Identification of Johnson-Cook, Thermal and Damage-related Material Parameters of S355MC Steel through Bayesian Inverse Analysis
Please login to view abstract download link
The material behavior of metals under highly nonlinear loading conditions remains insufficiently understood, limiting the predictive capability of simulations for complex manufacturing processes such as shear cutting of metal sheets. In this work, we investigate the identification of thermo-mechanical, damage-related and coupling material parameters for S355MC steel based on tensile test simulations conducted up to the onset of failure. To identify valid material parameters across the process parameter space, experiments and simulations are performed at varying ambient temperatures up to 400 degrees Celsius and testing speeds up to 15 mm/s. The simulations employ a highly efficient, thermodynamically consistent viscoplastic Johnson-Cook constitutive model with implicit time integration, coupled thermal effects including thermal expansion and softening, viscoplastic dissipation via the Taylor-Quinney coefficient, and a Lemaitre-type ductile damage model. The in-situ measured experimental data include force-elongation curves and temperature values at discrete points. Viscoplastic, thermal, and damage-related material parameters are identified through a Bayesian inverse analysis employing an adaptive sampling strategy tailored to computationally expensive forward models. The results demonstrate that measurements covering the full process parameter space are essential for resolving otherwise ambiguous material parameters. In particular, we explore how parameter identifiability can be enhanced by incorporating multiphysics information. Moreover, based on synthetic data, we investigate how parameter identification can be improved by considering experimental setups that induce more complex strain and stress states compared to standard uniaxial tensile tests.
