Data-driven multiscale damage and failure prediction

The anticipated outcome of the proposed work is a predictive computational theory for damage and failure (e.g. ductile fracture, fatigue, and creep) of complex, hierarchical materials such as metal alloys. The effort builds on data-driven, reduced order, and multiscale principles within mechanics with the potential for a transformative new theory. This theory will be used to enable design of materials and systems with a focus on quantification and prediction of unique mechanical properties exhibited by components and materials fabricated with metal additive manufacturing (AM). If funded, the project would consist of three primary thrust areas. First, an experimental characterization effort with synchrotron x-ray computed tomography and diffraction with in-situ mechanical testing would provide fundamental material information. Once this information was used to calibrate micromechanical models, simulation would be used to populate a database of microstructures and their mechanical response. Second, a new concurrent multiscale theory based on reduced order methods and applicable to a wide range of material classes will be developed, capable of capturing severe nonlinearity both in geometric and material response. This method will query the database constructed in the first phase for mechanical information, using that data to predict damage and failure. Third, verification and validation of the data-driven models will be conducted. This will include evaluation against reliable experimental and computational results for both global (macroscale) and local (microscale) predictions. Once validation and verification is completed, remaining time in the final stage of the proposal would be used to explore further applications of the method, such as soft materials.