Search

We consider well defined systems involving various subclasses of aggregation and/or fragmentation processes. These systems distinctly describe 1) the aggregation of ash, soot and raindrops in physical studies 2) oscillatory behaviour in aggregation-fragmentation activities and 3) the coagulation (addition) and disintegration (complete or incomplete) of clusters. Each system involves formulating nonlinear differential equations (ordinary or partial) which describe the physical processes at hand. In addition to that, kinetic rates which convey how clusters merge and/or fragment are applied and conditions present at the start of the event are set. The universal task is to analytically and/or numerically study the behaviour of each system for theoretical and/or physical applications. We begin by first microscopically deriving the rate coefficients of space - nonhomogeneous Smoluchowski equations by considering rainfall coagulation. We assume raindrops fall at the rainfall speed Uk = ̶ u0kµ and monodisperse initial conditions are administered. Results show that the total size distributions N(z; t) and single droplets c1(z; t) monotonically decline to the equilibrium state at each height z. For other raindrop size distributions ck(z,t) (k > 2), ck(z, t) either monotonically/nonmonotonically reach the equilibrium state. Next, we study the oscillatory behaviour in aggregation-fragmentation systems with aggregation kernel Kij = i ͮ j ͧ + j ͮ i ͧ and fragmentation kernel Fij = λKij (according to Matveev et al. [1, 2]). We find that oscillations occur for the condition v ̶ µ > 1. A disintegration rate λc(v, µ) exists such that λ < λc(v, µ) yields never ending oscillations while decaying oscillations occur when λ > λc(v, µ). For other conditions v ̶ µ < 1, an analytic approximation involving the generating function (and other techniques) is used to find the steady state solution. Lastly, we investigate the process of addition with disintegration. We first consider systems involving monomer-cluster interactions with power law dependent addition rate Ak = kª and shattering rate Sk = λ ks. A cluster grows by an addition of a monomer onto the cluster and completely shatters into monomers when an energetic monomer collides into it. The interplay between addition and shattering leads to the existence of phase transitions which separate different evolutionary behaviour. The phase transition can be discontinuous or continuous depending on the exponents (a, s). For discontinuous phase transitions, a critical shattering intensity λc exists which separates the jammed state from the equilibrium state. For continuous phase transitions, two critical points λc and λc2 separate the jammed state, super-cluster state (a new state) and equilibrium state. The same qualitative behaviour occur in addition with complete/incomplete disintegration with rates Ak = ka and Rk = λkr. Quantitative differences appear in the magnitude of the critical points λc,1 and λc,2 in continuous phase transitions and ?c in discontinuous phase transitions.

The present thesis is examining the forced vibrations of a circular cylinder in the low Reynolds number flow of 200. A numerical study is performed that employs an already existing algorithm developed by (Breuer 1998) and enhanced with the characteristic of the cylinder's motion by (MadaniKermani 2014) who employed the moving frame of reference method of (L. Li, Sherwin et al. 2002). The algorithm was extensively assessed for the benchmark studies of flow around a stationary circular cylinder. A new observation was made on the effect of the aspect ratio of the computational cells in the mid region of the wake. The studies so far are emphasizing on the characteristic of a dense mesh, with a small aspect ratio, in the high divergence areas in the near region of the cylinder surface, neglecting the effect of the regions away from the surface. The present study on a stationary circular cylinder flow, proved that the aspect ratio of the distant cells has a significant effect on the St number and the force coefficients. The main study of the thesis emphasizes on the lock-in region where the wake oscillates in unison with the harmonic motion of the cylinder. The study makes a new observation on the qualitative and quantitative description of the lock-in conditions. In particular, it reveals two regions of resonance and non-resonance lock-in. Despite the fact that the lock-in is achieved, when the frequency ratio is in the first part of the region away from the unity ratio, the forces are not greatly magnified. As the ratio approaches the unity the forces experience a resonance that reaches the highest value after the unity. Furthermore, the adaptation time of the flow to the motion of the cylinder is examined and extends the results of (Anagnostopoulos 2000) to the full extent of the lock-in and the non-lock-in regions. More precisely the flow strives to reach a steady state when it is in the lock-in region rather in the non-lock in cases it reaches the steady state faster. It is postulated that the adaptation time depends on both the numerical and the physical adaptation. Moreover, the force coefficients characteristic of sinusoidal behaviour in the lock-in region is attempted to be approximated by a Newton polynomial that is built by making use of the divided differences method. The amplitude of the forces is approximated by a third degree Netwon polynomial built from the results of the present thesis simulations. The use of an approximation is providing faster results ignoring the need for a full resolution of the Navier-Stokes equation.

