<p>The main aim of this PhD dissertation is to investigate the inherent and induced anisotropy in granular materials. The study includes both the experimental and theoretical aspects and provides a methodology for characterizing the mechanical response of granular materials that display anisotropy.</p> <p>The content of this thesis is divided into two main parts. The first part is focused on investigating the mechanical properties of materials with inherent anisotropy. In particular, an experimental program designed to investigate the mechanical properties of Ottawa standard sand (C109), with inherent anisotropy that is generated by the initial densification process, is described. The program involves a series of direct shear as well as triaxial axial tests. Its primary objective is to demonstrate that anisotropy may occur in sands that have nearly spherical particles (i.e. are typically considered as isotropic) provided the distribution of pore space has a preferred orientation due to the initial densification process. Following the experimental part, the mathematical formulation based on the Critical Plane Approach (CPA) is presented for describing anisotropic mechanical behavior of the material. The procedure for identification of parameters embedded in the constitutive model is outlined and an extensive numerical analysis is conducted simulating the experimental tests.</p> <p>The second part of this thesis deals with induced anisotropy and its focus is on developing an evolution law for the fabric of particulate materials as a function of continuing deformation. The microstructure descriptors are based on lineal intercept measurements and include the areal porosity and the mean intercept length distribution. The methodology involves performing a series of Discrete Element simulations for a granular assembly under evolving directions of the principal stress/strain and defining a correlation with the evolution of material axes. It is demonstrated that granular materials with spherical particles may become anisotropic due to the initial compaction process and that the induced anisotropy is characterized by the coaxiality between the microstructure and the total strain tensors. The proposed evolution law is incorporated into the constitutive framework for anisotropic materials, as discussed in the first part, and some numerical simulations are conducted. It is demonstrated that the proposed approach can describe, at least in a qualitative manner, several manifestations of induced anisotropy in granular materials.</p> / Doctor of Philosophy (PhD)

Fatigue in concrete is a complex phenomenon involving formation of microcracks, their coalescence into major crack and simultaneous formation of the fracture process zone ahead of the crack tip. Complex phenomena are best dealt through an energy approach and hence it is reasonable to use the theory of thermodynamics. Fracture mechanics and damage mechanics are two theories that are based on physically sound principles and are used to describe failure processes in materials. The former deals with the study of macroscopic cracks, whereas the latter defines the state of microcracking. In this study, the concepts from these theories are utilized to improve our understanding and modeling of fatigue process in concrete. In this thesis, a closed form expression for the thermodynamic function entropy is proposed and examined for its size independency and its use as a material property to characterize failure of concrete under fatigue. In the thermodynamic formalism, dissipative phenomena are described by a dissipation potential or its dual, from which evolution laws for internal variables could be defined. In this work, closed form expressions for dual of dissipation potential are derived using concepts of dimensional analysis and self-similarity within the framework of fracture mechanics and damage mechanics. Consequently, a fatigue crack propagation law and a fatigue damage evolution law are proposed respectively. A method is proposed in this study to correlate fracture mechanics and damage mechanics theories by equating the potentials obtained in each theory. Through this equivalence, a crack could be transformed into an equivalent damage zone and vice versa. Also, damage state corresponding to a given crack in a member can be quantified in terms of a damage index. An analytical way of computing size independent S-N curves is proposed, using a nonlocal damage theory by including aggregate size and specimen size in the formulation. It is realized from this study that fracture mechanics and damage mechanics theories should be used in a unified manner in order to accurately model the process of fatigue in concrete. Furthermore, based on the models developed in this study, several damage indicators for fatigue of concrete are proposed. The advantages and limitations of each of these indices are presented such that, the relevant damage index could be used, based on available parameters. Additionally, deterministic sensitivity studies are carried out to determine the most important parameters influencing fatigue life of a concrete member.

