Investigations of Turbulent Flow in a Centrifugal Compressor Vaned Diffuser by 3-Component Laser Velocimetry

The unsteady 3D impeller exit and vaned diffuser flow of a high-subsonic centrifugal compressor has been investigated with an LDV system. Time-resolved 3D velocity measurements were taken along a streampath at 8 positions from impeller exit downwards through the vaned diffuser and at 18 positions from hub to casing at each station. The compressor was operated at its best point at a rim Mach number of Mu = 0.75. Time-resolved (phase averaged) angle and velocity profiles are presented for 2 positions along the streampath. The time-averaged velocity, deterministic fluctuation intensity, turbulence intensity, and in-plane Reynolds sheer stress profiles, presented for all stations, show the evolution of flow and permit comparisons to in-house CFD calculations to be made. The flow leaving the impeller enters the diffuser with an asymmetric and distorted velocity profile. It is shown that the deterministic fluctuations caused by the jet/wake are quickly damped along the streampath. The results illustrate the deceleration of the flow arriving near the hub in the diffuser channel. The deceleration is accompanied by a sharp increase of turbulence. Near the casing, where the approach velocity is low, no deceleration occurs and the Reynolds stresses are high. Turbulence in the in-plane flow can be regarded as isotropic whereas the axial fluctuations clearly show a high amount of anisotropicity. The narrow diffuser passage required special optical measures for permitting close-to-wall LDV measurements. The experiences are described.

3D Flow Field Measurement Around a Rotating Stall Cell

Rotating stall is an unsteady phenomenon that arises in axial and radial flow compressors. Under certain operating conditions a more or less regular cell of turbulent flow develops and propagates around the annulus at a speed lower than rotor speed. Recently little work has been devoted to the understanding of the flow field pattern inside a rotating cell. However, this knowledge could be of help in the understanding of the interaction between the cell and the surrounding flow. Such information could be extremely important during the modelling process when some hypothesis have to be made about the cell behaviour. A detailed experimental investigation has been conducted during one cell operation of an isolated low-speed axial flow compressor rotor using a slanted hot wire and an ensemble average technique based on the cell revolution time. The three flow field components have been measured on 9 axial section for 800 circumferential points and on 21 radial stations to give a complete description of the flow field upstream and downstream of the rotor. Interpretation of data can give a description of the mean flow field patterns inside and around the rotating cell.

Flow Measurements Behind the Inlet Guide Vane of a Centrifugal Compressor

In a centrifugal compressor adjustable inlet guide vanes (IGV) in front of the impeller are used to regulate the pressure ratio and the mass flow. The stationary measurement of the velocity profile in front of the impeller with different angles of the IGV displays shock losses at the inlet edge of blade of the impeller. In the partial-load region (e.g. partial-load efficiency) the radial distribution of the flow influences considerably the performance of the impeller. The tested compressor consists of an adjustable IGV with straight vanes, a shrouded impeller and a vaneless, parallel diffuser. In the first measurement location, behind the IGV, total pressure, static pressure and flow angle were measured with a 5-hole cylinder probe. In the second measurement location, in front of the impeller, the measurement of the total pressure was carried out with a Kiel probe and the flow angle with a Cobra probe accordingly the static wall pressure was measured. Taking into consideration the fundamental thermodynamical equations it was possible to determine the velocity profiles because of the measured distributions of the flow angle in these two measurement locations. For different angles of the IGV and with various mass flows the distributions of the deflection defect behind the IGV are described. Starting with the measured distributions of the flow in front of the impeller the flow angles at the impeller inlet are calculated and the distributions of the incidence angle at the impeller inlet are figured out.

