Search

In dieser Arbeit werden neuartige methodische Erweiterungen der Lattice-Boltzmann-Methode (LBM) entwickelt, die effizientere Simulationen inkompressibler Wirbelströmungen ermöglichen. Diese Erweiterungen beheben zwei Hauptprobleme der Standard-LBM: ihre Instabilität in unteraufgelösten turbulenten Simulationen und ihre Beschränkung auf reguläre Rechengitter. Dazu wird zunächst eine Pseudo-Entropische Stabilisierung (PES) entwickelt. Diese kombiniert Ansätze der Multiple-Relaxation-Time (MRT)-Modelle und der Entropischen LBM zu einem expliziten, lokalen und flexiblen Stabilisierungsoperator. Diese Modifikation des Kollisionsschritts erlaubt selbst auf stark unteraufgelösten Gittern stabile und qualitativ korrekte Simulationen. Zur Erweiterung der LBM auf irreguläre Rechengitter wird zunächst eine moderne Discontinuous-Galerkin-LBM untersucht und um stabilere Zeitintegratoren ergänzt. Diese Studie demonstriert die drastischen Schwächen existierender LBMAnsätze auf irregulären Gittern. Basierend auf den gewonnenen Erkenntnissen gelingt die Formulierung einer neuartigen Semi-Lagrangeschen LBM (SLLBM). Diese ermöglicht in einzigartigerWeise sowohl die Verwendung irregulärer Gitter und großer Zeitschritte als auch eine hohe räumliche Konvergenzordnung. Anhand von Beispielsimulationen wird demonstriert, wieso dieser Ansatz anderen aktuellen Off-Lattice-Boltzmann-Methoden (OLBMs) in Effizienz und Genauigkeit überlegen ist. Weitere neuartige Aspekte dieser Arbeit sind die Entwicklung eines modularen Off-Lattice-Boltzmann-Codes und die Erweiterung der LBM um implizite Mehrschrittverfahren, mit denen eine Erhöhung der zeitlichen Konvergenzordnung gelingt. The present work develops novel methodological extensions to the lattice Boltzmann method (LBM). These extensions enable the method to efficiently simulate incompressible vortical flows. They cure two major drawbacks of the standard LBM: its instability in under-resolved turbulence and its restriction to regular computational grids. At first, a pseudo-entropic stabilizer (PES) is developed, which combines ideas from multiple-relaxation-time (MRT) models and entropic models. The new PES is local, explicit, and flexible. It modifies the collision step in a way that enables stable simulations and produces qualitatively matching results, even on strongly underresolved grids. To extend the LBM towards simulations on irregular grids, a recent discontinuous Galerkin lattice Boltzmann method is studied and enhanced by more stable time integrators. This study illustrates the severe shortcomings of existing off-lattice Boltzmann methods (OLBMs). Based on these findings, the present work succeeds in developing a semi-Lagrangian lattice Boltzmann method (SLLBM). This novel approach allows unstructured grids, large time steps, and high-order accurate representations of the solution to be used in a unique way. Applications to exemplary flows demonstrate how and why the new method outperforms other recent OLBMs in both efficiency and accuracy. Additionally, this work describes the development of a modular off-lattice Boltzmann code and shows that the method’s convergence order can be increased by implicit multistep methods.

This paper reports, for the first time, the computational performance of SequenceL for mesoscale simulations of large numbers of particles in a microfluidic device via the lattice-Boltzmann method. The performance of SequenceL simulations was assessed against the optimized serial and parallelized (via OpenMP directives) FORTRAN90 simulations. At present, OpenMP directives were not included in interparticle and particle-wall repulsive (steric) interaction calculations due to difficulties that arose from inter-iteration dependencies between consecutive iterations of the do-loops. SequenceL simulations, on the other hand, relied on built-in automatic parallelism. Under these conditions, numerical simulations revealed that the parallelized FORTRAN90 outran the performance of SequenceL by a factor of 2.5 or more when the number of particles was 100 or less. SequenceL, however, outran the performance of the parallelized FORTRAN90 by a factor of 1.3 when the number of particles was 300. Our results show that when the number of particles increased by 30-fold, the computational time of SequenceL simulations increased linearly by a factor of 1.5, as compared to a 3.2-fold increase in serial and a 7.7-fold increase in parallelized FORTRAN90 simulations. Considering SequenceL's efficient built-in parallelism that led to a relatively small increase in computational time with increased number of particles, it could be a promising programming language for computationally-efficient mesoscale simulations of large numbers of particles in microfluidic experiments. (C) 2016 Elsevier B.V. All rights reserved.

