Submitted by admin tede (tede@pucgoias.edu.br) on 2016-11-10T13:42:48Z No. of bitstreams: 1 SUZANA ALVES MENDANHA.pdf: 4085845 bytes, checksum: ed8772928c3151a3f2d34b1b727c09b3 (MD5) / Made available in DSpace on 2016-11-10T13:42:48Z (GMT). No. of bitstreams: 1 SUZANA ALVES MENDANHA.pdf: 4085845 bytes, checksum: ed8772928c3151a3f2d34b1b727c09b3 (MD5) Previous issue date: 2015-05-14 / The Planning and Production Control Department influence directly in the management of the organization of the production control, if this department does not perform adequately its functions will directly affect other sectors of the company, such as stocks, production and financial sector. This research aims to structure and monitor the implementation of the Kanban production model, integrating the existing MRP system in a food industry to meet the peculiarities of lack and scarcity of semi-finished products existing in the reality of the company. Using the references found in the literature and adding the acquired knowledge by practical study from the research-action, this research shows the development and deployment of a hybrid model. With this deployment, after the cycles of action research, it was possible to see a reduction in the stock of semifinished products at 25% and the value of capital employed at 24%, plus a reduction of finished product shortage cuts at 73%, even with an increase of 15% in this volume. From these results the hybrid model has become satisfactory to the company, since the MRP system and the Kanban production model was integrated complementary, even with the difficulty of the factory floor staff to follow the flow of the hybrid model of PCP. / O departamento de Planejamento e Controle da Produção influencia diretamente na gestão da organização do controle da produção, caso este departamento não desempenhar adequadamente suas funções afetará diretamente outros setores da empresa, como os estoques, a produção e o setor financeiro. Esta pesquisa tem como objetivo estruturar e acompanhar a implantação do modelo Kanban de produção, integrando ao sistema MRP já existente em uma indústria alimentícia, a fim de atender as peculiaridades de falta e escassez de produtos semiacabados existentes na realidade da empresa. Utilizando as referências encontradas na literatura e acrescentando o conhecimento adquirido pelo estudo prático oriundo da pesquisa-ação, esta pesquisa apresenta o desenvolvimento e a implantação de um modelo híbrido. Com esta implantação, após os ciclos da pesquisa-ação, foi possível verificar uma redução do estoque de semiacabados em 25% e o valor de capital empregado em 24%, além de uma redução de cortes de faltas de produto acabado em 73%, mesmo com um aumento de 15% em seu volume. A partir destes resultados o modelo híbrido se tornou satisfatório para a empresa, visto que o sistema MRP e o modelo Kanban de produção se integraram de forma complementar, mesmo apresentando dificuldades pela equipe de chão de fábrica de seguir o fluxo do modelo híbrido de PCP.

