Steel column ultra low-cycle fracture considering test protocol, strain-aging, and weld type

Thirty-two universal column sections with end plates, connected by fillet or butt welds, were subjected to large cyclic axial strains to study the effects of weld type, test protocol, strain aging, and buckling on ductile fracture. Test protocols included constant amplitude, incremental, and random displacements. Some specimens were preloaded and artificially aged to assess strain-aging effects. Three life-prediction approaches were evaluated: (i) standard Manson-Coffin (SMC), (ii) modified Manson-Coffin (MMC), and (iii) cumulative inelastic deformation (CID). For non-strain-aged specimens, behaviour was independent of weld type, with fractures occurring at the flange exterior rather than near welds. Buckling appeared either at the ends or mid-length. Strain-aged specimens fractured at both flange exteriors and heat-affected zones. No significant differences were observed for pre-strains beyond yield. At low amplitudes, fractures occurred later than in non-aged specimens, but at large amplitudes, earlier fracture was observed. In 8 of 9 cases, predicted damage indices matched experiments within 0.9–1.2. • M2 buckling reduced fracture capacity by up to 20% and accelerated crack initiation compared to M1. • Strain aging increased strength but lowered ductility; it improved fracture at small strains and reduced it at large. • Fillet and butt welds showed negligible fracture differences for non-aged specimens under the tested conditions. • Low end plate stiffness cut fracture capacity in aged and unaged specimens, often causing early HAZ cracks. • SMC, MMC, and CID gave consistent results; SMC was most conservative, while CID was simplest to apply.