Utilização de experimentos com relés para identificação de pontos da resposta em freqüência de sistemas multivariáveis quadrados e sintonia de controladores PID descentralizados. / It use relay experiments for square multivariable system frequency response and decentralized PID controller tunning.

L’objectif principal de cette thèse est de proposer une analyse des spécificités de la transduction microonde dans le cadre d’une application capteur d’ammoniac. Les deux spécificités principales sont la caractérisation large bande (1 et 8 GHz) et les matériaux sensibles (diélectriques). Cette méthode de transduction repose sur l’interaction entre un gaz polluant et un matériau sensible déposé à la surface d’une structure propagative spécifique aux microondes. La réponse du capteur n’est pas directement induite par les propriétés diélectriques de la molécule gazeuse, mais plutôt par celles de l’espèce cible adsorbée à la surface du matériau sensible. Cette adsorption provoque une modification des paramètres du capteur mesurés par un analyseur de réseau vectoriel. Contrairement à des transductions conventionnelles comme la conductimétrie, ce principe fonctionne à température ambiante avec tout type de matériau, y compris les isolants électriques.Les premiers travaux réalisés durant cette thèse ont conduit au développement d’un nouveau banc expérimental. Il est spécifiquement adapté à l’étude de capteurs microondes par mesures des coefficients de réflexion et de transmission. Ce développement comprend la conception de deux nouvelles générations de capteurs, recouverts d’oxydes métalliques (fer ou titane) commerciaux ou synthétisés dans le cadre de l’étude. Le premier capteur comporte des circuits interdigités tandis que le second est un résonateur trapézoïdal. Ce dernier est caractérisé par une série de fréquences d’intérêt distribuées régulièrement entre 1 et 8 GHz. L’association d’un spectromètre de masse au banc de mesure a permis de suivre les adsorptions/désorptions de l’espèce cible qui est l’ammoniac (10-100 ppm), mais également le comportement du gaz vecteur utilisé, l’argon, et de l’eau adsorbée initialement sur le matériau sensible ou volontairement ajoutée en cours de l’expérience. L’objectif est d’étudier le rôle de l’eau comme interférent vis-à-vis de la détection de l’ammoniac. Une troisième molécule d’intérêt, l’éthanol, a été également été utilisée durant les expériences afin d’estimer de potentielles différences de comportement entre les molécules détectées. Les résultats expérimentaux ont été exploités au moyen de protocoles de traitement de données spécifiquement établis durant cette thèse. Des traitements temporels ont été conduits afin d’étudier le comportement cinétique du capteur, tandis que des traitements spectraux ont permis d’appréhender l’aspect large bande de la réponse du capteur en présence de polluants.Le premier résultat majeur est l’augmentation significative de la sensibilité à l’ammoniac, qui a permis d’abaisser le seuil de détection à des concentrations d’ammoniac de l’ordre de 10 ppm. Le dioxyde de titane a été identifié comme un bon candidat pour la détection d’ammoniac, avec des variations de coefficient de réflexion allant jusqu’à 6 dB pour 300 ppm d’ammoniac. Le rôle de l’eau initialement adsorbée sur le matériau sensible a été élucidé, montrant que son influence n’est significative que lors des premières minutes des expérimentations. Ainsi, il est possible de détecter l’ammoniac en présence d’humidité. Les processus liés aux expositions aux flux gazeux et particulièrement au gaz vecteur ont été identifiés et ont confirmé que la réponse du capteur était uniquement due à son interaction avec les molécules cibles. Un autre résultat majeur est la définition des conditions opératoires nécessaires à l’établissement de la sélectivité. Notre analyse théorique a clairement démontré l’intérêt des mesures large bande en termes de discrimination de molécules cibles. Cette analyse a été testée dans le cadre d’expériences multicibles utilisant l’ammoniac, l’eau et l’éthanol. Ces observations ont permis d’établir un cahier des charges d’une nouvelle génération de capteurs microondes, garantissant une discrimination systématique entre ces trois molécules. / The main objective of this thesis is to propose an analysis of the microwave transduction specificities in the framework of ammonia sensing applications. The two main features of this transduction are its broadband characterization (1 to 8 GHz) as well as its sensitive materials (dielectrics). This transduction method is based on the interaction between a polluting gas and a sensitive material deposited on the surface of a microwave-specific propagating structure. The response of the sensor is not directly induced by the dielectric properties of the gaseous target molecule, but rather by those of the target species adsorbed on the surface of the sensitive material. This adsorption causes a modification of the sensor parameters measured by a vector network analyzer. Unlike more conventional transducers such as conductimetry, this principle works at room temperature with any type of material, including electrical insulators.The first work carried out during this thesis led to the development of a new experimental bench adapted specifically for the study of microwave gas sensors by measuring the S-parameters in reflection and transmission modes. This development includes the design of two new generations of sensors, coated with metal oxides (iron or titanium oxides) commercially available or synthesized during the study. The first sensor comprises interdigital circuits while the second sensor is a trapezoidal resonator. The latter is characterized by a series of frequencies of interest regularly distributed between 1 and 8 GHz. The association of a mass spectrometer with the measurement bench allowed to follow the adsorption and desorption behavior of the target species which is ammonia (10-100 ppm), but also the behavior of the vector gas conventionally used, argon, and water initially adsorbed on the sensitive material or intentionally added during the experiment. The objective is to study the role of water as interfering with the detection of ammonia, the main target species. A third molecule of interest, ethanol, was also used during the experiments in order to estimate the possible differences in the detected molecules behaviors. The experimental results were exploited using specific data processing protocols established during this thesis. Temporal treatments were carried out to study the kinetic behavior of the sensor, while spectral treatments allowed to apprehend the broadband aspect of the sensor response in the presence of pollutants.The first major result is the significant increase in sensitivity to ammonia, which significantly lowered the detection threshold to ammonia concentrations in the 10 ppm range. Titanium dioxide has been identified as a good candidate for ammonia detection, with reflection coefficient variations up to 6 dB for 300 ppm. The role of the water initially adsorbed on the sensitive material has been elucidated, showing that its influence is significant only during the first few minutes of the experiments. Thus, it is possible to detect ammonia in the presence of residual moisture. The processes induced by the gaseous exposures and particularly by the carrier gas were identified, and confirmed that the sensor response was solely due to its interaction with the target molecules. Another major result is the definition of the operating conditions that are necessary for the establishment of the selectivity. Our theoretical analysis clearly demonstrated the interest of broadband measurements in terms of discrimination of target molecules. This analysis has been tested in multitarget experiments using ammonia, water and ethanol. These observations allowed to establish the specifications of a new generation of microwave sensors, guaranteeing systematic discrimination between these three molecules.

