Made available in DSpace on 2014-06-11T19:22:18Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-02-27Bitstream added on 2014-06-13T19:27:07Z : No. of bitstreams: 1 lopes_jh_me_sjrp.pdf: 490230 bytes, checksum: f1439f3747d9808ad9df33bfcbfc16d3 (MD5) / Neste trabalho, estudaremos as equações de Navier-Stokes em Rn e mostraremos a existência de solução global, quando a velocidade inicial u0(x) pertence ao espaço Lp-fraco e tem norma suficientemente pequena. A análise da evolução da solução é realizada em espaços funcionais de Kato-Fujita, invariantes pelo scaling de Navier-Stokes. Mostraremos também que se u0(x) é homogênea de grau −1, as soluções também são invariantes por este scaling, ou seja, elas são auto-similares. Além disso, mostraremos a estabilidade assintótica das soluções mild / In this work, we study the Navier-Stokes equations in Rn and show the existence of global solution, when the initial velocity u0(x) belongs to weak Lp space with a sufficiently small norm. The evolution of the solution is analyzed in function spaces with Kato-Fujita type norms invariant by scaling of Navier-Stokes. We also show that if u0 is an homogeneous function of degree −1, the solutions are also invariant by that scaling, i.e., they are self-similar. Moreover, we show the asymptotic stability of mild solutions

This thesis deals with the problem of the asymptotic behavior of solutions to several nonlinear equations from fluid dynamics on both mesoscopic and macroscopic levels, including Boltzmann equation, compressible Navier-Stokes equations and the system of viscous conservation laws with positive definite viscosity matrix. The main purpose is to study the asymptotic behavior of solutions to those equations towards linear and nonlinear waves, such as shock waves, rarefaction waves and contact discontinuities as either the times goes to infinity, or the viscosity and heat conductivity go to zero for the macroscopic equations or the mean free path goes to zero for the mesoscopic equations. Those limit processes are singular. For the system of viscous conservation laws, we show the large time asymptotic nonlinear stability of a superposition of viscous shock waves and viscous contact waves for the system of viscous conservation laws with small initial perturbations, provided that the strengths of these viscous waves are small and of the same order. The results are obtained by elementary weighted energy estimates based on the underlying wave structure and a new estimate on the heat equation. For the Boltzmann equation, the main purpose is to study the asymptotic equivalence for the hard-sphere collision model to its corresponding Euler equations of compressible gas dynamics in the limit of small mean free path. When the fluid flow is a smooth rarefaction (or centered-rarefaction) wave with finite strength, the corresponding Boltzmann solution exists globally in time, and the solution converges to the rarefaction wave uniformly for all time (or away from t = 0) as the mean free path epsilon → 0. A decomposition of a Boltzmann solution into its macroscopic (fluid) part and microscopic (kinetic) part is adopted to rewrite the Boltzmann equation in a form of compressible Navier-Stokes equations with source terms. As a by-product, the same asymptotic equivalence of the full compressible Navier-Stokes equations to its corresponding Euler equations in the limit of small viscosity and heat-conductivity (depending on the viscosity) is also obtained. / Zeng, Huihui. / Adviser: Zhouping Xin. / Source: Dissertation Abstracts International, Volume: 70-09, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2009. / Includes bibliographical references (leaves 102-110). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.

