Ce travail a été consacré à la simulation numérique des équations de la mécanique des fluides par des méthodes de volumes finis implicites. Tout d’abord, nous avons étudié et mis en place une version implicite du schéma de Roe pour les écoulements monophasiques et diphasiques compressibles. Grâce à la méthode de Newton utilisée pour résoudre les systèmes nonlinéaires, nos schémas sont conservatifs. Malheureusement, la résolution de ces systèmes est très coûteuse. Il est donc impératif d’utiliser des algorithmes de résolution performants. Pour des matrices de grande taille, on utilise souvent des méthodes itératives dont la convergence dépend de leur spectre. Nous avons donc étudié le spectre du système linéaire et proposé une stratégie de Scaling pour améliorer le conditionnement de la matrice. Combinée avec le préconditionneur classique ILU, notre stratégie de Scaling a réduit de façon significative le nombre d’itérations GMRES du système local et le temps de calcul. Nous avons également montré l’intérêt du schéma centré pour la simulation de certains écoulements à faible nombre de Mach. Nous avons ensuite étudié et implémenté la méthode de décomposition de domaine pour les écoulements compressibles. Nous avons proposé une nouvelle variable interface qui rend la méthode du complément de Schur plus facile à construire et nous permet de traiter les termes de diffusion. L’utilisation du solveur itératif GMRES plutôt que Richardson pour le système interface apporte aussi une amélioration des performances par rapport aux autres méthodes. Nous pouvons également découper notre domaine de calcul en un nombre quelconque de sous-domaines. En utilisant la stratégie de Scaling pour le système interface, nous avons amélioré le conditionnement de la matrice et réduit le nombre d’itérations GMRES de ce système. En comparaison avec le calcul distribué classique, nous avons montré que notre méthode est robuste et efficace. / This thesis deals with numerical simulations of compressible fluid flows by implicit finite volume methods. Firstly, we studied and implemented an implicit version of the Roe scheme for compressible single-phase and two-phase flows. Thanks to Newton method for solving nonlinear systems, our schemes are conservative. Unfortunately, the resolution of nonlinear systems is very expensive. It is therefore essential to use an efficient algorithm to solve these systems. For large size matrices, we often use iterative methods whose convergence depends on the spectrum. We have studied the spectrum of the linear system and proposed a strategy, called Scaling, to improve the condition number of the matrix. Combined with the classical ILU preconditioner, our strategy has reduced significantly the GMRES iterations for local systems and the computation time. We also show some satisfactory results for low Mach-number flows using the implicit centered scheme. We then studied and implemented a domain decomposition method for compressible fluid flows. We have proposed a new interface variable which makes the Schur complement method easy to build and allows us to treat diffusion terms. Using GMRES iterative solver rather than Richardson for the interface system also provides a better performance compared to other methods. We can also decompose the computational domain into any number of subdomains. Moreover, the Scaling strategy for the interface system has improved the condition number of the matrix and reduced the number of GMRES iterations. In comparison with the classical distributed computing, we have shown that our method is more robust and efficient.

O método ISPH (do inglês, Incompressible Smoothed Particle Hydrodynamics) é um método de aproximação livre de malha que, através de um conjunto finito de partículas e uma formulação completamente Lagrangeana, permite a solução de diversos tipos de escoamentos. Entretanto, sua aplicação para escoamentos bifásicos ainda é um desafio, principalmente no que refere-se à manutenabilidade da interface entre fluidos. Diante disso, nesta tese é apresentado o desenvolvimento de um código numérico baseado no método ISPH, sendo propostas duas técnicas de tratamento de interface. Para tanto é realizado um estudo a cerca do método, considerando diferentes metodologias, e analisando pontos específicos, tais como a solução do campo de pressões. São apresentados resultados que mostram a eficácia do método, tanto em escoamentos monofásicos, quanto em escoamentos multifásicos, onde, neste caso, são destacadas as melhorias obtidas através das técnicas de tratamento de interface propostas. Por fim, é realizado um estudo do comportamento de misturas bifásicas, com referência ao fenômeno da inversão de fase / Incompressible Smoothed Particle Hydrodynamics (ISPH) method is a meshless approximation that has been used to simulate several types of fluid flows, through a finite particle set and fully lagrangian formulation. The application of ISPH method in two-fluid flow simulations however, has presented many challenges, specially related to the presence of the interface between different fluids. Thus, we present in this study the development of a numerical code based on ISPH, introducing novel interface treatment techniques. A thorough study about this method is provided, considering different methodologies and analysing specific points such as the position of the interface and the obtained pressure field. Results have been presented to show the methods developed in this thesis efficiently simulate two-fluid flows, illustrating the improvements achieved by the proposed interface treatment techniques. Finally, a study of biphasic mixture behavior is carried out with reference to phase inversion phenomena

