Thickness debit effect in Ni-based superalloys under different experimental conditions: A crack evolution perspective

This study comprehensively employed finite element analysis, experimental statistics, and machine learning to systematically investigate the influence of experimental conditions on the thickness debit effect and its underlying mechanisms. Simulation results show that under relatively low-temperature conditions, the samples exhibit a more pronounced thickness debit effect, while the influence of applied stress is minimal. Analysis of experimental data also supports this result. The relationship between temperature, applied stress, and the thickness debit effect was established using machine learning. Furthermore, results based on permutation importance indicate that temperature has a higher feature importance compared to applied stress. The crack analysis also indicates that sample thickness has a negligible effect on crack initiation but a substantial effect on crack propagation. At relatively low temperatures, limited stress relaxation keeps the local stress peak high, so that less damage accumulates in the matrix before propagation and the crack initiation stage occupies a smaller fraction of the rupture life; the less-damaged matrix slows subsequent crack growth, and the thickness-controlled propagation stage therefore accounts for a larger fraction of the rupture life, giving a more pronounced thickness debit effect. This behavior originates from the stress relaxation coefficient ( α ): while the elastic modulus ( E ), yield strength ( σ 0.2 ), and stress rupture strength have minimal impact, a lower α at low temperature limits stress relaxation and confines damage near the crack, whereas a higher α at high temperature relaxes the stress peak and spreads damage through the matrix. The α therefore plays a critical role in thickness debit effect.