Load cells are very important components in a robot system. They help the robot to get feedback from the environment around it and generate control signals accordingly. However, like every other sensor, load cells need to be calibrated over time to maintain their accuracy and precision. In the current method, they need to be detached from the robot. Then known weights are hung below the load cells to get the raw signal from the load cells. These two types of values will then be used to generate the equations that convert the raw signal to the force values. This is a challenge as not many robots are maintenance-friendly so detaching the load cells can take a lot of time, not to mention the process can damage the load cells if not conducted carefully. This research project utilizes mechanical simulation to calculate the known force values acting on the load cells without taking them out of the robot system. Then these force values are used for the calibration process. In this thesis, the in situ calibration method will be conducted on the actuator-controlled pendulum, and a bipedal robot when it is hanging on the gantry and standing on the ground. Also, since mechanical simulation requires a lot of computational power, a geometry simplification method will also be introduced so this in situ calibration method can be used for ordinary personal computers. The results show that the new calibration method is easy to work with, the force values still meet the requirements for calibration, and the computer only needs 10-12 seconds to run each simulation. / Master of Science / A robotic system usually need the load cell to generate the correct control signal. However, the load cell needs to be calibrated over time for maintenance. The current calibration method requires the load cell to be detached from the robot so the user can apply known forces to the load cell. This thesis introduces an in situ calibration method that can calculate forces that are applied to the load cell so the user does not need to detach the load cell from the robot. An optimization method is also introduced to make the calibration process can be done on an ordinary personal computer.

The life time of mechanical components can be increased by the presence of compressive residual stresses. Inherent to most production processes, residual stresses play a critical role on the mechanical parts behaviour. The knowledge and mastering of residual stresses and linked processes are thus fundamental. The development of efficient numerical methods to predict these residual stresses will allow to save costly experiments and to study the influence of the main parameters. This PhD presents the development and application of semi analytical methods (SAM) to the modelling of mechanical processes of compressive residual stress generation. The SAMs, initially developed for the simulation of elasto-plastic contacts, have the advantage of significantly low computation time compared to classical numerical methods. This method is first used to simulate the low plasticity burnishing process, with a rolling loading. Then, it is used for the simulation of impacts, first unique and then repeated. The frictionless rolling contact between two elasto-plastic bodies is first studied. The influence of plasticity, of the hardening model (isotropic or kinematic), of the geometry of the bodies in contact (spheres or ellipsoids) and of the loading type (indentation or rolling) on the contact pressure and plastic strains are analysed. Impacts simulation is then addressed. The developed method is first validated numerically then confronted to experimentations. Three materials have been particularly studied: 316L, AA 7010 and Inconel 600. The impacts dimensions and the generated strains, measured by digital image correlation, are used to validate experimentally the method. The ultrasonic shotpeening process has been specifically studied. The description of the kinematics of the shots put in movement by a sonotrode in a closed peening chamber has first been studied. The use of analytical formulae for the estimation of the coefficients of restitution, during the numerous impacts between shots and with the chamber's walls, allowed refining the calculation of the average impact velocity as a function of the process parameters. The SAM is the used to determine the plastic strain field induced by the impacts. At last a projection method is proposed to finally determine the residual stress field in thick or thin structures.

