Les méthodes d’évaluation environnementale sont de plus en plus utilisées pour mieux apprécier les impacts environnementaux générés par les activités humaines (produits, services, systèmes). L’analyse de cycle de vie (ACV) est sans doute la méthode d’évaluation la plus répandue. Elle est aussi souvent considérée comme la plus avancée, bien qu’elle comporte de sérieuses limites (elle n’intègre pas de réels indicateurs économiques ou sociaux, par exemple). Dans cette thèse de doctorat, j’examine plus particulièrement la problématique de l’intégration du temps dans les modèles d’inventaire et les calculs d’impact tout au long du cycle de vie. J’y présente une évolution de la méthodologie d’ACV traditionnelle pour le cas du réchauffement climatique, qui incorpore des aspects dynamiques et cumulatifs exprimés directement en équivalent-CO2. Cette perspective orientée vers une meilleure prise en compte dans les pratiques de reporting et/ou les politiques publiques est ensuite déclinée sur trois cas d’application de complexité croissante pour l’analyse. L’hypothèse centrale de ce travail est que le passage d’une ACV traditionnelle à une ACV dite « dynamique » permettrait d’obtenir des résultats d’évaluation d’impacts plus proches de la réalité des phénomènes environnementaux / Environmental impact assessment methods are now widely used in order to measure environmental impacts associated with human activities (for products, services, and systems). Life-cycle assessment (LCA) is without doubt the foremost assessment method. LCA is also often thought of as the more advanced one, despite serious limitations (e.g. LCA does not include properly economical or social dimensions). In this PhD report, I explore more specifically the issue of integrating time in both inventory models and impact assessments along the life-cycle. In the case of climate change, I offer an evolution of traditional LCA towards a framework that includes dynamic and cumulative aspects as expressed directly in CO2-equivalent. This approach, which is oriented towards reporting practices and/or public policies, is afterwards applied to three different case studies of growing complexity. The central hypothesis of this work is that switching from traditional to so-called “dynamic” LCA would allow for better results with regards to one reality of environmental processes

Contribution à l’étude du phénomène d’oscillation argumentaire. L’oscillateur argumentaire a un mouvement stable périodique, à une fréquence proche de sa fréquence fondamentale, lorsqu’il est soumis à une excitation provenant d’une source de type harmonique, à une fréquence qui est un multiple de ladite fréquence fondamentale, et agissant de manière telle que son interaction avec le système dépende des coordonnées d’espace du système. La présente thèse étudie quelques systèmes argumentaires et essaie de mettre en évidence des relations symboliques entre les paramètres de ces systèmes et leur comportement observé ou calculé. C’est la représentation de Van der Pol qui a été utilisée la plupart du temps pour représenter l’état du système, car elle est bien adaptée à la méthode de centrage, où l'on cherche une solution sous forme d’un signal de type sinusoïdal, d’amplitude et de phase lentement variables. L’originalité de la présente thèse vis-à-vis des publications antérieures est dans la modélisation, plus proche des systèmes physiques réels, dans les développements symboliques qui donnent des représentations inédites, dans le mode de réalisation des expériences, qui utilisent toutes une visualisation de Van der Pol en temps réel, et dans l’objet de l’expérience de la poutre excitée axialement de manière argumentaire. Au cours de cette thèse, des systèmes simples à un DDL ont été modélisés, construits et expérimentés. Des relations symboliques, notamment concernant les probabilités de capture par des attracteurs, ainsi que des critères de stabilité et une solution symbolique approchée, ont été mis en évidence. Un système continu constitué d’une poutre élancée excitée axialement a ensuite été modélisé à l’aide de deux modèles et expérimenté ; toujours dans le domaine symbolique, des propriétés ont été étudiées, notamment concernant des combinaisons de plages de paramètres permettant au phénomène argumentaire d’exister / Contribution to the study of the argumental oscillation phenomenon. The argumental oscillator has a stable periodic motion at a frequency close to its fundamental frequency when it is subjected to an excitation from a harmonic source at a frequency which is a multiple of said fundamental frequency, and acting in such a way that its interaction with the system depends on the space coordinates of the system. This thesis studies some argumental systems and tries to demonstrate symbolic relations between the parameters of these systems and their observed or calculated behavior. The Van der Pol representation was used most of the time to represent the state of the system, as it is well adapted to the averaging method, where a solution is sought as a signal of sinusoidal type, with slowly varying amplitude and phase. The originality of this thesis with respect to previous publications is in the modeling, closer to real physical systems, in the symbolic developments that give new representations, in the embodiment of the experiments, which all use a real-time Van der Pol visualization, and in the object of the experiment of the beam axially excited in an argumental way. During this thesis, simple systems with one DDL have been modeled, built and tested. Symbolic relationships have been highlighted, in particular with regard to the probabilities of capture by attractors, as well as stability criteria and an approximate symbolic solution. A continuous system consisting of an axially excited slender beam was then modeled using two models, and tested; still in the symbolic domain, properties have been studied, especially concerning combinations of parameter ranges allowing the argumental phenomenon to occur

No description available.

