Kyoto University (京都大学) / 0048 / 新制・課程博士 / 博士(工学) / 甲第16376号 / 工博第3457号 / 新制||工||1523(附属図書館) / 29007 / 京都大学大学院工学研究科都市社会工学専攻 / (主査)教授 細田 尚, 教授 後藤 仁志, 准教授 米山 望 / 学位規則第4条第1項該当

M.Ing. / This thesis considers in detail the applicability to hydro power systems of the theories of steady and unsteady flow in pipelines. In doing so.it highlights some of the shortcomings of these theories. An attempt is made by way of experimentation on a high pressure pipeline, to model some of the conditions which could occur in a full size future hydro power system. These experiments provide some quantitative data about the performance of some typical hydro power components such as pipes, orifices and valves, under steady and unsteady conditions. A computer program is included which was used to provide theoretical data to compare with the experimental results. The program was found to be limited in its capacity to provide accurate simulation of the experimental pipeline, but this was thought to be due to the pipeline not correctly modelling a hydro power system. Conclusions presented in this thesis will be of assistance to designers of future hydro power systems and to researchers continuing this work.

Controlled cooling is an important part of steel production industry that affects the properties of the outcome steel. Many of the researches done in controlled cooling are experimental. Due to progress in the numerical techniques and high cost of experimental works in this field the numerical work seems more feasible. Heat transfer analysis is the necessary element of successful controlled cooling and ultimately achievement of novel properties in steel. Heat transfer on the surface of the plate normally contains different regimes such as film boiling, nucleate boiling, transition boiling and radiation heat transfer. This makes the analysis more complicated. In order to perform the heat transfer analysis often empirical correlations are being used. In these correlations the velocity and pressure within the fluid domain is involved. Therefore in order to obtain a better understanding of heat transfer process, study of hydrodynamics of the fluid becomes necessary. Circular jet due to its high efficiency has been used vastly in the industry. Although some experimental studies of round jet arrays have been done, yet the characteristics of a single jet with industrial geometric and flow parameters on the surface of a flat plate is not fully understood. Study of hydrodynamics of the jet impingement is the first step to achieve better understanding of heat transfer process. Finite element method as a popular numerical method has been used vastly to simulate different domains. Traditional approaches of finite element method, Lagrangian and Eulerian, each has its own benefits and drawbacks. Lagrangian approach has been used widely in solid domains and Eulerian approach has been widely used in fluid fields. Jet impingement problem, due to its unknown free surface and the change in the boundary, falls in the category of special problems and none of the traditional approaches is suitable for this application. The Arbitrary Lagrangian Eulerian (ALE) formulation has emerged as a technique that can alleviate many of the shortcomings of the traditional Lagrangian and Eulerian formulations in handling these types of problems. Using the ALE formulation the computational grid need not adhere to the material (Lagrangian) nor be fixed in space (Eulerian) but can be moved arbitrarily. Two distinct techniques are being used to implement the ALE formulation, namely the operator split approach and the fully coupled approach. This thesis presents a fully coupled ALE formulation for the simulation of flow field. ALE form of Navier-Stokes equations are derived from the basic principles of continuum mechanics and conservation laws in the fluid. These formulations are then converted in to ALE finite element equations for the fluid flow. The axi-symmetric form of these equations are then derived in order to be used for jet impingement application. In the ALE Formulation as the mesh or the computational grid can move independent of the material and space, an additional set of unknowns representing mesh movement appears in the equations. Prescribing a mesh motion scheme in order to define these unknowns is problem-dependent and has not been yet generalized for all applications. After investigating different methods, the Winslow method is chosen for jet impingement application. This method is based on adding a specific set of partial differential Equations(Laplace equations) to the existing equations in order to obtain enough equations for the unknowns. Then these set of PDEs are converted to finite element equations and derived in axi-symmetric form to be used in jet impingement application. These equations together with the field equations are then applied to jet impingement problem. Due to the number of equations and nonlinearity of the field equations the solution of the problem faces some challenges in terms of convergence characteristics and modeling strategies. Some suggestions are made to deal with these challenges and convergence problems. Finally the numerical treatment and results of analyzing hydrodynamics of the Jet Impingement is presented. The work in this thesis is confined to the numerical simulation of the jet impingement and the specifications of an industrial test setup only have been used in order to obtain the parameters of the numerical model. / Applied Science, Faculty of / Mechanical Engineering, Department of / Graduate

