Lakštinėse spausdinimo mašinose spausdinant spalvotus atvaizdus svarbu mašinos atskirose sekcijose atspaudų sutapimą išlaikyti šimtųjų milimetro dalių tikslumu. Be to, spausdinamų atvaizdų geometrinis nesutapimas abiejose lapo pusėse neturi viršyti 0,1 mm. Viena iš pagrindinių nesutapimo priežaščių yra mechaninio krumpliaratinio perdavimo mechanizmo iš vienos sekcijos į kitą sukamieji virpesiai, kuriuos sukelia krumpliaračių paklaidos ir pasipriešinimo momentų pokyčiai tarp ofsetinių ir forminių cilindrų. Darbe analiziniu – kompiuteriniu būdu nagrinėjama šių virpesių įtaka lapo perdavimo tikslumui. Konkrečiam tyrimui pasirinkta „Heidelberg“ spausdinimo mašina GTO 52, tačiau darbe taikoma tyrimo metodika tinka ir kitoms mašinoms. Sudarytas elektros pavara valdomo lapų perdavimo mechanizmo dinaminis modelis su 16 laisvės laipsnių. Sistemos matematinis modelis (virpesių lygtys ir struktūrinė schema) sudarytas kompiuterizuotu būdu MATLAB/SIMULINK programų terpėje, kurioje atliktas sistemoje vykstančių dinaminių procesų modeliavimas. Pavaroje kylančios paklaidos apibūdinamos sukamųjų virpesių amplitudinėmis – dažninėmis charakteristikomis. Nagrinėjama šių charakteristikų pobūdžio, rezonansinių virpesių ir jų dažnių įtaka lapo perdavimo paklaidos dydžiui. / It is important in printing machines to maintain very high print quality (hundredths of a millimeter) in printing colorful prints. Also mismatches of prints in both sides of sheet must not exceed 0,1 millimeter. One of the reasons of mismatching is mechanical gear transfers mechanisms, from one printing unit to another, rotational oscillations, which are generated by gear errors and resistance moment changes between offset and print cylinder. These oscillations influence to sheet transfer accuracy are examined in analytical – computer way. For particular examination there was selected “Heidelberg” printing press GTO 52, but the examination method can be used for other presses. There was formed dynamic model with 16 degrees of freedom. Mathematical model of the system (oscillation equations and flowchart) was formed with MATLAB/SIMULINK program, in which dynamical process modeling was done. Errors in drive described as an amplitude frequency response. There was examined these response nature, resonant oscillations and its frequency influence to sheets transfer errors size.

The thermal control of a die is crucial for the development of high efficiency injection moulds. For successful thermal management, this research provides an effective control strategy to find sensor locations, identify thermal dynamic models, and design controllers. By applying a clustering method and sensitivity analysis, sensor locations are identified. The neural network and finite element analysis techniques enable the modeling to deal with various cycle-times for the moulding process and uncertain dynamics of a die. A combination of off-line training through finite element analysis and training using on-line learning algorithms and experimental data is used for the system identification. Based on the system identification which is experimentally validated using a real system, controllers are designed using fuzzy-logic and self-adaptive PID methods with backpropagation (BP) and radial basis function (RBF) neural networks to tune control parameters. Direct adaptive inverse control and additive feedforward control by adding direct adaptive inverse control to self-adaptive PID controllers are also provided. Through a comparative study, each controller’s performance is verified in terms of response time and tracking accuracy under different moulding processes with multiple cycle-times. Additionally, the improved cooling effectiveness of the conformal cooling channel designed in this study is presented by comparing with a conventional straight channel.

