L'effet de serre lié à l'émission de CO2 a conduit à des projets de stockage de ce gaz dans des formations réservoirs. Ces formations peuvent être traversées de failles et l'examen de la sûreté du stockage nécessite alors l'étude du risque de réactivation et de propagation de ces failles. Cette étude passe par des investigations approfondies portant sur des conditions de propagation des fractures sous sollicitations hydromécaniques. Cette thèse a pour objectif l'étude théorique et numérique de ces conditions ainsi que la simulation numérique de la propagation. La modélisation numérique des processus thermo-hydro-mécaniques dans les milieux poreux fracturés par la méthode des éléments finis (MEF) permet de simuler des phénomènes complexes et non linéaires. Les difficultés liées à l'intégration des équations d'échanges de fluide entre la fracture et la matrice environnante avec la MEF ont été résolues dans des travaux récents et nos simulations numériques ont pu être basées sur cette méthode. Dans un premier temps, nous avons modélisé l'écoulement transitoire dans et au voisinage d'une fracture soumise à une injection de fluide et nous avons étudié le facteur d'intensité des contraintes (FIC) à l'extrémité de la fracture dans le cadre de la théorie de la poroélasticité linéaire. Si les conditions d'injection sont maintenues constantes et la fracture n'évolue pas, l'écoulement tend vers un état stationnaire. Le FIC évolue au cours de la phase transitoire pour atteindre une valeur limite dans l'état stationnaire. La modélisation de l'écoulement transitoire est très coûteuse en temps de calcul et il est intéressant de trouver un moyen d'exploiter au mieux les résultats d'un calcul en état stationnaire. L'analyse théorique et les résultats des simulations numériques montrent en effet que le FIC calculé à l'état stationnaire peut fournir certaines bornes pour la propagation des fractures sous l'écoulement transitoire. Dans le cadre de la poroélasticité linéaire et de l'écoulement de Poiseuille dans les fractures, des expressions semi-analytiques pour le FIC à l'état d'écoulement stationnaire ont pu être dérivées. Pour des géométries simples, ces formules approximatives se révèlent efficaces pour discuter des conditions de propagation des fractures pour des cas typiques et simples de géométrie de la fracture et des conditions d'injection de fluide. Dans un deuxième temps, un Modèle de Fracture Cohésive (MFC) a été utilisé pour modéliser la propagation de fracture sur la base de l'endommagent. Ce modèle, basé sur un critère de rupture de Mohr–Coulomb modifié, permet de simuler l'endommagement de l'interface à la fois sous sollicitations en mode I et II. Une relation d'équivalence entre les paramètres de ce modèle et du modèle de Mécanique Linéaire de la Rupture (MLR) a été établie sur la base de la longueur de propagation de fracture sous des charges similaires. Cette relation permet l'extension de l'équivalence théorique entre MLR et MFC établie pour les matériaux fragiles et sur la base de critères énergétiques, à des matériaux quasi-fragiles et ductiles. On a d'ailleurs montré que le MFC permet de simuler certains phénomènes spécifiques tels qu'instabilités de propagation en mode I et II et le branchement de la fracture en mode II. Enfin, la prise en compte de la pression de fluide dans la fracture a permis d'obtenir un modèle de MFC couplé avec l'hydraulique qui a été implémenté dans un code numérique aux éléments finis en vue d'étudier la propagation des fractures sous sollicitations hydromécaniques. Des simulations numériques ont été réalisées afin d'étudier le risque de réactivation et de propagation des failles dans le contexte de stockage du CO2 en particulier dans une configuration de formation réservoir du type Bassin de Paris / Global warming effect related to CO2 emission has led to sequestration projects of this gas in reservoir formations. These formations can be crossed by faults and safety issue of storage requires the study of fault reactivation and propagation risk. This study goes through in-depth investigations of fracture propagation conditions under hydromechanical solicitations. This thesis aims at theoretical and numerical studies of these conditions and the numerical simulation of fracture propagation. Numerical modelling of thermo-hydro-mechanical processes in fractured porous media using Finite Element Method (FEM) allows the simulation of complex and nonlinear phenomena. Difficulties in integrating fluid mass exchange between fracture and surrounding matrix in the equations with FEM have been solved in recent works and our numerical simulations have been based on this progress. In a first step, we modelled transient flow subjected to a fluid injection and we studied the Stress Intensity Factor (SIF) at fracture tip in the framework of linear poroelasticity theory. If injection conditions are kept constant and the fracture does not evolve, the flow tends to a steady state. The SIF develops during transient phase to reach a limit value in the steady state. Modelling of transient flow is very time consuming and it is interesting to find a method to exploit the results of a calculation in steady state. Theoretical analysis and results of numerical simulations show that the SIF calculated at steady state can provide some bounds for fracture propagation under transient flow. In the framework of linear poroelasticity and Poiseuille flow in fractures, some semi-analytical expressions of SIF at steady state could be derived. For simple geometries, these approximate formulations are efficient to discuss fracture propagation conditions for typical and simple cases of fracture geometry and fluid injection conditions. In a second step, a Cohesive Zone/Fracture Model (CFM) was used to model fracture propagation on the basis of damage. This model, based on a modified Mohr-Coulomb failure criterion, simulates interface damage under both mode I and II loads. An equivalence relation between parameters of CFM and Linear Elastic Fracture Mechanics model (LEFM) was established on the basis of fracture propagation length under similar loads. This relationship allows the extension of theoretical equivalence between LEFM and CFM established for brittle materials and on the basis of energy criteria, for quasi-brittle and ductile materials. It has also shown that CFM can simulate specific phenomena such as propagation instabilities for mode I and II and fracture kinking under mode II. Finally, taking into account the fluid pressure in the fracture permitted to obtain a CFM coupled with hydraulic processes which has been implemented in a numerical finite element code to study fracture propagation under hydromechanical solicitations. Numerical simulations were performed to study the risk of fault reactivation and propagation in the context of CO2 injection in Paris Basin reservoir formation

