Shape optimization of cold-formed steel columns using generalized beam theory and genetic algorithms

Resumo

The search for lighter and more efficient design has put thin walled steel structures on the center of the
attention from civil engineers. Furthermore, with the necessity to fight its slenderness and improve its structural
worth we have a need to find its best optimal cross section. The newest way to achieve such is to wed both
Generalized Beam Theory (GBT) and Genetic Algorithms (GA) for an improved analysis that takes short periods
of processing times. Enhancing structural elements to endure local and distortional failures with longitudinal
stiffeners is not something new to engineers although could be even more efficient for a computer to do it. The
pursuit for the best stiffener geometry and its location based on the knowledge of the local and distortional
responses from multiple elements could save an abundance of resources and time. The goal of this work is to
implement a computational calculation using a genetic algorithm to find the most suitable solution within a limit
range of parameters for an optimal cross section design for channel and zed compressed columns. Genetic
Algorithms are a heuristic search that mimics natural evolution events. Learning by evolving generations populated
with random elements and combinations of the best cross sections the algorithm sets its goal to find the optimal
solutions for distortional and local strengths separately. After its goals are met is its job to try to cross both solutions
to turn into an optimal cross section design. With the local and distortional critical loads, the element could be
analyzed using the Direct Strength Method (DSM) that has been widely used to design Cold-Formed Steel (CFS)
elements. This implementation will be used in the future to accomplish the same on different structural elements
composed of CFS and improve the way these elements are fabricated thus resulting in better and lighter overall
structures.