Orientador: Maria Isabel Castreghini de Freitas / Banca: Silvia Selingardi Sampaio / Banca: Cláudia Maria de Almeida / Resumo: Esta pesquisa teve como objetivo analisar o processo de expansão urbano-industrial do município de Americana-SP, com o subsídio de geotecnologias. A ocupação urbana e industrial foi identificada e analisada ao longo de uma série temporal de aproximadamente 60 anos (1940 a 2005), apoiando-se na elaboração e análise do uso da terra do município. Essas etapas permitiram alcançar um diagnóstico da área de estudo atrelada a dados socioeconômicos e a identificação dos vetores de crescimento urbano-industrial para cada década da série temporal. A partir disso, foi desenvolvida a etapa de modelagem dinâmica espacial, utilizando-se um modelo de autômatos celulares para simular os cenários de uso da terra e obter prognósticos para a dinâmica urbana de Americana. Observou-se que a expansão urbano-industrial bem como a formação do contingente populacional do município acompanhou o processo de desenvolvimento do país, sendo o reflexo das políticas governamentais e do desempenho econômico e social que caracterizou cada período. Verificou-se também que o desenvolvimento industrial teve importante papel na ocupação urbana e no crescimento populacional, uma vez que impulsionou os fluxos migratórios que afluíram para o município. A ocupação da quase totalidade do território pela mancha urbana representa um entrave ao crescimento futuro da cidade. Entretanto, mantidas as características de uso e ocupação atuais, a dinâmica territorial deverá ser menor no curto e médio prazo, de acordo com a atual tendência de redução no ritmo de crescimento da população e, por conseqüência da área urbana, semelhante ao observado na maioria dos centros urbanos. Em suma, foram apontadas possíveis aplicações deste trabalho, além de sugestões, sobretudo relacionadas à etapa de modelagem, a qual constitui importante subsídio para o planejamento urbano. / Abstract: The aim of this research is to analyze the urban-industrial expansion process of Americana-SP town by means of geotechnologies. The urban and industrial occupation were identified and analyzed along a temporal series of approximately 60 years (1940 to 2005), based on the elaboration and analysis of land use. These stages allowed to achieve a diagnosis of studied area linked to socioeconomics information and the identification of vectors of urban-industrial growth for each decade of the temporal series. Next, the spatial dynamic modeling using a cellular automata model was developed to simulate land use scenarios and to obtain prognosis for the urban dynamics of Americana. It was observed that the urban-industrial expansion as well as the formation of population contingent followed the process of national development, reflecting governmental politicies and the economic and social performance that characterized each period. Furthermore, the industrial development had an important role on urban occupation and on population growth, since that impelled the rural-urban flows. The spread of urban areas in Americana municipality represents a hindrance to the future city growth. However, if the current land use and occupation trends are maintained the spatial dynamic would be small for the short and medium terms, according to actual tendency of decrease on population growth rates and consequently of urban areas, similar to that observed on the majority of urban centers worldwide. Finally, possible applications this work are pointed out, besides suggestions related to the advances in spatial modeling, since it constitutes an important tool for urban planning. / Mestre

Orientador: Rubens Maciel Filho / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Quimica / Made available in DSpace on 2018-08-09T11:12:44Z (GMT). No. of bitstreams: 1 Lima_NadsonMuriloNascimento_M.pdf: 1887705 bytes, checksum: e0761ce0e30dd05a47c9cfbe9b0968ef (MD5) Previous issue date: 2006 / Resumo: A síntese de controladores representa uma importante vertente dos desenvolvimentos atuais no campo da pesquisa acadêmica e industrial. Um controlador bem projetado pode significar sucesso no que se refere aos objetivos de produção, sendo gerados materiais com as especificações desejadas e proporcionando que o sistema opere sob certas restrições, levando em consideração aspectos relativos à operabilidade, segurança e minimização de resíduos. Para tanto, sabe-se que as etapas de modelagem são fundamentais para a delineação de estratégias de controle. No entanto, a obtenção de representações matemáticas precisas e, ao mesmo tempo, aplicáveis para controle da maioria dos processos de interesse da engenharia química é uma tarefa árdua, devido à presença de comportamentos dinâmicos não lineares e variantes ao longo do espaço e do tempo. Deste modo, busca-se a obtenção de modelos mais simples, porém dotados da imprescindível representatividade inerente aos sistemas de produção, a fim de serem projetadas estruturas de controle adequadas para cada necessidade específica. Este trabalho enfoca o desenvolvimento de um controlador híbrido preditivo baseado em modelos nebulosos (fuzzy) tipo Takagi-Sugeno para processos de polimerização, os quais apresentam dinâmicas altamente complexas e de difícil modelagem matemática, dificultando assim a aplicação de metodologias convencionais de controle. Foram considerados dois casos de estudo para análise de desempenho do controlador proposto: o processo de copolimerização em solução do metacrilato de metila e acetato