Automated structural design of shear wall structures based on modified genetic algorithm and prior knowledge

Abstract

Shear wall layout requires repetitive revisions and checks by an experienced designer, which is very time-consuming and inefficient. Therefore, an automated and effective approach to structural design of shear wall structures is needed. An automated generation method of structural model from architectural computer-aided design drawings is described, where the semantic segmentation of buildings based on the connected component analysis, graph structure and Procrustes analysis and an effective parametric modeling approach of shear wall structures integrated with prior knowledge are included. Besides, a modified genetic algorithm with greedy strategy is suggested for shear wall layout optimization. The proposed approach is shown to be effective and reliable through two real-world examples, demonstrating the promising future of automated structural design. This study addressed only the earthquake-resistant design of buildings, and other relevant issues such as beam layout optimization, multi-objective optimization, and economic feasibility of construction should also be addressed in the future.