3D reinforced composites are favored for aerospace, automotive and wind turbine applicationbecause of their high specific stiffness and strength in the in-plane and out-ofplanedirections. In these composites, pins, stitching yarns and binder yarns are insertedthrough-the-thickness of the in-plane fiber-reinforced regions. Binder parameters as diameter,content, pattern and tensioning can further be varied to regulate the out-of-planeproperties. However, the insertion of these binders distorts the reinforcement which furthercan affect the global and local mechanical behaviour. Unit-cell models offered avaluable approach to assess the effect of the distortions on these mechanical features.An approach is presented to include the main geometrical features of pinned, stitchedand 3D woven composites into mesoscopic unit-cell models. Discretised lines, whichrepresent the main geometrical features, are hereby gradually shaped by geometrical operationswhile a geometrical contact treatment account for line interactions. The localfiber volume fraction and fiber direction distributions are afterwards modelled on crosssectionsin a post-processing step. Tools are further proposed to automatically transformthe geometrical models into finite element models. The effect of distortions, local fibervolume fraction and fiber direction, and typical geometrical features for each 3D reinforcedcomposite, on the stiffness and damage initiation stress levels is investigated bymeans of elastic finite element (FE)-computations.The shape of geometrical features corresponding to the different binder parameters couldautomatically be generated and the dimensions of features could be controlled by the parametersof the geometrical operations. The stiffness of a 3D reinforced composite havebeen observed to be either decreased or increased (dependent on the stacking sequence,the binder type and the loading direction). Early damage initiation in the FE-modelswas observed to take place near the binder locations, which was mainly caused by transverseand shear cracking in the fiber-reinforced regions. Local fiber volume fraction andfiber direction have shown to affect damage initation mechanisms and stress levels, andshould therefore be properly included in the models. In future work, the possibility ofthe framework to generate unit-cells including voids and micro-vascular networks canbe investigated and the finite element models can be extended with damage and crackpropagation mechanisms for damage and failure computations. / Doctorat en Sciences de l'ingénieur et technologie / info:eu-repo/semantics/nonPublished

Dans le cadre de l'étude de la morphogénèse des plantes, bien que la génétique soit intensivement étudiée, le concept clé de rhéologie à différentes échelles : cellulaire, tissulaire ou de l'organisme entier, qui gouverne directement la croissance des organismes et devrait permettre la connaissance de l'évolution de la plante reste encore très pauvrement analysé.Ce travail porte ainsi sur la description des paramètres mécaniques des cellules de plantes, à travers la réalisation de modèles numériques physiques et réalistes de différents sous-domaines de l'organisme Arabidopsis thaliana. Dans ce projet, une approche systématique sera développée pour la réalisation de ces modèles où nos structures s'appuieront sur des données expérimentales (images des différents organes de la plante) décrivant précisément la structure interne d'A. thaliana, ce qui permettra l'obtention de structures et maillages réalistes.Ensuite l'étape de modélisation permettra d'une part de retrouver, à travers le champs de déformation de l'organe décrit, les caractéristiques mécaniques sous-jacentes au modèle, ce qui constitue une forme de problème inverse biologique complexe. Nous appliquerons cette méthodologie à différents sous-domaine de notre plante archétype, commençant par les jeunes tiges, puis les racines et enfin le meristème qui constitue la zone première de croissance de la plante.D'autre part, nous développerons un cadre théorique, sur lequel pourront s'appuyer les biologistes, décrivant un modèle d'organe d'A. thaliana réaliste, typiquement le meristème, et qui permettra de valider ou d'infirmer certaines hypothèses de manipulations encore sujettes à débat, notamment dans le cadre d'utilisation de microscope à force atomique afin d'extraire expérimentalement les propriétés mécaniques des tissus.Enfin, nous présenterons une nouvelle approche mécanique hiérarchique couplant des modèles de simulation à différentes échelles : Multifarious Hierarchy of Mechanical Models. Cette nouvelle approche est à la fois éclectique, car elle permet de coupler des modèles physiques variés (sans maillages, physique modale, éléments finis...), mais aussi flexible, car elle autorise la modularité des différents domaines des modèles sous-jacents, et enfin efficace par l'obtention de gains de performance et de temps face à des simulations éléments finis s'appuyant sur le même niveau de détail. Cette