Facilitators: Muge Fermen-Coker (USARL), Amiee Hungerford (LANL)
Modeling capabilities that can handle multi-disciplinary engineering challenges, and that can bridge multiple length and time scales, are required for cutting edge materials research and applications. This includes development of new materials and systems that operate in extreme environments (i.e. extreme temperatures, radiation fields, ballistic impact conditions, explosion, oxidation, corrosion, etc.) Among the desired outcomes of these future, multiscale, multiphysics codes are the ability to computationally predict material / component performance due to chemistry or microstructural modifications, and potentially allow for manufacturing optimization of structures that can mitigate negative environmental conditions. The ultimate goal is to select manufacturing techniques through computationally-informed decisions that predict whether component or system level performance will meet the application requirements. Enthusiasm for multiscale computation is further elevated by the promise that such algorithms will make efficient use of emerging HPC architectures. Current approaches to multiscale modeling typically improve single scale modeling capabilities and transfer information between various scales (usually bottom up) to bridge length and time scales. Will this approach be adequate?
Discussion in this group will aim to lay out a path for developing codes that can ultimately enable tailoring of the microstructure to meet application-specific requirements.
- Overall strategy for code development
- What are the intermediate steps to this ultimate goal?
- What collaborations are needed to drive progress in the next 5 years?
- Promising physics based models
- Verification/validation suite data that link length and time scales
- Optimization methods and metrics for deciding if application requirements are met
- Methods to handle uncertainties in the process
- Computational efficiency
- Visualization needs