Lattice-Boltzmann models have been very powerful tools to simulate fluid dynamics, difusion processes, mechanical waves and electrodynamics. Nevertheless, their applicability has been restricted due to the fact that most of them are build on Cartesian coordinates, which hinders them to take advantage of system’s symmetries to reduce the dimensions of the computational domain or to naturally addressing complex systems with curved boundaries, from which the Cochlea, the main auditory organ in mammals, is a paradigmatic example. This work designs and implements a novel lattice-Boltzmann model for the three-dimensional simulation of acoustic waves in general curvilinear coordinates. The method keeps in the computer the standard structure of a Cartesian system with the same velocity vectors in all cells, but it rescales the macroscopic fields and adds forcing terms to reproduce in the continuous limit the wave equation on general coordinates. The resulting second order method perfectly finds the vibrational normal modes of a Cylinder, a Trumpet and a Torus, keeps the isotropic wave propagation in real space and can be applied to any coordinate system. With this method in hand the simulation of the Cochlea was addressed. The Cochlea was parametrized as a tampered coiled tube with a cardioid as cross section and, it was scaled to have the real dimensions of a human Cochlea. As result, we found that the geometry of the Cochlea itself is enough to reproduce the effect of spatial frequency segregation: At high frequencies the net pressure on the Basilar membrane oscillates with larger amplitude close to the windows, and lower frequencies shift the location of maximal amplitude to the apex. The lower the frequency is the closer to the apex that maximum is located. Those results illustrate the high performance, flexibility and reliability of the proposed method, which constitutes a valuable contribution to the development of more powerful lattice-Boltzmann schemes. Los modelos de lattice-Boltzmann han sido una herramienta muy poderosa para simular dinámica de fluidos, difusión, ondas mecánicas y electrodinámica. Sin embargo, su aplicabilidad ha estado restringida debido a que la mayoría de ellos han sido construidos en coordenadas cartesianas, lo que les impide aprovechar las simetrías del sistema para reducir las dimensiones del dominio computacional o considerar de manera natural geometrías complejas con fronteras curvas, de las cuales la Cóclea, el órgano auditivo principal en los mamíferos, es un ejemplo paradigmático. Este trabajo diseña e implementa un modelo de lattice-Boltzmann novedoso para la simulación de ondas acústicas en coordenadas curvilíneas generales. El método mantiene en el computador la estructura estándar de un sistema cartesiano con los mismos vectores velocidad en todas las celdas, pero reescala los campos macroscópicos y añade forzamientos para reproducir en el límite continuo la ecuación de ondas en coordenadas generalizadas. El modelo de segundo orden que resulta calcula perfectamente los modos normales de vibración de un cilindro, una trompeta y un toro, mantiene en el espacio real la isotropía en la propagación de las ondas y puede ser aplicado a cualquier sistema de coordenadas. Con este método en mano se realizó la simulación de la Cóclea. Esta ´ fue parametrizada como un tubo de sección transversal variable en forma de cardioide que se enrolla en espiral, y fue ajustada para tener las dimensiones reales de una Cóclea humana. Como resultado, se encontró que la geometría de la Cóclea es suficiente para reproducir el efecto de separación espacial de las frecuencias. A altas frecuencias, la presión neta sobre la membrana Basilar oscila con mayor amplitud cerca de las ventanas, y frecuencias más bajas desplazan la ubicación de la amplitud máxima hacia el ápex. Entre más baja sea la frecuencia, más cerca del ápex se ubica este máximo de amplitud. Estos resultados ilustran el alto desempeño, flexibilidad y confiabilidad del método propuesto, que constituye una valiosa contribución al desarrollo de esquemas de lattice-Boltzmann más poderosos. Maestría

Flow and solute transport modeling in porous media is an important issue for the monitoring of pollutant contaminations in soils or retention ponds. It is often necessary to take into account macroporal type heterogeneities. For this purpose, several numerical models use double or multi permeability concepts in order to characterize different permeabilities of a porous medium. However classical models seem to underestimate the impact of the preferential path induced by a macropore. This impact has to be extended in the surrounding porous matrix for the good fitting of experimental data. This study proposes to understand hydraulic and mass exchange mechanisms between the macropore and the porous matrix, by using a model porous media made of glass beads, and by using the lattice-Boltzmann method for flow simulations. First results show the influence of flow rate on the apparition of an inflection on breakthrough curves. Results show that molecular diffusion, higher in this case, causes higher mass transfer from the macropore to the porous matrix. Moreover, simulations show that the preferential flow is extended to the porous matrix in a zone having the same dimension than the macropore. These results emphasize that the macropore influence has to be extented in the porous matrix for flow and the diffusive exchange zone depends on the molecular diffusion coefficient of the tracer.; La modélisation des écoulements et du transport dans les milieux poreux est un domaine actif pour le suivi de la contamination dans les sols ou les bassins de rétention notamment. La prise en compte d'hétérogénéité de type macroporale peut souvent être nécessaire, et pour cela de nombreux modèles numériques utilisent des concepts de double ou multi-perméabilités afin de rendre compte des différentes perméabilités d'un même milieu. Cependant, les modèles classiques semblent sous-estimer le rôle de chemin préférentiel d'un macropore, car le champ d'action de celui-ci doit être élargi dans la matrice poreuse pour ajuster au mieux les données expérimentales. Cette étude se propose de comprendre les mécanismes d'échange hydraulique et de masse entre le macropore et la matrice poreuse, en utilisant un milieu poreux modèle de billes de verre, et en utilisant la méthode de Boltzmann sur réseau pour simuler des écoulements. Les premiers résultats montrent que le débit d'injection influe sur l'apparition d'une inflexion des courbes de percée d'un traceur. Les résultats indiquent que la diffusion, plus forte dans ce cas, provoque des transferts de masse importants vers la matrice poreuse. De plus, les simulations montrent que l'écoulement préférentiel est étendu dans la matrice poreuse sur une zone de même dimension que le diamètre moyen des grains. Ces résultats montrent que l'influence du macropore doit être étendue dans la matrice poreuse pour l'écoulement et que la zone d'échanges diffusifs latéraux dépend du coefficient de diffusion du traceur.

