Rinaldo L. Miorini

The Internal Structure of the Tip Leakage Vortex Within the Rotor of an Axial Waterjet Pump

The complex flow field in the tip region of a turbomachine rotor, including the tip leakage flow and tip leakage vortex (TLV), has been studied for decades. Yet many associated phenomena are still not understood. This paper provides detailed data on the instantaneous and phase-averaged inner structures of the tip flow and evolution of the TLV. Observations are based on series of high resolution planar particle image velocimetry measurements performed in a transparent waterjet pump fitted into an optical refractive index-matched test facility. Velocity distributions and turbulence statistics are obtained in several meridional planes inside the rotor. We observe that the instantaneous TLV structure is composed of unsteady vortex filaments that propagate into the tip region of the blade passage. These filaments are first embedded into a vortex sheet, which is generated at the suction side of the blade tip, and then they wrap around each other and roll up into the TLV. We also find that the leakage vortex induces flow separation at the casing endwall and entrains the casing boundary layer with its counter-rotating vorticity. As it propagates in the rotor passage, the TLV migrates toward the pressure side of the neighboring blade. Unsteadiness associated with vortical structures is also investigated. We notice that, at early stages of the TLV evolution, turbulence is elevated in the vortex sheet, in the flow entrained from the endwall, and near the vortex core. Interestingly, the turbulence observed around the core is not consistent with the local distribution of turbulent kinetic energy production rate. This mismatch indicates that, given a TLV section, production likely occurs at preceding stages of the vortex evolution. Then, the turbulence is convected to the core of the TLV, and we suggest that this transport has substantial component along the vortex. We observe that the meandering of vortex filaments dominates the flow in the passage and we decompose the unsteadiness surrounding the TLV core to contributions from interlaced vortices and broadband turbulence. The two contributions are of the same order of magnitude. During late stages of its evolution, TLV breakdown occurs, causing rapid broadening of the phase-averaged core, with little change in overall circulation. Associated turbulence occupies almost half the width of the tip region of blade passage and turbulence production there is also broadly distributed. Proximity of the TLV to the pressure side of the neighboring blade also affects entrainment of flow into the incoming tip region.

Turbulence Within the Tip-Leakage Vortex of an Axial Waterjet Pump

Stereoscopic particle image velocimetry measurements are performed in an optical refractive index matched facility to investigate the evolution of turbulence in the tip region of an axial waterjet pump rotor. Presented analysis of mean flow velocity, vorticity, Reynolds stresses, and turbulence production/transport within the rotor passage focus on the tip-leakage vortex and associated flows. Turbulence production peaks in the shear layer that connects the blade-tip suction side with the vortex as well as in a region of flow contraction situated at the casing wall. Flow separation occurring there, as the leakage backflow meets the throughflow, detaches the boundary-layer vorticity, which is entrained into the tip-vortex perimeter. Upon the inclusion of turbulence transport in the analysis, a discrepancybetweendistributions of turbulent stresses andassociated production vanishes, except at the vortex core. There, the elevated turbulent energy (but relatively low production of Reynolds stresses) is presumably due to low dissipation. Within the aft part of the rotor passage, shortly after vortex bursting, the tip-leakage backflow reaches the neighboring blade. There, radial motion induced by the tip-vortex residual swirl detaches the pressure-side boundary-layer vorticity and injects it into the rotor passage.

Flow Visualization Using Cavitation Within Blade Passage of an Axial Waterjet Pump Rotor

