A Cartesian Mesh Method For Turbulent Flow Simulation

This STTR Phase I project proposes to develop a Cartesian mesh solver for high Reynolds number turbulent flow simulations. Due to its ease of grid generation, simplicity of flow solver, lower computational storage requirement, significantly less operational count per cell, and also due to rapid growth of computer power, the Cartesian mesh approach gets revitalized recently with grid adaptation. These reinvented Cartesian mesh methods use body surfaces to perform cell cutting to preserve the geometric fidelity. With a robust cell-cutting algorithm, the grid generation process can be completely automated. Coupled with a tree-based data structure and solution-based grid adaptation, these methods have made great success for inviscid flow simulations. On the other hand, the extension of the Cartesian mesh approach to viscous flows meets tremendous technical challenges. This is because a very fine mesh is usually required near the body surface, leading to largely varying cell sizes associated with the mesh refinement and the small cut cells associated with the body surface. This STTR effort proposes to use subgrid near the body surfaces to alleviate those issues.