Le travail de cette thèse s'inscrit dans le cadre du développement de la méthodologie CRONE (Commande Robuste d'Ordre Non Entier) multivariable. Il porte plus précisément sur la simplification de sa mise en oeuvre pour la synthèse de régulateurs multivariables robustes, avec une application au contrôle de bancs d'essais moteur haute dynamique, notamment ceux développés par l'entreprise D2T. Le premier chapitre présente les différents types de bancs d'essais, leur fonctionnement et leur problématique. Le deuxième chapitre présente tout d'abord une modélisation physique simple des bancs d'essais qui en permet ensuite leur simulation, puis est dédié à leur identification fréquentielle avec une application à un banc d'essais équipé d'un moteur Diesel. Le troisième chapitre présente la méthodologie CRONE multivariable et différents développements permettant d'en simplifier la mise en oeuvre notamment grâce à l'optimisation d'un nouveau paramètre. Illustrant ces développements, la boite à outils CRONE multivariable qui a été développée est alors utilisée pour la synthèse de la loi de commande d'une maquette de banc d'essais constituée de deux moteurs asynchrones. Finalement, le quatrième chapitre est dédié à la validation des développements présentés par leur application à un banc d'essais haute dynamique équipé d'un moteur Essence. Ce chapitre présente également les différents outils logiciel développés pour faciliter l'intégration de la méthodologie CRONE full MIMO aux produits D2T. / The work presented in this Thesis is part of the CRONE (robust control with fractional integration order) multivariable methodology development. It deals with the simplification of the robust multivariable control system implementation with application to the control of high dynamic engine test-benches, including those developed by the D2T company. The first chapter introduces different kind of test-benches and gives a general idea about how those systems work while explaining their problems. While the first part of chapter II presents a simple method for test-benches modeling which then allows their simulation, the second part is dedicated to their frequency-domain system identification. The defined identification procedure is applied to a test-bench equipped with a Diesel engine. The third chapter presents the CRONE multivariable methodology and different developments to simplify its use. Reflecting these developments, the developed CRONE multivariable toolbox is used to design a control system for a test-bench consists of two asynchronous motors. The fourth chapter is dedicated to validate the presented work by applying these developments to a high dynamic test-bench with a spark ignition engine. This chapter presents the various software tools developed to simplify the implementation to D2T products of the CRONE full MIMO methodology.

