An Improved Second-Order Finite-Volume Algorithm for Detached-Eddy Simulation Based on Hybrid Grids. (21st July 2016)

• Record Type: Journal Article
• Authors: Zhang, Yang; Zhang, Laiping; He, Xin; Deng, Xiaogang
• Is Part Of: Communications in computational physics. Volume 20:Number 2(2016:Aug.)
• Page Start: 459
• Page End: 485
• Publication Date: 2016-07-21
• Subjects: 76F99, -- 76M12, -- 65L80; Detached-Eddy simulation, -- second order scheme, -- self-adaptive dissipation, -- hybrid grid, -- finite-volume method; Mathematical physics -- Data processing -- Periodicals; Physics -- Data processing -- Periodicals; 530.150285
• DOI: 10.4208/cicp.190915.240216a
• Languages: English
• ISSNs: 1815-2406

Abstract

Abstract: A hybrid grid based second-order finite volume algorithm has been developed for Detached-Eddy Simulation (DES) of turbulent flows. To alleviate the effect caused by the numerical dissipation of the commonly used second order upwind schemes in implementing DES with unstructured computational fluid dynamics (CFD) algorithms, an improved second-order hybrid scheme is established through modifying the dissipation term of the standard Roe's flux-difference splitting scheme and the numerical dissipation of the scheme can be self-adapted according to the DES flow field information. By Fourier analysis, the dissipative and dispersive features of the new scheme are discussed. To validate the numerical method, DES formulations based on the two most popular background turbulence models, namely, the one equation Spalart-Allmaras (SA) turbulence model and the two equation k – ω Shear Stress Transport model (SST), have been calibrated and tested with three typical numerical examples (decay of isotropic turbulence, NACA0021 airfoil at 60° incidence and 65° swept delta wing). Computational results indicate that the issue of numerical dissipation in implementing DES can be alleviated with the hybrid scheme, the resolution for turbulence structures is significantly improved and the corresponding solutions match the experimental data better. The results demonstrate the potentiality of the present DES solver for complex geometries.