Comparative study on ultra-low-cycle-fatigue behaviour of three Indian structural steel grades

Predicting ductile fracture initiation is vital for deformation-controlled elements (DCEs) in seismic force- resistant systems (SFRS) to prevent premature failure and maximize structural members' plastic capacity. Seventy-three notched specimens from plates of structural steel grades E250, E350, and E450, commonly used for SFRS construction in India, are machined into ten notch configurations to simulate stress conditions similar to those faced by DCEs. Machined specimens are designed to mimic moderate to high-stress triaxialities ( 0 . 7 <= T <= 1 . 4 ) in three ultra-low-cyclic-fatigue (ULCF) scenarios: cycles-to-failure, cycles-pull-to-failure, and cycles-to-failure-tension-compression. Load versus notch elongation is precisely measured until ductile fracture initiates, using digital image correlation and fatigue-controlled actuators. Combined hardening parameters are calibrated to validate numerical models with experimental results. Validated models are utilized to derive stress triaxiality, Lode angle functions, and equivalent plastic strain, which are then employed to construct the ductile fracture initiation loci for all three grades. Proposed model (CELVGM) along with three existing models has been evaluated using DDPC approach. This approach considers skewness, decreasing prediction error (PE) range from 5.36%-43.19% (conventional method- average error dataset) to 0.60%-24.04% (median error dataset). Further, proposed model (PE=0.60%) outperforms other models for E350, while CSWDFM (PE=0.83%) and MKK (PE=0.98%) perform better for E250 and E450, respectively.