MEASUREMENT AND APPLICATION OF GEOMETRIC IMPERFECTIONS IN COLD-FORMED STEEL MEMBERS
MetadataShow full item record
Cold-formed steel members provide a unique solution to framing walls and floors of buildings and are popular in low and mid-rise construction in North America. Members are formed from coils of thin sheet steel and cold bent into useful shapes. Advantages of cold-formed steel construction include ease of transportation and erection, thanks to its light-weight properties. This same light-weight property; however, creates a thin-walled member and such members are vulnerable to geometric imperfections. Strength and stiffness of thin-walled structural members can be impaired due to imperfections, which has inspired studies on imperfection sensitivity of cold-formed steel, including: measurement techniques, imperfection characterization, and numerical simulations. In this study, an innovative imperfection measurement rig employing a laser triangulation technique is used to scan along targeted cold-formed steel members. The scan results in abundant measurement point clouds and these allow further exploration into the impact of geometric imperfections, especially for cross-section imperfections, which are constrained by conventional measurement techniques. In addition, the point clouds enable additional applications such as dimensional characterization and an ability to study the impact of different imperfection characterization approaches as compared to actual measured imperfections. In this dissertation, a newly developed imperfection measurement rig is carefully detailed and the results illustrated with examples. Collected data from the laser-based imperfection measurement rig requires specific post-processing to achieve useful and reliable geometric information. The post-processing procedures generally comprise data trimming, surface registration, and feature recognition algorithms. Modifications to the surface registration algorithm, i.e. iterative closest point), are described along with development of the feature recognition algorithm. These processing effects are demonstrated with three different section shapes of cold-formed steel members, i.e. C, Z, and Built-up C studs. Dimensional variations, which are considered as primary geometric imperfections, are statistically summarized for future research studies. For the fist-time, correlation matrices for the dimensions are estimated from measured data. These correlation matrices can be used to improve simulations of realized member geometry that underpin reliability analysis of cold-formed steel members. Based on measured point clouds, geometric imperfections are characterized into five classes when following conventional imperfection classification, i.e. bow (major axis bending), camber (minor axis bending), twist, flare/overbend, and crown. This method is carried out in this dissertation and compared with another characterization method known as modal imperfection characterization. Instead of using flare/overbend and crown as the cross-section imperfections, the cross-section buckling mode shapes in distortional and local buckling are utilized as the reference to the cross-section imperfections. Imperfections based on MID measures exist along the longitudinal length of a specimen and can be interpreted as power spectrums in the frequency domain, which provide an insightful understanding on classified imperfections involved with multiple frequency (reciprocals of half-wave lengths) content. These results enable another imperfection modeling approach, the 1D spectral approach, and potentially improve prediction accuracy for member strength. Numerical modeling of imperfections comprising both traditional and 1D spectral approach has been carried out and is validated through material and geometric nonlinear shell finite element analysis on the true geometry. Taken together this thesis provides a new platform for measuring imperfections and dimensions of structural members, and demonstrates the use of this platform in enabling advanced analysis methods on imperfect models of cold-formed steel members and improving the strength reliability predictions for cold-formed steel members.