construction
a Multi-objective indicator. This study establishes a comprehensive evaluation system for the configuration, incorporating architectural evaluation indicators, such as space occupancy(V) and material usage(M) and structural indicators, including structural displacement (Tf), tensile force (Ft), and compressive force (Fc). b AI surrogate model optimization process. Starting with a hemispherical with a 15 m diameter, the structural performance of the configuration is analyzed under varying design variables (section radii R and frame distances D) based on FEA. Surrogate model analyzes trends between variables and disciplinary objectives and establishes approximate functional relationships, based on AI deep learning. c Multi-objective trend variations with variables (Ry represents sectional radii of components along the y axis, and Dx represents the distance between individual frame units along the x axis). The z-axis represents the weighted normalized objectives, where a smaller value indicates higher efficiency.