We develop a two-fluid model (TFM) for heat transfer in dense non-Brownian suspensions. Specifically, we propose closure relations for the inter-phase heat transfer coefficient and the thermal diffusivity of the particle phase based on calibration against experimental data. The model is then employed to simulate non-isothermal flow in an annular Couette cell. We find that, when the shear rate is controlled by the rotation of the inner cylinder, both the shear and thermal gradients are responsible for particle migration. Within the TFM framework, we identify the origin and functional form of a "thermo-rheological" migration force that rationalizes our observations. Furthermore, we apply our model to flow in eccentric Couette cells. Our simulations reveal that the system's heat transfer coefficient is affected by both the classic shear-induced migration of particles and the newly identified thermo-rheological migration effect. Finally, we employed the proposed computational TFM framework to analyze electronics cooling by forced convection for microchannel cooling. We used a suspensions of high thermal conductivity (Boron Nitride) particles in a 3M Fluorinert FC-43 cooling fluid. Three-dimensional simulations were run to quantify the temperature distributions under uniform heating (5 W) and under hot-spot heating (2 W/cm^2) conditions. A 100 K junction level temperature improvement (enhanced thermal spreading) was seen for hot-spot heating and 15 K was observed for uniform heating, demonstrating the enhanced cooling capabilities of dense particulate suspensions of high-conductivity particles, over a clear FC-43 fluid.

abstract: The applications utilizing nanoparticles have grown in both industrial and academic areas because of the very large surface area to volume ratios of these particles. One of the best ways to process and control these nanoparticles is fluidization. In this work, a new microjet and vibration assisted (MVA) fluidized bed system was developed in order to fluidize nanoparticles. The system was tested and the parameters optimized using two commercially available TiO2 nanoparticles: P25 and P90. The fluidization quality was assessed by determining the non-dimensional bed height as well as the non-dimensional pressure drop. The non-dimensional bed height for the nanosized TiO2 in the MVA system optimized at about 5 and 7 for P25 and P90 TiO2, respectively, at a resonance frequency of 50 Hz. The non-dimensional pressure drop was also determined and showed that the MVA system exhibited a lower minimum fluidization velocity for both of the TiO2 types as compared to fluidization that employed only vibration assistance. Additional experiments were performed with the MVA to characterize the synergistic effects of vibrational intensity and gas velocity on the TiO2 P25 and P90 fluidized bed heights. Mathematical relationships were developed to correlate vibrational intensity, gas velocity, and fluidized bed height in the MVA. The non-dimensional bed height in the MVA system is comparable to previously published P25 TiO2 fluidization work that employed an alcohol in order to minimize the electrostatic attractions within the bed. However, the MVA system achieved similar results without the addition of a chemical, thereby expanding the potential chemical reaction engineering and environmental remediation opportunities for fluidized nanoparticle systems. In order to aid future scaling up of the MVA process, the agglomerate size distribution in the MVA system was predicted by utilizing a force balance model coupled with a two-fluid model (TFM) simulation. The particle agglomerate size that was predicted using the computer simulation was validated with experimental data and found to be in good agreement. Lastly, in order to demonstrate the utility of the MVA system in an air revitalization application, the capture of CO2 was examined. CO2 breakthrough time and adsorption capacities were tested in the MVA system and compared to a vibrating fluidized bed (VFB) system. Experimental results showed that the improved fluidity in the MVA system enhanced CO2 adsorption capacity. / Dissertation/Thesis / Doctoral Dissertation Mechanical Engineering 2019

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第18590号 / 工博第3951号 / 新制||工||1607(附属図書館) / 31490 / 京都大学大学院工学研究科原子核工学専攻 / (主査)教授 功刀 資彰, 教授 中部 主敬, 准教授 横峯 健彦 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

No description available.