de vinila, e a copolimerização industrial do eteno/1-buteno com catalisador Ziegler-Natta solúvel. Os modelos fenomenológicos de ambos os processos já se encontram descritos na literatura, sendo considerados como plantas virtuais para geração de dados dinâmicos e implementação do controlador. A partir de simulações computacionais, os modelos dinâmicos nebulosos funcionais foram construídos ¿ os quais demonstraram excelentes capacidades para predição das saídas dos processos como uma função dos dados dinâmicos de entrada ¿ sendo, posteriormente, inseridos na estrutura interna do controle preditivo DMC (Dynamic Matrix Control). A escolha do controlador DMC como base para o desenvolvimento da estrutura proposta deve-se ao fato de sua notória aplicabilidade industrial, aliada à simplicidade de projeto e execução, além de possibilitar a incorporação de restrições nas variáveis controladas e manipuladas. Por fim, foram comparados os desempenhos entre os controladores híbrido e DMC convencional para os problemas regulatório e servo, fornecendo resultados satisfatórios em ambas as situações. Isto demonstra o alto potencial do algoritmo proposto para o controle de sistemas não lineares / Abstract: Controller synthesis represents an important slope of recent development in the field of academic and industry research. A well-projected controller may express the success in reference to the aim of production, and it also creates materials with desirable specification and it allows that the system operates under certain restrictions, considering aspects related to operability, safety and minimization of residue. Then, it¿s known that modeling stage is fundamental to delineation of controller strategies. However, the obtaining of precise and applicable mathematical representation to control most of relevant process in chemist engineering is an arduous task, because of the presence of nonlinear dynamic behavior and space-time variation. Therefore, it searches obtaining of simplier models, but gifted of essential representativity inherent to production system, because of this suitable control structure to each specific necessity projected. This work focus on the development of predictive hybrid controller based on fuzzy models, type Takagi-Sugeno, to process of copolymerization, that present high complex dynamic and hard mathematical modeling, what is also difficult to apply conventional methodologies of control. Two study cases that present analyze of purpose controller were considered: the process of copolymerization in solution of methyl methacrylate and vinyl acetate, and industry copolymerization of ethene/1-butene with Ziegler-Natta catalyzation. The phenomenologic models of the two processes have already described in relevant literature, and they are considered as virtual plants to create dynamic data and controller implementation. Based on computer simulation, functional dynamic fuzzy models were made ¿ they demonstrate excellent capacity to output prediction in process as a function of input dynamic data ¿ and they are, subsequently, inserted in an internal structure of DMC (Dynamic Matrix Control) predictive control. The choice of DMC controller as the base to development of proposal structure is responsible to the fact of its well-known industry applicability, allied to simplicity of the project and execution, beyond it makes possible the incorporation of restrictions in controlled and manipulated variables. Finally, the performance between hybrid controller and conventional DMC to regulatory and servo problems were compared, and it supplies satisfactory results in both situation. It demonstrates high potential of propose algorithm to control nonlinear system / Mestrado / Desenvolvimento de Processos Químicos / Mestre em Engenharia Química

Les systèmes de sécurité active du véhicule de distribution sont utilisés pour réduire la survenue de situations accidentogènes ou aider le conducteur dans de telles situations pour éviter l'accident. Dans cette étude, nous nous intéressons à la prévention des accidents liés à la perte de contrôle du véhicule par le conducteur. Une partie de ces accidents peut être prévenue par l'utilisation de systèmes de sécurité active de type ESP, qui résout les problèmes d'instabilité du véhicule de type "porteur" en lacet et en roulis. Nous étudions la possibilité d'améliorer les performances de tels systèmes en combinant les actions de freins aux actions d'autres actionneurs disponibles sur le véhicule de distribution, tels que les directions actives avant et arrière. Le véhicule de distribution considéré dans notre cas est sur-actionné car le nombre d'actionneur dépasse le nombre de degrés de liberté contrôlés par ces actionneurs. Nous appliquons deux types de commande adaptés au contrôle des systèmes sur-actionnés: le premier est basé sur la structure hiérarchisée (connu sous le nom de structure avec "allocation de contrôle") et le deuxième est basé sur la structure non-hiérarchisée par "commande prédictive à base de modèle". Dans ce travail, nous adaptons ces deux approches à la problématique du contrôle du véhicule poids lourd de type "porteur". Le développement de la commande passe par la modélisation de la dynamique du véhicule, la validation des modèles et l'analyse de sa stabilité. Les résultats de simulation de prestations ISO, reproduisant des situations réelles et déstabilisantes pour le véhicule, sur le modèle représentatif du comportement réel du véhicule poids lourd démontrent l'efficacité de la commande développée