A comprehensive computational and experimental analysis has been conducted to characterize the flow dynamics and periodic structures formed in the wake of complex turning vanes. The vane packs were designed by the StreamVane swirl distortion generator technology, a design system that can efficiently reproduce swirl distortion for compressor rig and full turbofan engine testing. StreamVanes consist of an array of turning vanes that commonly contain variations in turning angle along their span, a nonaxisymmetric profile about the centerline, and vane-to-vane intersections or junctions to accurately generate the desired distortion. In this study, vane packs are considered complex if they contain two out of three of these features, a combination seen in other turbomachinery components outside of StreamVane design. Similar to all stator vanes or rotor blades, StreamVane vane packs are constructed using a series of cross-sectional airfoil profiles with blunt trailing edges and finite thicknesses. This, in turn, introduces periodic vortex structures in the wake, commonly known as trailing edge vortex shedding. To fully understand how the dynamics and coherent wake formations within vortex shedding impact both the flow distortion and structural durability of StreamVanes, it is first necessary to characterize the corresponding wakes in three dimensions. The current study provides an in-depth analysis to predict and measure the trailing edge vortex development using high-fidelity computational fluid dynamics and stereoscopic time-resolved particle image velocimetry experiments. Two testcase StreamVane geometries were specifically designed with complex features to evaluate their influence on the dynamics and coherence of the respective vane wakes. Fully three-dimensional, unsteady computational fluid dynamics simulations were performed using a Reynolds-Averaged Navier-Stokes solver coupled with a standard two-equation turbulence model and a hybrid, scale-resolving turbulence model. Both models predicted large-scale wake frequencies within 1—14% of experiment, with a mean difference of less than 3.2%. These comparisons indicated that lower fidelity simulations were capable of accurately capturing such flows for complex vane packs. Additionally, structural and modal analyses were conducted using finite element models to determine the correlations between dominant structural modes and dominant wake (flow) modes. The simulations predicted that vortex shedding modes generally contained frequencies 300% larger than dominant structural modes, and therefore, vortex induced vibrations were unlikely to occur. Lastly, mode decomposition methods were applied to the experimental results to extract energy ratios and reveal dynamic content across high-order wake modes. The vortex shedding modes generated more than 80% of the total wake energy for both complex vane packs, and dynamic decomposition methods revealed unique structures within the vane junction wake. In all analyses, comparisons were made between different vane parameters, such as trailing edge thickness and turning angle, where it was found that trailing edge thickness was the dominant vortex shedding parameter. The motivation, methodology, and results of the following research is presented to better understand the wake interactions, computational predictive capabilities, and structural dynamics associated with vortex shedding from complex vane packs. Although the results directly relate to StreamVane distortion generator technology, the qualitative and quantitative comparisons between the selected methods, geometry parameters, and flow conditions can be extrapolated to modern turbomachinery components in general. Therefore, this dissertation aims to benefit distortion generator and turbomachinery designers by providing insight into the underlying physics and overall modeling techniques of the wake dynamics in highly three-dimensional, complex components. / Doctor of Philosophy / A comprehensive analysis has been completed to characterize the unsteady wake flow produced by complex turning vane systems in three dimensions. Turning vanes are a common component utilized in the field of fluid dynamics and aerospace propulsion to effectively turn and manipulate the working fluid to the desired condition. For propulsion applications, similar vanes can alleviate performance losses by improving the overall aerodynamics and mitigating flow distortions entering the compressor of a jet engine. Conversely, complex turning vanes can also be used to reproduce the distortion for engineers to evaluate jet engine components when subjected to nonuniform flow ingestion. The distinct geometry features that make these vanes complex are also present in other turbomachinery systems outside of distortion generation. In any case, the cross-sectional profiles of the turning vanes commonly contain blunt ends or trailing edges due to engineering limitations and/or restrictions. This geometric feature introduces periodic wake structures, known as vortex shedding, that can negatively effect the performance of the overall system. It is therefore a necessity to characterize both the dynamics and coherence of vortex shedding to fully understand the flow features in highly three-dimensional flows. In the presented research, this is achieved by applying computational simulations and experimental measurements to extract the corresponding wake dynamics of complex vane packs. The selected testcases where designed using the StreamVane technology, a mature system that generates tailored turning vanes to reproduce flow distortion in jet engine or fan rig ground-testing facilities. The fluid simulations captured the expected wake flow and largescale structures convecting downstream of the vane packs. A comparison between two different flow models and the experimental results revealed minimal quantitative differences in the large-scale dynamics, which gave insight into the model selection to predict such flows. Additional structural simulations were performed to estimate the forcing and response of the vane packs when subjected to the aerodynamic loading. The results showed vortex shedding was highly unlikely to cause large amplitude vibrations and structural failures. In all analyses, the primary results were correlated with common vane parameters and operating conditions to evaluate their impact on the wake dynamics. The motivation, methodology, and results of the following research is presented to better understand the wake interactions, computational predictive capabilities, and structural dynamics associated with vortex shedding from complex vane packs. Although the results directly relate to StreamVane distortion generator technology, the qualitative and quantitative comparisons between the selected methods, geometry parameters, and flow conditions can be extrapolated to modern turbomachinery components in general. Therefore, this dissertation aims to benefit distortion generator and turbomachinery designers by providing insight into the underlying physics and overall modeling techniques of the wake dynamics in highly three-dimensional, complex components.

