Influence From Nozzle Guide Vane Wakes and Inlet End-Wall Boundary Layers on Turbine Rear Structure Aerodynamics

Abstract

CFD results can be improved by imposing accurate inlet boundary conditions. A previous paper presented extensive measurements from an engine realistic Turbine Rear Structure (TRS), complemented with CFD results using normal radial profiles (1D) at the inlet [10]. This paper extends stat study with two separate studies, investigating the effect of the upstream Nozzle Guide Vane (NGV) wakes and the inlet end-wall boundary layers. In the first study, simulations are done using full 2D inlet boundary condition, imposing the NGV wakes on the inlet. The second study investigates different inlet end-wall boundary layers. Comparisons with measurements show for some aspects that the upstream Nozzle Guide Vane (NGV) wakes and the inlet end-wall boundary layers are important. Predicting upstream forcing from the TRS on the LPT rotor requires a 2D inlet boundary condition. Also, a strong interaction between the incoming NGV wakes and the secondary flow loss-regions at the outlet is found. However, flow quantities like blade loading, outlet swirl, and OGV wakes are well predicted using regular 1D radial inlet profiles. How the inlet end-wall boundary layers are modelled have significant impact on secondary flows and outlet swirl. If the full boundary layers are prescribed, the secondary flows are over-predicted. This gives under turning and premature separations. With no inlet boundary layers, secondary flows are well captured and gives better separations, improved outlet swirl and wake predictions. The recommendation is therefore to remove the inlet boundary layers when using the Transition SST k-ω γ-Reθ model in TRS CFD simulations.