Numerical simulation of transitional turbomachinery flows

The numerical prediction of losses in turbomachinery flows can still be considered a challenge, due to the difficulty of modeling those flow regions which are strongly influenced by turbulence phenomena. Actually, a large part of numerical methods overestimate the magnitude of losses, whereas they can provide more reasonable loss-distribution profiles. In order to improve the loss-prediction capability of modern computational methods, improved turbulence models are needed, with particular attention to boundary-layer transition, which plays a fundamental role in turbomachinery flows. Commonly, such flows are characterized by transitional boundary layers, the transition being induced by: 1) the free-stream turbulence (bypass transition); 2) the formation of a separation bubble (separated-flow transition); 3) the wakes from upstream blades (wake-induced transition). Moreover, the curvature of the blade profile also influences the evolution of transition. Transitional boundary layers can be computed by introducing low-Reynolds-number modifications in the modeling of turbulence. Using the low-Reynolds-number k — ω model [1], the author has developed a methodology for the solution of the steady and unsteady Reynolds-averaged Navier-Stokes equations [2]. In the present work such a methodology is improved by using the Explicit Algebraic Stress Model (EASM) of [3], which accounts for mild nonequilibrium effects and turbulence anisotropy. Firstly, the subsonic flow through the T106 turbine cascade is considered as a suitable test case for validation in two space dimensions. In fact, such a test case is a formidable one due to the separated-flow transition which takes place in the rear part of the blade suction side. Then, the method is extended to three space dimensions and validated versus the experimental data available for the 3D flow through the T106 linear cascade. In particular, the loss-coefficient distribution downstream of the cascade is studied in order to demonstrate the effectiveness of the proposed approach.