Cavitation phenomena within an axial waterjet pump, AxWJ-2 [1,2] operating at and below the best efficiency point (BEP) are investigated using high-speed imaging. The purpose of these preliminary observations is to provide an overview of the physical appearance of several forms of cavitation under varying flow and pressure conditions. These observations provide a motivation for upcoming detailed velocity and turbulence measurements. The experiment is conducted using a transparent pump installed in an optically index-matched facility, which facilitates unobstructed visual access to the pressure and suction sides of the rotor and stator blade passages. By varying the cavitation index within the facility, the observations follow the gradual development of cavitation from inception level to conditions under which the cavitation covers the entire blade. Cavitation appears first in the tip gap, as the fluid is forced from the pressure side (PS) to the suction side (SS) of the rotor blade. Bubbly streaks start at the SS corner, and penetrate into the passage, and are subsequently entrained into the tip leakage vortex (TLV) propagating in the passage. Sheet cavitation also develops along the SS of the rotor leading edge and covers increasing fractions of the blade surface with decreasing cavitation number. At BEP conditions, the sheet is thin. Below BEP, the blade loading increases as a result of an increase in the incidence angle of the flow entering the passage relative to the blade. Consequently, the backward leakage flow also increases, further increasing the incidence angle in the tip region, and thickening the sheet cavitation there. Consistent with previous observations on swept hydrofoils, a re-entrant jet that flows radially outward develops at the trailing edge of the sheet cavitation. Only near the tip corner the trailing edge of the sheet cavitation is opened as the radial re-entrant flow is entrained into the TLV, forming an unstable and noisy spiraling pattern. Within a certain range of cavitation indices, when the sheet cavitation length at the blade tip extends to about 50–60% of the blade spacing, the sheet cavitation on every other blade begins to expand and contract rapidly, generating loud low-frequency noise. With further decrease in pressure, persistent alternating cavitation occurs, namely, the cavitating region on one blade becomes much larger than that in the neighboring one. The mechanisms involved and associated instabilities are discussed based on previous analyses performed for inducers. As the cavitation number is lowered even further, the sheet cavitation on the “heavily-cavitating” blade grows, and eventually passes the trailing edge of the rotor blade. At this condition, cavitation begins again to expand and contract rapidly on the “less-cavitating” blade, covering a significant portion of SS surface. At a lower pressure, all the blades cavitate, with the sheet cavitation covering the entire SS surface of the rotor blade. The large cavities on alternate rotor blade surfaces re-direct flow into the neighboring passages with the smaller cavities. As a result, there is a lower flow rate in the passage with the larger cavitation and higher flow rate in the neighboring passage. As the flow with the cavitating passage arrives to the leading edge of the stator flow rate, it increases the incidence angle at the entrance to the stator, causing intermittent sheet and cloud cavitation on the stator blade.Copyright

Analysis of Turbulence in the Tip Region of a Waterjet Pump Rotor

A series of high resolution planar particle image velocimetry measurements performed in a waterjet pump rotor reveal the inner structure of the tip leakage vortex (TLV) which dominates the entire flow field in the tip region. Turbulence generated by interactions among the TLV, the shear layer that develops as the backward leakage flow emerges from the tip clearance as a “wall jet”, the passage flow, and the endwall is highly inhomogeneous and anisotropic. We examine this turbulence in both RANS and LES modelling contexts. Spatially non-uniform distributions of Reynolds stress components are explained in terms of the local mean strain field and associated turbulence production. Characteristic length scales are also inferred from spectral analysis. Spatial filtering of instantaneous data enables the calculation of subgrid scale (SGS) stresses, along with the SGS energy flux (dissipation). The data show that the SGS energy flux differs from the turbulence production rate both in trends and magnitude. The latter is dominated by energy flux from the mean flow to the large scale turbulence, which is resolved in LES, whereas the former is dominated by energy flux from the mean flow to the SGS turbulence. The SGS dissipation rate is also used for calculating the static and dynamic Smagorinsky coefficients, the latter involving filtering at multiple scales; both vary substantially in the tip region, and neither is equal to values obtained in isotropic turbulence.Copyright

