Dynamic Adaptive Mesh Refinement for WMLES on Complex Configurations

View at publisher
Authors: Dimitri J. Mavriplis, Zhi Chun Yang, Andrew C. Kirby
Cited by: 0
References: 18

Abstract

This paper presents a dynamic adaptive mesh refinement (AMR) capability for wall-modeled large-eddy simulations (WMLES) of complex configurations using body-fitted, unstructured grids with hybrid element types. The approach links legacy computational fluid dynamics (CFD) codes to a highly scalable octree-based mesh-subdivision library capable of handling tetrahedral, prismatic, pyramidal, and hexahedral elements. By defining and implementing a general application programming interface (API), the framework enables large-scale dynamic AMR on high-performance computing (HPC) systems without requiring internal modifications to existing CFD codes. The development of the AMR interface is described in detail and demonstrated by coupling it with two CFD solvers, NSU3D and FUN3D. Results are presented for canonical test cases as well as RANS and WMLES applications of industrial relevance, including efficiency and scaling performance. The paper concludes with a discussion of future extensions, including support for GPU-accelerated architectures.

Fields of Study & Subjects

Computational Fluid Dynamics and Aerodynamics Advanced Numerical Methods in Computational Mathematics Fluid Dynamics and Turbulent Flows

Publication details

Journal / source
AIAA SciTech Forum
Publication date
8 January 2026