Neural network based rate- and temperature-dependent Hosford–Coulomb fracture initiation model

• DOI: 10.1016/j.ijmecsci.2023.108643
• Publication Date: 2023-07-21T07:11:48Z
• Authors: Xueyang Li; Christian C. Roth; Dirk Mohr
• Keywords: Necking; Strain hardening exponent; Hardening (computing)

The accurate description of the strain rate and temperature dependent response metals is a perpetual quest in crashworthiness forming applications. In present study, experiments are carried out to probe onset ductile fracture for an aluminum alloy AA7075-T6 136 combinations stress state, temperature. experimental campaign covers rates ranging from 0.001/s 100/s, temperatures 20°C 360°C. We combine YLD2000 yield surface with neural network based hardening law describe large deformation plasticity material. NN-based trained on data, achieving 3.9% accuracy force predictions including post-necking regime. loading paths extracted each simulation, showcasing non-proportionally evolving triaxiality, Lode angle parameter, A parameterized Hosford-Coulomb locus proposed, which trainable using these histories. proposed model validated against fracture, predicting at error 8%.