Cette thèse est une étude expérimentale et numérique du comportement au cisaillement de milieu granulaire humide sous l’effet de la quantité de liquide introduite et la contrainte normale appliquée. Les expériences ont été faites sur une cellule de cisaillement annulaire, pour une large gamme de contraintes appliquées allant de presque 0.3 kPa à 12 kPa. Les résultats donnent la variation de la contrainte de cisaillement en régime stationnaire en fonction de la contrainte normale pour une large variation de la quantité de liquide. Le liquide dans le milieu granulaire va de ponts liquides formés au point de contact jusqu’au remplissage totale de l'espace entre les grains. L’effet de liquide sur la résistance au cisaillement et la porosité de milieu granulaire a été analysé. Différents régimes du comportement de milieu granulaire humide ont été identifiés. Afin d’acquérir une compréhension microscopique du comportement au cisaillement de milieu granulaire sec et partiellement humide, la méthode des éléments discrets (DEM) a été utilisée. Des billes de verre de grande taille (2 mm de diamètre) ont été utilisées pour réduire le temps de simulation et faciliter la caractérisation à l’échelle de particule. Une première partie a été consacrée à l’étude de l’effet des propriétés microscopiques de particule (Module de Young et la friction de glissement) sur les propriétés macroscopiques de milieu granulaire sec et humide (le nombre de coordination, la porosité, le ratio de contraintes et la vitesse de particules). Une deuxième partie a été concernée par l’étude du comportement au cisaillement de milieu granulaire humide pour différentes fractions de liquide et différentes contraintes normales appliquées. En particulier, les forces capillaires et le nombre de ponts liquide ont été quantitativement analysés. / We study experimentally and numerically the shear behaviour of wet granular material. We investigate the effect of the liquid content and the applied normal stresses to this behaviour. An annular shear cell was used to carry out the experiments, for a large range of applied normal stress from about 0.3 kPa to 12 kPa. The results give the variation of the shear stress at steady-state as a function of the normal stress for a wide range of liquid fraction. The incorporated liquid goes from forming bridges at the contact point to completely filling the space between grains. The shear resistance and the voidage fraction variations with the liquid fraction were analysed. Depending on the applied normal stress and the liquid fraction, different regimes of the shear resistance were identified. The discrete element method (DEM) was used to gain a microscopic understanding of the shear behaviour of dry and partially wet granular material in the shear cell. Large size glass beads were used to speed up the computational time and to facilitate characterisation at the particle scale. First, the influence of the microscopic properties of the particle (The Young’s modulus and the sliding friction) on the macroscopic properties of dry and wet granular materials (the coordination number, the voidage fraction, the shear ratio and the velocity of particles) was investigated. Secondly, the shear behaviour of the partially wet granular material for different liquid fractions and normal stresses was studied. The capillary forces and the number of liquid bridges were quantitatively analysed.

Mit zunehmenden mineralischen Ressourcenbedarf steigen die Anforde-rungen an Aufbereitungsmaschinen wie den Center-Sizer. Um diesen An-forderungen gerecht zu werden, lag der Schwerpunkt bisher in der Ver-besserung der maschinenseitigen Modellbeschreibung, wobei für die Ma-terialmodelle zumeist stark vereinfacht blieben. Die vorliegende Arbeit behandelt die Entwicklung eines materialseitigen Modells basierend auf der Diskreten-Elemente-Methode, welches durch eine begründete Parametrierung sowie in Co-Simulation mit einem mehr-dimensionalen Maschinenmodell zur Belastungsprognose auf Center-Sizer dient. Sie leistet damit einen Beitrag zur Erweiterung der bestehenden Auslegungsmethode und bietet für weiterführende Forschungstätigkeiten eine substanzielle Grundlage. / As the demand for mineral resources increases, so do the requirements for processing machines such as the Center-Sizer. In order to meet these re-quirements, the focus has so far been on improving the model description on the machine side, whereby the material models mostly have remained simplified. The present dissertation deals with the development of a material model based on the discrete-element-method for the load prediction on center sizer by using determined parameterization method as well as the co-simulation with a multidimensional machine model. It contributes to the enhancement of the current method of structural design and it serves a substantial basis for further research projects.