Utilização de experimentos com relés para identificação de pontos da resposta em freqüência de sistemas multivariáveis quadrados e sintonia de controladores PID descentralizados. / It use relay experiments for square multivariable system frequency response and decentralized PID controller tunning.

International Telemetering Conference Proceedings / October 25-28, 1999 / Riviera Hotel and Convention Center, Las Vegas, Nevada / This paper describes the design and calibration of a four-channel, airborne, swept-tuned spectrum analyzer used in two hypersonic flight experiments for characterizing dynamic data up to 25 kHz. Built mainly from commercially available analog function modules, the analyzer proved useful for an application with limited telemetry bandwidth, physical weight and volume, and electrical power. The authors discuss considerations that affect the frequency and amplitude calibrations, limitations of the design, and example flight data.

Les systèmes d'énergie électrique embarqués ou semi autonomes connaissent des évolutions marquées. Parmi elles, on peut citer l'introduction massive d'électronique de puissance permettant d'améliorer significativement les performances via leur commande et l'interfaçage de sources de stockage. A contrario, de nombreux problèmes se posent lors du dimensionnement et du choix de l'architecture de commande du système : quelle stratégie de gestion énergétique doit-on mettre en œuvre au préalable pour optimiser les flux et le dimensionnement des sources ? Quelle commande permet d'intégrer à la fois des critères de performances dynamiques, de stabilité et de robustesse vis-à-vis des incertitudes paramétriques inhérentes à un processus de dimensionnement ? Enfin, quels outils et modèles doit-on mettre en œuvre pour garder un lien fort entre paramètres du système et critères de performances tout au long de ce processus ? Les travaux présentés dans ce mémoire proposent des premières réponses à ces questions, via l'utilisation de techniques de commande avancées comme la commande optimale ou la commande robuste. Des outils originaux de modélisation, basés sur les systèmes d'ordre non entier seront également présentés pour la modélisation de dispositifs diffusifs (machines électriques, générateurs électrochimiques) ou modélisés par des équations aux dérivées partielles.

The history of crude oil desalting/dehydration plant (DDP) has been marked in progressive phases-the simple gravity settling phase, the chemical treatment phase, the electrical enhancement phase and the dilution water phase. In recent times, the proper cachet would be the control-optimisation phase marked by terms such as "DDP process control", "desalter optimisation control" or "DDP automating technology". Another less perceptible aspect, but nonetheless important, has been both a punch listing of traditional plant boundaries and a grouping of factors that play the essential roles in a desalting/dehydration plant (DDP). Nowadays, modelling and optimising of a DDP performance has become more apparent in petroleum and chemical engineering, which has been traditionally concerned with production and refinery processing industries. Today's desalting/dehydration technology finds itself as an important factor in such diverse areas as petroleum engineering, environmental concerns, and advanced technology materials. The movement into these areas has created a need not only for sources useful for professionals but also for gathering relevant information essential in improving product quality and its impact on health, safety and environmental (HSE) aspects. All of the foregoing, clearly establishes the need for a comprehensive knowledge of DDP and emulsion theories, process modelling and optimisation techniques. The main objective of this work is to model and qualitatively optimise a desalting/dehydration plant. In due course, the contents of this thesis will cover in depth both the basic areas of emulsion treatment fundamentals, modelling desalting/dehydration processes and optimising the performance of desalting plants. In addition, emphasis is also placed on more advanced topics such as optimisation technology and process modifications. At the results and recommendation stage, the theme of this work-optimising desalting/dehydration plant will practically be furnished in an applicable scheme. Finally, a significant compendium of figures and experimental data are presented. This thesis, therefore, essentially presents the research and important principles of desalting/dehydration systems. It also gives the oil industry a wide breadth of important information presented in a concise and focused manner. In search of data quality and product on-line-improvement, this combination will be a powerful tool for operators and professionals in a decision support environment.

