Search

Geometrical uncertainties in a compressor (due to manufacturing tolerance and/or in-service degradation) often result in flow asymmetry around the annulus of a compressor that jeopardises compressor stability and performance. Usually, sensitivity of compressor stability and performance for any parametric variation is arrived at by considering all blades to have same dimension. In reality, an inherent blade-to-blade variation causes the blades to have a probability distribution. These blades can be redistributed circumferentially resulting in adjacent passage areas between different blades to be completely random and hence the performance variation. Surrogate model is preferred for quantifying the effects of parametric variation on compressor stability and performance given its quick turnaround time vis-a-vis CFD and experiments. In this thesis, uncertainties for three test cases were considered: each representative of fans on military aircraft engines, fans on civil aircraft engines and a 1-stage transonic compressor used in industrial gas turbine. This research establishes a rule of thumb to arrange blades of differing dimensions around the compressor to eke out maximum performance and stability margin. The parameters tip gap and stagger angle represent manufacturing tolerance while in-service degradation was represented by leading edge damage. For both random tip gap variation (0.15% to 0.94% span) and random leading edge damage (4% to 18% chord), the compressor performance and stability boundaries were found to be best with a zigzag pattern of blade arrangement and worst with a sinusoidal pattern of arrangement. The converse was found to be true for blades having random stagger angle variation (± 2.25% change in nominal stagger angle). The best/worst arrangement of blades with differing dimensions was ascertained using a mix of CFD and travelling salesman (TSP) analogy. The TSP analogy is handy for determining the best arrangement when two or more parameters vary simultaneously. Generalised surrogate model was developed to accurately predict the performance of compressors undergoing random tip gap and stagger angle variation. Due to its robustness, the surrogate model was combined with Monte Carlo technique to gauge the impact of parametric variation on quantities of interest (QoI). The mean absolute percentage error between CFD and surrogate models of stagger angle and tip gap (for different QoI) were found to be less than 0.14% and 1.5% respectively. This de novo analysis considers only the aerodynamic effect from geometric variations while neglecting the associated aeroelastic effects. Detailed analyses based on past experience and physical reasoning were used to validate the numerical simulations.

The design of modern gas-turbine engines is continuously being improved towards better performance, better efficiency and reduced cost. This trend of aero-engine design requires compression systems which produce higher pressure ratios and thus, have higher loaded blades and a closer spacing between blade rows. Such designs are more prone to aerodynamic instabilities and consequent stall and surge can be catastrophic. The majority of the research conducted on compressor stall and surge is limited to old designs with lower pressure ratios or single stage compression systems. In this thesis, the near and post stall behaviour of a modern multi-stage high-speed intermediate pressure compressor rig and an aero-engine three-shaft compression system are studied in detail. The main objective is to develop and validate reliable CFD models to predict surge and rotating stall and shed light on the underlying physical mechanism of the phenomena. CFD computations were performed to gain understanding of the current capability to model the flow behaviour of a multi-stage compressor rig near stall condition. Two turbulence models were tested and an extensive grid discretization study was performed. In order to improve the prediction of the compressor's stability boundary, a modification in the widely known Spalart-Allmaras turbulence model is proposed. Subsequently, unsteady CFD computations were carried out to evaluate the impact of flow unsteadiness in the performance prediction of this compressor rig. It was found that for operating conditions characterized by non-axisymmetric flow features, an unsteady full annulus model is required to predict the compressor performance. For low speeds, these flow features develop over a wide range of operating conditions. When the compressor operates at high speeds, these flow features are limited to operating conditions near the stability boundary. The above findings were validated against experimental results. Early stages of this research revealed that numerical calibration of a CFD surge computation in a three-shaft engine is a challenging task due to compressor matching. Hence, an iterative methodology for matching the compressors was introduced and validated against experimental data. This study considered a surge event where the engine was initially operating at mid power condition. When comparing the numerical result with measured data, it was found that the engine bleed system has a major impact on the aerodynamic loading predictions in the core system. Therefore, this system needs to be considered by component designers when accounting for robustness to surge loads. The post stall response of a three-shaft engine compression system which is initially operating at design was investigated. It was found that the maximum surge over-pressures are caused by a combined effect between a surge induced shock wave and high pressure gas travelling towards the core inlet during the surge blow-down period. Furthermore, it was demonstrated that the maximum surge loads are obtained for a surge event initiated by fuel-spike. Finally, a cheaper computational approach to model surge in axial compression systems is proposed. This approach consisted of using an unsteady single passage model to predict the flow behaviour during the surge event. After comparison with full annulus results for three different scenarios, it was concluded that the single passage is capable of predicting the blow-down period of surge which is characterized by a long period of flow reversal. This model fails to predict the correct time and length scales during surge onset and flow transition between reverse to forward flow at the beginning of recovery. These time instants are characterized by non-axisymmetric flow features. However, the single passage model shows a good correlation with the results obtained using a full annulus model for estimation of average values of static pressure and mass flows during surge. This can drastically reduce simulations times from months to days during compressor surge analysis.

