The presented work documents the influence of Voronoi block size and shape as well as internal mesh size on the calibrated fracture toughness KIC. It is documented that Voronoi based procedures have an inevitable error of up to ± 30%. On the other hand, this approach is able to reproduce complex fracture pattern in a realistic manner with reasonable computational power. The work propose a KIC calibration procedure and documents based on the comparison with lab tests, that crack propagation, fracture pattern as well as stress-strain behavior of brittle solids can be duplicated by calibrated Voronoi based DEM simulations. The thesis also documents a swelling law for the DEM code UDEC including parameter determination and validation on lab tests with swelling cement. Finally, calibrated concrete models with one or two holes under different boundary conditions are used to predict swelling induced cracking. Numerical predictions were compared with corresponding lab tests and showed satisfying results.

This study presents numerical modeling of geologic processes based on Discrete Element Method (DEM), including modeling of pull-apart basin development based on Particle Flow Code (PFC) and simulation of deformation and earthquake potential of Ordos Block (China) under present tectonic stress regime based on Universal Distinct Element Code (UDEC). A scale-independent modeling approach based on PFC2D has been established to simulate the development of pull-apart basins. The micro-scale PFC models are used to investigate crack propagation and basin development in releasing sidestep systems with pure strike-slip, transtensional, and transpressional master faults, respectively. In each system, three typical models including 30° underlapping, 90° non-overlapping, and 150° overlapping releasing sidesteps are chosen. The modeling results are compared with pull-apart basins in nature. The geometric differences of pull-apart basins result from both the initial strike-slip fault geometries and its various evolution stages. Rhomboidal basins which have larger basin length than the amount of motion form in overlapping systems and do not progress through the spindle-shaped and lazy-Z-shaped stages such as the Dead Sea basin. Rhomboidal basins with cross-basin faults tend to form in underlapping systems. Finally, the origin of rhomboidal pull-apart basins, depocenters of pull-apart basins, cross-basin faults and their significances, models for pull-apart basin development, and minimum displacements and ages to form pull-apart basins are discussed. A two-dimensional UDEC model involving Ordos Block and adjacent areas is set up. Boundary conditions based on present tectonic regime are assumed. Block rotations, shear stress and displacement on faults, ratio of shear to normal force are simulated. Slip tendency which represents the assessment of the potential for causing slip on individual faults and earthquake-prone of the faults is predicted. Modeling results are compared with geologic evidences.

<p>  </p> <p>Direct reduction ironmaking has gained popularity as a low carbon alternative to the typical blast furnace ironmaking route. A popular method of producing direct reduced iron is through the reduction of iron ore pellets in a reduction shaft furnace. Critical to this process is the use of a reliable continuous pellet feed system to provide a steady flow of pellets to the furnace. Therefore, any disruption in pellet flow can have a significant negative impact on the production rate of iron. </p> <p><br></p> <p>An iron ore pellet feed system for a direct reduction ironmaking furnace is jamming during winter operation. The pellets are jamming in a hopper at the top of the feed system above the furnace, and a hot gas, that seals off the furnace flue gas, flows counter to the pellets. A computational model of the feed system is built utilizing the discrete element method and computational fluid dynamics, using Siemen’s commercial multiphysics software Star-CCM+, to study the conditions that cause the jam to occur. The study is divided into six parts: pellet bulk flow calibration, computational cost reduction, modeling of the baseline operation, modeling the effect of moisture, development of a thermal model, and investigation of the minimal amount of icy and wet material to jam the system. The findings show that the location of jamming during operation matches the area in the simulation where it is most likely to occur, and that moisture alone is unlikely to result in jamming. Results indicate that the system will jam when charged with a minimum of 15% icy pellets, and when charged with 10% icy together with 5% wet pellets. Experimental work is recommended to validate the findings and to calibrate the simulations accordingly.</p>