Ce travail de thèse a été réalisé dans le cadre d'une collaboration internationale entre l'Université de Lorraine (France) et l'Université d'Oujda (Maroc). Les travaux réalisés concernent la modélisation de l'usinage par enlèvement de matière. Deux aspects importants de l'usinage ont été abordés : le processus de la formation de copeaux et les échanges thermiques à l'interface outil-copeau. Dans la première partie de la thèse, une modélisation par élément finis (EF) du processus de la coupe a été mise en place. La segmentation des copeaux a été particulièrement analysée grâce à l'introduction d'un nouveau paramètre, le Rapport d'Intensité de Segmentation, permettant de quantifier ce phénomène. Une corrélation entre la réduction de l'effort de coupe et l'intensité de segmentation a été établie. La deuxième partie de la thèse a été consacrée à l'étude des échanges thermiques à l'interface outil-copeau, qui contribuent entre autres à l'usure de l'outil de coupe. Un des points importants de l'étude est la mise en place d'une procédure d'identification hybride (analytique/numérique) permettant d'estimer le flux thermique transmis dans l'outil de coupe et de remonter au coefficient de partage de la chaleur à l'interface outil-copeau pour chaque vitesse de coupe. Avec les valeurs identifiées du coefficient de partage de la chaleur pour chaque vitesse de coupe, une loi d'échange thermique multi-branches a été proposée et ses paramètres identifiés. Cette loi donnant l'évolution du coefficient de partage de la chaleur en fonction de la vitesse de coupe a également été définie en fonction de la vitesse relative de glissement à l'interface outil-copeau dans le but de l'implanter dans un code de calcul EF. L'interface utilisateur VUINTER du code Abaqus/Explicit a été exploitée pour implanter la loi proposée, afin d'appréhender complètement le contact d'un point de vue mécanique et thermique. Il est désormais possible d'implanter via cette interface-utilisateur n'importe quelle autre loi de contact thermomécanique (frottement, coefficient de partage de la chaleur, etc.). L'implantation via la subroutine VUINTER a été validée sur des cas tests d'abord, et puis ensuite en usinage. Les résultats obtenus pour les flux thermiques avec cette nouvelle procédure sont en très bon accord avec les mesures expérimentales pour le couple outil-matière considéré : AA2024-T351/WC-Co / This PhD. thesis is realised in the framework of an international cooperation between the University of Lorraine (France) and the University of Oujda (Morocco). The work done concerns the modelling of machining process by material removal. Two important aspects of machining have been investigated: the chip formation process and the heat exchange at the tool-chip interface. In the first part of the thesis, a FE modelling of the cutting process has been established. Chips segmentation have been particularly analysed using à new parameter (Segmentation Intensity Ratio) allowing the quantification of the phenomenon. A correlation has been established between the cutting force reduction and the chip segmentation intensity. The second part of the thesis has been devoted to the study of heat exchange at the tool-chip interface, among other phenomena that contribute to the tool wear. One important point of the study is the establishment of a hybrid identification procedure (analytical/numerical) to estimate the heat flux transmitted into the cutting tool, and identification of the heat partition coefficient at the contact interface for each cutting speed. With identified values of the heat partition coefficient obtained by varying the cutting speed, a heat exchange multi-branch law has been proposed and parameters of this law have been identified. This law corresponds firstly to the evolution of the heat partition coefficient as a function of the cutting speed. Thereafter, it was defined in term of the relative sliding velocity at the tool-chip contact interface, in order to implement it in a FE code. The user interface VUINTER of Abaqus/Explicit has been used to implement the proposed law, to fully control the mechanical and thermal contact. It is henceforth possible to implement with this user interface any thermomechanical contact (friction, heat partition coefficient, etc.). The implementation via the user subroutine VUINTER was validated first on adequate tests, then on machining. The obtained results for heat fluxes with this new procedure are in good agreement with experimental measurements for the tool-workmaterial couple considered: AA2024-T351/WC-Co