et permettent d'évaluer le gain d'utilisation des actions combinées des actionneurs du véhicule par rapport au schéma classique. Ils permettent en outre de porter un jugement sur les deux types de commandes mises en œuvre. / The active safety systems for the delivery heavy vehicle are used to reduce the hazardous situations and assist the driver in such situations to avoid accidents. This study is focused on the prevention of accidents associated with the loss of control by the driver. Some of these accidents can be prevented by the use of active safety systems such as ESP, which solves the problems of the roll and yaw instability of the single unit heavy vehicles. We study the possibility of improving the performances of such systems by combining the braking actions to actions of other actuators available on the vehicle, such as active front and rear steering systems. The heavy vehicle considered in this study is over-actuated, since the number of actuators exceeds the number of degrees of freedom controlled by these actuators. We apply two control strategies to the over-actuated vehicle control: the first one is the hierarchical control also known as control structure with control allocation, the second one is the model predictive control. In this work, we adapt these two approaches to the single unit heavy vehicle control. The control development passes through vehicle dynamics modeling, models validation and stability analysis. Simulation results of the ISO features, representing real hazardous situations, performed on the high-fidelity model, demonstrate the effectiveness of the developed control and allow evaluating the gain of using the combined actions of actuators in comparison with the classical scheme. The obtained results also allow to judge on two implemented control strategies.

Cette thèse concerne la modélisation des dynamiques de maladies foliaires de cultures pérennes tropicales à différentes échelles spatiales. Cette approche de modélisation est appliquée à la cercosporiose noire du bananier. Il s’agit d’explorer et de déterminer les paramètres environnementaux et de résistance de l’hôte ayant une influence significative sur la dynamique spatiotemporelle de la maladie et d’apporter des éléments associés au contrôle de la cercosporiose noire.Deux modèles ont été développés dans le cadre de cette thèse. La dynamique épidémiologique au niveau de la plante est décrite par un modèle mécaniste décomposé en un modèle de croissance de la plante et un module épidémiologique décrivant le cycle épidémique du pathogène. L’architecture du bananier est prise en compte par le biais de compartiments foliaires positionnés dans l’espace. Le modèle a été validé par un jeu de données indépendant. Les expérimentations numériques et l’analyse de sensibilité du modèle réalisées par les méthodes de Morris et de e-FAST ont permis de mieux comprendre le fonctionnement épidémique de cette maladie et d’identifier les paramètres influant le plus la dynamique épidémique en particulier la vitesse d’extension des lésions, la durée d’incubation et l’efficacité d’infection .L’approche bayésienne a permis de prendre en compte l’information a priori disponible pour ces trois paramètres sur lesquels a porté l’inférence statistique. L’analyse de sensibilité du modèle a également permis d’identifier l’influence de deux paramètres liés à la croissance de la plante : le nombre de feuilles sur une plante et le rythme d’émission foliaire Un modèle de dynamique spatio-temporelle de la cercosporiose noire a été développé à l’échelle d’un territoire à partir d’enquêtes effectuées enMartinique pendant la période d’invasion de cette maladie de septembre 2010 à mai 2012. Les données récoltées étant censurées, l’inférence des paramètres du modèle a été réalisée dans un cadre bayésien en utilisant un algorithme d’augmentation de données. Le modèle et l’inférence développés permettent de reconstruire la dynamique spatio-temporelle de l’invasion et de prédire la fin d’invasion sur le territoire.Les deux modèles de dynamiques spatio-temporelles développés à deux échelles spatiales différentes ont permis d’acquérir des informations importantes pour construire des outils de conception de méthode de contrôle de la cercosporiose noire des bananiers. / This thesis concerns the modeling of the dynamics of foliar diseases of tropical tree crops at different spatial scales. This modeling approach is applied to black Sigatoka of banana. It is to explore and determine the environmental parameters and host resistance has a significant influence on the spatiotemporal dynamics of the disease and to provide elements associated with the control of Sigatoka noire.Deux models were developed as part of this thesis. The epidemiological dynamics at the plant is described by a mechanistic model decomposed into a growth model of the plant and an epidemiological unit describing the pathogen epidemic cycle. The architecture of the banana is taken into account through foliar compartments positioned in space. The model was validated by an independent data set. Numerical experiments and model sensitivity analysis performed by the methods of Morris and e-FAST enabled to better understand the functioning of this epidemic disease and identify the parameters affecting the most dynamic epidemic especially speed extension of the lesions, the incubation period and efficiency of Bayesian .L'approche infection allowed to take into account prior information available for these three parameters that were the statistical