We investigate the propagation of internal gravity waves in a rotating fluid with horizontal and vertical stratification. The modification of these waves by the presence of a vertically sheared geostrophic current is determined, and the rate of energy exchange between waves and current is estimated and compared to exchange rates of other interaction mechanisms. The effect of boundary conditions on the range of frequencies allowed for wave propagation is also considered. The wave amplitude has horizontal exponential dependence due to the horizontal density variation as well as to exchange of energy with the mean shear flow. The solution also shows a phase difference from surface to bottom. For waves propagating normally to a vertically sheared geostrophic current, the energy exchange mechanism is found to be weak when compared to other exchange mechanisms and is likely to be of little importance in the ocean. The imposition of boundary conditions on the wave solution alters the frequency range over which solutions may exist. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

Motions and stability of a typical B.C. fishing vessel were experimentally and numerically investigated in a longitudinal seaway condition. The experimental model was self-propelled, radio-controlled and equipped with an on-board data acquisition system. Pitch, roll, yaw, surge, and heave responses to regular waves of predetermined frequencies and amplitudes generated along a 220-ft model basin were obtained. Different displacement conditions and GM configurations were tested. The numerical model for the dynamic analysis of the fishing vessel motions has been implemented using strip theory. A computer program was developed to study the nonlinear motions of the vessel. The velocity dependent coupling terms, responsible for a major part of the nonlinear behavior, were included. A time dependent component analysis of the roll damping has been performed. Regular linear and nonlinear waves were used. A parametric study of the fishing vessel stability has been carried out by considering its dynamic response in waves of varying characteristics. Unstable behaviour was found to be closely related to waves of length of similar magnitude to the ship length. The effects of wave amplitude and rudder usage were found to be of capital importance in the capsizing process. Experimental and numerical results showed good agreement. / Applied Science, Faculty of / Mechanical Engineering, Department of / Graduate

The Experimental and Finite Element Investigation of Added Mass Effects on Ship Structures comprised three phases : 1) investigation of the fluid modelling capabilities of the Finite Element Program VAST, 2) experimental investigation to determine the effect of the fluid on the lowest natural frequencies and mode shapes of a ship model, and 3) comparison of these experimental results with numerical results obtained from VAST. The fluid modelling capabilities of VAST were compared with experimental results for submerged vibrating plates, and the effect of fluid element type and mesh discretization was considered. In general, VAST was able to accurately predict the frequency changes caused by the presence of the fluid. Experimental work both in air and water was performed on a ship model. The lowest four modes of vertical, horizontal, and torsional vibration were identified, and the effect of draught on the frequencies and mode shapes was recorded. When the experimentally obtained frequencies and mode shapes for the ship model were compared with the numerical predictions of VAST, good agreement was found in both air and water tests for the vertical vibration modes. / Applied Science, Faculty of / Mechanical Engineering, Department of / Graduate