Neste trabalho abordam-se alguns aspectos numéricos relacionados ao movimento de uma partícula carregada em um fluido ionizado, constituindo assim o fenômeno de eletroforese. A equação de Navier-Stokes governa a hidrodinâmica do sistema, sendo atribuída a esta um termo forçante relacionado a equação de Poisson-Boltzmann que descreve a interação com o campo elétrico. A simulação foi realizada utilizando a linguagem Fortran 90 e as discretizações foram feitas pelo método de diferenças finitas junto as equações governantes. Para descrever a posição da partícula foi usada a técnica de fronteira virtual, onde são criados pontos lagrangeanos sobre,o contorno da partícula formando uma espécie de segunda malha. Foram feitos testes com objetivo de validar as equações governantes e identificar a influencia das condições de contorno direção da partícula. / In this work we study some of the numerical aspects related to the motion of a charged particle in an ionized fluid, known as the phenomenon of electrophoresis. The Navier-Stokes equations models the ,hydrodynamics of the system, with the addition of a forcing term, related to the Poisson-Boltzmann equation, which describes the interaction of the electric field. The simulation was performed using Fortran 90 and discretizations were made by the method of finite differences applied to the governing equations. To describe the position of the particle the technique of virtual boundary is used, where Lagrangean points are created around the circumference of the particle forming a sort of second mesh. Tests were made with the objective of validating the governing equations and to identify the influence of boundary conditions in the direction of the particle.

Orientador: Juliana Conceição Precioso Pereira / Banca: Lucas Catão de Freitas Ferreira / Banca: Andréa Cristina Prokopczyk Arita / Resumo: Neste trabalho, estudaremos as equações de Navier-Stokes em Rn e mostraremos a existência de solução global, quando a velocidade inicial u0(x) pertence ao espaço Lp-fraco e tem norma suficientemente pequena. A análise da evolução da solução é realizada em espaços funcionais de Kato-Fujita, invariantes pelo scaling de Navier-Stokes. Mostraremos também que se u0(x) é homogênea de grau −1, as soluções também são invariantes por este scaling, ou seja, elas são auto-similares. Além disso, mostraremos a estabilidade assintótica das soluções mild / Abstract: In this work, we study the Navier-Stokes equations in Rn and show the existence of global solution, when the initial velocity u0(x) belongs to weak Lp space with a sufficiently small norm. The evolution of the solution is analyzed in function spaces with Kato-Fujita type norms invariant by scaling of Navier-Stokes. We also show that if u0 is an homogeneous function of degree −1, the solutions are also invariant by that scaling, i.e., they are self-similar. Moreover, we show the asymptotic stability of mild solutions / Mestre

In this work we study the existence of weak solutions for compressible Navier-Stokes equations in unbounded time dependent domains. Using the methods introduced in Feireisl E. Dynamics of Viscous Compressible Fluids we extend the results of article Feireisl E. Neustupa J. Stebel J., Convergence of a Brinkman-type penalization for compressible fluid flows, which studies the flow with a "no-slip" boundary condition on bounded domains. Next, we extend results of article Feireisl E. Kreml O. Nečasová Š. Neustupa J. Stebel J., Weak solutions to the barotropic Navier- Stokes system with slip boundary conditions in time dependent domains, which studies flow with compete Navier boundary condition. Finally, we discuss solutions for rotating fluid system. In this case, there are new members in momentum equation, representing the Coriolis and centrifugal force, which cause problems.

Title: Regularity criteria for instationary incompressible Navier-Stokes equations Author: Šimon Axmann Institute: Mathematical Institute of Charles University Supervisor: doc. Mgr. Milan Pokorný, Ph.D., Mathematical Institute of Charles University Abstract: In the present thesis we study the global conditional regularity of weak solutions to the Cauchy problem for instationary incompressible Navier-Stokes equations in three space dimensions. In the first section, we present an overview of known conditions implying the full regularity of the equations under conside- ration. For the sake of clarity, we expose only the regularity criteria on the scale of Lebesgue spaces, especially in terms of the velocity and its components, the gradient of the velocity and its components, the pressure and the vorticity. In the subsequent sections, we generalize four regularity criteria using two different techniques. We are able to replace one velocity component or its gradient, consi- dered in the known results, by a projection of the velocity into a general vector field. For the purpose of the second method, we also generalize the multiplicative Gagliardo-Nirenberg inequality.

