Session: 37-04 Unsteady Flow Modeling
Paper Number: 81677
81677 - Influence From Nozzle Guide Vane Wakes and Inlet End-Wall Boundary Layers on Turbine Rear Structure Aerodynamics
The Turbine Rear Structure (TRS) is situated in the rear part of a turbofan engine and is part of the Low Pressure Turbine (LPT) module. The TRS contains the rear engine mounts and holds the aft bearings of the low-pressure axis, and thereby provides structural support between the engine and the airplane. The TRS also contains the LPT Outlet Guide Vanes (OGVs). The OGVs should remove the swirl from the last LPT rotor with as low pressure loss as possible. Some OGV’s must also have a certain minimum thickness to provide the lead-through capability necessary to pass oil-tubes between the inner and outer parts of the engine. Hence, a TRS is a complex component with both structural and aerodynamic purpose, as well with several geometrical constraints.
A new state of the art engine realistic Turbine Rear Structure with a polygonal outer case, three recessed engine mounts with dedicated mount vanes, three tube vanes with increased thickness, and six regular vanes was presented in a previous paper. The configuration was experimentally investigated at three different operating conditions at an engine-representative Reynolds number. Extensive measurements of both inlet and outlet conditions, vane pressure distributions, and wakes were presented and compared with numerical steady-state RANS simulations. The numerical simulations in the previous paper used current industry best practice to impose circumferentially averaged radial profiles of flow quantities on the inlet boundary (1D profile). Some discrepancies between the numerical results and the experiments were found.
This paper extends the numerical results to also study the effect from the upstream Nozzle-Guide-Vane (NGV) wakes and the inlet end-wall boundary layers. Simulations are performed with a full 2D inlet boundary condition, which impose the incoming NGV wakes, and is compared with simulations using regular 1D radial inlet profiles. Numerical analyzes are also performed using different types of inlet end-wall boundary modeling. Both a sector with recessed engine mount and a sector with thicker tube-strut is analyzed. Steady state RANS simulations are performed where flow quantities like the upstream pressure forcing effect, vane pressure distributions, secondary flow structures, separation tendencies, outlet residual swirl angles and outlet wakes are compared to the experimental measurements.
The results clearly demonstrate that for some aerodynamic aspects it is essential to use the full 2D inlet profile in the TRS simulations. The most significant improvement is found in the predicted upstream forcing parameter. Simulations using 1D inlet radial profiles over-predicts the upstream forcing, whereas simulations with a 2D inlet matches the measured upstream forcing very well.
The inlet distortions from the upstream NGV wakes interact with the end-wall boundary layers and the secondary flow structures. To predict this interaction and to obtain a correct outlet total pressure contour, both downstream the engine-mount recesses and close to the hub, the full 2D inlet boundary profile must be iposed on the domain inlet. However, the TRS-vane pressure distributions, outlet vane-wakes, outlet swirl, and secondary flows are well predicted with a normal 1D inlet.
How the inlet end-wall boundary layers are set in the TRS simulations is very important. With full end-wall boundary layers, using all measurements and a physical extrapolation to the end-walls, both the 1D and 2D inlet simulations over-predict secondary flows and separation tendencies, both behind the engine recess and in the highly loaded 25% span region. Using only the measured part of the inlet boundaries, or completely excluding the boundary layers, improves the results significantly. However, if the inlet boundary layers are completely removed then the outlet hub boundary-layer becomes much too small.
Further details about the numerical and experimental comparisons, with some recommendations on how to set the inlet boundary conditions in a TRS CFD analysis, will be given in the full version of the paper, including an attempt to explain the physics behind the results.
Presenting Author: Pär Nylander GKN Aerospace Sweden
Presenting Author Biography: Pär Nylander has a Master degree in Space Engineering with a major in Aerospace Engineering at Luleå Technical University. He graduated in 2008.<br/>Pär Nylander has over 10 years of experience working as engineer within the field of CFD, with main focus on life science, automotive and aerospace industry.<br/>Pär is since 2014 working at GKN Aerospace Sweden as an aerodynamics engineer. Main focus is aero design of turbine rear structures for the civil jet engine market.
Authors:
Pär Nylander GKN Aerospace SwedenSrikanth Deshpande GKN Aerospace Sweden
Jonas Larsson GKN Aerospace Sweden
Influence From Nozzle Guide Vane Wakes and Inlet End-Wall Boundary Layers on Turbine Rear Structure Aerodynamics
Paper Type
Technical Paper Publication