Comparisons of High-Reynolds-Number EVM and DSM Models in the Prediction of Heat and Fluid Flow of Turbine Blade Cooling Passages

Abstract

This paper presents computations of flow and heat transfer through passages relevant to those used to internally cool gas-turbine blades, using high-Reynolds-number models of turbulence. Three types of internal flows are first examined, which between them contain all the main elements found in blade cooling passages; developing flow through a heated straight duct rotating orthogonally, repeating flow and heat transfer through a straight ribbed duct and flow and heat transfer through a round-ended U-bend of strong curvature square and of cross-section. Next, flows influenced by a combination of these elements are computed. The main objective is to establish how reliably, industry-standard high-Reynolds-number models can predict flow and wall-heat transfer in blade-cooling passages. Two high-Reynolds-number models have been used, the standard version of the high-Re k-ε (EVM) model and the basic high-Re model of stress transport (DSM). In all the cases the second-moment closure (DSM) consistently produced flow and thermal predictions that are closer to available measurements than those of the EVM model. Even the high-Re DSM predictions, however, are not in complete agreement with the experimental data. Comparisons with predictions of earlier studies that use low-Re models of turbulence show that at least some of the remaining differences between the current predictions and experimental data are due to the use of the wall-function approach.