The aim of this paper is to study the fluid flow behavior around a wedge shaped body with different wedge angles placed in a channel using Lattice-Boltzmann Method (LBM). The LBM has built up on the D2Q9 model and the single relaxation time method called the Lattice-BGK (Bhatnagar-Gross-Krook) model. The influence of the gap ratio G* = G/H, where H is the distance between two parallel walls, G is the gap between body and wall, on the flow field is illustrated. The gap ratio, G*, depends on the angle of wedge-shaped body (0° < < 180°). Streamlines, vorticity contours and pressure contours are provided to analyze the important characteristics of the flow field for a wide range of non dimensional parameters namely the Reynolds number (Re), Strouhal number (St) and the gap ratio(G*). However, it is seen that the flow can be characterized by three regions: (i) large gap ratio, 0.44 < G* < 0.50, with 0° < < 55°, (ii) intermediate gap ratio, 0.20 G* 0.44, with 55° 120° and (iii) small gap ratio, 0 < G* < 0.20, with 120° < < 180°. The simulation results are compared with experimental data and other numerical models and found to be very reasonable and satisfactory.

We investigate both drainage and spontaneous imbibition processes at the pore scale using a combination of micro- fluidic experiments and lattice-Boltzmann (LB) flow calculations. First, we have fabricated a range of specifically designed etched micro-models to investigate the role of pore shape and throat width on fluid displacement. These designs include junctions with both equal and unequal channel widths in order to achieve a range of capillary entry pressures. All models were etched into Poly(methyl methacrylate) (PMMA) and chemically treated to create a hydrophobic surface. The displacement process is captured via a high-speed video microscope under ambient conditions. The experimental results were then directly compared with LB simulations. For the drainage experiments, we observe that the fluid displacement in the junction follows the Young-Laplace Law. For the case of spontaneous imbibition, however, the models with unequal channel widths display different displacement behaviour. Our experimental observations are confirmed in detail by Lattice Boltzmann Method (LBM) simulations, lending credibility to our observations. Instead of following Young-Laplace filling rules, we observe that the throat in closest proximity fills up first. This has potentially important consequences for calculation of residual saturation of CO 2 at the core scale, which is determined by spontaneous imbibition of brine following CO 2 injection.

Flow and transport modeling through porous media is of primary concern nowadays, especially in order to progress in the understanding of pollutant transfers through soils. Soils present frequently heterogeneities such as macropores (caused by fauna, flora or cracks) and several numerical models use double or multi permeability concepts in order to take into account all flow types that may exist in such porous systems. Nevertheless, classical models seem underestimate the macropore effect on preferential flow and transport by restricting the preferential flow zone only to the volume occupied by the macroporosity. Various experimental studies prior to this thesis have questioned this hypothesis. This study proposes to understand the establishment of preferential flow and transport and in particular the mechanism of flow and solute exchanges between a synthetic macropore and a surrounding porous matrix in saturated condition. For this purpose, water tracing are realized for a model porous media constituted by glass beads, crossed by a synthetic macropore and implemented in laboratory columns. Breakthrough and transport in columns are characterized by monitoring the concentration at the end of the column by magnetic nuclear resonance. A numerical model developed on the basis of lattice-Boltzmann method is used to simul ate flow in macroporous system and identify preferential flow mechanisms at pore scale. Experimental data show that tracer transport is strongly dependent on injection flow rate and the diffusion coefficient in water. At high flow rate, the transport seems to occur exclusively in the macropore, with very little masse exchange with the porous matrix. At lower flow rates, the breakthrough exhibits an inflexion followed by a peak. The MRI images show a significant mass exchange of tracer between the macropore and the surrounding porous matrix. The numerical simulations are used to calculate the flow field in a porous system as a function of flow rate. They show that preferential flow is extended in porous matrix into a zone of same dimension the mean diameter of beads regardless of macropore size or injected flow rate, in the range of simulated flow rates. These experimental and numerical results show that macropore influence on transport should be extended through the surrounding porous matrix into a zone of the same size of grains diameter for flow and into a zone depending on diffusion coefficient as well as mean residence time of the studied tracer for solute transport, La modélisation des écoulements et du transport dans les milieux poreux est un domaine actif pour, notamment, progresser dans la compréhension du transfert des polluants dans les sols. Les sols présentent fréquemment des hétérogénéités comme des macropores (provoqués par la faune, la flore ou des fissures) et un certain nombre de modèles numériques utilisent les concepts de double ou de multi-perméabilité pour tenir compte de tous les types d’écoulements susceptibles de coexister dans de tels systèmes. Cependant, les modèles classiques semblent sous-estimer l’effet de la macroporosité sur l’écoulement et le transfert préférentiels et restreindre la zone d’écoulement préférentiel au seul volume occupé par la macroporosité. Diverses études expérimentales antérieures à cette thèse ont questionné cette hypothèse. Cette étude se propose de comprendre l’établissement de l’écoulement et du transport préférentiel et en particulier les mécanismes d’échange d’eau et de masse entre un macropore et une matrice poreuse environnante en condition saturée. Pour cela, des traçages de l’eau sont réalisés pour un milieu poreux modèle constitué de billes de verre, traversé par un macropore synthétique et mis en place en colonnes de laboratoire. Elution et transfert dans les colonnes sont caractérisés par suivi de la concentration en sortie et par imagerie par résonance magnétique. Un modèle numérique développé sur la base de la méthode de Boltzmann sur réseau est utilisé pour simuler numériquement des écoulements dans un système macroporé et identifier les mécanismes d’écoulements préférentiels à l’échelle de pores. Les données expérimentales montrent que le transfert du traceur est fortement dépendant du débit d’injection ainsi que du coefficient de diffusion dans l’eau. À fort débit, le transfert semble s’effectuer exclusivement dans le macropore, avec très peu d’échange avec la matrice. Pour des débits plus faibles, la percée présente une inflexion suivie d’un pic. Les images IRM montrent alors un échange significatif de traceur entre le macropore et la matrice poreuse environnante. Les simulations numériques sont utilisées pour calculer le champ de vitesse de l’écoulement dans le système en fonction du débit. Les modélisations numériques montrent que l’écoulement préférentiel est étendu dans la matrice poreuse sur une zone de même dimension que le diamètre moyen des grains indépendamment de la taille du macropore et du débit, dans la gamme de débits simulés. Ces résultats expérimentaux et numériques montrent que l’influence du macropore sur les transferts doit être étendue dans la matrice poreuse sur une zone de la taille des grains pour l’écoulement et sur une zone dépendant du coefficient de diffusion du traceur ainsi que du temps de séjour moyen de celui-ci pour le transfert des solutés