Combustion of Minto coal, a sub-bituminous eastern coal which is highly agglomerating and has a high sulphur content, was carried out in a 0.15 m internal diameter half-column spout-fluid bed combustor in inert beds of sand, with limestone addition for sulphur capture. The average bed temperature ranged from 800 to 900°C, flue gas oxygen level was 2.5 to 11.0%, auxiliary to total air was 0 to 0.50, and Ca/S molar ratio was 2.5. High vale coal was employed in hydrodynamic runs. Aspects studied included combustion efficiency, sulphur capture efficiency, axial and radial temperature profiles, axial O₂ and CO₂ concentration profiles, axial SO₂ concentration profiles, minimum spouting velocity, spouting stability, and maximum spoutable bed height. The principal problem encountered with Minto coal in this equipment was agglomeration during the heat-up period. A spout-fluid bed has proved to be great favourable for handling agglomerating coal relative to the standard spouted bed. When limestone was used as bed material, less sintering was encountered. However, limestone could not stand up to spouting for prolonged periods because of excessive attrition. Combustion efficiencies were found to be higher than 80% in the temperature range of 800 to 900°C without solid fines recycle. An increase of temperature between 800°C and 840°C was beneficial for combustion efficiency, while a further increase up to 885°C did not seem to have a significant effect on combustion efficiency. Increase of auxiliary/total air ratio was favourable to combustion efficiency at elevated temperatures. Sulphur capture efficiency passed through a maximum with increasing temperature between 800°C and 900°C The maximum value was obtained at around 830°C. NOx emission increased linearly with increasing flue gas oxygen level. No abrupt temperature increase above the bed surface was observed in both spouted and spout-fluid beds investigated in the present study. Temperature may increase above the bed surface for low excess oxygen runs in view of the substantial amount of combustion found to occur in the freeboard. Temperatures were more uniform after the introduction of auxiliary air. Most oxygen was consumed below the bed surface. Axial profiles showed a significant SO₂ jump in the spout over the bed height. Combustion and sulphation could be considered to occur in two main stages: (1) Combustion of carbon, at the same time as most of the sulphur is released. (2) Sulphation of the sorbent. The Mathur and Gishler (1955) and Wu et al. (1987) equations gave poor agreement with the minimum spouting velocity, Ums, over the entire range of temperature. For large particles Ums tended to increase with increasing temperature, while for small particles it decreased with increasing temperature. Gas viscosity should be taken into consideration for predicting Ums. A considerably greater effect of auxiliary to total air ratio, q/Qt, on total minimum spouting velocity was found at elevated temperatures than at room temperature. At the maximum spoutable bed height, the value of Um/Umf was found to decrease with increasing temperature and to be smaller than unity at elevated temperatures. The McNab and Bridgwater (1977) expression correctly predicted the observed trends of Hm and worked reasonably well at high temperatures, although it was found to over-predict Hm at lower temperatures. Hm decreased with increasing temperature for all particle sizes, with a faster decrease for smaller particles. Fluidization in the annulus was never observed as the termination mechanism of spouting at high temperatures. / Applied Science, Faculty of / Chemical and Biological Engineering, Department of / Graduate

Fluid mixing that is induced by hydrodynamic instability is ubiquitous in nature; the material interface between two fluids when perturbed even slightly, changes shape under the influence of hydrodynamic forces, and an additional zone called the mixing layer where the two fluids mix, develops and grows in size. This dissertation reports a study on the role of thermal fluctuations in fluid mixing at the interface separating two perfectly miscible fluids of different densities. Mixing under the influence of two types of instabilities is studied; the Rayleigh-Taylor (RTI) and Richtmyer-Meshkov (RMI) instabilities. The study was conducted using numerical simulations after verification of the simulation methodology. Specifically, fluctuating hydrodynamic simulations were used; the fluctuating compressible Navier-Stokes equations were the physical model of the system, and they were solved using numerical methods that were developed and implemented in-house. Our results indicate that thermal fluctuations can trigger the onset of RTI at an initially unperturbed fluid-fluid interface, which subsequently leads to mixing of multi-mode character. In addition we find that for both RMI and RTI, whether or not thermal fluctuations quantitatively affect the mixing behavior, depends on the magnitude of the dimensionless Boltzmann number of the hydrodynamic system in question, and not solely on its size. When the Boltzmann number is much smaller than unity, the quantitative effect of thermal fluctuations on the mixing behavior is negligible. Under this circumstance, we show that mixing behavior is the average of the outcome from several stochastic instances, with the ensemble of stochastic instances providing the bounds on mixing-related metrics such as the mixing width. Most macroscopic hydrodynamic systems fall in this category. However, when the system is such that the Boltzmann number is of order unity, we show that thermal fluctuations can significantly affect the mixing behavior; the ensemble-averaged solution shows a departure from the deterministic solution. We conclude that for such systems, it is important to account for thermal fluctuations in order to correctly capture their physical behavior.