Behavior of Attractors During Surge in Centrifugal Compression System

A dynamical system analysis method is presented, that permits the characterization of unsteady phenomena in a centrifugal compression system. The method maps one experimental time series of data into a state space in which behaviors of the compression system should be represented, and reconstructs an attractor that geometrically characterizes a state of the compression system. The time series of data were obtained by using a high response pressure transducer and an analog to digital converter at surge condition. For the reconstruction of attractors, a noise free differentiation method in time was employed. The differentiation was made by high order finite difference methods. To remove the influence of noise, the data were passed through a filter using a third order spline interpolation. In this study, the dimension of the state space was restricted to three. The measured data itself and their first and second derivatives in time are used to represent an attractor in the state space. The modeling of the system behavior from the time series of data by second order ordinary differential equations was attempted. It is assumed that the data and their derivatives satisfy the equations at each time. Then, appropriate coefficients are determined by a least square method. The reconstructed attractor showed complex cyclic trajectories at a first glance. However, by applying a band pass filter to the original signal, it was found that the attractor consisted of three independent wave forms and formed an attractor with torus-like behavior. In contrast, the solution by the modeled equations showed a type of limit cycle.

The Effect of the Inlet Boundary Layer on the Tip Flow Field in a Transonic Fan Rotor

Experimental and numerical investigations have been carried out to understand the effects of the inlet boundary layer (IBL) on the tip flow field including the aerodynamic performance in a transonic fan rotor. Both the steady and the unsteady phenomena in the tip flow field have been investigated for operating conditions near peak efficiency and near stall with the two types of tip IBL. In order 10 study these phenomena, high response pressure data with Kulite transducers and laser doppler velocimeter (LDV) data have been acquired around the tip region. Furthermore, three-dimensional Navier-Stokes numerical simulations have been compared with the measured results. The results indicate that the tip IBL significantly influences the spanwise distribution of pressure ratio around the tip region and the stall characteristics including the passage shock / tip leakage vortex interaction, the blockage generation, the wake structure, and the unsteadiness of the tip flow field. In particular, at a near stall condition for the thick IBL with high turbulence intensity level, the tip diffusion level is increased due to a larger blockage, which is generated downstream of a much stronger interaction region. These phenomena are a consequence of the low momentum fluid in the tip IBL, and significantly reduce the stall margin. Furthermore, the unsteadiness drastically increases around the interaction region and around the pressure surface where the blockage migrates. These unsteady phenomena are distinctive features near stall. Downstream of the rotor, the larger and more unsteady blockage is discharged from the pressure surface side, and complicates the three-dimensional rotor exit flow field around the tip region.

The Influence of Endwall Contouring on the Performance of a Turbine Nozzle Guide Vane

The paper presents the results of a detailed investigation of the flow field in a gas turbine linear cascade. A comparison between a contoured and a planar configuration of the same cascade has been performed, and differences in the three-dimensional flow field are here analyzed and discussed. The flow evolution downstream of the trailing edge was surveyed by means of probe traversing while a 3-D Navier-Stokes solver was employed to obtain information on flow structures inside the vaned passages. The experimental measurements and the numerical simulation of the three-dimensional flow field has been performed for two cascades; one with planar endwalls, and the other with one planar and one profiled endwall, so as to present a reduction of the nozzle height. The investigation was carried out at an isentropic downstream Mach number of 0.6. Airfoils of both cascades were scaled from the same high pressure gas turbine inlet guide vane. Measurements of the three-dimensional flow field have been performed on five planes downstream of the cascades by means of a miniaturized five-hole pressure probe. The presence of endwall contouring strongly influences the secondary effects; the vortex generation and their development is inhibited by the stronger acceleration taking place throughout the cascade. The results show that the secondary effects on the contoured side of the passage are confined in the endwall region, while on the flat side the secondary vortices display characteristics similar to the ones occurring downstream of the planar cascade. The spanwise outlet angle distribution presents a linear variation for most of the nozzle height, with quite low values approaching the contoured endwall. The analysis of mass averaged losses shows a significant performance improvement in the contoured cascade. This has to be ascribed not only to lower secondary losses but also to a reduction of the profile losses.

Numerical Investigation of Wake Interaction in a Low Pressure Turbine

A time-accurate Reynolds-averaged Navier-Stokes solver has been extended for a phenomenological study of wake/bladerow interaction in a low pressure turbine near midspan. To qualitatively account for unsteady laminar-turbulent boundary layer transition, a variant of the Abu-Ghanam Shaw transition correlation has been coupled with the Spalart-Allmaras one-equation turbulence model. The method is shown to be capable of capturing separated-flow and wake-induced transition, as well as becalming and relaminarization effects. The model turbine investigated consists of three stator and two rotor rows. Instantaneous Mach number and eddy-viscosity plots are presented to monitor the wake migration and interaction with downstream boundary layers. Especially on the suction sides, very large fluctuations of the skin friction coefficient are observed. Effects of the near and far wakes are identified.