Granular materials are widely used as unbound layers within the infrastructure system playing a significant role on performance and maintenance. However, fields like pavement and railway engineering still heavily rely on empirically-based models owing to the complex behaviour of these materials, which partly stems from their discrete nature. In this sense, the discrete element method (DEM) presents a numerical alternative to study the behaviour of discrete systems with explicit consideration of the processes at particulate level governing the macroscopic response.  This thesis aims at providing micromechanical insight into the effect of different particle sizes on the load-bearing structure of granular materials and its influence on the resilient modulus and permanent deformation response, both of which are greatly influenced by the stress level. In order to accomplish this, binary mixtures of elastic spheres under axisymmetric stress are studied using DEM as the simplest expression for gap-graded materials, which in turn also can be seen as a simplification of more complex mixtures. First, the effect of the fines content on the force transmission at contact level was studied. Results were used to define a soil fabric classification system where the roles of the coarse and fine fractions were defined and quantified in terms of force transmission. A behavioural correspondence between numerical mixtures and granular materials was established, where the mixtures were able to reproduce some of the most significant features regarding the resilient modulus and permanent strain dependency on stress level for granular materials. A good correlation between soil fabric and performance was also found. Generally, higher resilient modulus and lower deformation values were observed for interactive fabrics, whereas the opposite held for instable fabrics. Mixtures of elastic spheres are far from granular materials, where numerous additional factors should be considered. Nevertheless, it is the author’s belief that this work provides insight into the soil fabric structure and its effect on the macroscopic response of granular materials. / Grus i form av krossat bergmaterial används i stor utsträckning som obundna bär- och förstärkningslager inom tranportinfrastrukturen och spelar där en viktig roll för verkningsätt, drift och underhåll. Det finns emellertid begränsad kunskap om de fundamentala mekanismerna på partikelnivå (d.v.s. enskilda gruskorn), mekanismer som styr det makromekaniska verkningssättet. Områden såsom väg- och järnvägsbyggnad bygger fortfarande väsentligen på empiriskta baserade modeller p.g.a. dessa materials komplexa uppträdande under belastning. Denna komplexitet beror delvis på den diskreta naturen hos problemet vilket innebär att traditionell matematisk modellering som vore materialen homogena och kontinuerliga, blir inadekvat. Mot denna bakgrund utgör den s.k. diskreta elementmetoden (DEM) ett numeriskt alternativ för att studera verkningssätt hos diskreta system där man explicit beaktar mekanismerna på partikelnivå. Denna avhandling, som baseras på tre vetenskapliga bidrag, syftar till att ge mikromekaniska insikter vad gäller effekten av olika partikelstorlekar på bärförmågan hos grusmateral och dess inverkan på styvhet och motstånd mot permanenta deformationer. Båda dessa parametrar påverkas kraftigt av spänningsnivån och kan studeras genom triaxialförsök. För att undersöka detta studerades med hjälp av DEM binära blandningar av elastiska kulor – den enklaste modellen av grusmaterial med språng i fördelningskurvan – som utsattes för axialsymmetrisk belastning. Denna modell kan i sin tur ses som en förenkling av mer komplexa blandningar. Inledningsvis studerades effekten av finpartikelinnehållet på partikelkontakternas kraftöverföring. Resultaten användes för att klassificera olika typer av skelettstrukturer i grusmaterialet där den finare och den grövre fraktionens roller kvantifierades med utgångspunkt från kraftöverföringen i stället för från det makromekaniska verkningssättet. Resultaten visade en korrelation vad gäller verkningssättet mellan numeriska blandningar och grusmaterial, där de numeriska blandningarna kunde reproducera några av grusmaterials viktigaste kännetecken vad gäller spänningsberoendet för styvheten vid avlastning och motståndet mot permanent deformation. Vidare visades att styvheten kunde bestämmas ur första belastningscykeln vilket underlättar att övervinna de begränsningar avseende beräkningstid som annars förknippas med DEM. God överensstämmelse mellan grusmaterialets skelettstruktur och verkningssätt kunde också observeras. Generellt observerades högre styvhet och mindre permanenta deformationer för interaktiva skelettstrukturer medan det motsatta gällde för instabila strukturer. Numeriska blandningar av elastiska kulor är långt från verkliga grusmaterial, för vilka ett stort antal ytterligare faktorer måste beaktas. Icke desto mindre är det författarens övertygelse att detta arbete ger insikter i grusmaterialets skelettstruktur och dess effekter på det makromekaniska verkningssättet hos grusmaterial. / <p>QC 20161116</p>

Le renforcement des sols par géosynthétique est appliqué dans de nombreux types d’ouvrage : remblais sur sol compressible, talus sur fondations stables, remblais sur des cavités et ouvrages de soutènement. La stabilité de ces ouvrages dépend entre autres de l’efficacité des ancrages des nappes géosynthétiques. Les ancrages droit et avec retour sont les plus couramment utilisés. Afin d'améliorer les connaissances actuelles sur le comportement des systèmes d'ancrage, des études expérimentales et numériques ont été développées conjointement. Ce travail de thèse concerne dans une première partie, la modélisation physique tridimensionnelle du comportement des géosynthétiques pour deux types ancrages (ancrage droit et ancrage avec retour). Ces essais ont été réalisés dans une chambre d’étalonnage en conditions contrôlées et instrumentées en laboratoire. Dans