approche devrait permettre le développement d'algorithmes rapides pour les simulations physiques où un niveau de détail local est requis. Par exemple, elle pourrait s'appliquer aux simulations de microscope à force atomique décrites dans la partie précédente. / Morphogenesis in plants is an active field of research. While genetics influence on plant shape has been extensively studied, rheology through scales: from cells to the whole organism through various tissues or organs, is still poorly analysed. Nevertheless, it should lead to the understanding of plants evolving, since it directly drives the organism growth.This work aims at understanding the mechanical caracteristics of plant cells, through several realistic and physically based models of various sub-domains of our plant archetype: Arabidopsis thaliana. In this project, a systematic approach has been developped, where the structures underlying such models rely on experimental data (images of different plant organs), describing precisely the inner structure of A. thaliana, which allows the use of realistic meshesfor our simulations.Then physical modelling allows us to retrieve, through the deformation field of different plant sub-domains, the mechanical properties underlying each type of structure described, which is a typical inverse problem form of a complex biological system. We apply this optimization methodology to several plant organs, beginning with embryonic stem, then with roots and finally with meristems, which constitute the zone where cells can divide and growthtypically takes place.Then, we develop a theoretical framework, on which biologists may rely, describing a realistic model of plant sub-system, typically a meristem. We hope that this conceptual framework will help experimentalists to validate some hypothesis regarding plant manipulations that are still subject to debate, as the use of Atomic Force Microscopy to experimentaly extract mechanical parameters from various plant tissues.Finally, we present a new approach coupling a coarse physically based simulation to a more detailed one : the Multifarious Hierarchiy of Mechanical Models. MHMMs are eclectics as they combine arbitrarily any type of physically based simulation (meshless, modal physics, Finite Elements, ...). Moreover, they are flexible as they allow the modularity of the various domainscontaining the underlying models. Finally MHMMs are much faster than full Finite Element simulation, at the same level of detail. This should allow the development of fast algorithm for local detailed simulation, as was the case for the numerical Atomic Force Microscope in previous part.

The objective of the research presented in this thesis is to validate the parallel-bonded modeling method in the context of coupled thermo-mechanical simulations. The simulation results were compared with analytical and experimental data, in the attempt to assess the usability of this particular modeling method. Previous studies of numerical approaches that related to the thermal fracturing of hard rock had used continuum-based models with constitutive relations. The simulations in the thesis were conducted using Particle Flow Code (PFC) which was chosen for the research because of its several benefits. The code has unique features such as spontaneous damage development without imposed conditions, and emergent properties such as material heterogeneity, and dynamic behavior giving possibility to monitor synthetic seismic events. The basic code has been available since 1995 and research using the code has produced hundreds of publications. The thermal option for the code is a recent addition and lacked verification, validation and applications. The thesis is the answer for that. In the course of the research work new particle clustering and grouping routines were developed and tested. Three modeling studies were conducted varying from laboratory to field scales. The 2D modeling study of the heated cylinder experiment yielded similar results both in fracture-behavioral and acoustic emission (AE) magnitude ranges when compared with the laboratory data. The 3D cubic numerical specimens, created with breakable particle clusters, were heated, and the induced damage was observed by P wave velocity measurements. The results showed trends comparable to the laboratory data: P wave velocity decreases with rising temperatures of up to 250°C and cluster-boundary cracking occurs, comparable to grain-boundary cracking in the heated rock samples. The large 2D tunnel models captured the phenomena observed in-situ displaying the difference in the damage to the roof and floor regions, respectively. This damage was due to the filling material confinement of about 100 kPa on the tunnel floor. In general, the results of the thermo-mechanical simulations were in accordance with the experimental data. The modeled temperature evolutions during the heating and cooling periods were also in accordance with the experimental and analytical data.