Background: This study seeks to model aspects of the regeneration of radiata pine (Pinus radiata D.Don) seedlings under a range of environmental conditions. This study investigated whether “hybrid” mechanistic models, which predict plant growth and development using empirical representations of plant physiological responses to the environment, could provide a realistic alternative to conventional empirical regeneration models. Objectives: The objectives of this study were to 1) identify the functional relationships between the environmental conditions controlling germination, establishment and growth of radiata pine seedlings, under a range of those environmental conditions as specified by temperature and available light and soil water; and 2) specify those functional relationships in hybrid mechanistic (“hybrid”) models. Methods: Radiata pine seedling germination and growth were measured under controlled environmental conditions (incubators for seed germination, growth cabinets for seedlings), and results used to adapt, parameterise and test two published hybrid models; one for germination (the hydrothermal time model); and one for seedling growth in the first six months after germination, based on plant radiation use efficiency (RUE). The hydrothermal model was tested by incubating commercial radiata pine seeds under factorial combinations of temperature and water potentials where germination was likely to occur (12.5 ºC to 32.5 ºC and 0 MPa to –1.2 MPa.). 100 seeds were germinated for each factorial combination. The hydrothermal germination model was fitted to the germination data using non-linear regression modles, will allowed simultaneous estimation of all modle parameters. Seedlings were grown in controlled growth cabinets, and their RUE was calculated as the ratio of net primary production (NPP, specified in terms of an increase in oven dry biomass), to PAR intercepted or absorbed by a seedling. Estimation of seedling RUE required development of novel techniques for non-destructive estimation of seedling oven dry weight, and measurement of PAR interception by seedlings. The effect of varying PAR flux density on RUE was tested by measuring RUE of seedlings grown at 125, 250 and 500 µmol m⁻² s⁻¹. In a second experiment, the effect of deficits in available soil water on RUE was tested by measuring RUE of seedlings grown under 250 µmol m⁻² s⁻¹ PAR flux, and at different levels of available soil water. Available soil water was specified by a soil moisture modifier factor (ƒθ) which ranges between 1 for moist soils and 0 for soils where there is insufficient water for seedling growth. This soil moisture modifier had not previously been applied in studies of tree seedling growth. Temperatures for both seedling experiments were a constant 17.5 ºC (day) and 12.5 ºC (night). Results: Hydrothermal time models accurately described radiata pine seed germination. Model predictions were closely correlated with actual seed germination over the full range of temperature and water potentials where germination was likely to occur (12.5 ºC to 32.5 ºC and 0 MPa to –1.2 MPa. The minimum temperature for germination (base temperature) was 9.0 ºC. Optimum temperatures for germination ranged from ~20ºC for slow-germinating seeds to ~27 ºC for the fastest germinating seeds. The minimum water potential for seed germination varied within the seed population, with an approximately normal distribution (base water potential = –1.38 MPa, standard deviation of 0.48 MPa). In the process of developing the model, a novel explanation for the decline in germination rates at supra-optimal temperatures was developed (Section 3.4.6), based on earlier models proposed by Alvarado & Bradford (2002) and Rowse & Finch-Savage (2003). This explanation was that the decline in germination rate was not driven just by temperature, but by accumulated hydrothermal time above the base temperature for germination (T₀). This in turn raised the base soil water potential (Ψb) towards 0, so that the reduction in germination rate arose from a reduced accumulation of hydro-time, rather than from thermal denaturation of enzymes facilitating germination – the conventional explanation for non-linear accumulation of thermal time at supra-optimal temperatures for plant development. Upwards adjustment (towards 0 MPa) of base water potentials of germinating seeds occurred also at very cold temperatures in combination with high water potentials. In both cases (very cold or else supra-optimal temperatures) this upwards adjustment in base water potentials prevented germination of part of the seed population, and is proposed as a mechanism which enables seed populations to “hedge their bets” when germinating under less than ideal germination conditions. RUE of young germinated radiata pine seedlings growing in a controlled growth cabinet was not significantly different over a range of constant PAR flux densities. Mean RUE’s were 3.22, 2.82 and 2.58 g MJ⁻¹ at 125, 250 and 500 µmol m⁻² s⁻¹ respectively. In the second experiment, the novel use of a soil moisture modifier (ƒθ) to predict RUE of seedlings subjected to water stress proved successful within a limited range of soil water stress conditions. Measured seedling transpiration and stomatal conductance were closely correlated but seedling photosynthesis was less correlated with available soil water. This result suggests that photosynthesis was not coupled with stomatal conductance when PAR flux was 250 µmol m⁻² s⁻¹, which is well below saturating irradiance for C₃ plants. Conclusions: The use of hybrid, quasi-mechanistic models to describe tree seedling growth has been seldom explored, which necessitated the development of novel experimental and analytical techniques for this study. These included a predictive model of germination decline at sub- and supra-optimal temperatures; a method for accurately estimating seedling dry weights under a range of PAR flux densities; and a novel method for estimating light interception by small seedlings. The work reported in this thesis showed that existing hybrid models (the hydrothermal time germination model and the RUE model) can be adapted to model germination and growth of radiata pine seedlings under controlled environmental conditions. Nonetheless, further research is needed before the models can be confidently used as an alternative to conventional empirical models to model regeneration in “real-world” forests. Research priorities are the performance of hydrothermal germination models under variable field conditions, and the use of the soil moisture modifier for seedlings growing on a range of soil textures and under a range of PAR fluxes.

This dissertation addresses the development of integrative modeling strategies capable of combining deterministic and stochastic, discrete and continuous, as well as multi-scale features. The first set of studies combines the purely deterministic modeling methodology of Biochemical Systems Theory (BST) with a hybrid approach, using Functional Petri Nets, which permits the account of discrete features or events, stochasticity, and different types of delays. The efficiency and significance of this combination is demonstrated with several examples, including generic biochemical networks with feedback controls, gene regulatory modules, and dopamine based neuronal signal transduction. A study expanding the use of stochasticity toward systems with small numbers of molecules proposes a rather general strategy for converting a deterministic process model into a corresponding stochastic model. The strategy characterizes the mathematical connection between a stochastic framework and the deterministic analog. The deterministic framework is assumed to be a generalized mass action system and the stochastic analogue is in the format of the chemical master equation. The analysis identifies situations where internal noise affecting the system needs to be taken into account for a valid conversion from a deterministic to a stochastic model. The conversion procedure is illustrated with several representative examples, including elemental reactions, Michaelis-Menten enzyme kinetics, a genetic regulatory motif, and stochastic focusing. The last study establishes two novel, particle-based methods to simulate biochemical diffusion-reaction systems within crowded environments. These simulation methods effectively simulate and quantify crowding effects, including reduced reaction volumes, reduced diffusion rates, and reduced accessibility between potentially reacting particles. The proposed methods account for fractal-like kinetics, where the reaction rate depends on the local concentrations of the molecules undergoing the reaction. Rooted in an agent based modeling framework, this aspect of the methods offers the capacity to address sophisticated intracellular spatial effects, such as macromolecular crowding, active transport along cytoskeleton structures, and reactions on heterogeneous surfaces, as well as in porous media. Taken together, the work in this dissertation successfully developed theories and simulation methods which extend the deterministic, continuous framework of Biochemical Systems Theory to allow the account of delays, stochasticity, discrete features or events, and spatial effects for the modeling of biological systems, which are hybrid and multiscale by nature.