une deuxième partie de cette thèse, les paramètres d’interaction sol/géosynthétique déduits à partir de l’étude expérimentale ont été implémentés dans le code de calcul numérique bidimensionnel en milieu continu FLAC2D pour une meilleure compréhension du comportement des géosynthétique en ancrage. L’influence de plusieurs paramètres sur le comportement du géosynthétique en ancrage avec et sans retour a été traitée dans cette étude numérique. Parallèlement à cette modélisation, une autre modélisation numérique (discontinue) par la méthode des éléments discrets (PFC2D) a été réalisée. Ces modélisations ont donné des résultats proches de ceux obtenus expérimentalement et confirme l'analyse faite au sujet des mécanismes d'ancrage. / The soil reinforcement by geosynthetic is used in many types of structures: embankments on compressible soil, slope on a stable foundation, embankments on cavities and retaining structures. The stability of these structures specially depends on the efficiency of the anchors holding the geosynthetic sheets. The simple run-out and anchorage with wrap around are most commonly used. In order to improve the available knowledge of the anchorage systems behaviour, experimental and numerical studies were developed jointly. This thesis work concerns in the first part a three-dimensional physical modelling of the behaviour of geosynthetics in two anchors (simple run-out and anchorge with wrap around). The pull-out tests were performed in a test tank under controlled conditions in the laboratory. In the second part, the parameters of the interaction soil/geosynthetic found from the experimental study were used into the numerical code “FLAC2D” (continuous) for a better understanding of the behaviour of geosynthetics in anchorage. The influence of several parameters on the behaviour of geosynthetic was treated in this numerical study. In parallel with this model, another numerical modelling (discontinuous) by the discrete element method (PFC2D) was carried out. The results of these models are closely to experimental results which confirm the analysis about the anchoring mechanisms.

Les matériaux architecturés attirent de plus en plus d’attentions de par leur capacité à combiner différentes propriétés ciblées. Dans ce contexte, les matériaux enchevêtrés, et plus particulièrement les matériaux monofilamentaires enchevêtrés, présentent des propriétés intéressantes en terme de légèreté, de ductilité, et de facteur de perte. En raison de l’architecture interne complexe de ces matériaux, leur caractérisation et la compréhension des mécanismes de déformation nécessitent une méthodologie adaptée. Dans cette étude, l’enchevêtrement est réalisé manuellement pour différents fils d’acier et soumis à une compression oedométrique. De manière à étudier le comportement sous charge de ce type de matériaux, un dispositif de compression uniaxiale guidée a été mis en place dans le tomographe. Il est ainsi possible de suivre, à l’aide de mesures quantitatives, la déformation de l’échantillon et l’évolution du nombre de contacts pour différentes fraction volumiques. L’utilisation de ces données microstructurales a permis un meilleure compréhension du comportement mécanique de tels enchevêtrements. Une rigidité pouvant varier de 20 à 200 MPa en fonction des paramètres de mise en forme (diamètre et forme du fil, fraction volumique, matériau constitutif) a été déterminé. Un matériau homogène de rigidité plus faible a pu être obtenu en pré-déformant le fil sous forme de ressort avant enchevêtrement. Le facteur de perte du matériau a ensuite été mesuré à la fois sous chargement statique et dynamique. L’analyse mécanique dynamique a mis en évidence la capacité de ce matériau à absorber de l’énergie avec une valeur de facteur de perte d’environ 0.25. Les propriétés mécaniques du matériau ont tout d’abord été modélisées analytiquement par un modèle de poutres et un bon accord avec les résultats expérimentaux a pu être obtenu en définissant un paramètre d’orientation equivalent, spécifique à la compression oedométrique de matériaux enchevêtrés. En parallèle, un modéle éléments discrets a été developé afin de simuler le comportement en compression de matériaux monofilamentaires enchevêtrés. Ce modèle s’appuie sur une discrétisation du fil en éléments sphériques, acquise à partir de données de tomographie. Bien que seul le comportement élastique du fil constitutif ait été pris en compte, une bonne adéquation entre résultats numériques et expérimentaux a été obtenu en ajustant les coefficients de frottement du modèle. / Playing with the architecture of a material is a clever way of tailoring its properties for multi-functional applications. A lot of research have been made, in the past few years, on what is now referred to as “architectured materials” (metal foams, entangled materials, steel wool, etc), mostly for their capacity to be engineered in order to present specific properties, inherent to their architecture. In this context, some studies have been carried out concerning entangled materials but only a few on monofilament entangled materials. Such a material, with no filament ends, could exhibit interesting properties for shock absorption, vibration damping and ductility. In this study, entanglements were manually produced, using different types of wire, and submitted to constrained (inside a PTFE die) in-situ compressive tests within the laboratory tomograph. This technique enabled a 3D, non destructive, microstructural characterization of the complex architecture of these materials, along with the analysis of their macroscopic mechanical properties. The stiffness of this material was found to be in a 20-200 MPa range and homogeneous samples could be obtained, while lowering their stiffness, by pre-deforming the initial wire as a spring. Damping measurements were performed using different types of entanglements (constitutive materials, volume fraction, wire diameter, wire shape) under both monotonic and dynamic loadings and directly linked to the measurements of the number of contacts. The Dynamic Mechanical Analysis underlined the great capacity of this material to absorb energy with a loss factor of about 0.25 and damping was found to decrease with the stiffness of the entanglement. The mechanical properties of this material were first modeled using an analytical “beam” model based on the experimental evolution of the mean distance between contacts and a good agreement was found with the experimental results. In parallel, a Discrete Element Method was used in order to model the compressive behaviour of Monofilament Entangled Materials. Although purely elastic properties were taken into account in the model, a very good agreement with the experimental results was obtained by adjusting the friction coefficients of the model. This tends to prove that the plasticity of these entangled materials is rather due to the structure (friction) than to the constitutive material itself.

Les phénomènes d'auto-assemblage sont au premier plan de la recherche en sciences des matériaux car ils comblent le vide laissé entre les procédés d'assemblage à l'échelle macroscopique et nanoscopique. L'auto-assemblage est basé sur l'organisation spontanée de composants individuels en motifs et structures. Contrôler l’agencement de la matière peut accroître les propriétés de matériaux en introduisant une certaine anisotropie. Cet agencement, comme de nombreux matériaux naturels le prouvent, peut même sous certaines conditions faire émerger de nouvelles caractéristiques. Au cours de ces trois années, nous avons utilisé l’ « ice templating » (texturation à la glace) pour déclencher l’alignement de plaquettes de dimensions microniques, le but final étant de répliquer la microstructure de la nacre. Cette technique induit la ségrégation des constituants d’une suspension à l’échelle du micron tout en obtenant des échantillons de quelques centimètres cubes. Ce procédé a permis la création de matériaux inorganique avec une microstructure semblable à la nacre, en additionnant trois niveaux de contrôles successifs : l’alignement local des plaquettes, l’alignement à longue distance des cristaux de glaces et enfin le contrôle de l’interface entre ces-mêmes plaquettes. L’utilisation d’une modélisation par éléments discrets nous a permis d’étudier la dynamique de l’auto-assemblage des particules anisotropes. Ce modèle, parce qu’il tient compte de la dynamique du procédé, nous a révélé comment l’organisation de ces particules se produit. La tomographie par rayon X a permis de visualiser les structures finales des échantillons et d’attester de la pertinence du modèle. L’alignement local des plaquettes dans les parois générées par la solidification de la glace peuvent accroître les propriétés fonctionnelles et structurales de composites. Dans ce cadres deux applications ont été étudiées : la conduction thermique dans des composites nitrure de bore hexagonal / silicone et les propriétés mécaniques d’alumine macroporeuses. Une adaptation du procédé a permis d’obtenir l’alignement à longue distance (quelques centimètres) des cristaux de glaces. Différents outils ont été développés pour caractériser la réponse fonctionnelle de ce type de composite en fonction de leurs architectures aux deux échelles considérées (celles des macropores et parois). Enfin, après la mise en place de ces deux niveaux de contrôle sur la structure, l’addition d’une phase vitreuse inorganique et de nanoparticules aux joints de grains des plaquettes a introduit, de façon similaire à la nacre, des interfaces pouvant dévier et arrêter la propagation de fissures. / Self-assembly phenomena are of prime interest in materials science, because they fill the gap between assembly of macrostructure and processing of nanomaterials. Self-assembly is based on the spontaneous organization of individual small constituents into patterns and structures. Controlling the spatial arrangement can possibly improve materials properties by maximizing its response in a given direction. Furthermore, particular types of spatial arrangement, such as found in natural structures, can even induce new properties. During the past three years, we have used ice templating process to trigger the assembly of platelet-shaped particles to replicate the hierarchical structure of nacre. Control over this technique allowed structural customization at different length-scales: local orientation of the platelets, ice crystal long range order, and the control if the interfaces between the platelets. This hierarchical process has set the ground for the creation of a new fully mineral nacre-like alumina. The local platelet self-assembly triggered by ice growth was investigated by Discrete Element Modelling which provided new insight