inference. The model sensitivity analysis also identified the influence of two parameters related to plant growth: the number of leaves on a plant and rate of leaf emission A dynamic model of space-time Black Sigatoka has been developed at the scale of a territory from surveys enMartinique during the period of invasion of the disease from September 2010 to May 2012. The data collected being censored, inference of model parameters has was performed in a Bayesian framework, using a data augmentation algorithm. The model developed and inference possible to reconstruct the spatiotemporal dynamics of the invasion and predict the end of the invasion territoire.Les two spatio-temporal dynamics models developed at two different spatial scales has been gained important information for build tools design method of control of black Sigatoka of banana.

In this Thesis a Unified Discrete State-Space Model for power converters in CCM is presented. Two main approaches to arriving at the discrete model are used. The first approach involves an impulse function approximation of the duty cycle modulations of the converter switches , and this approach results in a small signal discrete model. The Second approach is direct and does not involve any approximation of the modulations , this approach yields both a large signal nonlinear discrete model and a linear small signal model. Harmonic analysis of the converter states at steady-state is done for steady-state waveform acquisition , which increases the accuracy of the model especially for finding the control to inductor current frequency response. Finally the Discrete model is verified for the Half-Bridge DC/DC topology for its three main control schemes (Asymmetric , Symmetric , DCS). A GUI platform in MATLAB is presented as a wrapper that utilizes the models and analysis presented in this thesis.

Os padrões espaciais das alterações da cobertura da terra podem ser analisados a partir de modelagens da sua dinâmica espacial. Essas modelagens são capazes de descrever possíveis fatores responsáveis pelas mudanças territoriais, bem como viabilizar a estimativa de impactos futuros sobre o ambiente simulado. Considerando as pressões do setor de infraestrutura portuária sobre as configurações da forma de uso e planejamento do espaço na região portuária da Baixada Santista, tais como o aporte por áreas de retroporto, o armazenamento e o transporte de cargas, este estudo buscou desenvolver modelos dinâmicos da paisagem e investigar a relação entre as alterações na cobertura da terra e a infraestrutura de transportes na região portuária da Baixada Santista. A partir de produtos de sensoriamento remoto e da metodologia de modelagem dinâmica espacial baseada em autômatos celulares, foram gerados mapas simulados e cenários de prognósticos, com e sem projetos de expansão portuária para o ano de 2024. Os mapas simulados do ano de 1995 e 2005 foram semelhantes aos mapeamentos temáticos de cobertura da terra reais e apresentaram boa avaliação. Importante ressaltar que, apenas o mapa simulado de 2013 não previu as novas instalações portuárias, mesmo com a simulação rodando com a variável de novos empreendimentos operacionais em curto intervalo de tempo. O modelo estacionário em 2024 mostrou a expansão de áreas construídas em substituição, sobretudo, às áreas de campo. Não houve expansão de novas manchas de áreas construídas e isso condiz com a tendência de expansão de áreas urbanas seguindo as áreas já consolidadas e com infraestrutura implantada. Com relação à geração de modelos com os projetos de expansão portuária, tem-se a implantação de diversos empreendimentos ocupando parte da margem esquerda do estuário e áreas de retroporto próximas do arco viário da Baixada Santista. / Spatial patterns of change in land cover can be analyzed with spatial dynamic models. This modeling can indicate the driving factors of territorial changes as well as assess future impacts over the simulated environment. Considering the pressures of the port infrastructure sector over the land use and physical planning in the region of Baixada Santista, such as the demand for backport areas as well as for loads storage and transport, this study aims to develop dynamic models of the landscape and investigate the relationship between land cover change and the transport infrastructure in the port area of Santos. With the aid of remote sensing products and cellular automata based models, simulated maps of past landscapes and maps of future landscape scenarios were generated, both considering and disregarding port expansion projects for the year 2024.The simulated landscape maps of 1995 and 2005 were similar to the real ones, and hence, presented high fitness indices. It is important to mention that only the simulated map of 2013, could not foresee the new port facilities, even when the runs took into account the variables associated with the new operational ventures in a short period of time. The stationary model showed an expansion of built-up areas, mainly replacing former field areas. There was no occurrence of new patches of built-up areas and this trend is consistent with the expansion of urban areas nearby the already consolidated and established areas disposing of infrastructure facilities. Regarding the generation of future simulated scenarios taking into account the port expansion projects, they could well reproduce the implementation of several projects along part of the left bank of the estuary and nearby areas of the road ring of Baixada Santista.