We present two distinct projects in this article. In the first project, an experiment aiming to quantify the impacts of material acoustic impedance and thickness on single laser-induced cavitation bubble dynamics with measurements of exerted pressure on a specific material in order to identify the primary sources most responsible for material damages is presented in this article. Two types of major pressure sources have been identified. For bubble collapsing near a rigid wall, when standoff ratio γ < 0.6, the ring collapse is the most prominent pressure source. The jet takes the strongest effects at γ = 1.12. The pressure is minimal at γ = 0.913. After the first jet impingement, a second ring collapse will follow and input the maximum pressure to the wall. By further increasing γ, a similar pressure profile of the second collapse to the first collapse is achieved, during which the pressure for the second collapse is minimal at γ = 1.41 and for the jet is maximum at γ = 1.79. Compared with the maximum pressure dealt by the first jet, the second ring collapse and jet are increasing much faster with the bubble size and eventually overwhelm the first jet. However, the first ring collapse is still the most dominant pressure source responsible for material damages. For wall featuring smaller acoustic impedance or thickness that cannot be approximated to a rigid body, the ring collapse and jet occur at smaller standoff ratios. The cavity shrinking rate suggests the maximum pressure exerted on the wall at applicable standoff ratios should be smaller than that on a rigid wall. In the second project, a comprehensive collection of dynamics of one and two laser-induced cavitation bubbles collapsing near different boundaries is presented in this article by measuring the velocity fields using particle image velocimetry (PIV) techniques. Cases include a single bubble collapsing near the hard, medium, and soft walls characterized by acoustic impedance, free collapse of two bubbles, and two bubbles collapsing near the hard and soft walls. We implemented the most significant velocity and top velocity regions derived from each velocity field to analyze the features of these cases. Before converging to free collapse, the bubble near the hard wall experienced a significant velocity decrease before collapse, the bubble near the medium wall was severely damped at a specific standoff distance, and the bubble near the soft wall collapsed much earlier and preserved a linear velocity region at low speed. Free collapse of two same bubbles underwent a decrease of acceleration before collapse. Decreasing the size of one bubble caused a jet in the other. With the presence of a hard wall near two bubbles, the bubble closer to it may be stretched to a cavity with a high aspect ratio, leading to very mild collapse. With a bigger bubble between a smaller one and the soft wall, the merging cavity may suppress the tendency of jet formation, making the velocity stay at low levels throughout the lifetime. For configurations regarding single bubbles collapsing near a wall and free collapse of two same bubbles, we performed data scaling to study the velocity variations for different bubble sizes by controlling the standoff ratios and assessed the data quality aided by curving fitting and statistics. Results indicated measured velocity regarding a single bubble collapsing near the wall over its diameter remained the same given a standoff ratio, while measured velocity did not change given a standoff ratio for free collapse of two same bubbles within the scope of the experiment. In addition, we detailed the experimental setup and water treatment for better signal-to-noise ratios as well as validated the system from both the PIV and high speed imaging approaches using free collapse of a single bubble to ensure the reliability of this experiment. / Doctor of Philosophy / The phenomenon of cavitation extensively exists. These small and transient bubbles are observed typically in fast moving fluids, e.g., shaking a bottle of water. Each bubble experi- ences a process of growth, collapse, rebound, and collapse again before it is gone. Although the bubble is tiny, the collapse of a bubble releases considerable pressure, which is intense enough to damage nearby objects over time. This interaction between bubbles and objects depends highly on the types of objects such as the materials and thickness. To study how the bubble behaves near a wall (object) and explain how the wall is damaged, we present two projects in this article. In the first project, we created a bubble near a wall at differ- ent bubble-to-wall distances and tracked how the bubble changed its shape until collapse with a fast speed camera. This work was repeated for multiple different wall materials and thickness. We then measured the pressure exerted by a bubble at a series of different bubble-to-wall distances on a specific wall equipped with a sensor. By comparing and sum- marizing results from both the bubble shape changes near different walls and the pressure measurement, we found the relationship between the magnitude of pressure and the distance between the bubble and the wall. In the second project, we implemented the particle image velocimetry (PIV) techniques to measure the velocity fields. By feeding particles into the fluid, PIV tracks the location differences of particles in two subsequent frames to determine the velocity of every point. Based on that, we obtained a collection of velocity fields of interaction between single bubbles and walls, two bubbles, and two bubbles and walls.