Le Tien Shan est une chaîne de montagne active située à plus de 1000 km de la limite de plaque la plus proche, le front Himalayen. Elle possède une histoire ancienne qui va de la fin du Protérozoïque à la fin du Paléozoïque dans un contexte d’accrétions successives formant la plus grande chaîne d’accrétion du monde, la CAOB. Afin de comprendre comment l’histoire paléozoïque du Tien Shan influence la localisation de la déformation cénozoïque, nous avons dans un premier temps étudié la structuration de la chaîne pour en identifier les principales structures héritées. Nous avons mis en avant le partitionnement de la déformation entre les zones de sutures Sud et Nord au cours de la collision entre le Tarim et le Tien Shan au Carbonifère supérieur. Le front de collision est caractérisé par des chevauchements et des détachements au sein d’unités métamorphiques. Au Nord, se trouve une zone en décrochement caractérisée par une structure en fleur positive. Ensuite, ces résultats sont utilisés comme les conditions initiales de modèles numériques thermo-mécaniques 2D dont le but est de tester l’influence de ces zones héritées sur la localisation de la déformation cénozoïque. Ces modèles montrent que la déformation cénozoïque dans le Tien Shan se localise à la faveur de zones de faiblesses crustales et non mantelliques. De plus, nous avons pu établir que la bordure nord du Tarim avait une rhéologie proche de celle du Tien Shan. Finalement, à l’aide des modèles numériques une étude systématique a permis de mettre en évidence que le couplage entre l’érosion et le réseau de drainage jouait un rôle important dans la répartition et l’âge des roches de basse température exhumées. / The Tien Shan is an active mountain belt located at more than thousand kilometres of the closest plate boundary, the Himalayan front. Its Late Proterozoic to Late Paleozoic history takes place during the CAOB accretion which represents the largest accretionary belt in the world. In order to understand how the Paleozoic tectonics of the Tien Shan influences the Cenozoic strain localization, we aim at identifying the main inherited structures of the belt. We highlighted the strain partitioning between the North and South suture zones during the Tarim-Tien Shan Late Carboniferous collision. The collisional front is characterized by thrusts and detachments in metamorphic units while northward, a strike-slip zone is evidenced by a positive flower structure. Then, these results are used as variable inputs in 2D numerical thermo-mechanical models in order to assess the role of these inherited structures on the Cenozoic strain localization. These models show that the Cenozoic deformation in the Tien Shan is localized in favour of crustal weak zones instead of mantellic ones. Moreover, we are able to show that the northern border of the Tarim has a Tien Shan like rheology. Finally, we performed a systematic numerical modelling analyse in order to show that the coupling between erosion and drainage network plays an important role on low temperature rocks exhumation ages and repartition.

