Abstract

This study strives to provide a critical review and evaluation of influential constituent components in the two-equation k-ε turbulence model class, which are often employed in large full-fidelity gas turbine simulations in industry. All conjectures made, regarding the turbulence model behavior, will be compared to in-house experimental film effectiveness data from a novel film cooling array configuration. This experimental data set is comprised of high resolution adiabatic film cooling effectiveness measurements throughout two film cooling arrays comprised of diffused film holes of modern shape (streamwise inclination of α = 30°, expansion of Φ1,2,3 = 10°, AR = 3.4, and β = 0° relative to the crossflow direction). The difference between each film cooling array is only a staggered verses inline pattern for the film holes, whereby each has a uniform spacing of X/D = P/D = 8. Local coverage, laterally averaged film cooling effectiveness, and superposition analysis was quantified over a variety of testing conditions (M = 0.5, 0.75, 1.0, 1.5, 2.0, 2.5 at DR = 0.9).

First, rendered quantifications of the anisotropic nature of the turbulence are shown throughout the near-field injection region, by leveraging the Reynolds stresses to form anisotropic-invariant maps. Next, results from k-ε models using a linear, a cubic, and a quadratic constitutive relation are compared. Furthermore, effective and conservative scaling of the production and destruction terms in the turbulent transport equations was performed, within reasonable bounds, and the resulting impact on the adiabatic film effectiveness was quantified. This scaling encompasses the Cε2 coefficient, as well as the Durbin realizibility coefficient used in the turbulent viscosity definition. Finally, various formulations of the turbulent Prandtl number were compared, with the resulting adiabatic film effectiveness observed.

This content is only available via PDF.
You do not currently have access to this content.