into the dynamic phenomenon responsible for the particles alignment. Synchrotron X-ray tomography was used to validated the model results. The different architecture observed in the final samples are not the result of a percolation threshold, as one could expect, but is instead a consequence of the delicate balance between pushing and engulfment at the solidification front. The local alignment of platelets can be beneficial for the functional and structural characteristics of composites and relevant aspects for two potential applications were investigated: the thermal properties of the hexagonal boron nitride/silicon rubber composites and the mechanical properties of macroporous alumina. Further adaptation of the process allowed for long range ordering of the ice crystals (up to the centimeter scale). Different tools have also been developed in order to characterize the response of composites as a function of the architecture at the level of the macropores and particle organisation. Once those two levels of alignment were achieved, the addition of a glassy phase and nanoparticles to the grain boundaries of the platelets introduces, just like in nacre, interfaces capable of deflect and even stopping crack propagation.

Les risques naturels tels que les laves torrentielles constituent des menaces réelles pourles zones urbanisées de montagne. Les bâtiments et infrastructures peuvent être exposésà de grandes forces d’impact en cas d’évènement extrême. La réduction de cette menace,par des ouvrages de protection, impose de quantifier l’impact de ces écoulements sur lesstructures, qu’elles soient flexibles ou rigides.Tout d’abord, un écoulement granulaire sec, composé de particules non-sphériquesglissant sur un plan incliné, est modélisé en utilisant une loi de contact visco-élastiqueavec critère de rupture de Mohr-Coulomb. Des données expérimentales de la littératureont été utilisé pour calibrer et valider le modèle. À cette fin, la forme de la particule,l’épaisseur de l’écoulement et la forme finale du dépôt sur le mur sont considerés. Lavalidation est basée sur l’impact sur un mur rigide divisé en six segments. La principalecontribution de la force totale normale appliquée sur le mur est due à la composantedynamiques. La distribution hétérogène de la force normale sur chaque partie du murest due au développement des chaînes de force différent pour chaque arrangement desparticules.Ensuite, un filet est modélisé en utilisant des éléments cylindriques. L’impact sur lefilet est modélisé en utilisant le même modèle d’écoulement que précédemment. Le rôledes dissipateurs d’énergie apparaît essentiel pour réduire la force d’impact sur le filet etlimiter la force appliquée sur les points d’ancrage latéraux.Pour la première fois, des simulations montrent que pour un même écoulementgranulaire la force d’impact est plus élevée pour un obstacle rigide, avec une différencede 50% par rapport à un obstacle flexible. les simulations permettent de définir quelquesrecommandation pour le dimensionnement des filets. Il est constaté que l’utilisationviiid’un maillage de filet plus petit que D90 de l’écoulement est acceptable en termes decapacité à retenir les matériaux en écoulement. En plus, si le câble en bas du filet n’estpas fixé, le filet pourrait perdre totalement sa capacité de retenue. / Natural hazards such as debris flows are real threat to the urbanization of mountainousareas. Local communities and infrastructures can be exposed to large impact forces inextreme debris events. Mitigation of such threats requires, along other measures, theestimation of the impact of such flows on protection structures (rigid walls and flexiblebarriers). In this thesis, Discrete Element Method (DEM) is used to model the granularflow, the rigid walls and flexible barriers.First, a dry granular flow made of non-spherical particles flowing in inclined plane ismodeled using a visco-elastic contact law with Mohr-Coulomb failure criterion. Experimentaldata from the literature is used to calibrate and validate the model. The modelis calibrated based on the shape of the particle, the flow thickness and the final shapeof the deposit on the wall. Validation procedure is based on the impact on a rigid walldivided into six segments. The main contribution of total normal force applied on thewall is found to be due to the dynamic component. On the micro-scale, development offorce chains is believed to cause heterogeneous distribution of normal force on each partof the wall, for multiple same-test conditions.Next, a flexible barrier is modeled using cylindrical elements. The impact on thebarrier is modeled using the same flow model used for wall-impact problem. The use ofenergy dissipators is found to be essential for minimizing the impact force on the barrier,and thus controlling the force applied on the lateral anchors.By comparing a rigid wall and a flexible barrier for the same flow, we found thatthe rigid wall is exposed to higher impact force, due its high global stiffness comparedwith the flexible barrier. Next, different simulations are carried out to recommend designguidelines for the flexible barrier. It is found that using a mesh size as large as D90 of theviflow is acceptable in terms of mass retaining capacity. In addition, not fixing the bottomcable of flexible barriers might lead to the total loss of its retaining capacity in extremeevents.