The objective of the research presented in this thesis is to validate the parallel-bonded modeling method in the context of coupled thermo-mechanical simulations. The simulation results were compared with analytical and experimental data, in the attempt to assess the usability of this particular modeling method. Previous studies of numerical approaches that related to the thermal fracturing of hard rock had used continuum-based models with constitutive relations. The simulations in the thesis were conducted using Particle Flow Code (PFC) which was chosen for the research because of its several benefits. The code has unique features such as spontaneous damage development without imposed conditions, and emergent properties such as material heterogeneity, and dynamic behavior giving possibility to monitor synthetic seismic events. The basic code has been available since 1995 and research using the code has produced hundreds of publications. The thermal option for the code is a recent addition and lacked verification, validation and applications. The thesis is the answer for that. In the course of the research work new particle clustering and grouping routines were developed and tested. Three modeling studies were conducted varying from laboratory to field scales. The 2D modeling study of the heated cylinder experiment yielded similar results both in fracture-behavioral and acoustic emission (AE) magnitude ranges when compared with the laboratory data. The 3D cubic numerical specimens, created with breakable particle clusters, were heated, and the induced damage was observed by P wave velocity measurements. The results showed trends comparable to the laboratory data: P wave velocity decreases with rising temperatures of up to 250°C and cluster-boundary cracking occurs, comparable to grain-boundary cracking in the heated rock samples. The large 2D tunnel models captured the phenomena observed in-situ displaying the difference in the damage to the roof and floor regions, respectively. This damage was due to the filling material confinement of about 100 kPa on the tunnel floor. In general, the results of the thermo-mechanical simulations were in accordance with the experimental data. The modeled temperature evolutions during the heating and cooling periods were also in accordance with the experimental and analytical data.

Considering the late in-vessel phase of an unlikely core melt down scenario in a light water reactor (LWR) with the formation of a corium pool in the lower head of the re-actor pressure vessel (RPV) the possible failure modes of the RPV and the time to failure have to be investigated to assess the possible loadings on the containment. In this work, an integral model was developed to describe the processes in the lower plenum of the RPV. Two principal model domains have to be distinguished: The temperature field within the melt and RPV is calculated with a thermodynamic model, while a mechanical model is used for the structural analysis of the vessel wall. In the introducing chapters a description is given of the considered accident scenario and the relevant analytical, experimental, and numerical investigations are discussed which were performed worldwide during the last three decades. Following, the occur-ring physical phenomena are analysed and the scaling differences are evaluated between the FOREVER-experiments and a prototypical scenario. The thermodynamic and the mechanical model can be coupled recursively to take into account the mutual influence. This approach not only allows to consider the tem-perature dependence of the material parameters and the thermally induced stress in the mechanical model, it also takes into account the response of the temperature field itself upon the changing vessel geometry. New approaches are applied in this work for the simulation of creep and damage. Using a creep data base, the application of single creep laws could be avoided which is especially advantageous if large temperature, stress, and strain ranges have to be covered. Based on experimental investigations, the creep data base has been de-veloped for an RPV-steel and has been validated against creep tests with different scalings and geometries. It can be stated, that the coupled model is able to exactly describe and predict the vessel deformation in the scaled integral FOREVER-tests. There are uncertainties concerning the time to failure which are related to inexactly known material parame-ters and boundary conditions. The main results of this work can be summarised as follows: Due to the thermody-namic behaviour of the large melt pool with internal heat sources, the upper third of the lower RPV head is exposed to the highest thermo-mechanical loads. This region is called hot focus. Contrary to that, the pole part of the lower head has a higher strength and therefore relocates almost vertically downwards under the combined thermal, weight and internal pressure load of the RPV. On the one hand, it will be possible by external flooding to retain the corium within the RPV even at increased pressures and even in reactors with high power (as e.g. KONVOI). On the other hand, there is no chance for melt retention in the considered scenario if neither internal nor external flooding of the RPV can be achieved. Two patents have been derived from the gained insights. Both are related to pas-sively working devices for accident mitigation: The first one is a support of the RPV lower head pole part. It reduces the maximum mechanical load in the highly stressed area of the hot focus. In this way, it can prevent failure or at least extend the time to failure of the vessel. The second device implements a passive accident mitigation measure by making use of the downward movement of the lower head. Through this, a valve or a flap can be opened to flood the reactor pit with water from a storage res-ervoir located at a higher position in the reactor building. With regard to future plant designs it can be stated - differing from former presump-tions - that an In-Vessel-Retention (IVR) of a molten core is possible within the reac-tor pressure vessel even for reactors with higher power.

