The Sod gasdynamics problem as a tool for benchmarking face flux construction in the finite volume method

TL;DR

This paper evaluates twenty-two flux construction methods in the finite volume method for solving the Sod gasdynamics problem, comparing their accuracy and computational performance in modeling inviscid fluid flow.

Contribution

It provides a comprehensive benchmark of various flux methods using the Sod problem, highlighting their accuracy and runtime characteristics.

Findings

Exact Riemann solver and van Leer flux methods performed best in accuracy.

All methods had similar runtimes, with the exact Riemann solver being the slowest.

The study offers guidance on flux method selection for CFD simulations.

Abstract

The Finite Volume Method in Computational Fluid Dynamics to numerically model a fluid flow problem involves the process of formulating the numerical flux at the faces of the control volume. This process is important in deciding the resolution of the numerical solution, thus its quality. In the current work, the performance of different flux construction methods when solving the one-dimensional Euler equations for an inviscid flow is analyzed through a test problem in the literature having an exact (analytical) solution, which is the Sod problem. The work considered twenty two flux methods, which are: exact Riemann solver (Godunov), Roe, Kurganov-Noelle-Petrova, Kurganov-Tadmor, Steger-Warming Flux Vector Splitting, van Leer Flux Vector Splitting, AUSM, AUSM+, AUSM+-up, AUFS, five variants of the Harten-Lax-van Leer (HLL) family, and their corresponding five variants of the Harten-Lax-van Leer-Contact (HLLC) family, Lax-Friedrichs (Lax), and Rusanov. The methods of exact Riemann solver and van Leer showed excellent performance. The Riemann exact method took the longest runtime, but there was no significant difference in the runtime among all methods.