Les structures en enrochements sont parmi les ouvrages les plus usuels de génie civil (barrages, murs de soutènement,. . . ). Des tassements importants peuvent apparaître tout au long de leur durée de vie et sont principalement dus à la rupture des blocs rocheux. Cette thèse propose un modèle numérique permettant de simuler le comportement de matériaux granulaires présentant des ruptures de grains. Afin de prendre en compte la nature discontinue de ces milieux, la méthode des éléments discrets est utilisée. La modélisation adoptée est de type "Non-Smooth Contact Dynamics", où les grains et particules sont supposés rigides. Afin de générer des blocs ayant des formes complexes, un modèle de grain 3D est proposé. Ce modèle de grain est ensuite discrétisé en sous-éléments de forme tétraédrique liés par des liaisons cohésives afin de pouvoir représenter la rupture. Un critère de rupture de Mohr-Coulomb est utilisé. Le modèle est implémenté sur la plateforme logicielle LMGC90. Le modèle est d’abord éprouvé lors de simulations d’écrasement de blocs cassables entre deux plaques. Plusieurs paramètres contrôlant la résistance du grain sont étudiés : cohésion intergranulaire, taille, discrétisation, forme et orientation du grain. L’effet d’échelle observé sur ce type de matériau est vérifié. Le modèle est ensuite testé lors de simulations numériques de compression œdométrique d’enrochements. L’effet des paramètres du modèle et de l’assemblage du milieu granulaire est également étudié. Les simulations œdométriques sont confrontées à des résultats expérimentaux et présentent une bonne concordance. Enfin, des expérimentations numériques sont menées afin d’étudier les énergies mises en jeu dans ces essais. L’énergie de création de surface est estimée pour ce type de matériau. Les résultats sont proches des données de la littérature. / Rockfill structures are very popular among civil engineering structures (dams, retaining walls, . . . ). Important settlements can take place during the lifetime of these structures, settlements mainly caused by the breakage of rockfill grains. This thesis proposes a numerical model that allows the simulation of the behavior of granular materials exhibiting grain breakage. To take into account the discrete nature of these media, the discrete element method is chosen. The adopted strategy is the Non-Smooth Contact Dynamics method, where grains are considered to be rigid. To generate blocks having complex shapes, a 3D grain model is suggested. This grain model is then discretized into tetrahedral subgrains, joined together using cohesive bonds so that breakage can be simulated. A Mohr-Coulomb failure criterion is used for the cohesive bonds. The model is implemented into the LMGC90 software platform. At first, the model is tested in single grain crushing simulations between two plates. Multiple parameters controling the strength of the grain are studied : the intra-granular cohesion, the size, the discretization and the orientation of the grain. The scale effect that characterizes this type of material is verified. Then the model is tested in numerical simulations of œdometric compression of rockfill. The influence of the parameters of the model and of those of the granular medium are studied. The results of œdometric simulations are compared to experimental results, and present a good agreement. Lastly, numerical experimentations are conducted in order to study the energies that are brought into play in the simulations. The surface creation energy is estimated for this type of material. Results are close to the data provided in the literature.

