Experimental and Numerical Analysis of Cold Formed Steel Columns and Beams

Abstract

Cold-formed Steel (CFS) buildings are gaining popularity in the construction industry due to seismic performance and rapid construction process. CFS buildings provide a significant advantage from a sustainability perspective because they can be disassembled and reused and are 100% recyclable. Production and partial assembly of CFS panels in the factory and the rest of the assembly part in the construction site reduce the structures' construction time and environmental impact. Despite the widespread use of light steel structures, the design of the CFS structures is still a current research topic, and it also contains many design and analysis questions that need to be answered. In connection with this, research on the CFS has increased to understand and improve CFS properties. Because of the thin member thickness of the CFS members, they are oversensitive to geometric imperfections compared with concrete and conventional steel members. Therefore, it is crucial to understand the effects of the geometric imperfections on axial and bending capacities in consideration of geometric distribution and magnitude of geometric imperfections. This thesis aims to understand the effects of geometric imperfection on the CFS columns and beams. Firstly, geometric imperfection distribution and magnitude of the test specimens were extracted from textured-based point clouds obtained with the help of the 3D optic scanners. In previous studies, measurements of geometric imperfections have been generally made from a limited number of selected points on the element. Therefore, the collected geometric imperfection data are not sufficient to represent actual three-dimensional geometry. Therefore, geometric imperfection data extracted from texture-mapped point clouds is also more reliable as it is representative of the actual geometry. Secondly, the CFS test specimens' behavior was investigated using the axial loading and four-point bending tests. The obtained test results were compared with the numerical models and design code predictions. Geometric imperfection magnitudes were integrated into the numerical models using the mode shapes obtained from linear buckling analysis. The obtained results showed that the numerical models predicted the observed behaviors of the test specimens well. However, our results also showed that the numerical model geometry should match the actual geometry to obtain reliable numerical models.