No description available.

Les avalanches de plaque représentent un risque naturel majeur dont la prévision demeure très difficile. Le manque de lois constitutives fiables à l’échelle du matériau rend difficiles les tentatives de modélisation de ce phénomène. Plus spécifiquement, la réponse mécanique de la neige durant et après la rupture, dans des régimes de chargements rapides , demeure relativement méconnue. La nature particulièrement fragile du matériau au sein de ce régime de déformation rend ardue la réalisation d’expériences et complique l’observation à l’échelle microstructurale.Dans ce travail de thèse, un modèle numérique de neige fondé sur la Méthode des Éléments Discrets a été développé en tant qu’alternative aux expériences. Le modèle nous permet de simuler la réponse de la neige à des chargements mécaniques en tenant compte de la microstructure réelle du matériau grâce à l’intégration d’images acquises par microtomographie à rayons X en entrée du modèle. La neige est considérée comme un matériau granulaire cohesif, et une méthode originale a été développée afin de modéliser la forme de chaque grain. Les grains individuels sont ensuite assemblés pour reconstituer la matrice de la neige grâce à la prise en compte de lois de contact cohésives.Le modèle a été utilisé afin d’explorer la réponse mécanique macroscopique de différent échantillons de neige à un chargement mixte normal-cisaillant. Trois modes de rupture ont été observés dans tous les échantillons de neige testés, en fonction du niveau de contrainte normale appliquée : une rupture en cisaillement localisée pour des niveaux de contrainte normale faibles (mode A), un effondrement normal induit par rupture en cisaillement à des niveaux intermédiaires de contrainte normale (mode B) et un effondrement normal pour des valeurs de contrainte normale élevées (mode C). Ces différents modes de rupture produisent une enveloppe de rupture fermée dans l’espace des contraintes, ce pour les différents types de neige étudiés.Les mécanismes internes conduisant à l’effondrement normal des échantillons ont été étudiés plus en détail à l’échelle microscopique. Il a été montré que ce mode de rupture était associé à un mécanisme de flambement des chaînes de force. En outre, la stabilité de ces chaînes de force semble être contrôlée par les contacts entre les éléments des chaînes et les grains environnants. La rupture de ces contacts, observée dans les modes B et C, autorise le développement du flambement des chaînes de force et aboutit à l’effondrement volumique. / Dry slab snow avalanches represent a major natural hazard that is extremely difficult to manage. Attempts to model this phenomenon are hindered by the lack of a constitutive law that would describe the mechanical behaviour of snow on a material scale. In particular, relatively little is known on the failure and post-failure response of snow at high loading-rates. The highly fragile nature of the material in this deformation regimerenders experimental investigation difficult and complicates observation at the microstructural level.As an alternative to experiments, a Discrete Element Method-based numerical model of snow is developed in this thesis. The model enables us to simulate the response of snow to mechanical loading, while accounting for actual snow microstructure by using X-ray attenuation images of snow microstructure as input. Snow is considered as a cohesive granular material and an original methodology is developed in order to model the shape of each grain. Individual grains are bound into the snow matrix by modelling cohesion between neighbouring grains.The model is then used to explore the macroscopic mechanical response of different snow samples to mixed-mode loading. Three typical modes of failure are observed in all tested snow samples, depending on the level of applied normal stress: a localized shear failure at low normal stress (mode A), a shear failure-induced volumetric collapse at intermediate levels of normal stress (mode B), and a normal failure and collapse for high values of normal stress (mode C). The observed failure modes result in closed failure envelopes and no qualitative difference is observed between the mechanical responses of different snow types.The internal mechanisms that lead to volumetric collapse are further examined on the microscale. Force chain buckling is identified as a trigger of this material instability. Additionally, force chain stability appears to be controlled by the contacts between the force chain members and the surrounding grains. The failure in these contacts, which is evidenced in modes B and C, allows force chain buckling to develop and results in subsequent volumetric collapse.