Granular materials are characterized by its discrete nature which makes their behaviour very complex to understand when subjected to various loading situations. Comparing other materials, the understanding of granular materials is poor. This is because experimental analysis provides the macroscopic responses of the considered assembly whereas the discrete nature of the particles point to the fact of understanding the micro scale details and correlating it with the overall behaviour. Among the various modeling tools viz. analytical, physical or numerical, Discrete Element Method (DEM) a numerical technique, originally developed by Cundall (1971, 1974) and modified by Cundall and Strack (1979a, 1979b) is widely used for granular materials. Later a thorough description of DEM was given by Cundall (1988) and Hart et al (1988). Moreover Cundall & Hart (1992) reported discrete element code as one which allows finite displacements and rotations of discrete bodies along with recognition of new contacts as the calculation progresses which is followed in particle flow code and is used for this study. Generally the discrete particles are modeled as discs or spheres in 2-D and 3-D simulations respectively. The discs or spheres were considered as it is very easy to characterize the grain interactions and the contact detection. Even though the significance of particle shape has been reported in literature, a comprehensive 3-D study of the effect of particle shape on the various aspects of soil behaviour is lacking and is not reported. Particle shape is generally defined in terms of elongation, roundness and texture. Elongation is an indication of the particle aspect ratio whereas roundness measures the sharpness or angularity of particle’s edges and corners. Texture is related to the roughness of the surface. Particle gradation also plays a role in the mechanical behaviour. The influence of each of these factors on the mechanical behaviour of the assembly is important. Hence the major factors like elongation, texture etc which are used to define the particle shape are incorporated in this study. Evaluating the particle shape is another hurdle. In this study, the particle shape is analyzed using a 3D laser scanner which helps to identify the major and minor axis lengths of the particle. Additionally, the effective use of 3D DEM on large scale real life applications incorporating the particle shape effect is also not dealt with very extensively. Hence in this research, a comprehensive study on the calibration of DEM using glass beads, effect of particle shape on drained and undrained monotonic behaviour, liquefaction, post liquefaction and dynamic properties and the application of DEM to a grain polishing machine and an underground tunnel assembly is presented. In the present study, a set of drained triaxial tests were done on glass bead assembly using a laboratory triaxial set up. The glass beads used for the test were spherical and ellipsoidal in shape. The shape of glass beads was characterized through a sophisticated method of 3D laser scanning. In this scanning, the shape of the image of the glass bead is captured through an array of digitized points. These images obtained as unstructured 3D triangular meshes on processing will render the long and short axes of the particle which can be used for proper modeling of the particle shape. After obtaining the long and short axes for the particles, the same is used for the numerical modeling of the glass beads. The numerical simulation results have shown that the assembly modeled with clumped particles gives results qualitatively and quantitatively similar to the observed experimental macro responses. Hence this is used to demonstrate the power of DEM to realistically model the granular behaviour by incorporating the particle shape effect. In addition, undrained simulation of granular material has been numerically predicted from drained triaxial tests which are similar to the approach proposed by Norris et al (1997). An excellent correlation between undrained results predicted from drained triaxial test and undrained test (performed under constant volume conditions) has been observed. This further underlines the fact that the constant volume simulations are equivalent to undrained tests. Having validated the DEM results to the values obtained from the experiments on glass beads, a series of monotonic drained and undrained triaxial tests were performed on cylindrical assemblies of height to diameter ratio 2:1. Four different sets of assemblies were prepared which consists of particles of different aspect ratio to study the influence of particle shape. The behaviour of these assemblies under drained shearing indicates that the strength of the clumped assemblies is higher than that of the spherical assembly at all confining pressures. This has been explained from the magnitude of the anisotropic vi coefficients associated with the fabric and normal contact force tensors. It is also noted that eventhough both assemblies reach the peak strength at the same axial strain, the strain softening associated with the clumped assembly is very rapid which is due to the fact that clumps try to push each other apart as it offers more resistance to rotation resulting in dilation. Another significant observation is that a general increase in aspect ratio will not keep on increasing the strength. As the aspect ratio increases, the particles have a tendency to orient along their larger dimensions. This helps them to attain the lowest potential energy leading to a stable equilibrium and resulting in inherent fabric anisotropy. But when the particles try to align along the larger dimension, the formation of strong contact forces along the direction of loading is hindered. In addition, the lower strength associated with the higher aspect ratio particle assembly can also be attributed to the formation of unexpected void spaces when these longer particles bridge gaps over the underlying