Self-shape optimisation of cold-formed steel closed profiles using genetic algorithms

For economical benefits, optimisation of mass-produced structural steel products is widely researched. The objective is to minimise the quantity of material used without sacrificing the strength and practicality of the structural members. Current research focuses on optimising the dimensions of conventional cross-sectional shapes but rarely considers discovering new optimum shapes. This report introduces the concepts of a new optimisation method which enables the crosssection to self-shape to an optimum by using the evolution and adaptation benefits of Genetic Algorithm. The feasibility and accuracy of the method are verified by implementing it to find optimum thin-walled profiles against simple parameters for which analytical solutions are known, namely the optimisation of doubly-symmetric closed profiles. Results show that the cross-section accurately self-shapes to its optimum in a low number of generations. Factors influencing the convergence are presented and future challenges to applying the method to optimisation of cold-formed steel profiles with practical applications are discussed.