De nos jours, les exigences en productivité dans le monde industriel imposent aux robots un comportement optimal en termes de précision géométrique et dynamique, et en termes de temps de réponse. Ainsi, la présence des flexibilités dans les liaisons pivots des structures mécaniques légères se déplaçant à grande vitesse et sous charges importantes peut limiter dynamiquement la précision et le temps de stabilisation sur la pose finale du robot. La problématique traitée dans ces travaux concerne la maîtrise des vibrations des robots sériels à liaisons flexibles durant les opérations de prise et dépose (Pick and Place).Dans ces travaux, nous effectuons une modélisation et une identification expérimentale des paramètres géométriques et dynamique d’un robot à liaisons flexible. Ce modèle sera utilisé dans la synthèse d’une loi de commande basée modèle dédiée aux robots à flexibilité articulaire. Cette stratégie permet de réduire les vibrations lors des phases exigeantes dynamiquement. Des simulations sur un robot Scara sont alors conduites pour valider la pertinence de cette loi de commande qui intègre un modèle des flexibilités présentes dans les liaisons pivots dans le schéma de commande. Nous appliquons sur le même simulateur du robot à liaisons flexibles trois autres stratégies de commande afin de faire une comparaison (commande PD, commande dédiée aux robots rigides et commande ne considérant pas les amortissements). Le schéma de la loi de commande basée modèle permet de respecter la précision de pose finale avec une diminution du temps de stabilisation. Finalement, Le calcul de l’erreur d’asservissement nous a permis de constater l’influence des erreurs de modélisation de la flexibilité sur la précision de la tâche. Dans ce contexte, une analyse de sensibilité aux paramètres influents est établie. / Nowadays, the demand of productivity in the industrial world of robotics require robots to behave optimally in terms of geometric and dynamic accuracy and response time. Thus, the presence of flexibilities in rotational joints can dynamically limit the position control of manipulators having lighter arms, higher payload-to-weight ratio and doing tasks at high speed. The problem addressed in this work concerns the vibration control of serial robots with flexible joints performing Pick and Place tasks. In this work, we carry out modelling and experimental identification of the geometric and dynamic parameters of a robot with flexible joints. This model is then used in the synthesis of a model-based control law dedicated to manipulators with flexible joints. This strategy reduces vibrations resulting from joints sensitivity during dynamically demanding phases. Simulations on a Scara robot are then conducted to validate the relevance of the proposed control law which integrates joint flexibilities in the form of a feedback loop in the control diagram. To this end, three other control strategies (PD control, control dedicated to rigid structures and control not considering damping) are applied to the same simulator in order to make a comparative analysis. The diagram of the model-based control law allows to respect the set point with a reduction in the stabilization time.Finally, the calculation of the servo error allowed us to see the influence of flexibility modeling errors on the accuracy of the task. In this context, the sensitivity of this control law is evaluated through a sensitivity analysis.