Active fluids comprise a variety of systems composed of elements immersed in a fluid environment which can convert some form of energy into directed motion; as such they are intrinsically out-of-equilibrium in the absence of any external forcing. A fundamental problem in the physics of active matter concerns the understanding of how the characteristics of the autonomous propulsion and agent-agent interactions determine the collective dynamics of the system. We study numerically suspensions of self-propelled diffusiophoretic colloids, in (quasi)-2d configurations, accounting for both dynamically resolved solute-mediated phoretic interactions and solvent-mediated hydrodynamic interactions. Our results show that the system displays different scenarios at changing the colloid-solute affinity and it develops a cluster phase in the chemoattractive case. We study the statistics of cluster sizes and cluster morphologies for different magnitudes of colloidal activity. Finally, we provide evidences that hydrodynamics plays a relevant role in the aggregation kinetics and cluster morphology, significantly hindering the cluster growth.

La determinació de la geometria i posició d’un stent és una de les tasques més sensibles quan es planifica una intervenció per estenosi de l’artèria pulmonar. L’objectiu d’aquest projecte és de determinar automàticament la posició i geometria del stent a implantar a partir d’imatges CT i utilitzant paràmetres hemodinàmics calculats a partir de la dinàmica de fluids digital (Computer Fluid Dynamics). La utilització de la mecànica de fluids com a mètode d’optimització s’explica pel fet que una estenosi introdueix un gradient de pressió al llarg de l’artèria on es troba la lesió. Així doncs, la implantació d’un stent té com a objectiu el restabliment del flux sanguini, és a dir, la reducció de la diferència de pressió entre els dos extrems de l’estenosi. La primera etapa del mètode desenvolupat és la segmentació i la reconstrucció en 3D de l’artèria pulmonar a partir de la tomografia del pacient (CT). Seguidament, s’extreu la línia central de l’artèria i s’identifica la posició de l’estenosi. La següent etapa consisteix en simular un conjunt de solucions no òptimes a partir de diferents diàmetres d’expansió del stent. Finalment, es calcula l’hemodinàmica de cada una d’aquestes solucions i se selecciona com a òptima aquella que ens dona una diferència de pressió mínima. Aquest mètode ha estat testejat amb dos casos diferents d’estenosi. Paral·lelament, s’ha mesurat la geometria del stent implantat en l’operació real i se n’ha calculat el gradient de pressió associat. Els resultats d’aquest estudi mostren una millora del 60% i el 23% en el gradient en relació a la intervenció real.

La presente nota descrive sinteticamente lo stato d'avanzamento nell'applicazione di modelli lattice-Boltzmann per la simulazione di processi di flusso e trasporto alla scala del poro. Tali simulazioni permettono di superare molti limiti del tradizionale approccio macroscopico. La combinazione della generazione sintetica di mezzi porosi, della simulazione del flusso e del trasporto alla scala del poro riassunto in questa nota offre un ampio spettro di applicazioni nel campo delle bonifiche in matrici ambientali.

This paper describes a massively parallel code for a state-of-the art thermal lattice- Boltzmann method. Our code has been carefully optimized for performance on one GPU and to have a good scaling behavior extending to a large number of GPUs. Versions of this code have been already used for large-scale studies of convective turbulence. GPUs are becoming increasingly popular in HPC applications, as they are able to deliver higher performance than traditional processors. Writing efficient programs for large clusters is not an easy task as codes must adapt to increasingly parallel architectures, and the overheads of node-to-node communications must be properly handled. We describe the structure of our code, discussing several key design choices that were guided by theoretical models of performance and experimental benchmarks. We present an extensive set of performance measurements and identify the corresponding main bot- tlenecks; finally we compare the results of our GPU code with those measured on other currently available high performance processors. Our results are a production-grade code able to deliver a sustained performance of several tens of Tflops as well as a design and op- timization methodology that can be used for the development of other high performance applications for computational physics.

