Abstract
Compliant contacting filament seals such as brush seals are well known to give improved leakage performance and hence specific fuel consumption benefit compared to labyrinth seals. The design of the brush seal must be robust across a range of operating pressures, rotor speeds and radial build-offset tolerances. Importantly the wear characteristics of the seal must be well understood to allow a secondary air system suitable for operation over the entire engine life to be designed. A test rig at the University of Oxford is described which was developed for the testing of brush seals at engine-representative speeds, pressures and seal housing eccentricities. The test rig allows the leakage, torque and temperature rise in the rotor to be characterized as functions of the differential pressure(s) across the seal and the speed of rotation. Tests were run on two different geometries of bristle-pack with conventional, passive and active pressure-balanced backing ring configurations. Comparison of the experimental results indicates that the hysteresis inherent in conventional brush seal design could compromise performance (due to increased leakage) or life (due to exacerbated wear) as a result of reduced compliance. The inclusion of active pressure-balanced backing rings in the seal designs are shown to alleviate the problem of bristle-backing ring friction, but this is associated with increased blow-down forces which could result in a significant seal-life penalty. The best performing seal was concluded to be the passive pressure-balanced configuration, which achieves the best compromise between leakage and seal torque. Seals incorporating passive pressure-balanced backing rings are also shown to have improved heat transfer performance in comparison to other designs.