<p> The object of this work was to evaluate the Supernova ninicomputer as a process simulation tool. To accomplish this obje c tive the GEMCS ste2dy-state simulation system wa.s programmed for the Supernova. This system was programmed to run in five different modes, BASIC , two modes which require that al.L progra1ns are written in Assembly Language, and two modes in which programs may be called from the disk. Of these latter two modes, one requires all programs to be written in Assembly Language; the other mode allows programs to be written in Assembly Language or FORTRAN IV. </p> <p> Two case studies which had previously been run on the CDC 6400 were run on the Supernova to test and evaluate f our of the modes in which GEMCS was proerammed for the Supernova. </p> <p> The studies have also provided an a ssessment of the hardware and softws.re facilities of the Supernova and the results of thjs assessment were documented. In addition the opera tion of the various strategies used to impl ement GEMCS have provided an indication of the feasibility of the impl ementation of the dynamic system simulation program, DYNSYS. </p> / Thesis / Doctor of Philosophy (PhD)

There is an increasing interest in automated mobile equipment in the construction, agriculture and mining industries to improve productivity, efficiency and operator safety. In general, these machines belong to a class of mobile vehicles with a tool for manipulating its environment to accomplish a repetitive task. Forces and motions are inherently coupled between the tool (<em>e.g.</em> bucket or blade) and the means of vehicle propulsion (<em>e.g</em>. wheels or tracks). Furthermore, they are often operated within uncertain and unstructured environments. A particularly challenging case involves the use of a bulldozer for the removal of excavated material. Modeling and control of mobile robots that interact forcibly with their environment, such as robotic excavation machinery, is a challenging problem that has not been adequately addressed in prior research. This thesis investigates the low-level modeling and control of a 3-DOF robotic bulldozing operation. Motivated by a bulldozing process in an underground mining application, a theoretical nonlinear hybrid dynamic model was developed. The model includes discrete operation modes contained within a hybrid dynamic model framework. The dynamics of the individual modes are represented by a set of linear and nonlinear differential equations. An instrumented scaled-down bulldozer and environment were developed to emulate the full scale operation. Model parameter estimation and validation was completed using experimental data from this system. The model was refined based on a global sensitivity analysis. The refined model was found to be suitable for simulation and design of robotic bulldozing control laws. Optimal blade position control laws were designed based on the hybrid dynamic model to maximize the predicted material removal rate of the bulldozing process. A stability analysis of the underlying deterministic closed-loop process dynamics was performed using Lyapunov’s second method. Monte Carlo simulation was used for further performance and stability analysis of the closed-loop process dynamics including stochastic state disturbances and input constraints. Results of the Monte Carlo simulation were also used for tuning the blade position control laws. Experiments were conducted with the scaled-down robotic bulldozing system. The control laws were implemented with various tuning values. As a comparison, a rule-based blade control algorithm was also designed and implemented. The experimental results with the optimal control laws demonstrated a 33% increase in the average material removal rate compared to the rule-based controller. / Doctor of Philosophy (PhD)

This thesis essentially deals with the development and numerical explorations of a few improved Monte Carlo filters for nonlinear dynamical systems with a view to estimating the associated states and parameters (i.e. the hidden states appearing in the system or process model) based on the available noisy partial observations. The hidden states are characterized, subject to modelling errors, by the weak solutions of the process model, which is typically in the form of a system of stochastic ordinary differential equations (SDEs). The unknown system parameters, when included as pseudo-states within the process model, are made to evolve as Wiener processes. The observations may also be modelled by a set of measurement SDEs or, when collected at discrete time instants, their temporally discretized maps. The proposed Monte Carlo filters aim at achieving robustness (i.e. insensitivity to variations in the noise parameters) and higher accuracy in the estimates whilst retaining the important feature of applicability to large dimensional nonlinear filtering problems. The thesis begins with a brief review of the literature in Chapter 1. The first development, reported in Chapter 2, is that of a nearly exact, semi-analytical, weak and explicit linearization scheme called Girsanov Corrected Linearization Method (GCLM) for nonlinear mechanical oscillators under additive stochastic excitations. At the heart of the linearization is a temporally localized rejection sampling strategy that, combined with a resampling scheme, enables