Modelling the vascular transport and adhesion of man-made particles is crucial for optimizing their efficacy in the detection and treatment of diseases. Here, a Lattice Boltzmann and Immersed Boundary methods are combined together for predicting the near wall dynamics of particles with different shapes in a laminar flow. For the lattice Boltzmann modelling, a Gauss-Hermite projection is used to derive the lattice equation; wall boundary conditions are imposed through the Zou-He framework; and a moving least squares algorithm accurately reconstructs the forcing term accounting for the immersed boundary. First, the computational code is validated against two well-known test cases: the sedimentation of circular and elliptical cylinders in a quiescent fluid. A very good agreement is observed between the present results and those available in the literature. Then, the transport of circular, elliptical, rectangular, square and triangular particles is analyzed in a Couette flow, at Re=20. All particles drifted laterally across the stream lines reaching an equilibrium position, independently of the initial conditions. For this large Reynolds number, the particle shape has no significant effect on the final equilibrium position but it does affect the absolute value and periodicity of the angular velocity. Specifically, elongated particles show longer oscillation periods and, most interestingly, larger variations in angular velocity. The longest particles exhibit a zero angular velocity for almost the whole rotational period. Collectively, this data demonstrates that the proposed approach can be efficiently used for predicting complex particle dynamics in biologically relevant flows. This computational strategy could have significant impact in the field of computational nanomedicine for optimizing the specific delivery of therapeutic and imaging agents.

Lithium-air batteries have the potential to become the future energy source for electric vehicles. Typically, the battery consists of a lithium metal negative electrode, a porous separator soaked with liquid electrolyte, and a porous air electrode where oxygen from the surrounding atmosphere is reduced during battery discharge. This configuration yields the highest theoretical capacity of all Li batteries [1]. In our approach we focus on systems employing aqueous electrolytes [2]. O2 is fed to the battery via so-called gas diffusion electrodes (GDEs) in which gas and liquid phase coexist. The use of hydrophobic binders ensures a fast transport of O2 in the gas phase. The distribution of the liquid electrolyte in the porous structure has a strong influence on effective transport parameters and active surface areas which finally determine the performance of the battery. In our contribution we present results of 2D and 3D lattice-Boltzmann simulations [3], [4] which are conducted on tomographic reconstructions of Ag-GDEs. In our simulations we explicitly take into account the heterogeneous wetting properties of hydrophilic Ag substrate and hydrophobic polymeric binder. The 3D simulations are used to extract effective transport parameters and specific surface areas at various saturations of the GDE with liquid electrolyte. The calculated pc-sw curves of the 3D simulations are compared to a series of independent 2D simulations. We show that the computationally less demanding 2D simulations can be used as an effective tool for the screening of wetting properties of different new electrode geometries. Finally, we present a short outlook on how the results of the LBM simulations can be used to predict the electrochemical performance of GDEs and full virtual battery cells, demonstrating the capabilities of our multi-scale approach for the development of novel battery materials.

Obtaining a good load balance is a significant challenge in scaling up lattice-Boltzmann simulations of realistic sparse problems to the exascale. Here we analyze the effect of weighted decomposition on the performance of the HemeLB lattice-Boltzmann simulation environment, when applied to sparse domains. Prior to domain decomposition, we assign wall and in/outlet sites with increased weights which reflect their increased computational cost. We combine our weighted decomposition with a second optimization, which is to sort the lattice sites according to a space filling curve. We tested these strategies on a sparse bifurcation and very sparse aneurysm geometry, and find that using weights reduces calculation load imbalance by up to 85%, although the overall communication overhead is higher than some of our runs., 11 pages, 8 figures, 1 table, accepted for the EASC2014 conference

We explore possible applications of the Lattice-Boltzmann Method for the simulation of geophysical flows. This fluid solver, while successful in other fields, is still rarely used for geotechnical applications. We show how the standard method can be modified to represent free-surface realization of mudflows, debris flows, and in general any plastic flow, through the implementation of a Bingham constitutive model. The chapter is completed by an example of a full-scale simulation of a plastic fluid flowing down an inclined channel and depositing on a flat surface. An application is given, where the fluid interacts with a vertical obstacle in the channel.

For applications where droplet breakup and surfactant adsorption are strongly coupled, a diffuse interface model is developed. The model is based on a free energy functional, partly adapted from the sharp interface model of [Diamant and Andelman 34(8):575–580, (1996)]. The model is implemented as a 2D Lattice Boltzmann scheme, similar to existing microemulsion models, which are coupled to hydrodynamics. Contrary to these microemulsion models, we can describe realistic adsorption isotherms, such as the Langmuir isotherm. From the free energy, functional analytical expressions of equilibrium properties are derived, which compare reasonably with numerical results. Interfacial tension lowering scales with the logarithm of the area fraction of the interface unloaded with a surfactant: $$\Delta \sigma \sim \ln(1-\psi_0)$$ . Furthermore, we show that adsorption kinetics are close to the classical relations of Ward and Tordai. Prelimary simulations of droplets in shear flow show promising results, with surfactants migrating to interfacial regions with highest curvature. We conclude that our diffuse interface model is very promising for apprehending the above-mentioned applications as membrane emulsification.

We have simulated intrusion of a non-wetting liquid into pores of varying shape and size. Simulations were based on the lattice-Boltzmann method and the Shan–Chen multiphase model. The liquid–solid contact angle for pores with circular cross-section was found to be equal to that for pores with square cross-section, and constant even for small pore sizes if the discretised shape of the circular cross-section was taken into account. For comparison, contact angle was also determined for a liquid column descending in a capillary tube, and the results were found to be consistent. Application of the method to mercury intrusion porosimetry is discussed.

International audience; In order to account for the specific filtration properties of a suspension of yeast cells, a mechanical model has been developed that describes the permeability of a compressible cake of cells as a function of the compressive pressure that it experiences. This model is found to agree well with data obtained during static filtration of a bed of cells, but agreement is not as good for data obtained during cake formation.

