Two-Phase bubbly flows are encountered in a wide range of industrial applications, particularly where phase changes occur as seen in high performance heat exchangers and boiling reactors for power generation. These flows have been extensively studied in channels with circular geometries using air-water flows, though little data exists for flows through narrow rectangular channels. Measurements in thin geometries are particularly challenging since large bubbles bridge the gap, and it is difficult to compare point measurements with photographic techniques. The objective of this study is to explore the abilities of hot-film anemometry and high speed photography for taking measurements in a narrow vertical rectangular channel for a range of volume fractions, with particular attention on the narrow dimension. Hot-film anemometry (HFA) is a measurement technique originally developed for the measurement of fluid velocities, but has since been found to have applications for broader measurements in multiphase flow. With the sensor operating on the principle of heat loss, the method takes advantage of the differing abilities of the phases to transport heat, with each phase leaving its own signature in the signal response. The linchpin of this method lies in the ability to accurately distinguish between the two phases within the signal, and to execute this operation, various algorithms and techniques have been developed and used with some success for a wide range of flow conditions. This thesis is a study of the various methods of analysis such as amplitude threshold for triggering, and small slope threshold for finely tuning the edges of the bubble interactions, and demonstrates the capabilities of the hot-film sensor in a narrow rectangular vertical duct with a high aspect ratio. A vertical acrylic test section was fabricated for the purposes of this study, inset with a rectangular channel 38.1mm in width and 3.125mm in depth. Experiments were conducted for volume fractions ranging from 2% to 35%, which remained within the limits of the bubbly flow regime, but ranged from small uniform bubbles to larger bubbles coalescing into a transition regime. The hot-film signal was analyzed for void fraction, bubble speed, and bubble size. An in-depth study of the various methods of phase discrimination was performed and the effect of threshold selection was examined. High-speed video footage was taken in conjunction with the anemometer data for a detailed comparison between methods. The bubble speed was found to be in close agreement between the HFA and high-speed video, staying within 10% for volume fractions above 10%, but still remaining under a 30% difference for even as low as the 2% volume fraction, where measurements have been found to be historically difficult. The trends with volume fraction between the HFA and high-speed results were very similar. A correlation for narrow rectangular channels employing a simple drift flux model was found to compare with the void fraction data where appropriate. Good agreement was found between the methods using a hybrid phase discrimination technique for the HFA data for the void fraction and bubble speed results, with the high-speed video results showing a slight over-estimation in regards to the bubble size.

Determination of particle slip is a major concern for particle based measurements in un- heated supersonic facilities, as it is a limiting factor for the instruments' frequency response. For the purpose of determining the particle deceleration through a stationary shock wave in a super sonic windtunnel, a novel 1-D Laser Doppler probe with an unique spatial range (~1.5 mm) is presented. The study first gives a short review of the physics of particle motion with respect to different drag models and flow regime encountered in super sonic flows. In the second part, the focus lies on the development of a new Laser Doppler probe using non Gaussian beams to obtain a prolonged measurement volume. This volume covers a major part of the particle lag after a shock wave. An experimental investigation on particle acceleration and drag, using different types and sizes of seeding material, including standardized microspheres is carried out in the Mâ = 2.0 super sonic facility. Three different types of particles with four different sizes are experimentally investigated. The experimental data provides mean velocity as a function of distance from the shock and reveals significant agglomeration and evaporation problems with Titanium Oxide and Polystyrene Latex spheres. Particle acceleration measurements are presented, proving the unique concept of the new Laser Doppler probe. Mean and instantaneous acceleration data is extracted from high SNR signals. The acceleration data obtained is consistent in magnitude and trend with the physical phenomena expected and shows the feasibility of the new instrument. / Master of Science