selecting from and appropriately modifying an ensemble of locally linearized trajectories whilst weakly applying the Girsanov correction (the Radon- Nikodym derivative) for the linearization errors. Through their numeric implementations for a few workhorse nonlinear oscillators, the proposed variants of the scheme are shown to exhibit significantly higher numerical accuracy over a much larger range of the time step size than is possible with the local drift-linearization schemes on their own. The above scheme for linearization correction is exploited and extended in Chapter 3, wherein novel variations within a particle filtering algorithm are proposed to weakly correct for the linearization or integration errors that occur while numerically propagating the process dynamics. Specifically, the correction for linearization, provided by the likelihood or the Radon-Nikodym derivative, is incorporated in two steps. Once the likelihood, an exponential martingale, is split into a product of two factors, correction owing to the first factor is implemented via rejection sampling in the first step. The second factor, being directly computable, is accounted for via two schemes, one employing resampling and the other, a gain-weighted innovation term added to the drift field of the process SDE thereby overcoming excessive sample dispersion by resampling. The proposed strategies, employed as add-ons to existing particle filters, the bootstrap and auxiliary SIR filters in this work, are found to non-trivially improve the convergence and accuracy of the estimates and also yield reduced mean square errors of such estimates visà-vis those obtained through the parent filtering schemes. In Chapter 4, we explore the possibility of unscented transformation on Gaussian random variables, as employed within a scaled Gaussian sum stochastic filter, as a means of applying the nonlinear stochastic filtering theory to higher dimensional system identification problems. As an additional strategy to reconcile the evolving process dynamics with the observation history, the proposed filtering scheme also modifies the process model via the incorporation of gain-weighted innovation terms. The reported numerical work on the identification of dynamic models of dimension up to 100 is indicative of the potential of the proposed filter in realizing the stated aim of successfully treating relatively larger dimensional filtering problems. We propose in Chapter 5 an iterated gain-based particle filter that is consistent with the form of the nonlinear filtering (Kushner-Stratonovich) equation in our attempt to treat larger dimensional filtering problems with enhanced estimation accuracy. A crucial aspect of the proposed filtering set-up is that it retains the simplicity of implementation of the ensemble Kalman filter (EnKF). The numerical results obtained via EnKF-like simulations with or without a reduced-rank unscented transformation also indicate substantively improved filter convergence. The final contribution, reported in Chapter 6, is an iterative, gain-based filter bank incorporating an artificial diffusion parameter and may be viewed as an extension of the iterative filter in Chapter 5. While the filter bank helps in exploring the phase space of the state variables better, the iterative strategy based on the artificial diffusion parameter, which is lowered to zero over successive iterations, helps improve the mixing property of the associated iterative update kernels and these are aspects that gather importance for highly nonlinear filtering problems, including those involving significant initial mismatch of the process states and the measured ones. Numerical evidence of remarkably enhanced filter performance is exemplified by target tracking and structural health assessment applications. The thesis is finally wound up in Chapter 7 by summarizing these developments and briefly outlining the future research directions

Estudamos o conceito de energia na teoria física a nível histórico e a nível de ensino-aprendizagem, isto é, do ponto de vista de sua criação, representação conceitual, matemática e experimental, na constituição de seu significado na teoria científica e na linguagem comum. Como um exercício preliminar, analisamos alguns dados sobre o ensino de conceitos de física na exposição científica do IFUSP \"Ondas, Campos e Partículas\", de 1989 a 1990, e, em 1991, com o nome de \"Do Caleidoscópio ao Carrossel\", no prédio da Engenharia Mecânica (Escola Politécnica). A exposição apresenta fenômenos das teorias da física nos campos da mecânica, eletricidade e magnetismo, termodinâmica e ótica, fenômenos que implicam ações em escalas macroscópicas e a nível atômico. Descrevemos uma entrevista de 13 minutos com quatro alunos de 1o grau que visitaram a exposição com monitoramento. Durante a situação da entrevista, que chamamos de pedagógica, esses alunos constróem o conceito de energia. Essa entrevista é analisada em termos dos contextos fenomenológico e epistemológico, contidos na exposição, ainda que os monitores que acompanharam a visita não falassem em energia. Descrevemos também uma entrevista com um monitor sobre ensino e aprendizagem. Enfocamos o conceito de \"conhecimento prévio do aluno como requisito para ensino\", e analisamos usando o conceito de zona de desenvolvimento proximal (ZDP) de Vygotsky. Usamos a moldura