Automated optimization for steel mold of precast components based on multi-objective evolutionary algorithm.

Designing an economical and reliable steel mold for precast concrete (PC) components can accelerate the promotion of prefabricated buildings. This paper presents an automated design framework for a steel mold of PC components to generate manufacturing information, including dimensions and location of each plate. The profile and dimensions of PC components are extracted from the shop drawings, and a finite element model of the steel mold is established and initialized. A multi-objective optimization model considering material usage and manufacturing convenience is developed to determine the geometry and quantity of plates in steel molds. The non-dominated sorting genetic algorithm II (NSGA-II) is applied to obtain a Pareto solution set for steel mold design. The production information in DXF format can be automatically generated by parsing Industry Foundation Classes (IFC) model. The cost-effectivity and practicality of the proposed framework were verified via the design case for a real PC stair mold. • An automated design framework is developed for steel mold of precast concrete components. • The mold design is modeled as multi-objective optimization model considering material and manufacturing cost. • Design parameters are extracted from shop drawings to automatically establish a parametric finite element geometric model. • The NSGA-II method performs better than MOEA/D and SPEA2 algorithms in terms of solution diversity and computing efficiency. [ABSTRACT FROM AUTHOR]