Indiana University-Purdue University Indianapolis (IUPUI) / In this thesis, a high- fidelity multibody dynamics model of a backhoe for simulating the digging operation is developed using the DIS (Dynamic Interactions Simulator)multibody dynamics software. Sand is used as a sample digging material to illustrate the model. The backhoe components (such as frame, manipulators links,track segments, wheels and sprockets) are modeled as rigid bodies. The geometry of the major moving components of the backhoe is created using the Pro/E solid modeling software. The components of the backhoe are imported to DIS and connected using joints (revolute, cylindrical and prismatic joints). Rotary and linear actuators along with PD (Proportional-Derivative) controllers are used to move and steer the backhoe and to move the backhoes manipulator in the desired trajectory. Sand is modeled using cubic shaped particles that can come into contact with each other, the backhoes bucket and ground. A cubical sand particle contact surface is modeled using eight spheres that are rigidly glued to each other to form a cubical shaped particle, The backhoe and ground surfaces are modeled as polygonal surfaces. A penalty technique is used to impose both joint and normal contact constraints (including track-wheels, track-terrain, bucket-particles and particles-particles contact). An asperity-based friction model is used to model joint and contact friction. A Cartesian Eulerian grid contact search algorithm is used to allow fast contact detection between particles. A recursive bounding box contact search algorithm is used to allow fast contact detection for polygonal contact surfaces and is used to detect contact between: track and ground; track and wheels; bucket and particles; and ground and particles. The governing equations of motion are solved along with joint/constraint equations using a time-accurate explicit solution procedure. The sand model is validated using a conical hopper sand flow experiment in which the sand flow rate during discharge and the angle of repose of the resulting sand pile are experimentally measured. The results of the conical hopper simulation are compared with previously published experimental results. Parameter studies are performed using the sand model to study the e ffects of the particle size and the orifi ces diameter of the hopper on the sand pile angle of repose and sand flow rate. The sand model is integrated with the backhoe model to simulate a typical digging operation. The model is used to predict the manipulators actuator forces needed to dig through a pile of sand. Integrating the sand model and backhoe model can help improving the performance of construction equipment by predicting, for various vehicle design alternatives: the actuator and joint forces, and the vehicle stability during digging.

This thesis reports recent numerical investigation of the rheological behavior of fresh concrete using the Distinct Element Method (DEM). Some relevant questions of the concrete rheology e.g. the influence of the concrete composition on the rheological behavior of the fresh concrete, the experimental determination of the Bingham rheological constants as well as the use of these constants in the numerical simulation were discussed thoroughly. An important topic of the performed investigation was the development of the numerical model for fresh concrete which enables simple, fast and stable predictive simulation of different technological operations with fresh concrete. Firstly, in a literature survey, the state-of-the-art of the numerical simulation of fresh concrete was presented and critically discussed in order to show advantages and disadvantages of other methods and modeling approaches. Open (unsolved) questions were highlighted and the basis for their investigation is created within this thesis. Fundamental concepts of the rheology were then presented and the basic rheological models of viscoelastic materials were considered; the rheological behaviors of different types of concretes were presented and its influencing factors were discussed. Additionally main methods for scientific investigation and testing of the fresh concrete were shown. The test methods were critically discussed in order to select the test, which has been used as a reference experimental test for the numerical simulations. Chosen reference experimental test was the slump flow test. The slump flow test was thoroughly analyzed and an analytical solution was developed which helps to interpret the results of measurements and provides a link between rheological constants and measured quantities. In a further step an extensive experimental program was carried out in order to investigate the rheological behavior of fresh concrete and get the input data