This dissertation focuses on the development and application of numerical algorithms for solving compressible multiphase fluid-material interaction problems. The first part of this dissertation is motivated by the extraordinary shock-resisting ability of elastomer coating materials (e.g., polyurea) under explosive loading conditions. Their performance, however, highly depends on their dynamic interaction with the substrate (e.g., metal) and ambient fluid (e.g., air or liquid); and the detailed interaction process is still unclear. Therefore, to certify the application of these materials, a fluid-structure coupled computational framework is needed. The first part of this dissertation developes such a framework. In particualr, the hyper-viscoelastic constitutive relation of polyurea is incorporated into a high-fidelity computational framework which couples a finite volume compressible multiphase fluid dynamics solver and a nonlinear finite element structural dynamics solver. Within this framework, the fluid-structure and liquid-gas interfaces are tracked using embedded boundary and level set methods. Then, the developed computational framework is applied to study the behavior a bilayer coating–substrate (i.e., polyurea-aluminum) system under various loading conditions. The observed two-way coupling between the structure and the bubble generated in a near-field underwater explosion motivates the next part of this dissertation. The second part of this dissertation investigates the yielding and collapse of an underwater thin-walled aluminum cylinder in near-field explosions. As the explosion intensity varies by two orders of magnitude, three different modes of collapse are discovered, including one that appears counterintuitive (i.e., one lobe extending towards the explosive charge), yet has been observed in previous laboratory experiments. Because of the transition of modes, the time it takes for the structure to reach self-contact does not decrease monotonically as the explosion intensity increases. Detailed analysis of the bubble-structure interaction suggests that, in addition to the incident shock wave, the second pressure pulse resulting from the contraction of the explosion bubble also has a significant effect on the structure's collapse. The phase difference between the structural vibration and the bubble's expansion and contraction strongly influences the structure's mode of collapse. The third part focuses on the development of efficient solution algorithms for compressible multi-material flow simulations. In these simulations, an unresolved challenge is the computation of advective fluxes across material interfaces that separate drastically different thermodynamic states and relations. A popular class of methods in this regard is to locally construct bimaterial Riemann problems, and to apply their exact solutions in flux computation, such as the one used in the preceding parts of the dissertation. For general equations of state, however, finding the exact solution of a Riemann problem is expensive as it requires nested loops. Multiplied by the large number of Riemann problems constructed during a simulation, the computational cost often becomes prohibitive. This dissertation accelerates the solution of bimaterial Riemann problems without introducing approximations or offline precomputation tasks. The basic idea is to exploit some special properties of the Riemann problem equations, and to recycle previous solutions as much as possible. Following this idea, four acceleration methods are developed. The performance of these acceleration methods is assessed using four example problems that exhibit strong shock waves, large interface deformation, contact of multiple (>2) interfaces, and interaction between gases and condensed matters. For all the problems, the solution of bimaterial Riemann problems is accelerated by 37 to 87 times. As a result, the total cost of advective flux computation, which includes the exact Riemann problem solution at material interfaces and the numerical flux calculation over the entire computational domain, is accelerated by 18 to 81 times. / Doctor of Philosophy / This dissertation focuses on the development and application of numerical methods for solving multiphase fluid-material interaction problems. The first part of this dissertation is motivated by the extraordinary shock-resisting ability of elastomer coating materials (e.g., polyurea) under explosive loading conditions. Their performance, however, highly depends on their dynamic interaction with the underlying structure and the ambient water or air; and the detailed interaction process is still unclear. Therefore, the first part of this dissertation developes a fluid-structure coupled computational framework to certify the application of these materials. In particular, the special material property of the coating material is incorparated into a state-of-the-art fluid-structure coupled computational framework that is able to model large deformation under extreme physical conditions. Then, the developed computational framework is applied to study how a thin-walled aluminum cylinder with polyurea coating responds to various loading conditions. The observed two-way coupling between the structure and the bubble generated in a near-field underwater explosion motivates the next part of this dissertation. The second part of this dissertation investigates the failure (i.e., yielding and collapse) of an underwater thin-walled aluminum cylinder in near-field explosions. As the explosion intensity varies by two orders of magnitude, three different modes of collapse are discovered, including one that appears counterintuitive (i.e., one lobe extending towards the explosive charge), yet has been observed in previous laboratory experiments. Via a detailed analysis of the interaction between the explosion gas bubble, the aluminum cylinder, and the ambient liquid water, this dissertation elucidated the role of bubble dynamics in the structure's different failure behaviors and revealed the transition mechanism between these behaviors. The third part of this dissertation presents efficient solution algorithms for the simulations of compressible multi-material flows. Many problems involving bubbles, droplets, phase transitions, and chemical reactions fall into this category. In these problems, discontinuities in fluid state variables (e.g., density) and material properties arise across the material interfaces, challenging numerical schemes' accuracy and robustness. In this regard, a promising class of methods that emerges in the recent decade is to resolve the exact wave structure at material interfaces, such as the one used in the preceding parts of the dissertation. However, the computational cost of these methods is prohibitive due to the nested loops invoked at every mesh edge along the material interface. To address this issue, the dissertation develops four efficient solution methods, following the idea of exploiting special properties of governing equations and recycling previous solutions. Then, the acceleration effect of these methods is assessed using various challenging multi-material flow problems. In different test cases, significant reduction in computational cost (acceleration of 18 to 81 times) is achieved, without sacrificing solver robustness and solution accuracy.