There is currently limited understanding of the role played by haemodynamic forces on the processes governing vascular development. One of many obstacles to be overcome is being able to measure those forces, at the required resolution level, on vessels only a few micrometres thick. In this paper, we present an in silico method for the computation of the haemodynamic forces experienced by murine retinal vasculature (a widely used vascular development animal model) beyond what is measurable experimentally. Our results show that it is possible to reconstruct high-resolution three-dimensional geometrical models directly from samples of retinal vasculature and that the lattice-Boltzmann algorithm can be used to obtain accurate estimates of the haemodynamics in these domains. We generate flow models from samples obtained at postnatal days (P) 5 and 6. Our simulations show important differences between the flow patterns recovered in both cases, including observations of regression occurring in areas where wall shear stress (WSS) gradients exist. We propose two possible mechanisms to account for the observed increase in velocity and WSS between P5 and P6: (i) the measured reduction in typical vessel diameter between both time points and (ii) the reduction in network density triggered by the pruning process. The methodology developed herein is applicable to other biomedical domains where microvasculature can be imaged but experimental flow measurements are unavailable or difficult to obtain.

To determine the characteristics and properties of cellular solids for an application, and to allow a systematic practical use by means of correlations and modelling approaches, we perform experimental investigations and develop numerical methods. In view of coupled multi-physics simulations, we employ the phase-field method. Finally, the applicability is demonstrated exemplarily for open-cell metal foams, providing qualitative and quantitative comparison with experimental data.

A lattice-Boltzmann scheme for natural convection in porous media is developed and applied to the heat transfer problem of a 1000 kg potato packaging. The scheme has features new to the field of LB schemes. It is mapped on a orthorhombic lattice instead of the traditional cubic lattice. Furthermore the boundary conditions are formulated with a single paradigm based upon the particle fluxes. Our scheme is well able to reproduce (1) the analytical solutions of simple model problems and (2) the results from cooling down experiments with potato packagings.

[EN] The final master thesis begins putting in place the LBM method among the most used CFD techniques. In following sections it is deeply explained how LBM internally works. After understanding, the main advantages and disadvantages of LBM are summarized. One of the LBM strengths is the capacity to solve very complex geometries, such as porous media. Therefore, after creating some porous domains, a study is done to examine the model performance, as well as the fluid mechanics inside the porous media., [ES] El documento del trabajo de fin de máster comienza posicionando el método Lattice-Boltzmann entre todos los métodos más utilizados en Dinámica de Fluidos Computacional. A partir de entonces se hace una explicación en profundidad acerca de como funciona para más tarde hacer una breve descripción de sus debilidades y fortalezas. Una de esas fortalezas es la solución de geometrías de gran complejidad, como los medios porosos. Finalmente, haciendo uso de algunas de estas geometrías, se pasa a estudiar las cualidades del modelo así como la mecánica de fluidos dentro del medio poroso.

We study the static and dynamical behavior of the contact line between two fluids and a solid plate by means of the Lattice Boltzmann method (LBM). The different fluid phases and their contact with the plate are simulated by means of standard Shan-Chen models. We investigate different regimes and compare the multicomponent vs. the multiphase LBM models near the contact line. A static interface profile is attained with the multiphase model just by balancing the hydrostatic pressure (due to gravity) with a pressure jump at the bottom. In order to study the same problem with the multicomponent case we propose and validate an idea of a body force acting only on one of the two fluid components. In order to reproduce results matching an infinite bath, boundary conditions at the bath side play a key role. We quantitatively compare open and wall boundary conditions and study their influence on the shape of the meniscus against static and lubrication theory solution.

Fine coal filtration and dewatering are of great importance to the coal industry due to the significant impact in the quality, shipping, and handling of clean coal. High moisture content in the clean coal product reduces its heating value, increases costs, and reduces the coke yield in the case of metallurgical coal. In this regard, it is of significant importance to improve our fundamental understanding of water removal from the pore network structures present in filtration cakes. This thesis research presents the results obtained from the analysis regarding fluid flow through packed particle beds such as those occurring during filtration in an effort to expand the studies of particle characterization and its influence on coal dewatering. The study shows the importance of high resolution x-ray microtomography (HRXMT) as an important analytical tool for the three-dimensional study of particle beds. The multiphase flow and dewatering that occurs during fine coal filtration is described, and important factors that influence the efficiency of filtration, including the particle size distribution, pressure drop, shape, and wetting characteristics of the coal particles are considered. The experiments are designed to simulate the process of coal filtration using the Lattice-Boltzmann methodology, and identify the conditions that lead to the improved water removal and moisture reduction. The thesis and research reported herein demonstrate how HRXMT and the Lattice-Boltzmann Method (LBM) can help in the short-term prediction and understanding of water removal from the pore network structures present in coal filtration cakes. Based on the analysis of HRXMT images, it is shown that the pore network structure has a significant influence on the retention of water in the filter cake. Narrow capillaries were found in the filter cakes with hydrophilic particles, while wider capillaries were mostly found in the filter cakes with hydrophobic particles. In addition, tests with different pressure drops were performed. Although the pore network structure analysis showed that the capillaries were narrower at higher initial pressures, the increase in pressure drop decreased the amount of water retained in the filter cake. The pressure drop increase helped overcome the capillary forces that retain the water in the filter cakes.

