High-order simulation of NASA juncture flow with Reynolds-stress model

The prediction of junction flows, such as those formed around a wing-fuselage intersection, remains a challenge owing to the strong anisotropy. Traditional eddy viscosity models(EVMs) usually predict a non-physical separation. To solve this problem robustly by higher-order methods, this paper proposes alternative scale-providing Reynolds-stress models(RSMs) in which the specific dissipation rate ω transport equation is replaced by λ (λ = 1/8√ω = 8√τ) in the framework of the original SSG/LRR model . The new scale-providing variable could improve the calculation robustness of turbulence models especially in high-order numerical methods owing to its natural boundary conditions and small gradient near the wall. Besides it is also combined with γ-Reθt transition model to extend its applications. In addition, some EVMs such as γ-Reθt SST , SST k-w and SST k-w with quadratic constitutive relation(QCR) are also implemented as comparations. The solution polynomials of the mean-flow and turbulence-model equations are both discretized by five-order weighted compact nonlinear schemes(WCNSs). NASA Juncture-Flow experimental data is used for validation. Finally, both the velocity components and Reynolds stresses near and upstream of the junction region predicted by the new RSM are better than the EVMs. The calculated range of junction separation zone is also well aligned with experiments. Therefore, the scale-transformed RSM model has successfully improved prediction accuracy in the juncture region compared to eddy viscosity models.