In nucleate pool boiling, a thin film is formed under the nucleating bubble in the region where the solid, liquid and vapor phases meet. This is called the three-phase contact line region. High rates of heat transfer have been observed in this region. Researchers have developed a pool boiling apparatus with laser interferometry and artificial nucleation sites to observe the changes in thickness, frequency, velocity and the dynamic contact angle of the microlayer. Similarly, multiple meniscus studies have been conducted to observe the characteristics of microlayer in pool boiling and various other applications such as ink jet printing, miniature heat pipes, spray cooling and drop wise condensation. In this research we look at the effect of temperature on dynamic micro contact angle and contact line velocities on a heated and polished copper surface. An apparatus was designed to oscillate a meniscus on the same copper surface. The meniscus was fed liquid to compensate for the evaporation thus making the meniscus stable. A laser source and an optical camera were incorporated to analyze the three-phase contact line region. The effect of the perimeter of the base of the meniscus and frequency of oscillating meniscus was studied. Contribution of contact line heat transfer was reported, and characteristics of microlayer formed in the contact line area of an oscillating meniscus were observed over one oscillation period.

The effects of oscillating and static retort thermal processing on heat penetration using an institutional size pouch were evaluated. A literature review of general information on retort thermal processing, process modes, processing mediums, retort pouches and starch was presented. Frigex-W starch (National Starch Food Innovation) and water mixtures were prepared at three different concentrations (1%, 3% and 5%). Five replications of starch and water mixtures were processed using water spray as the processing medium in institutional size retort pouches (29.2 cm. x 38.1 cm.) using both oscillating and static processing modes. Viscosity (by Brookfield viscometer and Bostwick consistometer), residual air (destructive technique), color (by Minolta L*a*b* colorimeter), processing time (slope of heat penetration curve) and process time to achieve 5 log reduction of Clostridium botulinum were determined for each starch and water mixture concentration level. The mean slope values of the two processing modes (oscillating and static) were not different (p>0.05) for the 1% and 5% concentrations. The mean slope values of the processing modes (oscillating and static) were different (p The results of this research were significant to the food industry because the oscillating technology may not be beneficial to companies processing low viscosity (1%) and/or high viscosity (5%) products since there was no difference between the two (oscillating and static) modes when evaluating processing time. Companies processing medium viscosity (3%) could benefit from investing in the oscillating technology because of the decreased processing time. The decreased processing time could result in increased production yields and retention of important nutrients. However, companies processing low viscosity (1%) and/or high viscosity (5%) products may still benefit from investing in oscillating technology because of other benefits. Other potential benefits (for example: retention of important nutrients) need to be investigated to determine the effectiveness of the oscillating technology.

