High resolution optical cameras and computer vision algorithms are employed in this research to create full field digital state models (3D+time) of buckling deformation over the full length of six built-up cold-formed steel column tests, from initial loading to collapse to the post-buckling response. Corresponding buckling and load-deformation response are always of interest in a thin-walled structure experiment because capacity and stiffness degradation can be related to elastic buckling modes that serve as the basis for capacity prediction equations in codes and standards. In a typical experiment, these modes are only coarsely tracked at discrete points with displacement transducers, or they can be measured with accurate laser-based systems down to 0.025 mm (0.001 in.) that typically requires equipment moving along a rail which is challenging to implement many times during an actual experiment. The image-based reconstruction approach in this research is different because it uses a series of cameras mounted around a cold-formed steel column that are controlled together to take pictures from strategically located vantage points six times per minute over the course of the test. Pixels common to multiple images are identified and relative positions are used to calculate their 3D coordinates in a point cloud, providing a full field time record of the experiment. These digital state models provide valuable information that can be used for buckling mode decomposition and to validate and inspire evolving analysis and design methods for thin-walled structures.
展开▼
