2019 CSCE Annual Conference - Laval (Greater Montreal) Conference
Dr. Robert Driver, University of Alberta
Dr. Leijun Li
Lateral-torsional buckling (LTB) is a potential mode of failure for steel beams involving a combination of weak-axis and torsional buckling resulting from strong-axis bending.
Canadian design standards CSA S6 and S16 specify checks that originated from the study of rolled steel beams to ensure they are designed to resist this failure mode. Although some research has been done to assess the applicability of these checks to welded sections, it was primarily conducted during the 1980s when controls on welding procedures were less stringent. Recent studies suggest that the Canadian standards overestimate the LTB capacity of welded beams; however, it is not clear that the residual stress distributions assumed are representative of girders in use today. For this reason, data from modern girders must be collected and examined.
The difference in LTB capacity between rolled and welded sections has been attributed to internal residual stresses. Because the distribution of these stresses differs considerably between welded and rolled beams, the buckling capacity of a welded beam could be substantially different from that of a rolled beam of identical cross-section.
The focus of this study is the measurement and analysis of residual stresses; another parallel study is carrying out large-scale buckling tests to determine girder LTB capacities experimentally. Consideration of residual stresses in conjunction with LTB capacities will allow for a better understanding of the effects of welding on LTB behaviour.
The testing program for this study comprises a comprehensive series of measurements on a set of reduced-scale lab test girders and a set of real bridge girders in the fabrication shop of a Steel Centre industry partner, Supreme Steel LP. Non-destructive ultrasonic testing (UT) is used to measure stresses in the latter, while the former are subjected to UT in conjunction with destructive sectioning tests to confirm the accuracy of the UT system. Past studies have shown UT to yield accurate results when used on welded steel structures; results agreed with those from x-ray diffraction, hole drilling and finite element modelling. The sectioning method has seen widespread use on welded steel sections and is widely considered to be an accurate baseline which non-destructive tests results can be compared to.
Knowledge of residual stress distributions and their impact on LTB behaviour will facilitate a broader evaluation of the current design procedure. Should the procedure prove unsafe, new provisions will be developed and proposed for Canadian design standards.