Turbulent boundary layers over rough surfaces are a common, yet often overlooked, problem of practical engineering importance. Development of correlations between boundary layer parameters that can be used in turbulence models and the surface geometry is the only practical option for solving these problems. Experiments have been performed on a two-dimensional zero pressure gradient turbulent boundary layer over sparsely spaced hemispherical roughness elements of 2 mm diameter. Laser Doppler velocimetry was used to measure all three components of velocity. The friction velocity was calculated using an integral momentum balance. Comparisons were made with various fitting methods that assume the von Kármán constant is appropriate for rough walls. Results indicate that this is not the case, and that the slope of the semi-logarithmic portion of the mean streamwise profile may be a function of the ratio of inner and outer length scales. Comparisons were also made between various correlations that relate the surface geometry to the behavior of the mean velocity profile. In general, the existing correlations achieved a reasonable agreement with the data within the estimated uncertainties. A detailed study of the local turbulent structure around the roughness elements was performed. It was found that, in contrast to `sharper-edged' elements such as cylinders, an elevated region of TKE and Reynolds shear stress was found downstream of the element below the peak. This can be explained by the delay in separation of the flow coming over the top of the element due to the smooth curvature of the element. Surface pressure fluctuation measurements were made as well using a dual microphone noise reduction technique. There have only been a few past experiments on the surface pressure fluctuations under rough wall boundary layers. However, it has been shown that the spectra of the wall fluctuations can be used to predict the far-field noise spectrum [1,2]. Therefore it is been the goal of this research to verify existing correlations between the surface pressure fluctuation spectrum and the surface geometry as well as develop new correlations that provide insight into the interactions between the turbulent motions in the flow surface pressure. / Ph. D.

Measurements were made in two-dimensional zero pressure gradient turbulent boundary layers over 5 geometries of three-dimensional densely distributed surface roughness. A 3-velocity component laser Doppler velocimeter was used to measure instantaneous velocities. These measurements permitted an independent estimate of skin friction on the surfaces using a momentum balance approach, and the validity of the von Karman constant for rough walls was tested. Five roughness fetches were evaluated: three sandpaper roughness fetches of varying grit size and two cases of uniformly distributed hemispheres of different spacing. Optical surface profilometry was used to characterize the geometry of the sandgrain surfaces. It was found that the smooth wall von Karman constant can not be assumed for densely distributed rough wall flows in order to determine the skin friction for these flows. This requires an independent measure of skin friction using more than a single boundary layer profile. Near wall flow structure measurements found that the hemispherical elements do not have high TKE or Reynolds shearing stress regions at the trailing edge of elements as had been shown for sparsely spaced cylindrical elements. This is likely due to the sharp trailing corner of the cylindrical elements, as opposed to an effect of spacing. Rather, hemispherical roughness has a periodically occurring high stress and TKE region located between two element centers in the stream-wise direction at a height of approximately 1.5 times the roughness element height. The periodic nature of the near wall flow extends to approximately 4 roughness element heights. The traditional roughness function f(&#955;) did not correlate well with &#955; or the modified &#923; for the experimental data. However, it was found that the friction coefficient for the current dense roughness cases is a constant 0.004, within the experimental uncertainty. Traditional inner wall scalings, outer wall scalings, and roughness scalings were not able to collapse surface pressure fluctuation spectra for the various rough wall surfaces tested. However, the data do collapse for individual geometries based on Reynolds number. This gives rise to the ability to predict pressure fluctuation spectra at other Reynolds numbers. / Ph. D.

