Bring together industry users, software providers, and applied researchers to define:
- Best practices for essential metadata for surrogate-ready CAE datasets: geometry parameterization, mesh and discretization, scalar vs field outputs, initial and boundary conditions, parameter space / DOE description, data quality and validation status.
- Dataset structure, formats, and interoperability conventions; large-dataset storage and data-reduction practice (lossless compression vs subsampling, with trade-offs made explicit).
The aim is to produce a NAFEMS best-practices specification defining the metadata needed to make engineering simulation datasets reusable for training and validating surrogate models — across solvers, organizations, and physics domains.
Focus team Chair – Asparuh Stoyanov, Key Ward