PIV Measurements of the Flow in the Tip Region of a Compressor Rotor

An axial turbomachine, adapted from the NASA Glenn Low-Speed Axial Compressor (LSAC), has been assembled for detailed flow and turbulence measurements in the JHU optical refractive index matched facility. The test section consists of a row of twenty inlet guide vanes (IGV), followed by fifteen rotor blades, then twenty stator blades. The blades have the same geometry, but lower aspect ratio as the inlet guide vanes and the first stage of the LSAC facility at NASA Glenn. Although smaller in scale, the Reynolds number based on the tip speed and rotor blade chordlength are comparable to those of the LSAC. The casing, rotor blades, as well as half the IGVs and stator blades are made of transparent acrylic, matched with the refractive index of the working fluid, a concentrated solution of sodium iodide and water. The facility is designed to allow optical flow measurements in all three blade rows and from all directions. Results presented in this paper are based on 2D PIV measurements focusing on the flow structure in the tip region of the rotor blade for two flow rates, one of them just above the stall level. Included are phase-averaged distributions of velocity, circumferential vorticity, and turbulent kinetic energy in several meridional planes dissecting the tip at different chordwise locations. They follow the evolution of the Tip Leakage Vortex (TLV) as it rolls up near the blade suction side, migrates across the rotor passage, and subsequently bursts. Upon bursting, the distinct high vorticity core is replaced by a broad region of elevated vorticity, which occupy a substantial fraction of the passage. Sample instantaneous realizations and higher resolution stereo-PIV data are also provided. The turbulent kinetic energy is high near the vortex core, in the shear layer connecting the vortex to the blade tip, and around the point of flow separation on the endwall casing, where the leakage flow meets the main passage flow.© 2014 ASME

Visualization and Time Resolved PIV Measurements of the Flow in the Tip Region of a Subsonic Compressor Rotor

A new optically index matched facility has been constructed to investigate tip flows in compressor-like settings. The blades of the one and a half stage have the same geometry, but lower aspect ratio as the inlet guide vanes and the first stage of the LSAC facility at NASA Glenn. With transparent blades and casings, the new setup enables unobstructed velocity measurements at any point within the tip region, and is designed to facilitate direct measurements of effects of casing treatments on the flow structure. We start with a smooth endwall casing. High speed movies of cavitation and time-resolved PIV measurements have been used to characterize the location, trajectory, and behavior of the Tip Leakage Vortex (TLV) for two flow rates, the lower one representing pre-stall conditions. Results of both methods show consistent trends. As the flow rate is reduced, TLV rollup occurs further upstream, and its initial orientation becomes more circumferential. At pre-stall conditions, the TLV is initially aligned slightly upstream of the rotor passage, and subsequently forced downstream. Within the passage, the TLV breaks up into a large number of vortex fragments, which occupy a broad area. Consequently, the cavitation in the TLV core disappears. With decreasing flow rate, this phenomenon becomes more abrupt, occurs further upstream, and the fragments occupy a larger area.Copyright

Three-Dimensional Structure and Turbulence Within the Tip Leakage Vortex of an Axial Waterjet Pump

The flow structure and dynamics of turbulence are investigated by means of three-dimensional stereo particle image velocimetry (Stereo-PIV) measurements within the tip leakage vortex (TLV) of an axial waterjet pump rotor. Both the blades and casing of the pump are transparent and their optical refractive indices are matched with that of the pumped fluid, providing unobstructed optical access to the sample area without image distortion. Data are acquired on selected meridional planes in the rotor passage as well as in three-dimensional domains obtained by stacking closely-spaced planes situated within the rotor passage. Presented data have been sampled in one of these 3D regions, at 67% of the blade tip chordlength. All components of velocity and vorticity are calculated, together with the whole strain-rate and Reynolds stress tensors. The entire set of contributors to the turbulence production-rate is also available. The TLV and associated flow structures are completely 3D and change significantly along the blade tip chordwise direction. The vortex originates from the rollup of a multi-layered tip leakage flow, and propagates within the rotor passage towards the neighboring blade. Because of layered backflow rollup, vorticity entrained in the TLV is convected along different paths and re-oriented several times within the vortex. As a result, the TLV consists of a core surrounded by a tube of three-dimensional vorticity that wraps around it helically. Propagation of tip leakage backflow into the passage and subsequent TLV rollup also cause flow separation at the casing endwall with ejection of boundary layer vorticity that is finally entrained into the outer perimeter of the TLV. This complex TLV flow dominates the tip region of the rotor and involves non-uniform distributions of strain-rate and Reynolds stresses resulting in well-defined peaks of turbulence production-rate. For instance, turbulence is produced locally both at the flow contraction point near the region of aforementioned endwall separation and in the shear layer that connects the vortex with the suction side corner of the blade tip. The spatial inhomogeneity of turbulent kinetic energy (TKE) distribution within the TLV, and the mismatch between locations of TKE and production-rate peaks can be explained by analyzing the 3D mean flow advection of turbulence, for example from the region of endwall boundary layer separation towards the outer region of the TLV. In addition to being spatially non-uniform, turbulence is also anisotropic in both the shear layer and periphery of the TLV. Conversely, turbulence is intense and relatively isotropic near the TLV core, as well as monotonically increasing along the vortex centerline. This trend cannot be described solely in terms of local production of turbulence; it must also involve slow turbulence dissipation associated with the meandering of relatively large-size, interlaced vortex filaments in the TLV core region.Copyright