for numerical simulation. Firstly, the experiments on macrolevel were performed. Here the rheological behavior of the fresh concrete flow in different tests was investigated (slump and slump flow tests, L-Box). Further, the experiments on mesolevel with polymer on Carbopol basis and mortar were developed and performed in order to investigate the interaction between distinct particles suspended in a fluid matrix. The necessary material parameters, especially those representative of the fluid suspension micromechanical behavior, i.e. the force-displacement relationship, yield force and bond strength, were determined by these experiments. The slump flow test was used as the basic test to calibrate the model for fresh concrete (key data: slump value, slump flow diameter (for concretes with a soft consistency) and the time of spreading). Thus, the decisive phenomena of the fresh concrete flow were highlighted, control points for a contact model were selected and the initial input data for the development of the contact model was obtained. Next, the user-defined contact model was developed and implemented into the Particle Flow Code ITASCA. The contact model was completely described and its limitations discussed. Then, the set of numerical tools was developed, which enable simplified and stable numerical simulation of the fresh concrete with particular behavior, i.e. automatic generation of the concrete with given particle grading, amount of fibers and air, automatic recalculation of the micromechanical parameters of the contact model from given initial yield stress and plastic viscosity. The model was calibrated by slump flow test simulations and validated by corresponding analytical approach. Further, the role of different model parameters was investigated by simulating the slump flow test. Furthermore, for verification of the model several additional experiments were simulated, i.e. L-Box and LCPC-box test. The results of modeling were compared with experimental results and discussed in detail. All numerical simulations provide qualitatively as well as quantitatively correct results and hence adequately represent the phenomena observed in real experiments. The thesis closes with general conclusions and outlook of the work. In the future, the developed contact model and tools of the “Virtual concrete laboratory” could be modified in order to extend the potential of the laboratory to cover such properties as thixotropic behavior of fresh concrete or simulating hardening of the concrete and behavior of the hardened concrete.

[pt] O Método dos Elementos Discretos (Discrete Element Method - DEM) é uma técnica numérico computacional capaz de simular o comportamento macroscópico de um material via solução das equações do movimento de seus constituintes. Para uma correta predição deste comportamento são informados, como dados de entrada, as características mecânicas dos elementos: os chamados microparâmetros. Contudo, não existe uma receita que determine estes microparâmetros baseados somente nas respostas macroscópicas do material simulado, necessitando de um passo adicional conhecido como Calibração. Tentativa e erro, um método ineficiente por conta de seu fator de escala desfavorável, é o mais comumente utilizado nesta etapa. Este trabalho propõe uma nova abordagem utilizando-se do método de otimização global Generalized Simulated Annealing, minimizando-se a área quadrática normalizada entre as curvas experimentais e calculadas de tensão-deformação axial e deformações volumétrica-axial simultaneamente. Foram efetuadas comparações via ensaio triaxial para dados sintéticos e reais cujos resultados demonstram o aproveitamento e aplicabilidade da técnica proposta. / [en] The Discrete Element Method (DEM) is a numerical computational technique that simulates the macroscopic material behaviour by solving the equations of motion of its constituents. For a correct prediction of this behaviour, are set as input data the mechanical characteristics of the elements, the so-called microparameters. However, there is no recipe for determining these microparameters based solely on the macroscopic responses of the simulated material. It is required an additional step known as Calibration. The method widely used in this calibration is trial and error, although is an inefficient method due its unfavorable scale factor. This work proposes a new approach using the Generalized Simulated Annealing global optimization method, minimizing the normalized quadratic area between the experimental and calculated curves of the axial stress-strain and volumetric-axial deformations curves simultaneously. Comparison is done using triaxial tests for both synthetic and real data whose results demonstrate the usefulness and applicability of the proposed approach.