grains. The critical state studies indicate that the clumped assembly is having a higher residual strength compared to that of the spherical assembly. In the case of clumped assemblies also, irrespective of the initial loose or dense state of the assembly and the confining pressure applied, the samples reached the same critical state which underlines that the critical state is unique for a granular material. Moreover, the critical state line is non-linear for both the spherical and clumped assemblies. The studies conducted on the liquefaction behaviour indicates that at lower confining pressures the assemblies with particles consisting of lower aspect ratios loses its strength at less number of cycles which can be attributed to the interlocking of non-spherical particles resulting in higher number of contacts per particle. Moreover, during the initial cycles of loading, it is seen that the strong contacts are predominantly in the vertical direction or more precisely along the direction of maximum axial strain. But the plots extracted at higher cycles indicated that the strong contacts along the vertical direction have diminished considerably. This reduction in contact force magnitude and force chain indicates a drop in the number of contacts and is clearly visible in the gradual decrease of average coordination number. Another significant observation is that as the confining stresses increases, the rate of pore pressure generation of the assembly vii consisting of only spherical particles is less compared to the other two samples. Furthermore, at higher confining pressures, when the load direction reverses, the fabric of the clumped assemblies fails to change to a new orientation immediately. But to retain equilibrium the force anisotropy will quickly adjust itself. This mismatch results in losing the contacts and resulting in lower strength and less resistance to liquefaction at higher stresses for assemblies consisting of clumped particles. The post liquefaction study of the numerically liquefied samples shows that the assembly consisting of clumped shaped particles gained strength at a much faster rate compared to the assembly consisting of only spheres. This may be attributed to the ability of the clumps to rearrange themselves on a faster rate compared to that of the spherical particles. The rate of development of average coordination number is very significant as it explains the ability of the assembly to build up the deviatoric stress from a complete collapsed structure. As the contacts develop, the average coordination number as well as the deviatoric stress starts increasing with both the values higher for the assembly consisting of clumped particles. The evaluation of the dynamic properties viz. shear modulus and damping ratio showed a trend similar to the experimental observations on real granular materials. It is observed that the normalized shear modulus reduces with an increase in shear strain and the rate of reduction is very high at low strains for all the samples. It can be seen that as the confining pressure increases, the normalized shear modulus value also increases and the rate of increment is higher for the assemblies consisting of non-spherical particles. Furthermore, for all the samples the threshold shear strain is about 0.001 up to which the behaviour is elastic. Beyond the threshold shear strain, the variation of the normalized shear modulus ratio is non-linear. At small shear strains, the energy dissipation is low resulting in smaller values of damping. As the strains increase, the non-linearity in the constitutive behaviour results in higher material damping leading to high damping value. The simulations of the food polishing machine helps to understand the pattern of hitting of clumped grains on the wall with due importance to the velocity of hit, angle of hit, force of hit, and the number of grains hitting the wall. The modeling and subsequent extraction of the data helped to identify that the wear and tear of the machine was not uniform and was clustered to specific regions due to the non-uniform distribution of the considered parameters. This helped to design a more sophisticated system such that the parts which are subjected to more deterioration are provided with additional support. To bring out the effect of the particle shape, simulations are performed using spherical particles and the results show that the pattern of variation is same, but the magnitudes are different owing to the less surface area associated with the spherical particles. The 3-D simulations of an underground tunnel assembly in a weak weathered rock helped to understand the variation in the stability of the system with and without lining. It was observed that the introduction of lining resulted in a more stable configuration and the circumferential stresses were found to be distributed uniformly around the tunnel. FEM simulations also show a similar trend of stress and strain variations but were unable to capture the ground loosening around the tunnel and the formation of the ground arch whereas DEM could realistically capture all these phenomena. It was observed that as the shape changes from sphere to non-spherical particles, circumferential stresses around the tunnels increased. In addition, as the distance from the tunnel face increases, the strains are reduced. The maximum vertical strain near the crown of the tunnel is observed for the assembly consisting of spherical particles. In short, this research is focused on a comprehensive understanding of the particle shape effect on the mechanical behaviour of granular mass. Numerical simulations incorporating the shape effect has been done on drained and undrained monotonic shear tests, critical state, liquefaction, post liquefaction and dynamic properties. Besides, the granular dynamics simulation of the movement of long food grains in a food polishing machine and the behaviour of an underground tunnel in a granular assembly is also reported.