Parametric and V&V study in a fundamental CFD process: revisiting the lid-driven cavity flow

• Published in: Aircraft engineering 2022-03, Vol.94 (4), p.515-530
• Main Authors: Ge, Mingming, Zhang, Xin-Lei, Brookshire, Kaleb, Coutier-Delgosha, Olivier
• Format: Article
• Language: English
• Subjects: Accuracy; Aerodynamics; Aircraft structures; Algorithms; Boundary conditions; Cavity flow; Computational fluid dynamics; Cost analysis; Discretization; Empirical analysis; Finite element method; Fluid dynamics; Fluid flow; Geometry; Iterative methods; Mathematical models; Methods; Navier-Stokes equations; Numerical analysis; Parameter sensitivity; Preconditioning; Reynolds number; Simulation; Velocity; Viscosity; Workflow
• ISSN: 1748-8842; 1758-4213
• DOI: 10.1108/AEAT-04-2021-0108

Purpose The openings on aircraft structures can be modeled from an aerodynamical point of view as lid-driven cavities (LDC). This paper aims to show the primary verification and validation (V&V) process in computational fluid dynamics (CFD, and to investigate the influences of numerical settings on the efficiency and accuracy for solving the LDC problem. Design/methodology/approach To dig into the details of CFD approaches, this paper outlines the design, implementation, V&V and results of an efficient explicit algorithm. The parametric study is performed thoroughly focusing on various iteration methods, grid density discretization terms and Reynolds number effects. Findings This study parameterized the numerical implementation which provides empirical insights into how computational accuracy and efficiency are affected by changing numerical settings. At a low Reynolds number (not over 1,000), the time-derivative preconditioning is necessary, and k = 0.1 can be the optimal value to guarantee the efficiency, as well as the stability. A larger artificial viscosity (c = 1/16) would relieve the calculating oscillation issue but proportionally increase the discretization error. Furthermore, the iteration method and the mesh quality are two key factors that affect the convergence efficiency, thus need to be selected “wisely”. Practical implications The study shows how numerical implementation can enhance an accurate and efficient solution. This workflow can be used to determine the best parameter settings whenever CFD researchers applying this LDC problem as a complementary design tool for testing newly developed solvers. Originality/value The studied LDC problem is representative of CFD analysis in real aircraft structures. These numerical simulations provide a cost-effective and convenient tool to understand the parameter sensitivity, solution receptivity and physics of the CFD process.