Due to manufacturing tolerance and deterioration during operation, fan blades in the same engine exhibit geometric variability. The absence of symmetry will inevitably exacerbate and contribute to the complexities of running geometry prediction as the blade variability is bound to be amplified by aerodynamic and centrifugal loading. In this study, the fan blade untwist (which is the blade deformation between its static condition and running condition) related phenomenon known as Alternate Passage Divergence (APD) is addressed. As the name suggests, APD manifests as alternating passage geometry (and hence alternating tip stagger pattern) when the fan stage is operating close to/at peak efficiency condition. APD can introduce adverse influence on fan performance, aeroacoustics behaviour, and high cycle fatigue characteristics of the blade. In this study, the APD behaviours of two transonic fan blade designs are compared. The main objective of the study is to identify the parameters contributing to the APD phenomenon. After the formation of alternating tip stagger pattern, APD's unsteady effect can cause the blades from one group (segmented by tip stagger angle) to switch to the other, creating a travelling wave pattern around the circumference. It was found from numerical assessment on a randomly mis-staggered assembly that real engines can potentially experience such travelling disturbance and suffer fatigue damage. The phenomenon is termed APD-induced Non-Synchronous Vibration (NSV) and is abbreviated as NSV in this study. An idealised case is used to capture the bulk behaviour from the more complex cases in real engines and to decipher the underlying mechanism of this travelling disturbance. The results indicate that the driving force originates from the interaction between passage shock displacement and the passage geometry. Based on the findings on APD & NSV, vibration attenuation methods are explored. Using machine learning techniques, a passive attenuation method is found to minimise the chance of NSV manifestation for a given set of fan blades. Alternatively, active attenuation method is implemented through blade redesign which modifies the passage geometry.

Inlet distortion can lead to loss in efficiency and stability margin of the fan which in return jeopardises flight safety. These aspects driven by inlet distortion are becoming increasingly challenging for the designers of next-generation turbofan engines. With the increase of computational capability and improvements in numerical models, computational fluid dynamics (CFD) is becoming increasingly powerful and favoured by scientific researchers and industrial engineers. Time-accurate, high-fidelity CFD simulations of the compressor behaviour at extreme operational conditions (such as during stall with inlet distortion) has become possible. In the present thesis, CFD is used to determine the effects of inlet distortion on fan aerodynamic stability and stall hysteresis. NASA stage 67 is used for this study. At the very beginning an appropriate numerical strategy was developed and validated with extensive experimental data. A good match was obtained for both the flow field variables at the peak efficiency point and the stall boundary. In this research, two types of inlet distortion were examined. One is the consistent distortion which mimics the setup in experiments with slow throttling; another type is the abrupt distortion (due to sudden maneuver) whose effect is poorly understood. It will be shown that abrupt distortion can result in larger stall margin loss than consistent distortion. Therefore, experimental tests based on consistent distortion tend to be more optimistic in stall margin prediction. Thereafter the stall and recovery process of a transonic fan with both types of inlet distortions were performed. The results showed that the stall process with inlet distortion can be very different from that in uniform inflow. However, distortion has minor effect on the recovery point (corrected mass flow) of the fan and the clean flow region plays the most important role in the recovery process. In the presence of abrupt distortion, it was found that the stall margin of a fan can be influenced by the length of exit duct. This phenomenon was explained using the wave propagation theory. A shorter exit duct reduces the time lag of expansion pressure wave reflected from the nozzle, which upon arriving at fan trailing edge can prevent the fan from stalling. A critical length ratio was proposed which provides useful guidelines on test rig and engine design. A preliminary study of the behaviour of the BLI fan with serpentine intake (S intake) at near stall condition and its stall process was performed. It was found that the distortion pattern upstream of the fan is complex and can be divided into different zones radially. The stall behaviour of the fan is similar to that with a circumferential distortion, but more complex because of the coupling with the swirl distortion near the casing. Although the present work is restricted to NASA stage 67, some of the conclusions gained are general and expected to be valid for modern fan and compressor designs. Finally, during this research it has become apparent that there is a significant lack of open published measured data for fans and compressors operating under inlet distortions, which is mainly due to the difficulties and costs involved in setting up such experimental campaigns. The above indicates that validated CFD codes are going to play an even more important role in development of distortion tolerant fans. The objective of this work is to show the suitability of CFD for the modelling of fan aerodynamic performance and stability with inlet distortion, which can provide an economical alternative strategy to subscale rig tests.