L'usage d'humains virtuels s'est démocratisé à de nombreuses activités ces dernières années.Au-delà de la chirurgie virtuelle, les corps virtuels sont de plus en plus utilisés pour concevoir des dispositifs médicaux, des véhicules et des outils de notre quotidien plus généralement.Ils se sont avérés être également d'extraordinaires supports à l'apprentissage de l'anatomie.De récents films (Avatar, Le seigneur des anneaux, etc) ont démontré que l'anatomie et la biomécanique peuvent être utilisées pour concevoir des personnages d'une grande qualité.Cependant, reproduire le comportement des structures anatomiques demeure une tâche complexe, et de nombreuses connaissances variées sont nécéssaires à la mise en place de simulation de qualité de nos organes. Ceci fait de la modélisation pour la simulation d'humains une problématique non résolue, une tâche fastidieuse, mais également un sujet de recherche fascinant.À travers ces travaux de thèse, nous abordons cette problématique de la construction de systèmes musculo-squeletiques pour ces domaines variés : animation, biomécanique et aide à l'apprentissage.Notre objectif est de simplifier le processus entier de création en le rendant plus intuitif et plus rapide.Notre approche consiste à pallier à chacune des difficultés, à savoir : la représentation et la manipulation de connaissances anatomiques, la modélisation géométrique et la simulation efficace de systèmes musculosquelettiques grâce à trois principalescontributions introduites durant ces travaux de recherche.Notre première contribution se focalise sur la construction biomécanique d'un modèle hybride du rachis lombaire.Dans ces travaux, nous montrons que les approches hybrides combinant des systèmes de corps rigides et des modèles éléments finis permettent d'obtenir des simulations en temps intéractifs, précises, et respectant les principes de l'anatomie et de la mécanique.Notre seconde contribution s'intéresse aux problématiques liées à la complexité des connaissances anatomiques, physiologiques et fonctionnelles. En se basant sur une ontologie de l'anatomie et une ontologie inédite de la physiologie humaine, nous introduisonsun pipeline pour la construction automatique de modèles simulant les fonctions de nos organes.Celles-ci permettent d'exploiter les connaissances anatomiques complexes via des requêtes simples.Les sorties de ces requêtes sont utilisées pour créer des modèles simulables retranscrivant les aspects fonctionnels tels qu'ils ont été formalisés et décrits par les anatomistes.Enfin, notre troisième contribution : le transfert d'anatomie, permet d'adapter les modèles géométriques et mécaniques à la morphologie de patients spécifiques.Cette nouvelle méthode de recalage permet de reconstruire automatiquement l'anatomie interne d'un personnage défini par sa peau en transférant les organes d'un personnage de référence.Elle permet de pallier à la nécessité de re-construire ces géometries pour chaque nouvelle simulation, et contribue ainsi à accélérer la mise en place de simulations spécifiques à une grande variété d'individus de morphologie différente. / The use of virtual humans has spread in various activities in recent years.Beyond virtual surgery, virtual bodies are increasingly used to design medical devices, vehicles, and daily life hardware more generally.They also turn out to be extraordinary supports to learn anatomy.Recent movies (Avatar, Lord of the Rings, etc) demonstrated that anatomy and biomechanics can be used to design high-quality characters.However, reproducing the behavior of anatomical structures remains a complex task, and a great amount and variety of knowledge is necessary for setting up high quality simulations.This makes the modeling of human body for simulation purposes an open problem, a tedious task, but also a fascinating research subject.Through this PhD, we address the problem of the construction of biomechanical models of the musculoskeletal systems for several domains : animation, biomechanics and teaching.Our goal is to simplify the entire process of model design by making it more intuitive and faster.Our approach is to address each difficulty : the representation and use of anatomical knowledge, the geometrical modeling and the efficient simulation of the musculoskeletal system thanks to three novel contributions introduced during these research works.Our first contribution focuses on the biomechanical construction of a hybrid model of lumbar spine.In this work, we show that hybrid approaches that combine both rigid body systems and finite element models allow interactive simulations, accurate, while respecting the principles of anatomy and mechanics.Our second contribution addresses the problem of the complexity of anatomical, physiological and functional knowledge.Based on a novel ontology of anatomical functions of the human body, we introduce a novel