Design and Testing of Swept and Leaned Outlet Guide Vanes to Reduce Stator-Strut-Splitter Aerodynamic Flow Interactions

Large circumferential varying pressure levels produced by aerodynamic flow interactions between downstream stators and struts present a potential noise and stability margin liability in a compression component. These interactions are presently controlled by tailoring the camber and/or stagger angles of vanes neighboring the fan frame struts. This paper reports on the design and testing of a unique set of swept and leaned fan outlet guide vanes (OGVs) that do not require this local tailoring even though the OGVs are closely coupled with the fan frame struts and splitter to reduce engine length. The swept and leaned OGVs not only reduce core-duct diffusion, but they also reduce the potential flow interaction between the stator and the strut relative to that produced by conventional radial OGVs. First, the design of the outlet guide vanes using a single bladerow three-dimensional viscous flow analysis is outlined. Next, a two-dimensional potential flow analysis was used for the coupled OGV-frame system to obtain a circumferentially non-uniform stator stagger angle distribution to further reduce the upstream static pressure disturbance. Recognizing the limitations of the two-dimensional potential flow analysis for this highly three-dimensional set of leaned OGVs, as a final evaluation of the OGV-strut system design, a full three-dimensional viscous analysis of a periodic circumferential sector of the OGVs, including the fan frame struts and splitter, was performed. The computer model was derived from a NASA-developed code used in simulating the flow field for external aerodynamic applications with complex geometries. The three-dimensional coupled OGV-frame analysis included the uniformly-staggered OGVs configuration and the variably-staggered OGVs configuration determined by the two-dimensional potential flow analysis. Contrary to the two-dimensional calculations, the three-dimensional analysis revealed significant flow problems with the variably-staggered OGVs configuration and showed less upstream flow non-uniformity with the uniformly-staggered OGVs configuration. The flow redistribution in both the radial and tangential directions, captured fully only in the three-dimensional analysis, was identified as the prime contributor to the lower flow non-uniformity with the uniformly-staggered OGVs configuration. The coupled three-dimensional analysis was also used to validate the design at off-design conditions. Engine test performance and stability measurements with both uniformly- and variably-staggered OGVs configurations with and without the presence of inlet distortion confirmed the conclusions from the three-dimensional analysis.

The Effect of Tip Clearance Gap Size and Wall Rotation on the Performance of a High-Speed Annular Compressor Cascade

The performance of a high speed annular compressor cascade for different clearance gap sizes, with stationary or rotating hub wall is investigated. Five hole probe measurements, conducted at the inlet and outlet of the cascade, are used to derive blade performance characteristics, in the form of loss and turning distributions. Characteristics are presented in the form of circumferentially mass averaged profiles, while distributions on the exit plane provide information useful to interpret the performance of the blading. Static pressure distributions on the surface of the blades as well as inside the tip clearance gap have also been measured. A set of four clearance gap sizes, in addition to zero clearance data for the stationary wall, gives the possibility to observe the dependence of performance characteristics on clearance size, and establish the influence of rotating the hub. Overall performance is related to features of the tip clearance flow. Increasing the clearance size is found to increase losses in the clearance region, while it affects the flow in the entire passage. Wall rotation is found to improve the performance of the cascade.

Surge in a Low-Speed Radial Compressor

Surge is measured in a low-speed radial compressor with a vaned diffuser. For this system, the flow coefficient at surge is determined. This coefficient is a measure for the inducer inlet flow angle and is found to increase with increasing rotational speed. Moreover, the frequency and amplitude of the pressure oscillations during fully-developed surge are compared with results obtained with the Greitzer lumped parameter model. The measured surge frequency increases when the compressor mass flow is throttled to a smaller flow rate. Simulations show that the Greitzer model describes this relation reasonably well except for low rotational speeds. The predicted amplitude of the pressure rise oscillations is approximately two times too small when deep surge is met in the simulations. For classic surge, the agreement is worse. The amplitude is found to depend strongly on the shape of the compressor and throttle characteristic, which are not accurately known.