Gas turbine engines are one of the most common technologies for power generation globally. Improving gas turbine efficiencies is a crucial challenge that engine designers face with the ever-increasing effects of climate change and global financial pressures driving this requirement. Rim seals are fitted between the turbine discs and their adjacent casings, at the periphery of the wheel-space. Rim seals are utilised within gas turbines to prevent damage to highly stressed components, they provide this by reducing the degree of ingestion of hot gases into the wheel space.

This thesis describes a computational study using the commercial Computational Fluid Dynamics (CFD) code StarCCM+ to investigate and evaluate the performance of unsteady Reynolds averaged Navier-Stokes (URANS) and wall-modelled large eddy simulation (WMLES) on rim seal flows over a range of operating conditions and rim seal configurations. Initially, both URANS and WMLES models were validated against a simple experimental rotor stator case and a closed axial rim seal model was compared to previously published wall resolved LES (WRLES). Both models show reasonable agreement with rotor torque coefficients and velocity profiles within the seal, WMLES was found to yield flow structures representative of the WRLES.

These modelling strategies were then implemented for axial turbine chute rim seal geometries tested at the Oxford rotor facility and the 1.5 stage axial turbine rig at the University of Bath. The sensitivity of predicted rim seal performance to turbulence modelling was investigated over different flow regimes and compared with previously published experimental and computational data. Evaluations for the Oxford chute seal within the rotationally driven ingestion regime show significant sensitivity to turbulence modelling. Both URANS and WMLES capture aspects of the flow physics, but the larger vortical structures prominent in present URANS models evidently drive ingress deeper into the cavity disagreeing with some previous computational findings. Conversely, improvement in sealing prediction is obtained with the WMLES model compared to the URANS model within the pressure driven regime. The present WMLES captures an improved radial distribution of ingestion within the cavity compared to previous WMLES. For the Bath Chute seal, WMLES improves prediction in seal performance compared to URANS. However, it fails to accurately capture ingestion within the inner cavity, underpredicting ingestion compared to previous URANS results and experimental data.

Next, a systematic study of sealing performance for an axial rim seal over a range of conditions is reported, comparing URANS and WMLES models. For the rotationally driven condition, URANS displays larger, more coherent vortical flow structures than the WMLES. The larger vortices in the URANS drive ingress into the wheel-space resulting in higher levels of ingestion than indicated by the WMLES. A circumferential pressure asymmetry within the annulus was generated using an approximate analytical solution for potential flow with a periodic circumferential variation. For the combined ingestion condition, at the higher purge flow rate WMLES show higher levels of ingestion and flow unsteadiness than II URANS. Whilst in the pressure driven regime reasonable agreement between the two models is obtained. The present results give some explanation for the mixed results reported for the performance of URANS models in previous studies.

The use of the simplified rim sealing models was shown to be an efficient method of ranking seal designs and investigating sensitivity to seal geometry. Four rim seal configurations, two chute seals, an axial seal and a radial seal were investigated. Additionally, two Hybrid seals combining geometric characteristics from both the chute and radial seal were considered. Significant sensitivities of sealing performance to turbulence modelling are identified, but URANS and WMLES show similar trends in the ranking of seal performance. The study was extended to include various geometric modifications including varying overlap, seal clearance, chute angle and the inclusion of a step into the annulus at the end of the chute seal. Both turbulence models indicated that lowering the chute seal angle and inclusion of small step gave significant improvement. The step seal was further evaluated at test rig conditions and showed improved sealing performance compared to the Oxford chute seal. Although rankings of seal performance vary based on the turbulence model, both agree on the best and worst performing seals.

Finally, investigations were performed on the effect of varying both the density and swirl of the annulus and purge flow. Rim seal performance was found to be relatively insensitive to the changes in density. However, a significant decrease in seal performance was found with the inclusion of swirl in the annulus due to the amplified levels of shear between the purge and annulus flow. The addition of swirl within the purge flow had a relatively small effect, improving seal performance slightly.