This thesis describes a research study relating naturalistic Global Positioning System (GPS) driving data with long-term traffic safety performance for two classes of roadways. These two classes are multilane arterial streets and limited access highways. GPS driving data used for this study was collected from 33 volunteer drivers from July 2012 to March 2013. The GPS devices used were custom GPS data loggers capable of recording speed, position, and other attributes at an average rate of 2.5 hertz. Linear Referencing in ESRI ArcMAP was performed to assign spatial and other roadway attributes to each GPS data point collected. GPS data was filtered to exclude data with high horizontal dilution of precision (HDOP), incorrect heading attributes or other GPS communication errors. For analysis of arterial roadways, the Two-Fluid model parameters were chosen as the measure for long-term traffic safety analysis. The Two-Fluid model was selected based on previous research which showed correlation between the Two-Fluid model parameters n and Tm and total crash rate along arterial roadways. Linearly referenced GPS data was utilized to obtain the total travel time and stop time for several half-mile long trips along two arterial roadways, Grand Avenue and California Boulevard, in San Luis Obispo. Regression between log transformed values of these variables (total travel time and stop time) were used to derive the parameters n and Tm. To estimate stop time for each trip, a vehicle “stop” was defined when the device was traveling at less than 2 miles per hour. Results showed that Grand Avenue had a higher value for n and a lower value for Tm, which suggests that Grand Avenue may have worse long-term safety performance as characterized by long-term crash rates. However, this was not verified with crash data due to incomplete crash data in the TIMS database. Analysis of arterial roadways concluded by verifying GPS data collected in the California Boulevard study with sample data collected utilizing a traditional “car chase” methodology, which showed that no significant difference in the two data sources existed when trips included noticeable stop times. For analysis of highways the derived measurement of vehicle jerk, or rate of change of acceleration, was calculated to explore its relationship with long-term traffic safety performance of highway segments. The decision to use jerk comes from previous research which utilized high magnitude jerk events as crash surrogate, or near-crash events. Instead of using jerk for near-crash analysis, the measurement of jerk was utilized to determine the percentage of GPS data observed below a certain negative jerk threshold for several highway segments. These segments were ¼-mile and ½-mile long. The preliminary exploration was conducted with 39 ¼-mile long segments of US Highway 101 within the city limits of San Luis Obispo. First, Pearson’s correlation coefficients were estimated for rate of ‘high’ jerk occurrences on these highway segments (with definitions of ‘high’ depending on varying jerk thresholds) and an estimate of crash rates based on long-term historical crash data. The trends in the correlation coefficients as the thresholds were varied led to conducting further analysis based on a jerk threshold of -2 ft./sec3 for the ¼-mile segment analysis and -1 ft./sec3 for the ¼-mile segment analysis. Through a negative binomial regression model, it was shown that utilizing the derived jerk percentage measure showed a significant correlation with the total number of historical crashes observed along US Highway 101. Analysis also showed that other characteristics of the roadway, including presences of a curve, presence of weaving (indicated by the presence of auxiliary lanes), and average daily traffic (ADT) did not have a significant correlation with observed crashes. Similar analysis was repeated for 19 ½-mile long segments in the same study area, and it was found the percentage of high negative jerk metric was again significant with historical crashes. The ½-mile negative binomial regression for the presence of curve was also a significant variable; however the standard error for this determination was very high due to a low sample size of analysis segments that did not contain curves. Results of this research show the potential benefit that naturalistic GPS driving data can provide for long-term traffic safety analysis, even if data is unaccompanied with any additional data (such as live video feed) collected with expensive vehicle instrumentation. The methodologies of this study are repeatable with many GPS devices found in certain consumer electronics, including many newer smartphones.

No description available.