de Vygotsky (VY 1896/1934) na análise da entrevista com os alunos, que é um exemplo de como considerar a ZDP. Ela não visa avaliação, e nem apenas coleta de dados. Propomos o desenvolvimento de entrevistas pedagógicas para já se constituir uma situação de ensino-aprendizagem, como uma interação desejável, dialógica, onde a ZDP, isto é, a capacidade do aluno \"vir a saber\", seja desenvolvida. A análise da emergência do conceito de energia no caso estudado da entrevista com alunos é feita com a perspectiva de sistemas dinâmicos, dada a natureza complexa dos fenômenos estudados. A História e Filosofia das Teorias Científicas são usadas como ferramenta heurística para identificar correlação na constituição do espaço de significação. Concluíremos que, se os monitores não enfocavam diretamente o conceito de energia e os alunos aprendiam mesmo assim, é porque a exposição é um bom contexto onde a entrevista pedagógica pode ser utilizada como instrumento de ensino, Por fim, falamos da necessidade de um treinamento para monitores, onde sua formação inclua estudos históricos e epistemológicos e uma preparação para suprir a deficiência que detetamos na exposição, isto é, a de não abordar os conceitos na sua forma matematizada. A pesquisa das condições para aprendizagem na Exposição Científica e em Entrevista se desenvolve com instrumentos próprios: um sistema de análise em vídeo foi construído para pesquisar a linguagem usada na interação entre professor e aluno. Objetivamos estabelecer condições propícias para interações dialógicas. Como resultado secundário, foi adquirida certa familiaridade em trabalhar com o vídeo como instrumento para, ao mesmo tempo, mostrar e desenvolver aquisição do conhecimento científico de conceitos de física. / We have studied the concept of energy in Physics in the historical and the Teaching-Learning contexts, i.e., from the point of view of its creation, its mathematical and experimental conceptual representation, of its scientific and ordinary meaning constitution. As a preliminary exercise, we analysed some data about the Physics concepts teaching at the \"Waves, Fields and Particles\" Science Exhibition Center located at that time, in the Engineering Building of the University of São Paulo. The content of the Exhibition was regular demonstrations of phenomena delineated by classical theories of Mechanics, Eletromagnetism, Thermodynamics and some topics in Modern Physics. An interpretation of the exhibition could be in terms of the concept of energy and its transformations, as was shown by an interview with four visiting students described here. During this interview set, which we named pedagogical, these students constructed the concept of energy. This interview is analysed in terms of phenomenological and epistemological implications in the exhibition. A monitor thinking about teaching and learning has also been situated out of an interview. This monitor considers the pupil\'s previous subject matter knowledge as a requirement for teaching. By means of the ZDP (zone of proximal development), we are warned by Vygotsky against the extreme pre-requisites, which can limit the learning. We also make use of Vygotsky\'s concept in the analysis of the interview with students. The interview described in the first part of this work is a ZDP application example. It neither aims evaluation nor just data gathering. It will be developed to be a teaching-learning instrument. We believe this is possible, for it will be an interaction where the ZDP or the pupil\'s ability to \"come to know\" is to be considered. The case study of the emergence of the concept of energy from the interview with pupils is analysed within the framework of dynamic systems given the complex nature of the considered phenomena. The interview was set out in the field of the interactions where information are exchanged between the interviewer and the group and between the peer to peer interaction. The meaning of the concept of energy in Physics was related to the context of the exhibition where the concept was implicitly present. We came to the conclusion that this science exhibition, where the pedagogical interview can be utilized as a tool for teaching, has a good context for learning, once the students showed to have learned even though the docents hadn\'t focused the concept of energy directly. Then we commented about the need of a monitor\'s training, where historical and epistemological studies should be included. This training should also account for the exhibition\'s detected deficiency in approaching the scientific concepts in their mathematical formulation. The research for the emergence of learning in the scientific exhibition center and in interviews is developed by proper tools: a video analysis system is devised to search for the most efficient language in the interaction between teacher and pupil, or in any other teaching-learning situations. Our aim is to establish propitious settings for dialogical interactions in the attempt of dealing with the Vygotsky\'s cognitive \"proximal zone of development\", which is important when we consider the \"process\" of learning a complex concept. As a parallel outcome, we got some familiarity with video recording and analysis as an instrument for both developing and showing the student\'s acquisition of scientific knowledge of physical concepts.