pipeline to automatically build models that simulate physiological functions of our bodies.The ontology allows us to extract detailed knowledge using simple queries.The outputs of these queries are used to set up simulation models of the functional aspects as they were formalized and described by anatomists.Finally our third contribution, the anatomy transfer, allows the mapping of available geometrical and mechanical models to the morphology of any specific individual.This novel registration method enables the automatic construction of the internal anatomy of any character defined by his skin, by transferring organs from a reference character.It allows to overcome the need to re-construct these geometries for each new simulation, and it contributes to accelerate the simulations setup for a range of people with different morph

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第22687号 / 医博第4631号 / 新制||医||1045(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 山下 潤, 教授 木村 剛, 教授 浅野 雅秀 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Considering the late in-vessel phase of an unlikely core melt down scenario in a light water reactor (LWR) with the formation of a corium pool in the lower head of the re-actor pressure vessel (RPV) the possible failure modes of the RPV and the time to failure have to be investigated to assess the possible loadings on the containment. In this work, an integral model was developed to describe the processes in the lower plenum of the RPV. Two principal model domains have to be distinguished: The temperature field within the melt and RPV is calculated with a thermodynamic model, while a mechanical model is used for the structural analysis of the vessel wall. In the introducing chapters a description is given of the considered accident scenario and the relevant analytical, experimental, and numerical investigations are discussed which were performed worldwide during the last three decades. Following, the occur-ring physical phenomena are analysed and the scaling differences are evaluated between the FOREVER-experiments and a prototypical scenario. The thermodynamic and the mechanical model can be coupled recursively to take into account the mutual influence. This approach not only allows to consider the tem-perature dependence of the material parameters and the thermally induced stress in the mechanical model, it also takes into account the response of the temperature field itself upon the changing vessel geometry. New approaches are applied in this work for the simulation of creep and damage. Using a creep data base, the application of single creep laws could be avoided which is especially advantageous if large temperature, stress, and strain ranges have to be covered. Based on experimental investigations, the creep data base has been de-veloped for an RPV-steel and has been validated against creep tests with different scalings and geometries. It can be stated, that the coupled model is able to exactly describe and predict the vessel deformation in the scaled integral FOREVER-tests. There are uncertainties concerning the time to failure which are related to inexactly known material parame-ters and boundary conditions. The main results of this work can be summarised as follows: Due to the thermody-namic behaviour of the large melt pool with internal heat sources, the upper third of the lower RPV head is exposed to the highest thermo-mechanical loads. This region is called hot focus. Contrary to that, the pole part of the lower head has a higher strength and therefore relocates almost vertically downwards under the combined thermal, weight and internal pressure load of the RPV. On the one hand, it will be possible by external flooding to retain the corium within the RPV even at increased pressures and even in reactors with high power (as e.g. KONVOI). On the other hand, there is no chance for melt retention in the considered scenario if neither internal nor external flooding of the RPV can be achieved. Two patents have been derived from the gained insights. Both are related to pas-sively working devices for accident mitigation: The first one is a support of the RPV lower head pole part. It reduces the maximum mechanical load in the highly stressed area of the hot focus. In this way, it can prevent failure or at least extend the time to failure of the vessel. The second device implements a passive accident mitigation measure by making use of the downward movement of the lower head. Through this, a valve or a flap can be opened to flood the reactor pit with water from a storage res-ervoir located at a higher position in the reactor building. With regard to future plant designs it can be stated - differing from former presump-tions - that an In-Vessel-Retention (IVR) of a molten core is possible within the reac-tor pressure vessel even for reactors with higher power.