.Dans cette thèse, il a été étudié la structure des milieux poreux, en particulier sur des sols sableux et un paquet virtuel de sphères; Sur la base de références bibliographiques, nos propres outils ont été créés pour calculer la porosité, la surface spécifique et la distribution de la taille des pores. Nous avons construit un algorithme pour résoudre l'équation de diffusion de l'advection directement sur la structure du milieu poreux (en utilisant un résultat d'image 3D du scan $ mu $ CT du support poreux). Nous avons utilisé l'opérateur de division pour calculer la partie advective avec une méthode de volume fini (FV), mettant en œuvre un schéma de réduction de la variation totale (TVD). La partie diffusion a été calculée en utilisant une méthode de FV et avec l’aide du logiciel MUMPS pour résoudre le système linéaire résultant. A partir du champ de concentration obtenu avec l’algorithme et suivant une méthode de moyenne volumique, nous avons calculé les propriétés macroscopiques de: dispersivité et coefficient de dispersion à Pe = 223,23,2.3,0,23 pour un échantillon de sable de Fontainbleau NE34. Nous avons observé que ces résultats dépendent de la qualité de l'image 3D. Les propriétés structurelles et de transport ont été étudiées à l'aide d'images 3D à différentes résolutions. Les images à différentes résolutions ont été appelées images redimensionnées, elles ont été générées numériquement et prises directement à partir du micro CT scan. Comme premier résultat, nous avons proposé un critère basé sur la distribution de la taille des pores pour déterminer si une résolution d'image 3D convient ou non au calcul de la perméabilité d'un matériau granulaire avec une méthode de volume fini (FV). Dans un deuxième résultat, nous avons montré comment les propriétés des macros de transport de soluté sont moins affectées par une détérioration de la résolution que la propriété d’écoulement de la perméabilité (les deux cas sont calculés par une méthode FV). Et comme troisième résultat, nous avons montré qu'une image numérique redimensionnée préserve davantage le comportement des propriétés macroscopiques qu'une image réelle redimensionnée. / In this thesis it was studied the structure of the porous media, particularly on a sandy soils and a virtual pack of spheres; based on bibliographic references here were generated our own tools to compute the porosity, specific surface and pore size distribution. We built an algorithm to solve advection diffusion equation directly on the porous media structure (using a 3D image result of the $mu $ CT scan of the porous media). We used the splitting operator to compute the advective part with a Finite Volume (FV) method, implementing a Total Variation Diminishing (TVD) scheme. The diffusion part was computed using with a FV method with the assistance of the MUMPS software to solve the resulting linear system. From the concentration field obtained with the algorithm and following a volume averaging method, we computed the macroscopic properties of: dispersivity and dispersion coefficient at Pe=223,23,2.3,0.23 for a sample of Fontainbleau NE34 sand. We observing that these results depend on the quality of the 3D image, structural and transport properties were studied using 3D images at different resolutions. The images at different resolutions were called rescaled images, and they were generated numerically and taken directly from the micro CT scan. As a first result we proposed a criterion based on the pore size distribution to assess if a 3D image resolution is suitable or not for permeability computation of a granular material with a finite volume (FV) method. As a second result we showed how the solute transport macro properties are less affected by a deterioration of the resolution than the flow property of permeability (both cases computed through a FV method). And as a third result we showed that a numerical rescaled image preserve the behavior of the macroscopic properties more than a real rescaled image

<p> As space exploration continues to advance, so does the drive to inhabit celestial bodies. In<br> order to expand our civilization to the Moon or even other planets requires an enormous amount of research and development. The Resilient Extra-Terrestrial Habitat Institute is a NASA funded project that aims to develop the technology needed to establish deep-space habitats. Deep-space inhabitation poses many challenges that are not present here on earth. The Moon, for example, has temperatures that range from -233−123°C. Aside from the extreme temperatures, a variety of thermal loads will need to be handled by the Environmental Control and Life Support System (ECLSS). Aside from the research and architecture of the International Space Station’s ECLSS, very little information is known about disturbances related to the thermal management of extra- terrestrial habitats.<br> </p> <p>RETHi is developing a Cyber-Physical Testbed (CPT) that represents a one-fifth scale<br> prototype of a deep space habitat. In order to answer difficult research questions regarding ECLSS and thermal management of a deep-space habitat, a heat pump was modeled and validated with the physical part of the CPT. Once validated, the heat pump model is able to accurately predict the steady state behavior given the indoor and outdoor conditions of the testbed. When coupled with the interior environment (IE) model, it gives insight into the system’s requirements and response. Experimental testing was conducted with the heat pump in order to validate the model. After the model was validated, a series of parametric studies were conducted in order to investigate the effects of varying thermal loads and dehumidification. Since the groundwork was laid through model development and experimentation, future work consists of designing a more versatile heat pump to test a variety of disturbance scenarios. Although the heat pump model is specifically designed for the CPT, it proves to be versatile for other closed and pressurized environments such as aircraft and clean rooms according to the analysis of dehumidification and dependence on pressure. </p>

