Numerical methods for solving mechanical problems, including finite element analysis, computational fluid dynamics, multiscale modeling, and the simulation of complex engineering systems
3 papersThis paper extends the framework of relaxation Runge-Kutta methods to ensure conservation or dissipation of general convex quantities, such as entropy. It enforces stability through a relaxation parameter, modifying existing Runge-Kutta implementations with minimal cost and maintaining accuracy and other desirable properties.
This study finds that the grid-point requirement for DNS of a spatially developing turbulent boundary layer is N ~ Re^2.05, less than previously thought. It also estimates time-step requirements for DNS and LES, suggesting that wall-modeled LES becomes increasingly more cost-efficient than wall-resolved LES and DNS as Reynolds number increases.
This paper describes Athena++, a new framework for adaptive mesh refinement (AMR) calculations, particularly suited for astrophysical fluid dynamics. The framework is designed to be flexible and modular, enabling various physics, including non-relativistic and relativistic MHD, and incorporates a "task list" model for dynamic scheduling to improve parallel performance and simplify multiphysics simulations.