Experimental Characterization of the Rim Sealing Effectiveness in a Stator-Rotor Cavity

The aim of this PhD thesis has been to design, commission and test a new Rotating Cavity Rig to experimentally investigate the problem of hot gas ingestion inside a stator-rotor cavity of a gas turbine. Such facility, designed to be operated under non-reactive conditions, is a single-stage test rig where the test section is constituted by a stator disk which features 44 integrated vanes and a rotor disk which features 66 integrated blades. Both the disks can accommodate replaceable cover plates that, despite some inevitable simplifications, have been specifically designed to replicate different engine rim seals as close as possible. The interchangeability and modularity of the covers allow replacing them with different geometries in order to evaluate the impact of multiple geometrical parameters on the sealing performance. In particular, seven different configurations of radial rim seals were experimentally investigated by performing CO2 gas sampling measurements on the stator surface. This approach showed that increasing the axial overlap or the distance between the vanes TE and the blades LE provides a modest reduction in the sealing flow rate necessary to fully seal the cavity. On the contrary, it was found that increasing the radial gap leads to an increase in the dimensional sealing flow rate which does not appear to be linear with the increase in the radial gap. Although gas sampling measurements represent a standard and robust approach to determine the effectiveness inside a stator-rotor cavity, the use of this technique inevitably results in single point measurements where any 2D effect is neglected. Therefore, an alternative approach based on the exploitation of Pressure Sensitive Paint is presented. Prior to this research, there were no documented cases in literature where PSP had been applied to investigate the phenomenon of hot gas ingestion. First, the PSP setup was validated through a comparison of the results obtained on the stator side with those obtained by performing standard gas sampling measurements. Subsequently, the validated PSP setup was used to obtain 2D effectiveness maps also on the rotor side. This allowed the investigation of the presence of a buffer layer formed by the purge flow pumped towards the outer radius by the rotation of the rotor disk which provides further protection to the rotor side with respect to the stator side. Finally, the results obtained on both sides of the cavity were correlated with each other by performing sealing effectiveness measurements across the cavity width through the use of a concentration probe. Furthermore, the PSP allowed performing the analysis of the effect of different values of Density Ratio on the sealing performance of the rim seal, an aspect usually neglected in the majority of the experimental facilities present in literature. Hence, two distinct values of Density Ratio were achieved by using either pure N2 (DR = 1) or CO2 (DR = 1.52) as purge flow. The resulting 2D effectiveness maps obtained on both sides of the cavity showed a similar qualitative behaviour for the two gases when the same dimensional quantity of purge flow was injected into the wheel-space. However, it was observed that the use of CO2 (higher purge flow density) consistently led to reduced levels of protection. Finally, the experimental data points gathered from both the stator and the rotor sides of the cavity were respectively fitted by using either a modified version of the Orifice Model that explicitly includes the effect of the Density Ratio or the Buffer Ratio model. In the end, this allowed the evaluation of the influence of the Density Ratio on the sealing performance of the rim seal and on the low-order models employed to perform the data analysis.