The investigation documented here aims to contribute to the understanding of secondary flows in modern high-pressure (HP) turbine blade passages. More specifically, it aims to improve the understanding of vortical structures near the endwall with respect to the presence of an upstream cavity that approximates the gap present in actual engines between the rotor and stator of an HP turbine. Further, it aims to assess the viability of using non-axisymmetric endwall contouring to reduce endwall losses, including those generated by the presence of an upstream cavity, using a modern airfoil, at HP turbine representative speeds and at off-design Mach numbers.

To attempt to achieve this, a combination of flow measurement, flow visualization, and computational fluid dynamics (CFD) with linear turbine cascades was used. The test matrix consisted of three cases all with one common blade cascade geometry (SL2P). SL2P was combined with one of three endwalls:​ a baseline flat endwall, a flat endwall with a cavity incorporated upstream of the blade row, and a contoured endwall with the same upstream cavity geometry. A combination of pressure probes, pressure taps, and flow visualization were used to collect quantitative and qualitative data in a blow-down type wind tunnel. Complementary CFD studies were also carried out using the commercial CFD code ANSYS CFX.

It was found that the presence of an upstream cavity can noticeably alter the structure and the strength of the secondary flow. When compared to the baseline flat endwall, measurements downstream of the trailing edge have shown a substantial increase in the size and strength of the passage vortex. The presence of the cavity also introduces a significant increase in the level of overturning. The secondary kinetic energy increased significantly due to the presence of the cavity, and a major effect on the overall losses was also evident, with the cavity endwall generating up to 14% higher mixed-out row losses relative to the baseline flat endwall at the design Mach number of 0.80. It was also found that the endwall contouring design used can successfully lower the losses by altering the structure and the strength of the secondary flows. When compared to the cavity endwall, the contoured endwall results have shown a significant reduction in the size and strength of the counter vortex, accompanied with a large reduction in the secondary kinetic energy in the vortex-vortex interaction regions between oppositely rotating vortices. Overall, the contoured endwall showed a drop of up to 10.7% in mixed-out row losses relative to the cavity endwall, with slightly higher losses than the baseline flat endwall. In assessing the off-design performance of the endwall contouring, the experimental results showed that the endwall contouring can continue to successfully lower the losses at small off-design Mach numbers (Mach 0.69, 0.75, 0.84, and 0.89).

The computations did not agree well with the experiment, and overestimated both the losses and the secondary kinetic energy. They predicted almost equal losses for both the flat and cavity endwalls, and no benefit for the endwall contouring design used in this study.