Robust and Efficient WMLES Unstructured Grid Generation with HeldenMesh

Mesh generation for wall-modeled large eddy simulations (WMLES) based CFD simulations represents a critical area of research as there are significant challenges which must be overcome before this technology can be adopted for widespread use.nbsp; Two critical challenges associated with these scale-resolving simulations is mesh size and mesh quality.nbsp; Mesh size represents a critical challenge as WMLES simulations require at least an order of magnitude increase in mesh size compared to current Reynolds-averaged Navier-Stokes (RANS) based simulations. This means required mesh sizes are measured in billions of nodes (exascale) rather than tens or hundreds of millions.nbsp; Existing commercial mesh generators can take several days just to generate these large meshes.nbsp; Mesh quality also represents a critical challenge as the WMLES solver is much more sensitive to element regularity, edge alignment, element type (hex, prism, pyramid, tet), maximum aspect ratio, and surface spacing than typical RANS simulations. The objective of our Phase II effort is to solve these significant challenges by developing an automated, scalable, and high-quality mesh generation capability for next generation CFD based WMLES applications.nbsp; Our approach develops enhancements to our industry leading time-to-mesh HeldenMesh grid generator to improve its support for WMLES applications while also reducing current Reynolds-averaged Navier-Stokes (RANS) based mesh generation times.nbsp; It also develops a new tool which rapidly generates the billions of nodes meshes needed for WMLES simulations using a robust and automated mesh refinement approach ndash; reducing WMLES mesh generation times from days to seconds.nbsp; Finally, it also proves the production readiness of these tools on several real-world WMLES applications while also establishing the best practices needed to ensure solution accuracy.nbsp; Our program represents a key enabler for widespread adoption of WMLES.