We investigate the performance of the HemeLB lattice-Boltzmann simulator for cerebrovascular blood flow, aimed at providing timely and clinically relevant assistance to neurosurgeons. HemeLB is optimised for sparse geometries, supports interactive use, and scales well to 32,768 cores for problems with ~81 million lattice sites. We obtain a maximum performance of 29.5 billion site updates per second, with only an 11% slowdown for highly sparse problems (5% fluid fraction). We present steering and visualisation performance measurements and provide a model which allows users to predict the performance, thereby determining how to run simulations with maximum accuracy within time constraints., Accepted by the Journal of Computational Science. 33 pages, 16 figures, 7 tables

[Notes_IRSTEA]WR011385 [Departement_IRSTEA]Eaux [TR1_IRSTEA]ARCEAU; International audience; The recent advances in 3-D imaging of porous structures have generated a tremendous interest in the simulation of complex single and two-phase flows. Lattice-Boltzmann (LB) schemes present a powerful tool to solve the flow field directly from the binarized 3-D images. However, as viscosity often plays an important role, the LB scheme should correctly treat viscosity effects. This is the case using a LB scheme with two relaxation times (TRT) unlike the broadly used, the single-relaxation rate, BGK, where the velocity of the modeled fluid does not vary as the inverse of the viscosity applying the bounce-back (no-slip) boundary rule. The aim of this work is to apply the LB-TRT approach to different types of porous media (straight channels, 2-D model porous media, sandstone) to solve for the flow field and to evaluate the approach in terms of parameter dependence, error and convergence time on the basis of permeability. We show that the variation of permeability with the free relaxation parameter L of the TRT scheme depends on the heterogeneity of the sample and on the numerical resolution. The convergence time depends on the applied viscosity and the parameter standing for the speed of sound, thus the computation time can be reduced by choosing appropriate values of those parameters. Two approaches to calculate permeability (Darcy’s law and viscous energy dissipation) are proposed and investigated. We recommend to use Darcy’s law, as dependence on L is less important. Periodic (in the presence of a driving body force) and pressure boundary conditions are evaluated in terms of the results.

In this thesis the basic properties of the lattice- Boltzmann method (LBM) are introduced. Also, a particle model used in combination with LBM is presented, and the combined model is then applied to ink propagation in samples of paper. Simulation geometries were acquired with two different methods, namely confocal microscopy and X-ray tomography. The effect of simulation parameters and paper properties on ink propagation was considered. Promising results were acquired with both image acquiring techniques, but for a better consistency between simulations and experiments the methods should be combined so as to get images that include the full thickness of the sample together with the ink distribution. Adjustment of simulation parameters in confocal microscopy geometries showed that diffusion coefficient was the main parameter that explained the form of the ink distribution curves. This suggests an intuitively expected result: ink tends to move with the major flow channels unless diffusion is strong enough to separate enough of ink particles from the flow. Simulations done in samples of varying sizing gave incoherent results. One simulation series was in agreement with experimental results, while another was not. Possible explanations to this dilemma includes effects of paper heterogeneity on the results, differencies in paper properties in macroscopically different parts of the paper, and chemical effects not taken into consideration in the model. We can conclude however that simulations with the methods applied here qualitatively capture the main features of the settling in paper. In part this conclusion is based in the fact that the experimentally observed division into two separate components in the distribution of attached ink pigments was also realised in the simulations.

Laboratory tests or measures on the field are widely carried out in order to evaluate the drainage capability of pavement. Such an approach is in general not efficient and it shows poor significance. In this paper The authors propose a novel method, very effective for mix design, based on simulation of unsteady-flow of water through open-graded mixture. In particular the authors have modelled internal microstructures of an open-graded asphalt sample, positioning any aggregate particle and creating the bitumen film around each of them. The unsteady-flow of water inside the sample is simulated to evaluate the expected permeability for different specimens and compaction.

We develop and test numerically a lattice-Boltzmann (LB) model for nonideal fluids that incorporates thermal fluctuations. The fluid model is a momentum-conserving thermostat, for which we demonstrate how the temperature can be made equal at all length scales present in the system by having noise both locally in the stress tensor and by shaking the whole system in accord with the local temperature. The validity of the model is extended to a broad range of sound velocities. Our model features a consistent coupling scheme between the fluid and solid molecular dynamics objects, allowing us to use the LB fluid as a heat bath for solutes evolving in time without external Langevin noise added to the solute. This property expands the applicability of LB models to dense, strongly correlated systems with thermal fluctuations and potentially nonideal equations of state. Tests on the fluid itself and on static and dynamic properties of a coarse-grained polymer chain under strong hydrodynamic interactions are used to benchmark the model. The model produces results for single-chain diffusion that are in quantitative agreement with theory.

Up to tenth-order finite difference schemes are proposed in this paper to solve one-dimensional advection-diffusion equation. The schemes based on high-order differences are presented using Taylor series expansion. To obtain the solutions, up to tenth-order finite difference schemes in space and a fourth-order Runge-Kutta scheme in time have been combined. The methods are implemented to solve two problems having exact solutions. Numerical experiments have been conducted to demonstrate the efficiency and high-order accuracy of the current methods. The techniques are seen to be very accurate in solving the advection-diffusion equation for Pe ? 5. The produced results are also seen to be more accurate than some available results given in the literature. © Association for Scientific Research.