Since their introduction in the late 20th century, lithium-ion batteries have become the leading battery technology for portable devices and electric vehicles due to their high energy density and rechargeability. However, the increasing demand for a longer battery life span is hindered by the fading of the battery’s charge capacity over prolonged use. This reduction in charge capacity stems from electrochemical and mechanical degradation of the battery cells. The main research focus in the literature has been on the chemical degradation of battery cells. However, the mechanical degradation also substantially contributes to the battery’s capacity degradation. Therefore, it is crucial to understand the mechanical properties of the battery cells to be able to mitigate mechanical degradation. The battery’s mechanical degradation stems from the electrode layers’ constituents. This thesis aims to model the positive electrode’s mechanical properties by recreating its granular microstructure using the discrete element method. In Papers 1 and 2, a discrete element method modelling framework is developed, which can reconstruct a positive electrode layer of a lithium-ion battery, simulate manufacturing processing steps, and determine the mechanical properties of the electrode layer. The framework uses two contact models, representing the positive electrode material in the form of particles and a binder agent, which gives the electrode layer its structural integrity. The framework is used to link the mechanical behaviour of the electrode particles and the binder agent to the mechanical behaviour of the entire electrode layer. The framework is able to capture the layer’s pressure sensitivity and relaxation behaviour, properties which have been displayed in the literature through experimental testing. / Sedan de introducerades i slutet av 1900-talet har litiumjonbatterier blivit den ledande batteriteknologin för portabla enheter samt elfordon på grund av deras höga energidensitet och återladdningförmåga. Den ökade efterfrågan på utökade batterilivslängder är dock hämmad av reduceringen av uppladdningskapacitet över längre användningstider. Denna reducering av laddningskapacitet kommer från elektrokemisk och mekanisk degradering av battericellerna. Det största forskningsintresset i litteraturen har varit på den kemiska degraderingen av battericellerna. Dock ger den mekaniska degraderingen ett betydande bidrag till batteriets kapacitetsdegradering. Därför är det viktigt att förstå battericellens mekaniska egenskaper för att kunna förhindra mekaniskdegradering. Batteriets mekaniska degradering beror på elektrodlagrets beståndsdelar. Denna avhandlings målsättning är att modellera den positiva elektrodens mekaniska egenskaper genom att återskapa dess granulära mikrostruktur med hjälp av diskret elementmetodik. I Artikel 1 och 2 utvecklades ett ramverk för modellering med användning av diskreta elementmetoden, vilket kan återskapa det aktiva lagret för en positiv elektrod, simulera tillverkningsprocesser, samt fastställa elektrodlagrets mekaniska egenskaper. Ramverket använder två kontaktmodeller som representerar det positiva elektrodmaterialet i form av partiklar samt ett bindemedel, som ger elektrodlagret dess strukturella integritet. Ramverket används för att undersöka hur de mekaniska egenskaperna för det hela elektrodlagret beror på egenskaperna för de aktiva partiklarna samt bindemedlet. Ramverket kan fånga lagrets tryckkänslighet samt dess relaxering, egenskaper som har påvisats i litteraturen genom experimentell provning. / <p>Qc240322</p>

A significant portion of the reinforced concrete railway bridges in Sweden are reaching their designed lifespan and are scheduled to be demolished and replaced in the upcoming years. To limit the econom-icand environmental impact related to the replacement of existing railway infrastructure, a comprehen-siveevaluation of their capacity is required with the aim of extending its lifespan. In fact, experimental evidence has shown that some of these bridges may have a higher capacity than previously determined due to the conservative assumptions used during their design. The proper stress distribution pattern at the ballast-concrete interface is among the factors that need to be studied, as research on the topic has shown that some of the available guidelines to calculate it can produce conservative results. In this paper, available analytical models for computing the internal forces in concrete bridges due to train axle loads are compared to a numerical model calibrated using the experimental results obtained from the test of ballasted reinforced concrete trough bridge, a typical structural type found in Sweden, and existing research. As a first step, a literature review of existing numerical modeling strategies for ballast-edconcrete railway structures (e.g., finite element models, discrete element models, and their combina-tion)is conducted. Then, the most appropriate numerical modelling strategy is identified and used to develop the numerical model of the bridge, including the ballast. Finally, results of contact pressure and vertical stresses in the numerical model are compared to those obtained analytically.

