CFD investigations of turbine rim seal physics

This PhD research project , involving collaborations with the University of Oxford and the Rolls-Royce ICORE program, has the main aim to improve understanding of fluid-dynamic phenomena associated with the sealing of axial turbine rims. This is important in the design of gas turbines for aeroengines and power generation. In particular, designers would like to accurately predict the effects of rim seal cavity phenomena on turbine aerodynamic spoiling and heat transfer at minimum computational cost. The employment in the last decades of computational fluid dynamics (CFD) models has shown mixed results when compared to experiments. In some cases good qualitative and/or encouraging quantitative agreement between CFD and measurement has been demonstrated. In studies of some test cases a clear mismatch between experimental and numerical data has been identified. This is often attributed to modeling errors but it should be noted that experimental uncertainties may also be a strong factor. This has been confirmed in recently published experimental work. Recent publications of both computational and experimental work also confirm that rim seal cavity flows are ruled by complex physics and mixed flow phenomena, challenging current understanding and design methods in industry.
In the first part of this research project, a systematic assessment of the main annulus flow effect through URANS simulations for three different configurations (rotor-stator system without blades and vanes, rotor-stator system with vanes and rotor-stator system with blades and vanes) will be carried out. A particular focus on sensitivity of the unsteadiness and the sealing effectiveness to the main annulus and inner cavity flow conditions, axial and annulus gap will be given. Species concentrations through the solution of a passive scalar concentration equation will be employed for the estimation of the sealing effectiveness and the effect of seal type will also be considered. Comparison will be made to measurements taken at the University of Oxford. Owing to the computational expense of large scale simulations no such systematic study has previously been attempted. The results will be of direct interest to Rolls-Royce engineers and in combination with the complimentary experimental programme will further clarify the strengths and weakness of CFD models.
The second part of the research may depend on the outcomes of the initial study and experimental programme. For example further CFD studies may be required to investigate experimental factors (such as rotor eccentricity) that could lead to differences with experiments, or to investigate geometrical changes that would be of particular interest in design. To investigate modeling effects, more computationally demanding techniques, such as large-eddy-simulation will be considered and developed in collaboration with other researchers in the Thermo-Fluid Systems University Technology Centre.