The performance of ebullated bed hydroprocessors depends on the fluids distribution system and liquid recycle pan. Given that bubbles do not readily coalesce in the bed, the original bubble size distribution generated at the bubble cap distributor likely impacts buoyancy-based phase separation at the recycle pan. Gas entrained in the liquid recycle increases bed gas holdup at the expense of liquid holdup and product yield. The aim of this work was to investigate the impact of gas-liquid distribution system on resulting bubble properties and dynamics and incorporate a distributor sub-model into an existing fluid dynamics model of the industrial hydroprocessor. The size of initial bubbles formed in the plenum chamber was found to have negligible impact on phase holdups above the distributor. However, resulting bubble properties were found to depend on distributor geometry, distributor power dissipation and gas-liquid velocity ratio. In addition, a new set of scaling laws for gas-liquid distributors, based on dimensional analysis and similitude, was proposed. Geometric scaling was based on matching distributor fractional open area and ratios of critical dimensions. Dynamic similarity was based on matching three dimensionless groups and bubble coalescence behaviour. A bubble size distribution model was then developed. Both pressure and distributor were found to have an impact on individual bubble drag coefficients, as they both altered bubble size distribution. A novel drag model was thus also developed at industrially relevant conditions. Finally, a new gas-liquid distributor sub-model, including bubble size distribution and drag models previously developed, was incorporated into an overall fluid dynamics model of the hydroprocessor. The bubble size distribution model was also coupled with existing gas-liquid separation sub-model to better predict recycled gas and liquid fractions. A sensitivity analysis performed with the overall model revealed distributor configurations with potential of improving the processing capacity of the hydroprocessor.

1D stellar evolution and nucleosynthesis simulations have traditionally modeled the mixing within convection zones as a diffusive process. The fluids within a convection zone are advecting and do not diffuse. However the diffusive approximation is valid when the burning timescale of an exothermic reaction is longer than the convective turn over timescale to which the mixing of those species is approximated over. Since it is 1D, it also assumes that the material is isotropically distributed within the convection zone. In the He-flash convection zones of rapidly accreting white dwarfs (RAWD) H is ingested and burned well within the convective turn over time of 38 minutes. The H is burned through the exothermic 12C(p,γ)13N reac- tion, Q = 1.944 MeV, and then the unstable 13N, with a half-life of 9.6 minutes, will decay to 13C which will undergo the 13C(α,n)16O reaction releasing neutrons. The neutron densities, depending on the H-ingestion rates and mixing details, reach Nn ≈ 1013 − 1015 cm−3 which starts the i-process within the convection zone. The H burning provides energy to the flow leading to the dynamic details of the flow being important for the mixing of the H and thus the i-process nucleosynthesis. This is a convective-reactive environment. The isotropic, well mixed over many convective turn over timescales, and long burning timescale assumptions for H in the diffusive approximation are broken in the convective-reactive environment of a He-shell flash convection zone in a RAWD. To more accurately model convective-reactive mixing environments, a 1D two stream advective mixing model is formulated. A downstream advects H-rich material from the top of the convection zone down to the H-burning region while the upstream advects H-poor material back up to the upper convective boundary. The mixing model includes a horizontal mass flux, γ, which describes the efficiency to which mass is mixed between the two streams. This predominately causes the homogenization of the material between the two streams. The radial mass flux, α, and the horizontal mass flux, γ, are calibrated from 3D hydrodynamic simulations of the RAWD in order to model the mixing within the He-flash shell convection zone. The downsampled 3D cartesian data output, the briquette data, from the 3D hy- drodynamic simulations is used to compute γ. This required using numerical tools to interpolate quantities onto spherical shells from 3D cartesian data and to decompose the radial velocity field into its spherical harmonic modes. Trilinear interpolation is the simplest 3D interpolation method that was tested and it was the interpolation method of choice due to the constraints it has on the interpolating function. The validity of using higher order methods on the briquette data was studied in detail but was determined to not be usable due to the computational effort and constraints of the methods. The two stream model post-processing of the H burning within the 3D hydro- dynamic simulations of the RAWD showed excellent agreement in the metrics of the total mass of H burned, the burning rate and burning location of H. This includes two models which undergo dramatic H-ingestion and burning events caused by a GOSH, Global Oscillations of Shell H-ingestion. By adding a network containing 1000’s of species to the 1D advective mixing model, the i-process from the RAWD is simulated and compared with a traditional 1D diffusive mixing model. The resulting neutron densities between the two models are comparable however the efficiency to which each produce the heaviest stable elements are different. To reproduce the elemental abun- dance distribution of the CEMP-r/s star CS31062-050, the diffusive model is run for 15 days of stellar time while the advective model is run for 20 days. The H-ingestion into the He-shell as predicted by the stellar evolution calculations lasts 30 days. The i-process material within the RAWD can be removed from it and participate in the galactic chemical evolution of the galaxy that it resides in. This is due to the RAWD possibly reaching the Chandrasekhar mass and from the loss of material through stellar winds and common envelope interactions with its nearby companion star. / Graduate