Modern ships are usually designed for a broad range of sea conditions. In particular,much of the effort during ship design is focused on reduction of ship roll due to rough sea. Ship roll has always been of significant concern as it largely affects the ship performance. In order to mitigate the detrimental effects of ship roll, many kinds of ship stabilization systems havebeen tried and tested. One such system that is currently investigated in this work is the active-fin ship stabilization system. It consists of a system of dynamically oscillating hydrofoils, responding to cancel ship roll by producing counter acting lift. Below the dynamic stall angle,the lift force produced by hydrofoils is proportional to angle of attack. So ship stabilization systems are often operated between large angles of attack. However, at large angle of attack, aregion of low pressure forms over the suction side of the hydrofoil. The low pressure results in localized increase in cavitation. Cavitation affects the performance of fin stabilization systems,and it is therefore necessary to study and understand the flow physics involving cavitation,occurring over on oscillating fin stabilizers. In this study, the active fin stabilization system issimplified to a 2D NACA0015 hydrofoil geometry, and the aperiodic fin oscillation is simplified to a periodic hydrofoil oscillation and, to stay close to reality, a near trapezoidal hydrofoil oscillation profile is used. The present work is a contribution to the numerical simulation of cavitating flows over 2D NACA 0015 stationary and oscillating hydrofoils. The primary objective of this research is to establish a methodology to accurately identify the onset of cavitation, andto study and quantify the deleterious effects of cavitation on performance. In order to investigate and quantify the detrimental effects of cavitation on the present ship stabilization system, the commercial flow solver FLUENT v6.3 is employed. A detailed formulation of the numerical model is described, and its application to simulate cavitating flow over stationary and oscillating 2D NACA 0015 hydrofoils is demonstrated. Validation studies for the turbulence viscosity modifications for the RNG k epsilon equation have been presented. For both thestationary and oscillating cases, good agreement between the predictions and measurements is obtained. Simulation of cavitating flow over stationary hydrofoils is carried out, to understand the different steady and unsteady cavitating regimes, and to identify an approximate cavitation inception angle. The scillating hydrofoil analysis is carried out using a dynamic hybrid grid where the boundaries of the outer domain are fixed, and the hydrofoil and the inner domain surrounding the hydrofoil oscillate periodically according to a prescribed profile non sinusoidal.A user defined function (UDF) is coupled with the flow solver to oscillate the hydrofoil with its prescribed oscillation profile. A comparative study on effects of grid topology, flow regime and cavitation has been carried out at 5 Reynolds numbers.

A comprehensive experimental program has been conducted on a LS(1)-0421MOD airfoil model in the 3'x5' OSU subsonic wind tunnel. Surface pressure distributions were obtained for 2D baseline and 3D configurations under clean and surface grit conditions. Several vortex generator configurations were evaluated. The data were taken for steady state and unsteady conditions. The steady state data included angles of attack from 0 to 30 degrees and Reynolds numbers of 1.0 million. The unsteady conditions were simulated using a face cam that provided a sinusoidal angle of attack variation with 10 degrees amplitude for three frequencies of 0.6 and 1.8 Hz at mean angles of attack of 8, 14 and 20 degrees. Surface pressure data were obtained from six spanwise stations, which were integrated to local coefficients. The maximum 2D lift coefficient obtained for the 1.0 million Reynolds number was 1.58 at 14.4 degrees angle of attack. For the 3D case the maximum lift coefficient at the wall was 1.58 at 19.5 degrees and at the tip was 1.20 at 18.3 degrees. The results showed that the application of the grit roughness reduces the maximum lift coefficients in all configurations by as much as 50%. The Flat and Curled vortex generators increased the maximum lift coefficient for both the 3D tip and wall stations, up to 1.6 and 1.92, respectively. The application of the vortex generators shifted the stall angle of attack by approximately 30%. A gritted model with the vortex generators saw an increase in both the maximum lift and stall angle of attack by approximately 25% in comparison to grit only. The unsteady maximum lift coefficients were always higher than those for the steady state up to 60% and showed, generally, large hysteresis loops. The hysteresis loops were smaller for the 3D wing configuration due to the tip vortex influence, therefore smallest hysteresis loops occurred at the tip. The Flat and Curled vortex generators removed the hysteresis loops for all frequencies at 14 degrees mean angle and significantly reduced the minimum value of the pitching moment and the pressure drag at stall.

An experimental study has been carried out to determine how laminar flow is distributed in branching networks. For low-pressure flows typical of the pulmonary airways, the partitioning of regional inspiratory flows in symmetric planar networks has been observed to deviate markedly from predicted distributions based on the concept of flow resistance. By contrast, expiratory flows are distributed uniformly among all the branches. A series of experiments has been performed to demonstrate that inspiratory flow distributions are determined largely by the axial velocity profile at the inlet and by the rate of convective mixing within branch segments joining adjacent bifurcations. These results have been incorporated into a proposed non-resistive model of regional flow regulation, based on changes in bronchial geometry. These flow principles have been extended to include asymmetric branching models of pulmonary airways, using regional flow irreversibility as a quantitative index of the non-resistive component of flow partitioning. In auxiliary experiments an oscillating jet technique has been developed to estimate the rate of mixing produced by flow through a bifurcation. An empirical model utilizes both the skewness of the inlet axial velocity profile and downstream convective mixing to predict the degree of non-uniform flow splitting in symmetric networks.

