Stencil selection algorithms for WENO schemes on unstructured meshes

• Published in: Journal of computational physics Vol. 475; p. 108840
• Main Author: Tsoutsanis, Panagiotis
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 15.02.2023
• Subjects: Finite-volume; Least-squares; Stencils; WENO; Finite-volume; Least-squares; Stencils; WENO
• ISSN: 0021-9991; 1090-2716; 1090-2716
• DOI: 10.1016/j.jcp.2019.07.039

In this paper, a family of stencil selection algorithms is presented for WENO schemes on unstructured meshes. The associated freedom of stencil selection for unstructured meshes, in the context of WENO schemes present a plethora of various stencil selection algorithms. The particular focus of this paper is to assess the performance of various stencil selection algorithm, investigate the parameters that dictate their robustness, accuracy and computational efficiency. Ultimately, efficient and robust stencils are pursued that can provide significant savings in computational performance, while retaining the non-oscillatory character of WENO schemes. This is achieved when making the stencil selection algorithms adaptive, based on the quality of the cells for unstructured meshes, that can in turn reduce the computational cost of WENO schemes. For assessing the performance of the developed algorithms well established test problems are employed. These include the least square approximation of polynomial functions, linear advection equation of smooth functions and solid body rotation test problem. Euler and Navier-Stokes equations test problems are also pursued such as the Shu-Osher test problem, the Double Mach Reflection, the supersonic Forward Facing step, the Kelvin-Helmholtz instability, the Taylor-Green Vortex, and the flow past a transonic circular cylinder. •Various central and directional stencils algorithms presented.•Uniqueness, Cover, Width and Compactness the most dominant parameters for directional stencils.•Algorithms suitability is dependent on the numerical framework.•Adaptive stencils reduce the computational cost of the WENO schemes.