The lattice Boltzmann (LB) method is a versatile way to model complex fluids with hydrodynamic interactions through solving the Navier-Stokes equations. It is well-known that the role of hydrodynamic interactions is ignorable in studying the Boltzmann equilibrium of colloidal (Brownian) particles. However, full hydrodynamic interactions play an important role in their dynamics. In the LB framework for moving colloids, the “bounce-back on links” method is used to calculate the hydrodynamic forces. In this thesis, three kinds of colloidal complex fluids with full hydrodynamic interactions are simulated by lattice Boltzmann methods: colloids in a binary fluid, magnetic colloids in a single fluid and magnetic colloids in a binary fluid. First, we have done extensive simulations of nanoparticles in a binary fluid, following up previous work[1] which predicted formation of a “bijel” (bicontinuous interfacially jammed emulsion gel) in symmetric fluid quenches. Our work in this thesis focuses on the analysis of the dynamics after nanoparticles become arrested on the fluid-fluid interfaces under conditions varying from a symmetric quench to a strongly asymmetric quench. Although these new simulations extend the time window studied by a factor of two, slow domain growth is still observed. Our new analyses address the mechanics of the slow residual dynamics which involves cooperative motion of the nanoparticles at the fluid-fluid interfaces. The second topic is the LB simulation of colloidal ferrofluids to see the effect of full hydrodynamic interactions among magnetic colloids. The main focus is on how the hydrodynamic interaction affects both the equilibrium dynamics of these dipolar systems and also their transient dynamics to form clusters. Numerically, magnetic colloids are implemented with the long-range dipolar interactions described by Ewald summation. To check the effect of full hydrodynamic interactions, Brownian dynamics without any hydrodynamic interaction has been done for comparison: Monte Carlo results are also reported. We confirm that our LB generates the Boltzmann distribution for static equilibrium properties, by comparison with these methods. However, the equilibrium dynamics is altered: hydrodynamic interactions make the structural relaxations slower in both the short-time and the long-time regime. This slow relaxation rate is also found for transient motions. The third topic addresses magnetic colloids in a binary fluid. In contrast with the preceding two systems which correspond directly to laboratory experiments, this last system is so far only predicted by the LB results in this thesis. To explore this hypothetical new material by the LB method, the basic structures are investigated in terms of both domain growth morphology and the arrangement of magnetic colloids. Under conditions varying from a symmetric quench to an asymmetric quench, a chainlike arrangement is observed for dipoles jammed on the surfaces, but the basic morphology of domains is still maintained regardless of the dipolar strength. In addition, applying external field affects the morphology of domains and the stability of domain structures.

This dissertation focuses on the study of spectral stability in traveling waves, with a special interest in planar detonations in the multidimensional reactive Navier-Stokes equations. The chief tool is the Evans function, combined with STABLAB, a numerical library devoted to calculating the Evans function. Properly constructed, the Evans function is an analytic function in the right half-plane whose zeros correspond in multiplicity and location to the spectrum of the traveling wave. Thus the Evans function can be used to verify stability, or to locate precisely any unstable eigenvalues. We introduce a new method that uses numerical continuation to follow unstable eigenvalues as system parameters vary. We also use the Evans function to track instabilities of viscous detonations in the multidimensional reactive Navier-Stokes equations, building on recent results for detonations in one dimension. Finally, we introduce a Python implementation of STABLAB, which we hope will improve the accessibility of STABLAB and aid the future study of large, multidimensional systems by providing easy-to-use parallel processing tools.