The main focus of this work is to investigate thermal performance of self-contained drum motor drive systems (SCDMDS). All components of a SCDMDS are contained inside a rotating drum including the electric motor, gearbox, and an air/oil multiphase flow. A considerable amount of heat is generated within the SCDMDS from various sources, namely, the electric motor losses, the oil viscous dissipation and the gearbox losses. In meantime, a limited amount of heat is dissipated through the surface of the rotating drum and the side flanges. Therefore, a SCDMDS sometimes encounters a serious overheating problem, which often results in electric motor failure. The different heat generation and dissipation mechanisms as well as the two-phase flow within the SCDMDS have been studied experimentally and numerically under different operating parameters, namely, the oil level (OV), the drum rotational speed (N), the torque (ζ), the number of motor poles (n) and the electric motor dimensions. The effects of rubber lagging material and thickness as well as the use of rubber belts have been investigated as well. The numerical part of the present study has been carried out using Ansys-CFX and was validated using experimental data. Results showed that the optimum oil level (OV) for the best thermal performance is about 65%. The increase in the rotational speed (N) enhanced the heat transfer within the SCDMDS due to the improved oil splashing. Viscous dissipation (VD) between the motor stator and the rotating drive drum was found to be almost negligible. However, oil viscous dissipation within the gap between the motor rotor and stator was found to have an important effect on the thermal performance. An analytical model has been developed and implemented using MATLAB to estimate VD within the motor. The losses from the gearbox were studied experimentally and numerically considering planetary and co-axial gear trains. The numerical work was carried out using the KISSsoft and KISSsys software. Results showed that the increase in the drum rotational speed (N) or the drum torque (ζ) increased the gearbox losses. In the planetary gearbox, any increase in the OV increases the churning losses, however, the increase in OV increased the losses in the co-axial gearbox up to OV = 31% beyond which the losses remained constant. After understanding the complex interplay between all the heat generating and dissipating mechanisms within the SCDMDS, a number of possible modifications have been proposed in order to resolve the overheating problem. The effect of cooling the electric motor by using an axial air flow has been investigated. The effect of adding fins along the inner surface of the outer rotating drum has also been studied. Correlations of the various contributing mechanisms have been developed. Based on a thermal resistance network, a SCDMDS sizing and performance assessment computer software tool in the form of a digital twin (DT) has been developed. A user-friendly interface has been developed using Visual Basics and Excel. The DT estimates temperature distribution and the amount of heat generated and dissipated from each component within the SCDMDS and hence it identifies whether the case is considered safe to operate or overheating is expected. In overheated cases, the DT also suggests several possible modifications the user could consider to resolve the overheating problem. The DT has been validated against several experimental case studies and found to be very reasonably accurate. / Thesis / Doctor of Philosophy (PhD) / This study is focused on investigating heat transfer and fluid flow inside a self-contained drum motor drive system (SCDMDS). The problem of interest involves multiple heat sources enclosed inside a tight space of the rotating drum. There is an electrical motor, gearbox and a multiphase (oil/air) flow inside the rotating drum of the SCDMDS. In this thesis, experimental test rigs were constructed to investigate the effect of a number of operating and geometrical parameters. In addition, numerical analysis of the multiphase oil/air flow was carried out using Ansys - CFX. The KISSsoft and KISSsys software packages were used to determine various types of heat losses within the geartrain. Due to the presence of multiple heat sources inside a confined space, overheating of a number of SCDMDS has been reported. The overheating problem worsened even more when rubber lagging is used to increase traction between the drive drum and the belt. Several correlations have been developed for various heat transfer mechanisms governing the overall thermal performance of the entire SCDMDS. An analytical model (a digital twin) has been developed using Visual Basics and Excel. The digital twin estimates the temperature distribution and the amount of heat generated and dissipated inside the SCDMDS. It has been validated against many case studies provided by the industrial partner. The model identifies the possibility of overheating and provides the user with several potential modifications to resolve it. Hence, the model can be used as a performance and design tool of various models of SCDMDS.