In recent years, the discontinuous Galerkin (DG) method has been successfully applied to solving hyperbolic conservation laws. Due to its compactness, high order accuracy, and versatility, the DG method has been extensively applied to convection-diffusion problems. In this dissertation, a numerical package, texttt{PHORCE}, is introduced to solve a number of convection-diffusion problems in neutral fluids and plasmas. Unstructured grids are used in order to randomize grid errors, which is especially important for complex geometries. texttt{PHORCE} is written in texttt{C++} and fully parallelized using the texttt{MPI} library. Memory optimization has been considered in this work to achieve improved efficiency. DG algorithms for hyperbolic terms are well studied. However, an accurate and efficient diffusion solver still constitutes ongoing research, especially for a nodal representation of the discontinuous Galerkin (NDG) method. An affine reconstructed discontinuous Galerkin (aRDG) algorithm is developed in this work to solve the diffusive operator using an unstructured NDG method. Unlike other reconstructed/recovery algorithms, all computations can be performed on a reference domain, which promotes efficiency in computation and storage. In addition, to the best of the authors' knowledge, this is the first practical guideline that has been proposed for applying the reconstruction algorithm on a nodal discontinuous Galerkin method. TVB type and WENO type limiters are also studied to deal with numerical oscillations in regions with strong physical gradients in state variables. A high-order positivity-preserving limiter is also extended in this work to prevent negative densities and pressure. A new interface tracking method, mass of fluid (MOF), along with its bound limiter has been proposed in this work to compute the mass fractions of different fluids over time. Hydrodynamic models, such as Euler and Navier-Stokes equations, and plasma models, such as ideal-magnetohydrodynamics (MHD) and two-fluid plasma equations, are studied and benchmarked with various applications using this DG framework. Numerical computations of Rayleigh-Taylor instability growth with experimentally relevant parameters are performed using hydrodynamic and MHD models on planar and radially converging domains. Discussions of the suppression mechanisms of Rayleigh-Taylor instabilities due to magnetic fields, viscosity, resistivity, and thermal conductivity are also included. This work was partially supported by the US Department of Energy under grant number DE-SC0016515. The author acknowledges Advanced Research Computing at Virginia Tech for providing computational resources and technical support that have contributed to the results reported within this work. URL: http://www.arc.vt.edu / Doctor of Philosophy / High-energy density (HED) plasma science is an important area in studying astrophysical phenomena as well as laboratory phenomena such as those applicable to inertial confinement fusion (ICF). ICF plasmas undergo radial compression, with an aim of achieving fusion ignition, and are subject to a number of hydrodynamic instabilities that can significantly alter the implosion and prevent sufficient fusion reactions. An understanding of these instabilities and their mitigation mechanisms is important allow for a stable implosion in ICF experiments. This work aims to provide a high order accurate and robust numerical framework that can be used to study these instabilities through simulations. The first half of this work aims to provide a detailed description of the numerical framework, texttt{PHORCE}. texttt{PHORCE} is a high order numerical package that can be used in solving convection-diffusion problems in neutral fluids and plasmas. Outstanding challenges exist in simulating high energy density (HED) hydrodynamics, where very large gradients exist in density, temperature, and transport coefficients (such as viscosity), and numerical instabilities arise from these region if there is no intervention. These instabilities may lead to inaccurate results or cause simulations to fail, especially for high-order numerical methods. Substantial work has been done in texttt{PHORCE} to improve its robustness in dealing with numerical instabilities. This includes the implementation and design of several high-order limiters. An novel algorithm is also proposed in this work to solve the diffusion term accurately and efficiently, which further enriches the physics that texttt{PHORCE} can investigate. The second half of this work involves rigorous benchmarks and experimentally relevant simulations of hydrodynamic instabilities. Both advection and diffusion solvers are well verified through convergence studies. Hydrodynamic and plasma models implemented are also validated against results in existing literature. Rayleigh-Taylor instability growth with experimentally relevant parameters are performed on both planar and radially converging domains. Although this work is motivated by physics in HED hydrodynamics, the emphasis is placed on numerical models that are generally applicable across a wide variety of fields and disciplines.

This paper describes the modelling of bubbly flow in a bubble column considering non-drag forces, polydispersity and bubble induced turbulence using the Eulerian two-fluid approach. The set of used closure models describing the momentum exchange between the phases was chosen on basis of broad experiences in modelling bubbly flows at the Helmholtz-Zentrum Dresden-Rossendorf. Polydispersity is modeled using the inhomogeneous multiple size group (iMUSIG) model, which was developed by ANSYS/CFX and Helmholtz-Zentrum Dresden-Rossendorf. Through the importance of a comprehensive turbulence modeling for coalescence and break-up models, bubble induced turbulence models are investigated. A baseline has been used which was chosen on the basis of our previous work without any adjustments. Several variants taken from the literature are shown for comparison. Transient CFD simulations are compared with the experimental measurements and Large Eddy Simulations of Akbar et al. (2012).