This thesis derives a first approximation model that will describe the flow of fluid in microfluidic devices such as in microchannels, microdiffusers and micronozzles using electrical network modelling. The important parameter that is of concern is the flow rates of these devices. The purpose of this work is to contribute to the physical component of our interactive Virtual Reality (VR)-prototyping tool for MEMS, with emphasis on fast calculations for interactive CAD design. Current calculations are too time consuming and not suitable for interactive CAD with dynamic animations. This work contributes to and fills the need for the development of MEMS dynamic visualisation, showing the movement of fluid within microdevices in time scale. Microfluidic MEMS devices are used in a wide range of applications, such as in chemical analysis, gene expression analysis, electronic cooling system and inkjet printers. Their success lies in their microdimensions, enabling the creation of systems that are considerably minute yet can contain many complex subsystems. With this reduction in size, the advantages of requiring less material for analysis, less power consumption, less wastage and an increase in portability becomes their selling point. Market size is in excess of US$50 billion in 2004, according to a study made by Nexus. New applications are constantly being developed leading to creation of new devices, such as the DNA and the protein chip. Applications are found in pharmaceuticals, diagnostic, biotechnology and the food industry. An example is the outcome of the mapping and sequencing of the human genome DNA in the late 1990's leading to greater understanding of our genetic makeup. Armed with this knowledge, doctors will be able to treat diseases that were deemed untreatable before, such as diabetes or cancer. Among the tools with which that can be achieved include the DNA chip which is used to analyse an individual's genetic makeup and the Gene chip used in the study of cancer. With this burgeoning influx of new devices and an increase in demand for them there is a need for better and more efficient designs. The MEMS design process is time consuming and costly. Many calculations rely on Finite Element Analysis, which has slow and time consuming algorithms, that make interactive CAD unworkable. This is because the iterative algorithms for calculating the animated images showing the ongoing proccess as they occur, are too slow. Faster computers do not solve the void of efficient algorithms, because with faster computer also comes the demand for a fasters response. A 40 - 90 minute FEA calculation will not be replaced by a faster computer in the next decades to an almost instant response. Efficient design tools are required to shorten this process. These interactive CAD tools need to be able to give quick yet accurate results. Current CAD tools involve time consuming numerical analysis technique which requires hours of numerous iterations for the device structure design followed by more calculations to achieve the required output specification. Although there is a need for a detailed analysis, especially in solving for a particular aspect of the design, having a tool to quickly get a first approximation will greatly shorten the guesswork involved in determining the overall requirement. The underlying theory for the fluid flow model is based on traditional continuum theory and the Navier-Stokes equation is used in the derivation of a layered flow model in which the flow region is segmented into layered sections, each having different flow rates. The flow characteristics of each sections are modeled as electrical components in an electrical circuit. Matlab 6.5 (MatlabTM) is used for the modelling aspect and Simulink is used for the simulation.

Cette thèse porte sur le développement de méthodes numériques pour le calcul d'écoulements incompressibles multi-fluides sur grille de calcul, s'inscrivant ainsi dans le cadre du projet MecaGrid.Une étude de la grille MecaGrid met en évidence son caractère hétérogène et ses conséquences. Plusieurs techniques d'optimisation sont présentées afin d'améliorer son utilisation : répartir la masse de calcul de façon adaptée, et privilégier le travail des processeurs vis-à-vis des communications réseaux pénalisantes.Des méthodes numériques sur maillage non structuré composé de tétraèdres sont choisies pour réaliser la simulation directe d'écoulements multi-fluides avec capture d'interface. Nous adoptons une approche unique qui rappelle celle du Multiscale, dans laquelle la condensation d'une fonction Bulle pyramidale est utilisée comme technique universelle de stabilisation.Les équations de Navier-Stokes incompressible sont résolues par une méthode éléments finis mixtes à interpolation P1+/P1. Un schéma temporel d'Euler implicite est appliqué, en association avec un algorithme de Newton pour linéariser le problème.Les interfaces sont capturées par une technique Level Set à interpolation continue P1 qui consiste à résoudre une équation de transport stabilisée par la condensation d'une Bulle (Residual-Free Bubbles). Le couplage avec une équation d'Hamilton-Jacobi permet de réinitialiser la fonction Level Set au court de son transport. La comparaison avec une méthode Galerkin discontinue proche du Volume of Fluid montre que le Level Set se distingue par sa simplicité et l'absence de diffusion numérique.Enfin, les simulations numériques sont validées par plusieurs cas test reconnus.