An Experimental and Numerical Study of Splitter-Vane Concept for Turbine Rear Structures

Paper i proceeding, 2026

In jet engines, sustainability drives researchers to continuously improve the performance of engine components through lightweight designs. Reducing the weight of components in jet engines, and in turn reducing specific fuel consumption, is an area of improvement that has challenged the research and engineering community for decades. In this context, the turbine rear structure (TRS), which is located aft of low-pressure turbine (LPT), is assessed for weight reduction possibilities in the present work. TRS is a considerable heavy component as it must fulfil several functions. Along with providing mechanical support for the engine, TRS also serves the aerodynamic purpose of de-swirling the flow from LPT. One option to achieve a lower weight in the TRS is to use a shorter component, thereby requiring less material for the structure itself. However, a shorter structure means a more demanding and aggressive aerodynamic design for de-swirling the flow. To achieve the objective of having both a reduced total pressure loss and an effective de-swirling, in this paper a splitter-vane design is explored and analyzed.  

In comparison to a state-of-the-art TRS design without splitter, a splitter vane design helps on reducing the vane chord length while retaining the number of primary Outlet Guide Vanes (OGVs). With a shorter chord, the diffusion on the suction side of vanes is expected to increase and hence may become a source of inducing losses. This reduction in chord negatively impacts the de-swirling of flow as well. To address this issue, splitter-vanes are added in-between primary OGVs to accompany the flow on the suction side of the latter in a TRS.

The aerodynamic design of the TRS protagonist of this study is performed by GKN aerospace, Sweden. The TRS prototype is then manufactured and tested in the experimental LPT-OGV rig at Chalmers University of Technology for its aerodynamic functionality. Experimental data are studied and compared to state-of-the-art engine-realistic TRS design without splitter running at the same operating conditions. It is observed that the de-swirling of flow, which is one of the required aerodynamic functionalities for TRS, could be optimally achieved by the splitter-vane concept. A remarkable experimental result is very much downstream shifted laminar-turbulent transition on the splitter vanes.

Numerical studies are also performed on the splitter-vane design at GKN Aerospace with an industrial state-of-the-art CFD tool and a mesh sensitivity study is performed to determine and choose an accurate and dependable domain. A turbulence model used with laminar-to-turbulent transition is used. A two-dimensional plane of measured stagnation properties is imposed at inlet as boundary conditions and numerical results are compared with the experimental data from the prototype testing.

The results from the experiments and numerical analysis clearly indicate region of higher diffusion on the suction side for splitter-vane configuration when compared to TRS design without splitter while the flow is effectively de-swirled. A good agreement is achieved when comparing the experimental data with the numerical results. Lastly, while one of the possibilities for the positioning of the splitter vane relative to the primary OGV is studied in this work, it is believed by the authors that there is margin for performance improvement through such design parameter.