This thesis is a study of the effects of distributed roughness elements on a two-dimensional turbulent boundary layer. Measurements were taken on a total of ten rough wall configurations: four involving Gaussian spikes, and six with circular cylindrical posts. Rough wall flows are particularly suited to study with Laser Doppler Velocimetry (LDV) due to the fact that measurements are required near a solid surface, as well has in highly turbulent fluid. The LDV system used in this study is a fine resolution (~50 micron), three-component, fiber optic system. All mean velocities, Reynolds stresses, and triple products are measured. This study is unique in the range and variety of roughness cases for which data was taken. The data show that the flow over a rough wall is characterized by high levels of turbulence near the roughness element peaks at the interface between low-speed, near-wall fluid and the higher speed fluid above. Behind an element, high-momentum fluid sweeps toward the wall, and there is a small region of ejection of low-momentum fluid. Cylindrical elements typically have larger magnitudes of turbulent stresses at their peaks compared to Gaussian elements. Trends in mean velocity profile parameters such as displacement height, roughness effect, and wake parameter are examined with respect to roughness element geometry and spacing. / Master of Science

This report documents the measurements of turbulence quantities resulting from vortices embedded in a zero pressure gradient turbulent boundary layer. Turbulent boundary layers are found in most flow regimes over large scale vehicles and have been studied for many years. Various systems to control separation of boundary layers have been proposed, but vortex generators have proven to be an economical choice as they are often used to fix deficiencies in a flow field after large scale production of a vehicle has commenced. In order to better understand the interaction between vortex generators and the boundary layer in which they are embedded, an experiment has been performed using through non-invasive Comprehensive Laser Doppler Velocimeter. The results show that normalization on edge velocity is appropriate for comparison with previous work. The 1/S parameter and vq^2 parameter were found to be most appropriate to correlate the Reynolds stresses and triple products, respectively. The higher inflow edge velocity and greater momentum thickness, creating a lower vortex generator to boundary layer height ratio, result in a more diffuse vortex as compared to previous work conducted in the same wind tunnel, with the same geometry, but with different inflow conditions. / Master of Science

No / River morphological processes are among the most complex and least understood phenomenon in nature. Recent research indicates that the braiding of marine waterways of the estuary zone occurs at an aspect ratio similar to the alluvial braided river. The instability of complex sporadic fluvial processes at river-sea interface is responsible for bar formation in alluvial as well as in marine waterbodies Due to the lack of knowledge of flow characteristics around bar, the flow structure around the sand bar is analyzed. The bursting events play the crucial role in understanding the fluvial characteristics in the vicinity of submerged structure. The study of bursting events around the mid-channel bar is only done by the present author. The effect of submergence ratio on the turbulence behavior in the proximity of bar is analyzed in this study. The flow turbulence generated by the mid-channel bar is also analyzed in detail. The extreme turbulent burst is segregated from low intensity turbulent events by using the hole size concept. The effect of hole size on the parameter Dominance Function is analysed which is not yet studied by any researcher for mid-channel bar. The Momentum Dominance Function (MDF) parameter increases with increase in the Hole Size. This indicates that the magnitude of upward flux increases with increase in the hole size. The effect of bar height on the turbulent burst which is not yet studied by any researchers is analyzed in the present research. The joint probability distribution of bursting events is modeled using the Gram-Charlier bivariate joint probability function. The joint probability distribution gives the details of probabilistic structure of flow in the vicinity of bar. The effect of bar is predominant only in the lower flow layer. The joint probability distribution graph becomes more eccentric toward the dominant quadrants with increase in the submergence ratio. This indicates that the probability of dominant events further increases with increase in the submergence ratio.

No description available.

