Code Verification of the HIGRAD Computational Fluid Dynamics Solver

Van Buren, Kendra L.¹  Canfield, Jesse M.¹  Hemez, Francois M.¹  Sauer, Jeremy A.¹ 

[1] Los Alamos National Laboratory

关键词: ADVECTION;    ANALYTICAL SOLUTION;    APPROXIMATIONS;    COMPUTERIZED SIMULATION;    CONVERGENCE;    DIFFUSION;    EVALUATION;    FLUID FLOW;    FLUID MECHANICS;    HYDRODYNAMICS;    IMPLEMENTATION;    LAGRANGIAN FUNCTION;    LANL;    NAVIER-STOKES EQUATIONS;    NUMERICAL SOLUTION;    PERFORMANCE;    VERIFICATION;    WIND TURBINES;   

DOI  :  10.2172/1040022
RP-ID  :  LA-UR-12-21165
PID  :  OSTI ID: 1040022
Others  :  TRN: US201211%%96

美国|英语

来源: SciTech Connect

【 摘 要 】

The purpose of this report is to outline code and solution verification activities applied to HIGRAD, a Computational Fluid Dynamics (CFD) solver of the compressible Navier-Stokes equations developed at the Los Alamos National Laboratory, and used to simulate various phenomena such as the propagation of wildfires and atmospheric hydrodynamics. Code verification efforts, as described in this report, are an important first step to establish the credibility of numerical simulations. They provide evidence that the mathematical formulation is properly implemented without significant mistakes that would adversely impact the application of interest. Highly accurate analytical solutions are derived for four code verification test problems that exercise different aspects of the code. These test problems are referred to as: (i) the quiet start, (ii) the passive advection, (iii) the passive diffusion, and (iv) the piston-like problem. These problems are simulated using HIGRAD with different levels of mesh discretization and the numerical solutions are compared to their analytical counterparts. In addition, the rates of convergence are estimated to verify the numerical performance of the solver. The first three test problems produce numerical approximations as expected. The fourth test problem (piston-like) indicates the extent to which the code is able to simulate a 'mild' discontinuity, which is a condition that would typically be better handled by a Lagrangian formulation. The current investigation concludes that the numerical implementation of the solver performs as expected. The quality of solutions is sufficient to provide credible simulations of fluid flows around wind turbines. The main caveat associated to these findings is the low coverage provided by these four problems, and somewhat limited verification activities. A more comprehensive evaluation of HIGRAD may be beneficial for future studies.