<p>Electric load forecasting is a critical tool for power system planning and the creation of sustainable energy systems. Precise and reliable load forecasting enables power system operators to make informed decisions regarding power generation and transmission, optimize energy efficiency, and reduce operational costs and extra power generation costs, to further reduce environment-related issues. However, achieving desirable forecasting performance remains challenging due to the irregular, nonstationary, nonlinear, and noisy nature of the observed data under unprecedented events. In recent years, deep learning and other artificial intelligence techniques have emerged as promising approaches for load forecasting. These techniques have the ability to capture complex patterns and relationships in the data and adapt to changing conditions, thereby enhancing forecasting accuracy. As such, the use of deep learning and other artificial intelligence techniques in load forecasting has become an increasingly popular research topic in the field of power systems. </p> <p>Although deep learning techniques have advanced load forecasting, the field still requires more accurate and efficient models. One promising approach is federated learning, which allows for distributed data analysis without exchanging data among multiple devices or cen- ters. This method is particularly relevant for load forecasting, where each power station’s data is sensitive and must be protected. In this study, a proposed approach utilizing Federated Deeponet for seven different power stations is introduced, which proposes a Federated Deep Operator Networks and a Lagevin Dynamics-based Federated Deep Operator Networks using Stochastic Gradient Langevin Dynamics as the optimizer for training the data daily for one-day and predicting for one-day frequencies by frequencies. The data evaluation methods include mean absolute percentage error and the percentage coverage under confidence interval. The findings demonstrate the potential of federated learning for secure and precise load forecasting, while also highlighting the challenges and opportunities of implementing this approach in real-world scenarios. </p>

The heat input to the tool has been investigated for friction stir welding (FSW) of aluminum alloy AL 7075-T7351 over a wide range of process operating parameters using a combined experimental/numerical approach. In a statistical Design of Experiments fashion, 54 experimental welds (bead-on-plate) were performed at 27 different parameter combinations. Measured outputs during each of the welds included forces in all three coordinate directions and internal temperature of the rotating tool at three locations near the tool/workpiece interface. The heat input to the tool was also identified for each weld using infrared imaging temperature measurement techniques and the portion of the total mechanical power entering the tool was calculated. These values were subsequently analyzed to identify the effect of process operating parameters. Two-dimensional, axisymmetric numerical heat conduction models of the tool were then produced and the approximate spatial distribution of the heat input to the tool along the tool/workpiece interface was identified. Experimental values for the heat input to the tool ranged from 155 W to 200 W, comprising 2.8% to 5.1% of the total mechanical power. Regression equations developed for the two values show that each is a function of the process operating parameters. Heat conduction models of the tool show that the approximate spatial distribution of the heat input to the tool along the tool/workpiece interface is one where the heat input is distributed non-uniformly along the interface, with 1% entering the tool at the pin, 20% entering at the base of the pin, and the remainder entering the flat portion of the shoulder. This distribution was valid for the majority of process operating parameter combinations tested. The maximum predicted temperature for the simulations occurred in the pin. This result was verified by the experimental tool temperature measurements. Insights gained into the FSW process from the combined experimental/numerical investigation were then discussed.

<p>Breast cancer affects 12.5\% of women over their life time and tissue expansion (TE) is the most common technique for breast reconstruction after mastectomy. However, the rate of complications with TE can be as high as 15\%. Even though the first documented case of TE happened in 1957, there has yet to be a standardized procedure established due to the variations among patients and the TE protocols are currently designed based on surgeon's experience. There are several studies of computational and theoretical framework modeling skin growth in TE but these tools are not used in the clinical setting. This dissertation focuses on bridging the gap between the already existing skin growth modeling efforts and it's potential application in the clinical setting.</p> <p><br></p> <p>We started with calibrating a skin growth model based on porcine skin expansions data. We built a predictive finite element model of tissue expansion. Two types of model were tested, isotropic and anisotropic models. Calibration was done in a probabilistic framework, allowing us to capture the inherent biological uncertainty of living tissue. We hypothesized that the skin growth rate was proportional to stretch. Indeed, the Bayesian calibration process confirmed that this conceptual model best explained the data. </p> <p><br></p> <p>Although the initial model described the macroscale response, it did not consider any activity on the cellular level. To account for the underlying cellular mechanisms at the microscopic scale, we have established a new system of differential equations that describe the dynamics of key mechanosensing pathways that we observed to be activated in the porcine model. We calibrated the parameters of the new model based on porcine skin data. The refined model is still able to reproduce the observed macroscale changes in tissue growth, but now based on mechanistic knowledge of the cell mechanobiology.  </p> <p><br></p> <p>Lastly, we demonstrated how our skin growth model can be used in a clinical setting. We created TE simulations matching the protocol used in human patients and compared the results with clinical data with good agreement. Then we established a personalized model built from 3D scans of a patient unique geometry. We verified our model by comparing the skin growth area with the area of the skin harvested in the procedure, again with good agreement.</p> <p><br></p> <p>Our work shows that skin growth modeling can be a powerful tool to aid surgeons design TE procedures before they are actually performed. The simulations can help with optimizing the protocol to guarantee the correct amount of skin is growth in the shortest time possible without subjecting the skin to deformations that can compromise the procedure.</p>

