Experience with the Johnson-King turbulence model in a transonic turbine cascade flow solver

Abstract

Experience is described of the application of the Dawes Navier-Stokes solver to a turbine test case involving shock-boundary-layer interaction. Modifying the implementation of the algebraic eddy-viscosity turbulence model gave improved agreement between the predictions and the experimental results. In order to improve the predictions still further, the nonequilibrium turbulence model of Johnson and King was incorporated into the code. This gave limited further improvement in the overall loss prediction, but the base pressure prediction was not improved. Only a minor shock-induced separation was predicted, compared with the increasing major separation observed in cascade tests with increasing outlet Mach number, although it is necessary to be aware of the possible differences between the real test flow and the two-dimensional steady flow that was modeled. Better prediction of the test flow may require the incorporation of higher-order turbulence models, in which the physics of the flow is more fully represented. Additionally, three-dimensional and unsteady capability may be needed.