Ces travaux de thèse ont été réalisés grâce à une collaboration entre Safran Helicopter Engines (anciennement Turbomeca) et le Laboratoire de Mécanique des Contacts et des Structures (LaMCoS) de l’INSA de Lyon (UMR CNRS 5259). Les boîtes de transmission par engrenages des moteurs d’hélicoptères convoient la puissance mécanique du turbomoteur aux accessoires (pompes, démarreur) et au rotor. Leur conception dépend des nécessités des équipements embarqués, en particulier l’allègement pour réduire la consommation en carburant. Les engrenages haute vitesse de la transmission sont allégés grâce à des enlèvements de matière dans les corps sous la denture, les voiles-minces. Un modèle dynamique d’engrenages a été développé pendant ce projet de recherche. Son approche modulaire permet l’inclusion conjointe des sollicitations dues aux vibrations de l’engrenage et de la nouvelle flexibilité des voiles-minces. Il dérive d’un modèle à paramètres concentrés, comprenant des arbres en poutre, des paliers et carters sous forme de raideurs additionnelles et un élément d’engrenage rigide inclus par son nœud central. Hypothèse est faite que tous les contacts sont situés sur les lignes de contact du plan d’action. Ces lignes sont discrétisées selon des tranches-minces dans les dents et la déviation normale des cellules est recalculée à chaque pas de temps selon la déflexion de la denture. Le nouveau modèle remplace l’engrenage rigide par une modélisation EF du pignon et/ou de la roue condensée sur les nœuds de jante. Une interface lie les raideurs du plan d’action discrétisé aux éléments finis du corps d’engrenage. L’élément prend donc en compte à la fois les sollicitations de l’engrenage et le comportement statique et modal des corps flexibles en dynamique. Des comparaisons sont faites avec des données numériques et expérimentales. Elles attestent de la capacité du nouveau modèle à prédire le comportement dynamique des engrenages flexibles à hauts régimes de rotation. Ces résultats intègrent entre autres des données locales et globales en dynamique. Finalement, le modèle est utilisé sur les deux cas académiques validés pour visualiser les effets des corps flexibles plus en détails. Un premier focus sera fait sur la déflexion statique due aux charges d’engrènement et sur l’optimisation sur le fonctionnement dynamique possible. Puis, les impacts des sollicitations de l’engrènement sur le voile en rotation seront étudiés. Enfin, le pignon et la roue seront affinés, afin de visualiser l’optimisation massique possible et son impact sur la dynamique de l’engrenage. / The research work presented in this manuscript was conducted in the Contact and Structural Mechanics Laboratory (LaMCoS) at INSA Lyon, in partnership with Safran Helicopter Engines (formerly-Turbomeca). In helicopters, the power from the turboshaft is transmitted to the rotor and the various accessories (pumps, starters etc…) via transmission gearboxes. In the context of high-speed, light-weight aeronautical applications, mechanical parts such as gears have to meet somehow contradictory design requirements in terms of reliability and mass reduction thus justifying precise dynamic simulations. The present work focuses on the definition of modular gear dynamic models, capable of integrating both the local phenomena associated with the instant contact conditions between the tooth flanks and the more global aspects related to shafts, bearings and particularly the contributions of light thin-rimmed /-webbed gear bodies. The proposed models rely on combinations of condensed sub-structures, lumped parameter and beam elements to simulate a pinion-gear pair, shafts, bearings and housing. Mesh elasticity is time-varying, possibly non-linear and is accounted for by Winkler foundations derived from a classic thin-slice model. The contact lines in the base plane are therefore discretised into elemental segments which are all attributed a mesh stiffness function and a normal deviation which are updated depending on the pinion and gear angular positions. The main originality in this PhD consists in inserting condensed finite elements models to simulate flexible gear bodies while keeping the simple and faster rigid-body approach for solid gears. To this end, a specific interface has been developed to connect the discretised tooth contact lines to the continuous finite element gear body models and avoid numerical spikes in the tooth load distributions for example. A number of comparisons with numerical and experimental results show that the proposed modelling is sound and can capture most of the quasi-static and dynamic behaviour of single stage reduction units with thin-webbed gears and/or pinions. The model is then applied to the analysis of academic and industrial gears with the objective of analysing the contributions of thin, flexible bodies. Results are presented which highlight the role of centrifugal effects and tooth shape modifications at high speeds. Finally, the possibility to further improve gear web design with regard to mass reduction is investigated and commented upon.