Validation of Numerical Computations and Turbulence Models Combinations for Gas Turbine Cascade Blade Flow

Abstract

The accuracy of computer codes for turbo-machinery turbulent flow fieldcalculations relies strongly on the type and behavior of the turbulence model usedin the computations. Analysis of different Reynolds Average Navier-StokesEquation (RANS) based turbulence models was applied to predict the flow field inthe linear first stage gas turbine cascade blade. The experimental investigation isalso introduced to validate the accuracy of turbulence models. This was done byusing five linear cascade blades tested in an open jet type low-speed subsonic windtunnel. The static pressure distribution was measured at the midspan of cascademiddle blade by using static pressure taps. The numerical results obtained fromdifferent turbulence model simulations is individually reviewed for the correctnessof its predictions and compared with the experimental data in terms of integratedflow parameters, such as static pressure coefficient distribution on both blade sides.The results show that RNG k-e turbulence model gave the best prediction ofpressure distribution when compared with the experimental data. Prediction ofstandard k-ε and k-ω turbulence models fail to predict accurately the flow fieldparameters in cascade passage. Prediction of (k-ε) turbulence model overestimatethe turbulence kinetic energy values, especially in the regions of high velocity atblade suction side, also not accurately predict the flow separation on the bladesuction side.