A utilização de Redes Neurais Artificias para fins de identificação e controle de sistemas dinâmicos têm recebido atenção especial de muitos pesquisadores, principalmente no que se refere a sistemas não lineares. Neste trabalho é apresentado um estudo sobre a utilização de um tipo em particular de Rede Neural Artificial, uma Perceptron Multicamadas com Atraso de Tempo, na estimação de estados da etapa fermentativa do processo de Reichstein para produção de vitamina C. A aplicação de Redes Neurais Artificiais a este processo pode ser justificada pela existência de problemas associados à esta etapa, como variáveis de estado não mensuráveis e com incertezas de medida e não linearidade do processo fermentativo, além da dificuldade em se obter um modelo convencional que contemple todas as fases do processo. É estudado também a eficácia do algoritmo de Levenberg-Marquadt, na aceleração do treinamento da Rede Neural Artificial, além de uma comparação do desempenho de estimação de estados das Redes Neurais Artificiais estudadas com o filtro estendido de Kalman, baseado em um modelo não estruturado do processo fermentativo. A análise do desempenho das Redes Neurais Artificiais estudadas é avaliada em termos de uma figura de mérito baseada no erro médio quadrático sendo feitas considerações quanto ao tipo da função de ativação e o número de unidades da camada oculta. Os dados utilizados para treinamento e avaliação da Redes Neurais Artificiais foram obtidos de um conjunto de ensaios interpolados para o intervalo de amostragem desejado. / ldentification and Control of dynamic systems using Artificial Neural Networks has been widely investigated by many researchers in the last few years, with special attention to the application of these in nonlinear systems. ls this works, a study on the utilization of a particular type of Artificial Neural Networks, a Time Delay Multi Layer Perceptron, in the state estimation of the fermentative phase of the Reichstein process of the C vitamin production. The use of Artificial Neural Networks can be justified by the presence of problems, such as uncertain and unmeasurable state variables and process non-linearity, and by the fact that a conventional model that works on all phases of the fermentative processes is very difficult to obtain. The efficiency of the Levenberg Marquadt algorithm on the acceleration of the training process is also studied. Also, a comparison is performed between the studied Artificial Neural Networks and an extended Kalman filter based on a non-structured model for this fermentative process. The analysis of lhe Artificial Neural Networks is carried out using lhe mean square errors taking into consideration lhe activation function and the number of units presents in the hidden layer. A set of batch experimental runs, interpolated to the desired time interval, is used for training and validating the Artificial Neural Networks.