Suffosion ist ein hydromechanischer Prozess, welcher die Umlagerung und den Transport von feinen Partikeln eines Bodens infolge einer Wasserströmung beschreibt. Als Folge des Materialverlustes vergrößert sich die Porenzahl und die Dichte verringert sich, wodurch sich auch die hydraulischen und bodenmechanischen Eigenschaften des Bodens verändern. Um die Auswirkungen einer Suffosion besser bewerten zu können, wurden die Einflüsse von geometrischen Faktoren (Anfangsfeinanteil und Probendichte), hydraulischen Faktoren (Filtergeschwindigkeit) sowie der effektiven Spannung untersucht. Anhand von 3D-DEM Simulationen mit binären Mischungen wurde die Struktur (Kontakt- und Kontaktkraftverteilung) eines suffosionsanfälligen Bodens und ihre Änderung infolge der genannten Einflüsse untersucht. Hierbei kann klar erkannt werden, dass die feinen Anteile einer Mischung weniger kontaktiert werden als die groben und auch weniger Kontaktkräfte übertragen. Mit steigender Dichte und steigendem Feinanteil steigen allerdings auch die Kontakte und übertragenen Kontaktkräfte der feinen Partikel deutlich an. Anhand von Laborversuchen mit isotroper Belastung und konstanter Durchströmung, wurden die Auswirkungen der verschiedenen Faktoren auf die ausgetragene Materialmenge, die hydraulischen Eigenschaften des Bodens und die infolgedessen auftretenden Verformungen untersucht. Es zeigt sich ein Anstieg der ausgetragenen Materialmenge für einen steigenden Feinanteil, eine steigende hydraulische Einwirkung, sowie eine geringere Probendichte und eine geringere isotrope Belastung. Mit steigender Materialdichte und abnehmender hydraulischer Einwirkung konnte eine Abnahme des Durchlässigkeitsbeiwertes gezeigt werden, welche auf eine Ablagerung von transportierten Partikeln hindeutet. Eine Dimensionsanalyse der Ergebnisse kann einen direkten Zusammenhang zwischen ausgetragenem Feinmaterial und volumetrischer Dehnung des Probekörpers infolge Suffosion zeigen, welcher durch einen Vergleich mit Ergebnissen aus der Literatur bestätigt werden kann. An erodierten Proben wurden drainierte Triaxialversuche durchgeführt. Die Ergebnisse zeigen, dass das qualitative Verhalten der erodierten, sowie der intakten Proben hauptsächlich von der relativen Lagerungsdichte des Grobmaterials bestimmt wird. Weiterhin konnte anhand eines Vergleiches mit Literaturergebnissen gezeigt werden, dass hauptsächlich der Bodenzustand vor der Scherung, beschrieben durch Porenzahl und Porenzahl des Grobmaterials, das Scherverhalten bestimmen. Die Art wie dieser Zustand erreicht wurde, ob durch Suffosion oder künstlich hergestellt, scheint keinen wesentlichen Einfluss auf den maximalen Spannungszustand η = q/p0 während der Scherung zu haben. Weiterhin wurde der Erosionsdurchbruch in feinkörnigen Böden infolge einer Kontakterosion an der Grenzfläche eines feinkörnigen Dichtmaterials zu einem grobkörnigen Filtermaterial untersucht. Hierfür wurden künstlich hergestellte, geschichtete Proben durchströmt, um den kritischen hydraulischen Gradienten für einen Erosionsdurchbruch zu ermitteln. Es wurde hierbei festgestellt, dass der Gradient infolge einer steigenden effektiven Spannung und einem steigenden Überkonsolidierungsgrad des Dichtmaterials sowie einem kleineren Porendurchmesser des Filtermaterials steigt. Anhand einer Dimensionsanalyse konnte gezeigt werden, dass der kritische Gradient auch von der Bodenart abhängen muss.