Multiphase flows are simultaneous flows of two or more immiscible fluids in a pipe or vessel. Multiphase flows occur in a wide variety of industrial applications, such as chemical reactors, power generation, oil and gas production or transportation, etc. In most of these applications, efficiency and process reliability depend not insignificantly on the composition and flow morphology of these multiphase flows. Therefore, accurate determination of parameters such as phase fractions and their spatial distribution, as well as measurement of volumetric or mass flow rates, is essential to optimize and ensure correct operation of the equipment. For a better prediction of flow characteristics of multiphase systems, the development and validation of analytical models and CFD codes for simulations of multiphase flows has been promoted for some time in thermofluid dynamics research. For this purpose, the in-depth analysis of multiphase flows with high spatial and temporal resolution is essential. However, to date, there is no universal sensor that can directly measure all the required flow parameters over the full range of all flow conditions. Therefore, several strategies have been developed to solve this problem. For pure measurement of fluid composition and mixture volume flow, for example, the fluid mixture is often conditioned before measurement by separation into individual phases or by homogenization. However, this does not allow any more information about the flow morphology. In situations where the fluid cannot be preconditioned, for example when investigating bubble size distributions or predicting plug flows, imaging techniques such as wire-mesh sensors therefore play an important role because they provide cross-sectional images of the flow in rapid succession. This information can be used to determine phase distributions and identify flow regimes, which in turn can serve as input to other sensors to find optimal operating points. In addition, such information is important for validating models and numerical simulations. Although wire-mesh sensors are very attractive and now widely used due to their high spatial and temporal resolution, the measurement signals obtained from the sensor can be corrupted by energy losses and channel crosstalk under certain conditions. Therefore, a better understanding of the real physical conditions when using wire-mesh sensors is essential to improve the measurement accuracy and to extend the range of applications, e.g., for the measurement of media with very high conductivities or for an accurate quantification of individual phases in three-phase flows. In the present work, the current limitations of existing wire-mesh sensor systems are investigated in detail, thus providing a basis for technical improvements and the development of new methods for better interpretation of the measured values of wire-mesh sensors. For this purpose, the electronic measurement principle and the real sensor geometries are first investigated with respect to inherent energy losses and channel crosstalk. Based on mixing models, a method for visualization and quantification of three-phase gas-oil-water flows even in the presence of dispersions is presented. In addition, nonlinearities of wire-mesh sensors are predicted for the first time by a hybrid model based on the finite element method, which also incorporates the real parameters of the electronic components of signal generation and measurement. This model is subsequently used to generate synthetic data and to test new correction methods. Finally, two methods are proposed to compensate for unavoidable energy losses. The first method allows inherent determination of energy losses that cannot be suppressed by further circuit optimization. The second method allows determination of the voltage drop caused by the impedance of the electrodes when measured in highly conductive liquids. Numerical and experimental analyses show an improvement in the measurement accuracy of wire-mesh sensors with respect to the average and local phase fractions. The deviations of the average phase fraction were reduced from more than 15% to less than 2% and the deviations in local measurements from more than 30% to less than 5%.:Abstract 3 Zusammenfassung 5 Statement of authorship 9 Acronyms 13 Symbols 15 1. Introduction 17 2. State of the science and technology 21 3. Wire-mesh sensor and experimental test facilities 43 4. Three-phase flow measurement based on dual-modality wire-mesh sensor 53 5. Wire-mesh sensor model based on finite-element method and circuit simulation 67 6. Analysis of non-linear effects in measurements of wire-mesh sensor 79 7. Methods for improving the measurement accuracy of wire-mesh sensors 87 8. Conclusions and outlook 97 Bibliography 101 Appendices 111 A. List of scientific publications 113

No description available.

No description available.

No description available.

No description available.

No description available.