Recherche des neutrinos de ultra haute energie avec le telescope ANTARES. Exploitations de 3 ans de donnees (2008-2011) pour la recherche d'un signal astrophysique, soit flux de Waxman Bahcall, soit neutrinos cosmogeniques. Utilization de technique d'analize multivariable. / Search for ultra high energy neutrinos with ANTARES telescope. Data analysis on 3 years of lifetime (2008-2011), to look for astrophysical signal, according to Waxman-Bahcall flux or cosmogenic one. Exploitation of multivariate analysis techniques.

Este trabalho faz uma introdução à evolução dos processos de laminação a frio, acionamento e controle dos mesmos, com foco na laminação de encruamento. O principal objetivo é estabelecer um modelo linear dos sistemas de acionamento, que permita o projeto de um controle multivariável. Pretende-se que a robustez do controle multivariável contribua para a eliminação de defeitos associados aos processos de bobinamento e desbobinamento. A eliminação destes defeitos traz um grande retorno econômico em função do grande volume de produção destas instalações. Para isto é desenvolvido um modelo matemático do sistema completo. No modelo são evidenciadas não linearidades associadas à variação dos diâmetros e inércias das bobinas na desenroladeira e enroladeira. Para solucionar esse problema as parcelas contendo as não linearidades são removidas do modelo. Posteriormente essas parcelas são repostas através de malhas feedforward no controle. Para validar o modelo são projetados controladores PID semelhantes aos utilizados no laminador real. Com os resultados obtidos através de simulação, discute-se a validade do modelo comparando-os com os dados do processo real. A partir desse modelo é desenvolvido o modelo linear em espaço de estados, a ser utilizado no projeto do controlador robusto multivariável. Em seguida é desenvolvido o projeto deste controlador multivariável utilizando a técnica LQG/LTR. Finalmente, são discutidos os resultados das simulações, comparando-os com os dados reais da planta. De um modo geral, os resultados apontam para uma boa aproximação entre a planta real e o modelo proposto. / This work is an introduction to the evolution of the processes of cold rolling, drive and control, focusing on skinpass rolling. The main objective is to establish a model of linear drive systems, allowing the design of a multivariable control. It is intended that the robustness of multivariable control contributes to the elimination of defects associated with processes of coiling and uncoiling. The elimination of these defects poses a great economic return due to the large volume of production of these facilities. To achieve this deal it was developed a mathematical model of the complete system. Nonlinearities associated with variation in diameters and inertia reels in desenroladeira and reel are evidenced in the model. To solve this problem the terms containing nonlinearities were removed from the model and, subsequently reintroduced by feedforward control loops. To validate the model, PID controllers were designed in a similar way to those used in real mill. With the results obtained by simulation, we discuss the validity of the model by comparing them with the actual process data. From this model it was developed a linear state space model to be used in the design of robust multivariable controller. After that it was developed the design of this multivariable controller using the technique LQG / LTR. Finally, it was discussed the simulation results, comparing them with actual data of the plant. Overall, the results point to a good approximation of the real plant and the proposed model.

