Recent experience with different turbulence models applied to the calculation of flow over aircraft components

An evaluation of ten turbulence models is made for compressible flows encountered in current aircraft applications. The Baldwin–Lomax and P.D. Thomas algebraic models, the Baldwin–Barth and Spalart–Allmaras one-equation models, five low-Reynolds-number k– models and the Menter SST blended k–/k–ω model are examined. A zonal, upwind, implicit, factored algorithm is used to solve both the mean flow equations and the turbulence model equations for three-dimensional, compressible turbulent flow. Calculations are presented for both internal and external flowfields including a two-stream mixing layer, a supersonic flat-plate boundary layer, a transonic supercritical airfoil, a shock wave/turbulent boundary layer interaction, an ejector nozzle, a highly offset diffuser, and a twin impinging jet flowfield. The influence of two modifications to the production of turbulent kinetic energy for the low-Reynolds-number k– models is evaluated, a vorticity-based strain rate and a production limiter. A compressibility correction for high speed shear layers is also examined. Comparisons of the results of the various turbulence models are made with experimental measurements. Significant differences are observed in the model predictions when applied to the same problem using the same computational mesh and mean flow solver. The algebraic models are unable to capture the physics of these complex flowfields, particularly for the internal flow calculations. The performance of each model is dependent on the application. No universal model is found for all flowfields examined. Each one-equation and two-equation model has specific strengths and weaknesses and the performance of each model is assessed.