Tunnels are increasingly used in high-speed rail projects to mitigate against issues such as environmental noise, land disputes, and unsuitable terrain. However, the trend for increasing train speeds will result in unacceptable noise emissions from tunnels without the use of effective countermeasures. Novel countermeasures for the propagation of pressure waves in tunnels and the emission of sound waves into the environment, commonly referred to as micro-pressure waves, were numerically investigated in this work. This following countermeasures were considered: (1) the design and optimisation of an array of Helmholtz resonators embedded in redundant tunnel space; (2) a preliminary parametric study on the effect of modifying the junction geometry between the tunnel and side branches (e.g. ventilation shafts) for noise emissions from side branches. Helmholtz resonators are used extensively in engineering disciplines where noise attenuation is an important factor (e.g. jet-engine liners). However, their ability to suppress noise emissions from tunnels has not been demonstrated. This work investigates the effectiveness of these countermeasures when applied to a representative tunnel system and compares their performance to existing ones (e.g. tunnel entrance hoods) using numerical techniques. One and two-dimensional models were developed to predict the performance of these countermeasures, subject to realistic geometric constraints and operating conditions. The geometry of the array is optimised to provide robust performance over a range of operating conditions. The numerical predictions are validated against experimental data, and are benchmarked against analytical predictions and CFD. Finally, the combination with existing countermeasures is studied and enhancements to the models are proposed. Both countermeasures were found to work effectively for a physically representative system.

The design and operation of gas-turbine engines is heavily influenced by the off-design stability of the compressor, which limits obtainable performance and may result in aeroelastic vibration issues. Whilst Variable Stator Vanes (VSVs) are widely used to mitigate this problem, research considering the mismatching effect of VSVs away from their optimal settings is limited. In this thesis, a high-speed variable geometry compressor is studied at part-speed conditions and VSV setting adjustments are made to deliberately trigger stable rotating stall and study its behaviour. Examination of unsteady measurements reveals two "families'' of rotating stall, each at different frequencies, where the dominant behaviour depends upon the VSV settings. Stall in the front stage is shown to consist of a spatially non-uniform and time-varying pattern of short lengthscale cells, which couple with rotor vibration and propagate as noise. Second stage stall is a longer lengthscale uniform disturbance consisting of fewer stall cells. The stalling pressure amplitudes are also found to correlate well to blade loading parameters from a one-dimensional meanline model. Steady and unsteady CFD simulations at these stalled conditions confirm that the behaviour is due to regions of stall in the front stage tip region together with the hub of the second stage. These CFD calculations naturally result in the formation of stall cells and give a credible match to the experiment. Inviscid reasoning explains how this flowfield is due to spanwise static pressure gradients arising from part-speed closure of the VSVs. Finally, the non-dimensional cell propagation speed (Vstall/U) for each family of stall is shown to be uniquely determined by the VSV settings. This appears to be linked to the axial flow velocity local to the cell and suggests that cell speed may be restated in a more universal non-dimensional form. Furthermore, simulations show the importance of flowfield coupling mechanisms in determining the number of stall cells, which are also driven largely by the VSV settings.

Modern gas turbine design continues to drive towards improved performance, reduced weight and reduced cost. This trend of aero-engine design results in thinned blade aerofoils which are more prone to aeroelastic problems such as flutter. Whilst extensive work has been conducted to study the flutter of isolated turbomachinery blades, the number of research concerning the unsteady interactions between the blade vibration, the resulting acoustic reflections and flutter is very limited. In this thesis, the flutter of such embedded blade rows is studied to gain understanding as for why and how such interactions can result in flutter. It is shown that this type of flutter instability can occur for single stage fan blades and multi-stage core compressors. Unsteady CFD computations are carried out to study the influence of acoustic reflections from the intake on flutter of a fan blade. It is shown that the accurate prediction of flutter boundary for a fan blade requires modelling of the intake. Different intakes can produce different flutter boundaries for the same fan blade and the resulting flutter boundary is a function of the intake geometry in front of it. The above finding, which has also been demonstrated experimentally, is a result of acoustic reflections from the intake. Through in-depth post-processing of the results obtained from wave-splitting of the unsteady CFD solutions, the relationship between the phase and amplitude of the reflected acoustic waves and flutter stability of the blade is established. By using an analytical approach to calculate the propagation and reflection of acoustic waves in the intake, a novel low- fidelity model capable of evaluating the susceptibility of a fan blade to flutter is proposed. The proposed model works in a similar fashion to the Campbell diagram, which allows one to identify the region (in compressor map) where flutter is likely to occur at early design stages of an engine. In the second part of this thesis, the influence of acoustic reflections from adjacent blade rows on flutter stability of an embedded rotor in a multi-stage compressor is studied using unsteady CFD computations. It is shown that reflections of acoustic waves, generated by the rotor blade vibration, from the adjacent blade rows have a significant impact on the flutter stability of the embedded rotor, and the computations using the isolated rotor can lead to significant over-optimistic predictions of the flutter boundary. Based on the understanding gained, an alternative strategy, aiming to reduce the computational cost, for the flutter analysis of such embedded blades is proposed. The method works by modelling the propagation and reflection of acoustic waves at the adjacent blade rows using an analytical method, whereby flutter computations of the embedded rotor can be performed in an isolated fashion by imposing the calculated reflected waves as unsteady plane sources. Computations using the proposed model can lead to two orders of magnitude reduction in computational cost compared with time domain full annulus multi-row computations. The computed results using the developed low-fidelity model show good correlation with the results obtained using full annulus multi-row models.