Experimental Validation of New Creep Failure Model for Stainless Steel 304 and Comparison with Established Models

The limitations of the existing creep failure models have led to the development of a new approach for creep life prediction. Traditional models such as Norton–Bailey and Omega often fall short in accurately capturing the tertiary phase of creep in engineering materials. Similarly, models like Kachanov–Rabotnov, Theta Projection, and the Sine Hyperbolic methods require detailed material-specific parameters to effectively predict damage. To overcome these challenges, a novel hybrid creep model has been proposed by integrating the Norton–Bailey and Kachanov–Rabotnov formulations. This combined framework harnesses the strengths of both models while compensating for their individual shortcomings. The model was incorporated into the finite element (FE) software ABAQUS through curve fitting—utilizing regression analysis to transition from the baseline Omega model to the Norton–Bailey model. Validation was carried out through a series of experimental tests, including ambient and high-temperature tensile tests, as well as creep tests on stainless steel 304. The results were then compared with finite element creep simulations using dog bone specimens. The new model demonstrated excellent predictive accuracy, outperforming traditional models like Omega and Norton–Bailey, achieving up to 91.79% accuracy in creep strain and creep strain rate over a 1000-hour test and 93.92% accuracy for 336 h test.