Abstract
Ingress is the ingestion of the hot gases from the main flow-path of a gas turbine engine into the cavity (wheel-space) formed between the rotor disc and the stationary components. A superposed sealing flow, typically bled from the compressor, and rim seals fitted to the periphery of the wheel-space offer protection to the highly stressed components within. However, as these measures can negatively impact the engine’s overall efficiency, design is critical. Parasitic leakage flows can enter the wheel-space via the stationary casing due to mating interfaces in the engine’s hardware. These flows can affect the fluids structure in the wheel-space impacting ingress and thus, the performance of the engines rim seals. Owing to the aggressive thermal and centrifugal loading experienced during the turbine operating cycle, the degree of leakage and its effect on ingress are difficult to predict.This thesis describes a series of experimental studies (some supported by computations) with simplified and engine realistic geometries. A new experimental facility is presented in order to conduct the engine realistic experiments and are assessed over a variety of sealing and leakage configurations. These experiments are conducted to consider the potential for leakage flows to be modified in order to minimise their parasitic effect on disc cooling, and ultimately engine performance. The performance of geometric wheel-space modifications is also considered. Measurements of static and total pressure, swirl and species concentration were used to assess the performance and flow structure of the wheels-spaces over a range of leakage-sealing ratios and non-dimensional flow-rates. Measurements of unsteady pressure are used to assess the presence of large scale structures in the engine realistic geometry.
In the geometrically simple wheel-space, with only leakage flows, data is presented to investigate the effects of swirling the leakage flow either with or against the disc rotation. When swirled with the rotor the sealing effectiveness was enhanced by up to 15%, due to the promotion of swirl near the rim gap compared to the axially-introduced baseline and counter-swirled introduction. The axially injected momentum from the leakage created a toroidal vortex in the outer part of the cavity. By modifying the axially injected leakage outlet port the momentum of the leakage flow could be manipulated. An increase in momentum resulted in an increase the influence of the toroidal vortex on the flow structure in the wheel-space, negatively impacting the sealing effectiveness.
An overview of the design, construction and commissioning of a new single stage facility based around realistic aircraft engine geometry provided by the programmes sponsors Safran Aircraft Engines is presented.
Experiments were conducted in the annulus and wheel-space of the aforementioned facility with no leakage flow. Measurements in the annulus fluid were found to be consistent with other experimental facilities. A Batchelor style flow regime in the upper section of the inner wheel-space and a mixing region in the outer wheel-space is proposed. A greater circumferential variation of ingress in the outer wheel-space as opposed to the inner wheel-space was observed. These results set a baseline for comparisons with further leakage configurations and geometries. Rotating structures were not definitively identified in the wheel-space but supporting computations suggested that the wheel-space geometry suppressed their propagation.
Leakage flows were introduced to the wheel-space of the engine realistic facility at a variety of leakage-sealing flow ratios. Measurements related to the sealing effectiveness suggested that leakage flow is detrimental to sealing performance, compared to the baseline, when expressed in terms of total secondary flow through the wheel-space. However, when expressed in terms sealing flow, increasing leakage-sealing flow ratio generally shows a performance benefit. Introducing leakage flow only had the most dramatic decrease in sealing performance. Leakage flows had no dramatic effect upon flow structure when compared to the baseline.
Geometric modifications were made to the wheel-space of the engine realistic facility with sealing flow only. The first two configurations examined the effect of either increasing or decreasing the axial gap of the wheel-space. Decreasing the axial gap resulted in a decrease the levels of ingestion whilst increasing the axial gap resulted in increased ingestion, when compared to the baseline. The effect could be accounted for by redefining the seal clearance term in the non-dimensional sealing flow parameter. These changes did not affect the flow structure when compared to the baseline. The final configuration involved decreasing the axial overlap of the fins on the rotor and stator cover plates. This modification had no effect in the inner wheel-space and only a marginal one in the outer in terms of sealing performance and no effect on flow structure when compared to the baseline.
| Date of Award | 28 Mar 2024 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Sponsors | Safran Aircraft Engines |
| Supervisor | Carl Sangan (Supervisor), James Scobie (Supervisor) & Gary Lock (Supervisor) |