[pt] Esta tese apresenta o estudo dinâmico de três sistemas pendulares e de um sistema de corpo livre no espaço com 3 graus de liberdade. O primeiro sistema pendular consiste de um pêndulo acoplado ao centro de uma roda, que rola sem escorregar na direção horizontal, enquanto o segundo, se baseia num pêndulo simples, porém com comprimento variável, que ao mudar seu tamanho consegue ganhar/perder energia para aumentar/diminuir a amplitude de seu movimento e finalmente o terceiro está baseado num pêndulo duplo que, a despeito de restrições impostas ao movimento consegue inserir/retirar energia do sistema de forma similar ao segundo. O modelo de corpo livre no espaço está baseado na suspensão cardânica de um giroscópio e se utiliza de um modelo didático real de um giroscópio para observação das características dinâmicas. A partir destes exemplos estudou-se formas de controle não-linear para movimentar os sistemas de maneira a utilizar-se da mudança de posição interna do centro de massa para injetar e retirar energia dos sistemas. Foram gerados modelos matemáticos simulados no Simulink valendo-se do Matlab para análise, e geradas animações também com o Matlab para melhor observação dos efeitos. Em paralelo, para dois destes sistemas foram construídos em bancada experimentos para comprovação dos resultados numéricos, e os resultados são comparados em cada caso, analisando as diferenças. Ao final, todas as observações sobre os estudos foram analisadas, e comentários feitos baseados nos resultados, além de sugerir trabalhos futuros. / [en] This thesis presents the study of the dynamics of three pendulum systems and a 3DoF free body in space. The first pendular system is based on a simple pendulum coupled to the center of a wheel that translates horizontally without slip; while the second system is based on a simple pendulum, with variable length, which is able to acquire/lose energy to grow/diminish the amplitude of its movement; and finally the third one is based on a double pendulum that, in spite of movement restrictions, can as well inject/drop energy like the second system. The free body in space is based on a real gyroscope for didactical use, which is helpful for the observation of the dynamic characteristics of the motion. Using these examples a non-linear control was designed to drive the system by using the property that changing the internal position of the center of mass it is possible to inject or to subtract energy from the systems. Mathematical models were simulated with Simulik software, Matlab was used for the analysis, and animations were created also with Matlab for a better sight of the effects. In parallel, there were developed 2 test rig systems for verification of the numerical results. In the conclusions all the considerations about the study were analyzed, and comments made on the results, as well also future developments are suggested.

Data Handover est une librairie de fonctions adaptée aux systèmes distribués à grande échelle. Dho offre des routines qui permettent d'acquérir des ressources en lecture ou en écriture de façon cohérente et transparente pour l'utilisateur. Nous avons modélisé le cycle de vie de Dho par un automate d'état fini puis, constaté expérimentalement, que notre approche produit un recouvrement entre le calcul de l'application et le contrôle de la donnée. Les expériences ont été menées en mode simulé en utilisant la libraire GRAS de SimGrid puis, en exploitant un environnement réel sur la plate-forme Grid'5000. Par la théorie des files d'attente, la stabilité du modèle a été démontrée dans un contexte centralisé. L'algorithme distribué d'exclusion mutuelle de Naimi et Tréhel a été enrichi pour offrir les fonctionnalités suivantes: (1) Permettre la connexion et la déconnexion des processus (ADEMLE), (2) admettre les locks partagés (AEMLEP) et enfin (3) associer les deux propriétés dans un algorithme récapitulatif (ADEMLEP). Les propriétés de sûreté et de vivacité ont été démontrées théoriquement. Le système peer-to-peer proposé combine nos algorithmes étendus et le modèle originel Dho. Les gestionnaires de verrou et de ressource opèrent et interagissent mutuellement dans une architecture à trois niveaux. Suite à l'étude expérimentale du système sous-jacent menée sur Grid'5000, et des résultats obtenus, nous avons démontré la performance et la stabilité du modèle Dho face à une multitude de paramètres / Data Handover is a library of functions adapted to large-scale distributed systems. It provides routines that allow acquiring resources in reading or writing in the ways that are coherent and transparent for users. We modelled the life cycle of Dho by a finite state automaton and through experiments; we have found that our approach produced an overlap between the calculation of the application and the control of the data. These experiments were conducted both in simulated mode and in real environment (Grid'5000). We exploited the GRAS library of the SimGrid toolkit. Several clients try to access the resource concurrently according the client-server paradigm. By the theory of queues, the stability of the model was demonstrated in a centralized environment. We improved, the distributed algorithm for mutual exclusion (of Naimi and Trehel), by introducing following features: (1) Allowing the mobility of processes (ADEMLE), (2) introducing shared locks (AEMLEP) and finally (3) merging both properties cited above into an algorithm summarising (ADEMLEP). We proved the properties, safety and liveliness, theoretically for all extended algorithms. The proposed peer-to-peer system combines our extended algorithms and original Data Handover model. Lock and resource managers operate and interact each other in an architecture based on three levels. Following the experimental study of the underlying system on Grid'5000, and the results obtained, we have proved the performance and stability of the model Dho over a multitude of parameters