Scaling Sealing Performance Across Engine Operating Conditions

Abstract The Secondary Air System in an aero-engine uses relatively cool purge from the compressor to limit the ingress of hot annulus gases into vulnerable turbine cavities through rim seals. Superfluous use of purge is inefficient, while insufficient use leads to thermal degradation of highly-stressed turbine components. This study establishes a predictive design tool to fully characterise a rim seal across the performance envelope of the engine. Physically-informed low-order models are important in the engine design process. The Ingress Wave Model uses a single, empirically-correlated parameter to physically link the shear-driven unsteadiness in the cavity with the swirl in the annulus. In this paper, newly-collected and existing experimental data from three facilities and rim seal geometries demonstrated that this unsteadiness is a linear function of the annulus swirl, with the superposition of purge creating a weak, secondary effect. The novel introduction of the linear correlation enables predictions of sealing effectiveness across the entire engine operating range from just two data points. A larger set of data is shown to improve the accuracy and robustness. Here the method is validated by data collected at low technology readiness level (TRL). The methodology could be applied to data collected from a high TRL demonstrator engine or computational fluid dynamics. This will reduce the number of demonstrator experiments (and associated costs) during design iterations. This paper provides an original scaling methodology in the practical context of the engine design process, including the effects of density. Aero-engines operate with a significant purge-mainstream density ratio (DR), due to differences in the temperatures of the two streams. The methodology incorporates predictions of DR and is further validated against data collected at DR = 1 and 1.5. The model demonstrates that neglecting DR will provide significantly underpredicted rim seal performance.