Abstract
The main objectives of the paper are to test widely used turbulence models against selected benchmark problems identifying range of applicability, and limitations of the models used and to help accurately, predict flow in turbine blade passages and disc cavities. The following models are considered: the k epsilon model with and without a Kato-Launder correction and a Richardson number correction, the two-layer k-epsilon k-l model, and the Spalart -Allmaras model with and without correction for rotation. Weaknesses in the models are identified and suggestions made for possible improvements. Numerical implementation of the wall functions approach is also considered. The test cases considered include flat plate boundary layer flat plate heat transfer: enclosed rotating disc, and a combined turbine blade/disc cavity, model. Comparisons art, made with experimental data and computations from different CFD codes.