Water quality in river systems is an important issue, and relies on various factors including our ability to predict how effluents from outfalls mix with river water. However, predicting mixing in rivers, and especially in river bends, remains a difficult problem to solve. The goal of this project is to develop a comprehensive picture of the mixing mechanisms of an effluent jet in a river bend. This is done with experiments in both bend flumes in the University of Ottawa Water Resources Engineering Laboratory. The large bend flume is 1 m in width, and contains a single 135° bend of radius 1.5 m, and the small flume has a channel width of 0.2 m with a 135° bend of radius 0.3 m. The experiments in the large flume used acoustic Doppler velocimeters to measure velocity, and the experiments in the small flume used particle image velocimetry to track flow fields. Large eddy simulation (LES) were also completed using the same channel geometry as the small flume. To complete the parametric analysis on mixing of a neutrally buoyant effluent jet in a channel bend, 35 flow conditions, from seven channel aspect ratios and five momentum ratios, are modelled using LES. Each flow condition is also modelled without the jet present. Particle image velocimetry data from the small bend flume validates the LES models. Additionally, acoustic Doppler velocimeter tests were completed in the large bend flume under two different flume flow rates, two jet flow rates, and two aspect ratios. These models and measurements provide a broad range of the parameters under investigation. The experiments in the large bend flume establish the shape of the jet's trajectory within the channel bend, and how it differs from a trajectory in a straight crossflow. From these experiments, it is established that the centre position of the secondary circulation cell is an important parameter for determining the position of the jet. Through the LES models, more details of the 3D velocity and effluent distributions are available, allowing for a detailed analysis of how the secondary circulation develops and how the jet vortices change the development patterns. A method for clustering instantaneous vortices to separate sub-cells of secondary circulation is established, and is used to set a baseline for the development of secondary flow in a channel bend without a jet. The effect of an added jet was investigated in detail for a single flow condition, and then with machine learning techniques to develop a parametrical model incorporating both channel and jet flow conditions. The best performing machine learning model for the parametrisation of secondary flow cells with the jet is the ANFIS model coupled with a decision tree classifying the presence of each sub-cell; without the jet, the best-performing model is the ANFIS model without any additional classification. The effluent distribution is well-characterised using multiple linear regression. The addition of a jet changes the relative strengths of secondary circulation sub-cells and their circulation development and retention characteristics, though the total circulation in the bend is not strongly affected by the jet.

This thesis introduces the Deterministic Disturbance Generator (DDG) and its development process. The DDG performs two motions and five pitch rates. The flap motion, which rotates the airfoil from 0◦ to 20◦ and back, and the ramp motion, which rotates it from 0◦ to 20◦ with a dwell of 1s before returning to 0◦. To determine the angle of attack, a Matlab function converted thrust rod displacement into the assumed angle, validated against true angle of attack measurements on the DDG. Mean angular displacements were plotted, and standard deviations of the 95% confidence intervals were calculated within ±1.3◦ for all motions. The mechanical force on the actuator was computed to be 77N. Aerodynamic forces on the DDG were determined to be 15N and 19N for flap and ramp motions respectively. The total force on the system did not exceed 100N in any case, staying below the peak force capacity, while acceleration reached its limit. Flow velocimetry in the Virginia Tech Stability Wind Tunnel (VTSWT) employed a time-resolved Particle Image Velocimetry (PIV) to study the effects of 20◦ flap and ramp motions, with mean actuation times of 63ms and 37ms. Flap motion showed a significant deficit in mean streamwise velocities, and the ramp motion exhibited similar behavior until its dwell position, generating a large wake region due to airfoil stall after its peak. Comparison of data from the Goodwin Hall Subsonic Tunnel (GHST) with VTSWT data for overlapping domains revealed similar flow field features when normalized based on the boundary layer velocity (43mm plane from wall) of the latter. Considering actuation time differences, the freestream normalized GHST data was combined with VTSWT data. The cohesive PIV domain offered a broader perspective on the missing flow features. / Master of Science / A Deterministic Disturbance Generator (DDG) was designed to generate consistent largescale transversal transient disturbances in the wall boundary layer of the Virginia Tech Stability Wind Tunnel. It comprises an airfoil connected to an actuator through a rotating mechanism. The rotating mechanism can be controlled by manipulating the actuator to induce motion. The rotational speed of the airfoil is regulated by a program provided to the actuator. The DDG motions were validated to achieve nearly identical motion profiles to ensure it produced consistent turbulence wakes. The linear displacement of the actuator and airfoil was measured using a laser sensor, and a code was developed to convert this data into the observed angle of attack. Tests were conducted to verify repeatability and fine-tune the system's motions. A comprehensive description of the fabrication process, hardware and software setup, and calibration procedures involved in developing the DDG are provided. Using aerodynamic models, a computational study is performed to determine the forces associated with the airfoil and actuator. Subsequently, the DDG was subjected to testing in two wind tunnels: the Goodwin Hall Subsonic Tunnel for preliminary characterization and error mitigation and the Virginia Tech Stability Wind Tunnel for final assessment of the DDG's performance. Flow velocimetry data obtained from both tests are analyzed, revealing similarities in the induced motions. Mean flow fields and turbulence values are determined, and the effects of different pitch rates are also assessed. Finally, the mean flow fields corresponding to identical motion types from both datasets were integrated into a cohesive plot. This resulted in a comprehensive understanding of the flow field.