A two-level optimization framework for new family of CFS sections

In this study, a two-level optimization framework for cold-formed steel (CFS) is proposed. A concise family with a minimum number of CFS sections is obtained to substitute the original 48 commonly used commercial sections, while satisfying the flexural capacity and ductility requirements. Firstly, the numerical models of the commercial sections are established according to the real dimensions, and the flexural capacity and rotation ductility factor can be obtained based on the equivalent energy elastic-plastic (EEEP) bi-linear model. Then, a first-level opti-mization scheme based on back-propagation neural network-genetic algorithm (BP-GA) is proposed, and the optimal dimensions of each section that can satisfy the optimization objectives of flexural capacity and ductility are determined. To obtain a concise optimized family, a second-level optimization scheme is further proposed. When the optimal fitness equals or exceeds the baseline fitness of the original 48 commercial sections, an optimal family is determined. The optimized family provides the same flexural capacity and ductility in a minimum number as covered by the original 48 commercial sections, which will significantly improve the standardization and efficiency of CFS production.