Experimental and Numerical Study of Laminar-Turbulent Transition on a Low-Pressure Turbine Outlet Guide Vane

Paper in proceeding, 2020

In modern commercial aviation engines, the low-pressure turbine (LPT) has a high outlet swirl to maximize turbine power to weight ratio. Downstream of the last LPT rotor is the turbine rear structure (TRS) that with relatively few low-aspect-ratio outlet guide vanes (OGV) de-swirls the flow to maximize the thrust. The performance of the TRS is strongly connected to secondary flow structures, which in turn are strongly influenced by the laminar-turbulent transition. Transition can be challenging to predict in turbomachinery due to the highly complex flow present. At the design point the TRS can have both by-pass transition and laminar separation with transition and a following turbulent reattachment. In addition, a TRS needs to perform well in a large off-design envelope, with large variations of the inlet swirl angle. Accurately predicting transition, both at the design point and in important off-design points, is critical for the development of future TRS modules. In modern geared and ultra-high by-pass engines the TRS swirl angle off-design requirements are also increasing.

There are several available transition models in RANS simulations and most of them need parameter tuning when introduced to new conditions. Evaluation of these models for different turbomachinery components is relatively well covered in the literature even though the model specifics often is a classified property of engine manufacturers. However, there are no cases in the literature of transition prediction with experimental verification in the TRS at engine-realistic conditions.

This work presents the first experimental verification of laminar-turbulent transition in a TRS module tested in the LPT-OGV experimental facility at Chalmers Laboratory of Thermal and Fluid Science. The facility is a semi-closed rig using a rotating 1.5 stage shrouded low-pressure turbine stage to create engine representative inlet conditions for the TRS downstream of the LPT stage. Transition was measured using differential IR-thermography (DIT) which is a non-intrusive two-dimensional measurement technique. The technique was specially developed at Chalmers for this particular purpose and validated by boundary layer hot-wire measurements. The numerical analysis was done using commercially available transition models in Fluent and Ansys CFX. Gamma-theta transition model was used with the k-omega SST turbulence model. Experiments and numerical simulations were performed at a chord Reynold number of 235000 and with LPT outlet swirl angles covering both the design point (ADP) and relevant off-design points.

Numerical and experimental results show that agreement between transition models and experiments can be achieved at these conditions. Boundary layers on the pressure side and suction side undergo laminar-turbulent transition for the selected test range. At decreased OGV aerodynamic load, the boundary layer on the pressure side near the leading edge is laminar along most of the span. At higher OGV loads the secondary flow is influencing the region near the shroud on the pressure side as well as near the hub on the suction side. The transition on the suction side midspan is significantly influenced by the vane load. The numerical analysis was used to better understand the involved flow mechanisms.