Simultaneous Measurements of 3D Flow and Wall Deformation in a Compliant Wall Turbulent Channel Flow

This study focuses on the interaction of a turbulent channel flow at Reτ=2310 over a flat, compliant boundary made of PDMS (Polydimethylsiloxane). Two noninvasive optical techniques, namely tomographic PIV (TPIV) and Mach-Zehnder Interferometry (MZI), are integrated to perform simultaneous measurements of the 3D flow field and the corresponding surface deformation. The measurements are performed in a refractive index-matched facility, where the working fluid is aqueous solution of sodium iodide (NaI). The TPIV measurement volume is 30×10×10 mm3 in the streamwise, wall-normal and spanwise directions, respectively. The MZI phase evaluation and unwrapping algorithms have been developed, calibrated and implemented. Preliminary results show qualitative correlation between wall deformation and flow structures near the wall.Copyright

Time-Resolved Simultaneous Measurements of Flow Field and Surface Deformation of Turbulent Channel Flow Over a Compliant Wall

As an initial step in our effort to investigate the interaction of a turbulent channel flow with a compliant wall, this paper focuses on the measurement techniques. Two noninvasive optical techniques, namely tomographic PIV (TPIV) and Mach-Zehnder Interferometry (MZI), are integrated to simultaneously measure the time-resolved, wall-normal deformation of the compliant transparent wall and the 3D velocity field of a turbulent channel flow above it. The two systems utilize the same laser, but different cameras. The paper provides a description of the optical setup, detailed information about calibration of the MZI system, as well as sample combined 3D velocity distributions and wall deformations. The measured wall deformation can be decomposed into low frequency structure modes, and higher frequency features that appear to advect with the flow.Copyright

On the Determination of the Transfer Function of Infinite Line Pressure Probes for Turbomachinery Applications

This paper deals with the use of the infinite line pressure probes (ILP) to measure fluctuating pressures in hot environments in turbomachinery applications. These probes, sometimes called waveguide measuring systems, and composed of a series of lines and cavities are using a remote pressure sensor. Ideally they should form a non-resonant system. This is however not always the case and the frequency response of these systems is of course limited by the tubing (diameter and length) but is also highly dependent on other geometrical parameters like sudden expansions or discontinuities in the tubing, or parasite cavities. The development of a new model for ILP simulation, based on the analogy between the propagation of the pressure waves in a line-cavity system and the electrical transmission line, is presented. Unlike the models based on the Bergh and Tijdeman equations, this approach allows the simulation of systems presenting parallel branches. This makes the model appropriate for the prediction of the frequency response of ILP. The model is validated by a comparison of the results with the theory of Bergh and Tijdeman, and with experimental results from the literature and from shock tube tests. Finally, the model is applied for the optimization of ILPs, representative of the systems used in the aeronautics industry, and compared to the experimental results performed on an axial compressor. In those tests, a typical ILP geometry is installed on the compressor casing to measure static pressure fluctuations in the rotor tip gap. Simultaneous measurements with a fast response flush-mounted sensor provided data for comparison and validation of the predicted transfer function.Copyright