Large-eddy simulation of the inflow turbulence transport and aerodynamics of a rectangular 5:1 cylinder with high-order numerical methods

Abstract

In this paper, the transport of inflow wind turbulence and the aerodynamics of the BARC (benchmark on the aerodynamics of a rectangular 5:1 cylinder) model are evaluated by large-eddy simulation (LES) based on high-order flux reconstruction (FR) numerical schemes in combination with an inflow turbulence generation method called DSRFG (discretizing and synthesizing of random flow generation). The results of this study show that (1) the key technique of high-order discretization accuracy lies in the high-order reconstruction of a polynomial and an autorefined LES filter scale, which ensures the consistency of the subgrid scale (SGS) model with high-order discretization schemes; (2) high-order FR discretization is proven to be able to attain a converged “discretization order independence” state for a fixed-grid arrangement with a log-linear increase in computational costs based on order promotion; (3) in striking a balance between accuracy and computational cost, the promotion discretization order of the FR scheme is more effective than refining the grid number with low-order discretization; and (4) the fifth-order discretization of the FR method together with a high-order-adaptable DSRFG inflow turbulence generator can generate realistic or targeted incoming turbulence features, maintain turbulence transporting characteristics with minimum dissipation, and capture the subtle turbulence interaction mechanisms of flows around the BARC model properly at different inflow turbulence conditions under reasonable computational costs.