Structural Response in Thin-Walled Steel Structures Subjected to Compressive Axial Dynamic Loading

Abstract

As Swedish cities densify, the proximity between transportation routes and buildings increases, raising concerns about the potential impact of exceptional accidents such as explosions from transportation routes on nearby buildings. In such events, corrugated sheet metal (CSM), frequently used for roofing applications in buildings, becomes susceptible. These thin-walled structures possess limited capabilities in supporting dynamic loads, thus endangering the structural integrity. This thesis aims to investigate the structural response of a compressed thin-walled steel structure. A simplified model involving thin steel rods subjected to axial loads under both quasi-static and dynamic conditions is employed to understand the fundamentals of dynamic buckling. The employed methodology comprises experimental testing and numerical analyses using the FE software Abaqus CAE. To form the foundation for the experiments, initial numerical analyses are conducted, incorporating the Cowper-Symonds material model for simulation of strain rate effects. In the experimental testing, two different rod configurations are subjected to axial velocities ranging from 0.013 mm/s (10-4 s-1) to 195 mm/s (1.5 s-1). Outcomes of these tests are then validated against subsequent numerical analyses. Findings from experiments and numerical analyses reveal that the structural response of the steel rods is governed by buckling. At increased loading rates, the material exhibits a strengthening effect, manifested through the increase in the critical buckling force. Additionally, inertia effects in the form of dynamic oscillations become apparent at specific strain rates. The initiation of these effects is further accelerated with higher slenderness. Moreover, amplified slenderness and/or initial imperfections cause reduction in load-bearing capacity. Lastly, the amplitude of the dynamic oscillations is damped with increased initial imperfections. Overall, strong correspondence between the outcomes of the two methods is found. While there are disparities in the predicted axial capacity, the dynamic behavior of the rods simulated in Abaqus CAE closely reflects the dynamic effects observed during experimental testing.