Made available in DSpace on 2015-03-04T18:51:07Z (GMT). No. of bitstreams: 1 DissertationDRGolbert_versao_final.pdf: 9706339 bytes, checksum: 45a86747e8469ad89e82dc19d8322037 (MD5) Previous issue date: 2009-03-25 Fundação Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro The goal of this work is to study de modeling of incompressible fluid flows through the Lattice-Boltzmann method (LBM). In this class of methods the equations based on mesoscopic kinetics allow us to model the macro-continuum behavior of the fluid dynamics. Therefore, a theoretical study of the LBM is performed including the analyses of different equilibrium distributions, lattice models, its relationships with the Boltzmann equation as well as its asymptotic approximation to the Navier-Stokes equations. On the other hand, aspects related to the imposition of boundary conditions are also studied, identifying adequate procedures to the problems here presented. Posteriorly, a detailed study of numerical nature about the performance of the LBM in the computational simulation of fluid flows is developed, involving stationary and transient problems, for cases in 2D and 3D. Once we have insight on the main characteristics of the model, techniques for the tuning of LBM's parameters are introduced with the purpose of attaining consistent and reliable results, according to the physical conditions of the problems under consideration. These techniques are employed with emphasis in 3D time dependent problems, whose characteristics are similar to those found in the blood flow modeling in arteries. O objetivo deste trabalho é estudar a modelagem do escoamento de fluidos incompressíveis mediante o método de Lattice-Boltzmann (LBM). Nesta classe de métodos as equações baseadas na cinética mesoscópica nos permitem modelar o comportamento macro-contínuo da dinâmica de fluidos. Desta forma, realiza-se um estudo teórico do LBM incluindo a análise de diferentes distribuições de equilíbrio, modelos de lattice, suas relações com a equação de Boltzmann assim como sua aproximação assintótica às equações de Navier-Stokes. Por outro lado, estudam-se os aspectos relacionados à imposição de condições de contorno identificando procedimentos adequados para os problemas aqui tratados. Posteriormente, realiza-se um estudo detalhado de caráter numérico sobre o desempenho do LBM na simulação computacional de escoamentos de fluidos, envolvendo problemas estacionários e transientes, para casos em 2D e 3D. A partir do conhecimento das características do modelo, desenvolvem-se técnicas para efetuar a calibração dos parâmetros do LBM visando à obtenção de resultados coerentes e confiáveis de acordo às condições físicas do problema. Estas técnicas são empregadas com ênfase em problemas 3D dependentes do tempo, e cujas características são similares às encontradas na modelagem do escoamento sanguíneo em artérias.

In den Ingenieurwissenschaften, so auch im Bauwesen, gewinnen Fragestellungen der Strömungsmechanik und deren Lösung mithilfe von Strömungssimulationen zunehmend an Bedeutung. Eine wesentliche Aufgabe besteht darin, Methoden und Ansätze zu deren Anwendung zu erforschen und zu validieren. Im ersten Teil dieser Arbeit wird ein Integrationskonzept vorgestellt, mit dem der Gesamtablauf von Strömungssimulationen effizienter gestaltet werden kann. Es wird dazu ein Rahmenwerk entwickelt, welches das Beziehungsgeflecht zwischen den geometrischen Modellen der Teilaufgaben von Strömungssimulationen darlegt und den Ablauf der Simulationen prinzipiell vorgibt. Als zentrale Datenstruktur wird ein Oktalbaum, ein dreidimensionaler Vertreter der Baumdatenstrukturen, verwendet. Dieser unterstützt einerseits das Abbilden der geometrischen Modelle und ermöglicht andererseits eine automatische Generierung kartesischer Gitter. Die angestrebte Kopplung von CAD und Strömungssimulation wird an zwei ausgewählten Beispielen demonstriert. Der zweite Teil der Arbeit beschäftigt sich mit der Erweiterung der Lattice-Boltzmann Methode für Simulationen auf nicht-uniformen Gittern für zweidimensionale Probleme. Hierfür werden zunächst die Grundlagen der Methode dargestellt und die entsprechenden Erweiterungen für die Nichtuniformität im Detail erläutert. Von essentieller Bedeutung sind dabei die das Gitter repräsentierenden Datenstrukturen. Auch hier finden Baumdatenstrukturen, für zweidimensionale Anwendungen als Quadtrees bezeichnet, optimale Einsatzmöglichkeiten. Die Besonderheiten, die sich in der prototypischen Implementierung ergeben, werden ausführlich dargestellt. Da der Prototyp die Modifikation der Gitterkonfiguration erlaubt, werden erstmalig adaptive Lattice-Bolzmann Simulationen durchgeführt und deren Effizienz gezeigt. Fluid mechanics and its applications using methods of computational fluid dynamics (CFD) gain an increasing interest in the field of engineering technologies. One of the main issues of sciences in this area is the exploration of methods and solutions dealing with CFD. In the first part of this work an integration concept, with the purpose of optimization of the application flow of fluid-flow simulations, is presented. Therefore, a framework, identifying the relationsships between the different geometric models and defining the sequence of the different tasks in flow simulations, is proposed. The central datastructure is based on an octree, a three-dimensional representative of the spacetrees. It is used to support the mapping of the different geometric models onto each other and to ensure the automation of the cartesian grid generation. The aspired coupling of CAD and flow simulation is demonstrated using specific examples. The second part of this work deals with the extension of the Lattice-Boltzmann method for simulations on non-uniform two-dimensional grids. Therefore, a brief introduction of the basics of the method is given and the corresponding enhancements for the non-uniform grids are explained in detail. The datastructure, representing the grid, is of essential interest. The two-dimensional version of the spacetrees, the quadtrees, find their optimal employment for this purpose. The main aspects, concerning the prototypic implementation, are described. As the prototyp allows a modification of the grid configuration during runtime, adaptive Lattice-Boltzmann simulations can be performed for the first time demonstrating their efficiency.