<p>The document describes the computational framework to optimize spring-driven Autoinjectors. It involves Bayesian Optimization for efficient and cost-effective design of Autoinjectors.</p>

The purpose of this research is to examine the behaviour of elliptical concrete-filled steel tubular stub columns under a combination of axial force and bending moment. Most of the research carried out to date involving concrete-filled steel sections has focussed on circular and rectangular tubes, with each shape exhibiting distinct behaviour. The degree of concrete confinement provided by the hollow section wall has been studied under pure compression but remains ambiguous for combined compressive and bending loads, with no current design provision for this loading combination. To explore the structural behaviour, laboratory tests were carried out using eight stub columns of two different tube wall thicknesses and applying axial compression under various eccentricities. Moment-rotation relationships were produced for each specimen to establish the influence of cross-section dimension and axis of bending on overall response. Full 3D finite element models were developed, comparing the effect of different material constitutive models, until good agreement was found. Finally, analytical interaction curves were generated assuming plastic behaviour and compared with the experimental and finite element results. Ground work provided from these tests paves the way for the development of future design guidelines on the member level.

Bending deformation characteristics of monolithic, bi-layer and tri-layer laminate sheet materials are studied using Analytical and FE models in this work. The analytical model, based on advanced theory of pure bending considers von Mises yielding, Ludwik hardening law and Bauschinger effect for various laminate constituent thickness ratios. The principal stresses and strains through the thickness and, change in relative thickness at specified bend curvatures are obtained as a function of increasing curvature during bending. Additionally, 2D and 3D finite element (FE) based models for bending are developed to overcome simplifications of the analytical models such as the effect of specimen width on strain distribution. Further, to experimentally assess and validate bending characteristics from the analytical models, a new experimental bend test-jig that is closer to pure bending is developed. The experimental set-up is an open concept design that allows access to the tensile surface as well as through-thickness region for recording and analyzing strains using an online strain mapping system based on digital image correction (DIC) method. Experimental bending is carried out on annealed AA2024 monolithic aluminum alloy sheet and Steel/Aluminum (SS400/AA1050) bi-layer laminate sheet at different thickness ratios. The model and experiments are studied in terms of stress and strain distribution as a function of relative thickness for different clad to matrix thickness ratios. Further the case of simultaneous bending and stretching over small radius bending is analyzed for limit strain prediction using an existing limit strain criterion based on major strain acceleration. An angular stretch bend test is used to subject an hour-glass shaped AA20240-O aluminum sheet specimen to simultaneous stretching and bending deformation while continuously imaging the critical tensile surface region using an optical camera. The strain development in the critical region is subsequently analyzed using digital image correlation (DIC) method. The effect of DIC parameters such as facet size, facet step, and effect of curve fitting procedures on limit strain are studied. An average limit strain of 0.2 is obtained for AA2024 for a facet size of 9x9 pixels, a facet step of 5 pixels and by applying a 5th order polynomial curve fit to the strain data. The results obtained are comparable with a limit strain of the material. The results are compared with a commercially available tri-layer laminate sheet material Alclad 2024 that has 80 μm thin layer of soft AA1100 on both surfaces of harder AA2024 core material. An improved stretch bendability limit strain of 0.24 for Alclad 2024 tri-layer specimen was predicted by utilizing the major strain acceleration criterion. The thin AA1100 protective layer produced a positive effect on the stretch bendability of Alclad 2024 when compared with monolithic AA2024 specimen. / Thesis / Doctor of Philosophy (PhD)