To determine a suitable boundary-condition model for the contact line in oscillatory flow, a vertical plate oscillated sinusoidally in dye-laden water with an air interface is considered experimentally. Constrained by the desirability of a two-dimensional flow field and the laboratory facility, eight individual frequencies (1 to 20 Hz) each with seven different stroke amplitudes (0.5 to 6 mm) are chosen for the experimental program. The corresponding Reynolds number for each experiments varies from 1.6 to 1878.3 which is large relative to the Reynolds number in the conventional uni-directional contact-line experiments (e.g. Dussan V.'s (1974) experiments). To facilitate prediction, a high-speed video system is used to record the plate displacement, the contact-line displacement, and the dynamic behavior of the contact angle. A particle-tracking-velocimetry (PTV) technique is used to detect particle trajectories near the plate such that the boundary condition along the entire plate can be modeled. Two sets of PTV experiments are conducted. One experiment is for the stick contact-line motion, the other experiment is for the stick-slip contact-line motion. Several interesting measured contact-line phenomena are shown in the present results. Several models are developed to explain these complicated phenomena. These models explain the contact-line and contact-angle behaviors well. An expression for $\lambda$, the dimensionless capillary coefficient (nondimensionalized by the maximum plate velocity), is formulated such that the dynamic behavior at the contact line is predicted accurately. The results from the PTV experiments show that a vortex is formed near the meniscus in the stick-slip contact-line experiment; however, in the stick contact-line experiment, no such vortex is present. Using the present experimental results, a model is developed for the boundary condition along the vertically oscillating plate. That is, slip occurs within a specific distance from the contact line while the flow obeys the no-slip condition outside this slip region. This model is more realistic and accurate than models based upon modified, uni-directional contact-line boundary-condition formulations.

Resonant vertical excitation of a Tension Leg Platform (TLP) is typically of high frequency and small amplitude. The hydrodynamic damping of structural members of a TLP in this mode of oscillation is studied in this thesis. Experimental flow visualization studies were conducted on oscillating cylinders and disks of different thickness. The Keulegan-Carpenter (KC) number was varied from 0.2-0.5, and frequency parameter $(\beta)$ in the range 2500-12500. A double-exposure variant of the Digital Particle Image Velocimetry technique was developed to obtain velocity and vorticity information of the flow. It was found that thin disks of finite edge thickness generated a flow that remained symmetric with respect to the mean position of oscillation. For disks of infinitesimal edge thickness, the initial start-up condition determined the emerging vortex shedding pattern. The relative effects of convection and viscous diffusion were dependent on a KC number based on disk thickness. Drag force measurements were conducted on an axially oscillating cylinder of 0.457 m (1.5 ft) diameter and 1.219 m (4.0 ft) draft, at zero and small forward speeds. Experiments were conducted at KC numbers in the range of 0.1-1.0, and $\beta$ in 89236-191700. From these experiments, a definite nonlinear trend was observed between the drag force and velocity. The drag coefficient $(C\sb{d})$ was approximately doubled upon addition of a disk to the bottom of the cylinder. Effects of forward speed were significant at low values of KC, but were diminished at larger values. The damping coefficients of individual structural components of a TLP follow different scaling laws. Rules are presented for scaling friction and form drag components from model to full scale. Results from experiments are used to obtain a scaling law for vertical columns of a TLP. Previously published results are used for horizontal pontoons. An example TLP calculation shows that the equivalent linear damping ratio of horizontal cylinders is approximately 0.049-0.078%, depending upon cylinder shape, and that for vertical cylinders is in the range 0.025-0.394%, depending upon KC and the size of the disk used.