O vírus da influenza tipo A (VIA) é um importante agente em rebanhos suínos ao redor do mundo. Alguns subtipos do vírus podem ser transmitidos entre espécies diferentes, como aves, homem e suínos, proporcionando o aumento de mutações genômicas e de novas cepas circulantes. Os suínos são considerados hospedeiros de "mixagem", uma vez que possuem receptores para cepas de aves, de humanos e suínos. A doença clínica em suínos é caracterizada por quadro respiratório agudo e brando, com duração de 5 a 7 dias. Em rebanhos de reprodutores – como as Unidade Produtoras de Leitões (UPL) – um surto epidêmico de influenza pode levar o estabelecimento de uma infecção endêmica com duração de semanas a meses, sem produção de sinais clínicos evidentes. Protocolos de biosseguridade vêm sendo incorporados e padronizados pelas agroindústrias, visando prevenir a introdução de doenças infecciosas em rebanhos. Entretanto, existem lacunas no conhecimento dos fatores associados à biosseguridade em rebanhos de matrizes suínas brasileiras e sua relação com doenças infecciosas. Por essa razão, um estudo transversal foi realizado para estimar a soroprevalência do VIA em matrizes de UPL e explorar práticas de biosseguridade associadas à presença de anticorpos contra o vírus da influenza. Ao todo, foram amostradas 404 matrizes em 21 granjas. O diagnóstico sorológico foi realizado pelo ELISA (protocolo in house). Todas as amostras positivas pelo ELISA foram testadas usando a inibição de hemaglutinação (IH) para diagnosticar a presença de H1N1pdm2009, H1N2 e H3N2 como subtipos de vírus influenza. As informações sobre práticas de biosseguridade foram obtidas através da aplicação de um questionário. A associação entre o resultado do diagnóstico do ELISA de cada uma das matrizes amostradas e as práticas de biosseguridade da propriedade foi feitas através de um modelo de Regressão de Poisson Robusta, estimando a Razão de Prevalência (RP) como medida da associação. A prevalência estimada de anticorpos anti-VIA nas matrizes foi de 63,9% (IC 95%: 55% - 72%), sendo que todas as granjas tiveram resultados positivos. A frequência dos subtipos nas matrizes usando IH foi 51,9% para H1N1, 27,8% H1N2 e 0,6% H3N2. Coinfecções entre H1N1 e H1N2 foram observadas em 19 granjas. As práticas de biosseguridade associadas significativamente com a presença de anticorpos (p-valor <0,05) foram a "presença de tela anti-pássaros" (RP = 0,75) e "local de aclimatação para leitoas" (RP = 0,57) como fatores protetivos e "reposição externa de leitoas" (RP = 1,38) como associada a uma maior prevalência do vírus da influenza suína. Foi possível verificar que a soroprevalência do VIA nas matrizes comerciais da população estudada é alta, indicando que os animais são frequentemente expostos ao patógeno, e que algumas medidas de biosseguridade estão associadas com a ocorrência da doença, fornecendo subsídios técnicos sobre a importância dos protocolos de biosseguridade para a promoção da saúde do plantel. / Influenza A virus (IAV) is an important infectious agent in pig herds across the globe. Some subtypes of this virus can be transmitted between different species, such as birds, human and pig, increasing genomic mutations and evolving of new circulating strains. Pigs are considered "mixed vessel" for influenza A, since they have cell receptors for birds, humans and pigs strains. The clinical disease in pigs is characterized by an acute and mild respiratory disease, lasting from 5 to 7 days. In breeding herds such as sow farms, an epidemic outbreak of influenza can lead to the establishment of an endemic infection lasting weeks to months without clinical signs. Biosecurity procedures were incorporated and standardized by the agroindustry in order to prevent both the introduction and dissemination of infectious diseases. However, there are gaps in the knowledge about what biosecurity factors are associated with infectious diseases in Brazilian herds. For this reason, a cross-sectional study was carried out to estimate IAV seroprevalence in sows and assess which biosecurity practices are associated with the prevalence of influenza virus antibodies. Four hundred forty-four sows were sampled from 21 farms. Serological assays were performed using an ELISA test (in-house protocol). All ELISA positive samples were tested using the hemagglutination inhibition test (HI) to identify the presence of H1N1pdm2009, H1N2 and H3N2 subtypes. Information of biosecurity practices was obtained through the application of a questionnaire. Association between ELISA diagnostic result of each sampled sow and biosecurity practices was assessed using a robust Poisson regression and the Prevalence Ratio (PR) was used as the measure of association. The estimated prevalence of anti-IAV antibodies in sows was 63.9% (95% CI: 55% - 72%), and all farms had at least one seropositive sow s. The frequency of subtypes using HI was 51.9% for H1N1, 27.8% for H1N2 and 0.6% for H3N2. Co-infections with H1N1 and H1N2 were observed in 19 farms. Biosecurity practices such as "presence of bird-proof" (PR = 0.75), and "presence of an acclimatization unit" (PR = 0.57), protective ones, and "external replacement of gilts" (PR = 1.38), which was positively associated with the IAV prevalence, were statistically significant in the final model (p-value <0.05). It was possible to verify that IAV seroprevalence is high and some biosecurity procedures were associated with the serologic status, offering technical subsidies about the importance of the biosecurity for the herd heath.