Experiment-informed models and simulation-informed experiments: optimizing residual stress diffraction measurements

Accurate residual stress predictions for thermal processing in engineering alloys with finite element approaches are highly valuable for addressing both research and manufacturing questions. Such models are often based on residual stress diffraction experiments, but the creation of models “after-the-fact” leave the burden of discovery squarely within the experimental space. Conversely, FEM-based simulations of residual stress without experiment calibration run the risk of insufficient materials property assumptions or unreasonable process step simplification. The research presented here provides two examples of the integration of modeling and experiment approaches for quantifying residual stress evolution during thermal processing. In one example, we demonstrate the usefulness of incremental model optimization (adjusting materials properties, phase transformation kinetics, process steps and mesh control), through comparison to surface residual stress measurement using Cr anode x-ray diffraction on AISI 9310 gear steel samples subjected to carburizing and hardening heat treatments. In the second example, neutron diffraction measurement of 70/30 CuNi weldments are guided by FEM predictions of residual stress distributions. These examples illustrate two frameworks for integration of residual stress modeling and experimental measurement that reduce computational error and optimize limited access to high fidelity experiment method (e.g., neutron beamtime allocation).