Models of the fate and transport of river plumes and the bacteria they carry into lakes are developed. They are needed to enable informed decisions about beach closures to avoid economic losses, and to help design water intakes and operate combined sewer overflow schemes to obviate exposure of the public to potential pathogens. This study advances our understanding of river plumes dynamics in coastal waters by means of field studies and numerical techniques. Extensive field measurements were carried out in the swimming seasons of 2006 and 2007 on the Grand River plume as it enters Lake Michigan. They included simultaneous aerial photography, measurements of lake physical properties, the addition of artificial tracers to track the plume, and bacterial sampling. Our observed results show more flow classes than included in previous studies (e.g. CORMIX). Onshore wind can have a significant effect on the plume and whether it impacts the shoreline. A new classification scheme based on the relative magnitude of plume-crossflow length scale and Richardson number based on the wind speed is devised. Previous studies on lateral spreading are complemented with a new relationship in the near field. The plume thickness decreased rapidly with distance from the river mouth and a new non-dimensional relationship to predict thickness is developed. Empirical near field models for surface buoyant plumes are reviewed and a near field trajectory and dilution model for large aspect ratio surface discharge channels is devised. Bacterial reductions due to dilution were generally small (less than 10:1) up to 4.5 km from the river mouth. E. coli decay rates were significantly affected by solar radiation and ranged from 0.2 to 2.2 day-1 which were within the range of previous studies in Lake Michigan. Total coliform survived longer than E. coli suggesting different die-off mechanisms. Mathematical models of the bacterial transport are developed that employ a nested modeling scheme to represent the 3D hydrodynamic processes of surface river discharges in the Great Lakes. A particle tracking model is used that provides the capability to track a decaying tracer and better quantify mixing due to turbulent diffusion. Particle tracking models have considerable advantages over gradient diffusion models in simulating bacterial behavior nearshore that results in an improved representation of bacteria diffusion, decay and transport. Due to the complexity and wide variation of the time and length scale of the hydrodynamic and turbulent processes in the near field (where plume mixing is dominated by initial momentum and buoyancy) and far field (where plume mixing is dominated by ambient turbulence), a coupling technique is adapted. The far field random walk particle tracking model incorporates the empirical near field model. It simulates the transport, diffusion and decay of bacteria as discrete particles and employs the near field output as the source and transports the particles based on ambient currents predicted by the 3D hydrodynamic model. The coupled model improves dilution predictions in the near field. The new techniques advance our knowledge of the nearshore fate and transport of bacteria in the Great Lakes and can be ultimately applied to the NOAA Great Lakes Coastal Forecasting System to provide a reliable prediction tool for bacterial transport in recreational waters.

[pt] A importância da previsão de carga a curto prazo (até uma semana à frente) em crescido recentemente. Com os processos de privatização e implantação de ompetição no setor elétrico brasileiro, a previsão de tarifas de energia vai se tornar extremamente importante. As previsões das cargas elétricas são fundamentais para alimentar as ferramentas analíticas utilizadas na sinalização das tarifas. Em conseqüência destas mudanças estruturais no setor, a variabilidade e a não-estacionaridade das cargas elétricas tendem a aumentar devido à dinâmica dos preços da energia. Em função das mudanças estruturais do setor elétrico, previsores mais autônomos são necessários para o novo cenário que se aproxima. As ferramentas disponíveis no mercado internacional para previsão de carga elétrica requerem uma quantidade significativa de informações on-line, principalmente no que se refere a dados meteorológicos. Como a realidade brasileira ainda não permite o acesso a essas informações será proposto um previsor de carga para o curto-prazo, considerando restrições na aquisição dos dados de temperatura. Logo, tem-se como proposta um modelo de previsão de carga horária de curto prazo (um dia a frente) empregando dados de carga elétrica e dados meteorológicos (temperatura) através de modelos de árvore de decisão. Decidiu-se pelo modelo de árvore de decisão, pois este modelo além de apresentar uma grande facilidade de interpretação dos resultados, apresenta pouquíssima ênfase em sua utilização na área de previsão de carga elétrica. / [en] The importance of load forecasting for the short term (up to one-week ahead) has been steadily growing in the last years. Load forecasts are the basis for the forecasting of energy prices, and the privatisation, and the introduction of competitiveness in the Brazilian electricity sector, have turned price forecasting into an extremely important task. As a consequence of structural changes in the electricity sector, the variability and the non-stationarity of the electrical loads have tended to increase, because of the dynamics of the energy prices. As a consequence of these structural changes, new forecasting methods are needed to meet the new scenarios. The tools that are available for load forecasting in the international market require a large amount of online information, specially information about weather data. Since this information is not yet readily available in Brazil, this thesis proposes a short-term load forecaster that takes into consideration the restrictions in the acquisition of temperature data. A short-term (one-day ahead) forecaster of hourly loads is proposed that combines load data and weather data (temperature), by means of decision tree models. Decision trees were chosen because those models, despite being easy to interpret, have been very rarely used for load forecasting.

La classification d'images s'articule généralement autour des deux étapes que sont l'étape d'extraction de signatures suivie de l'étape d'analyse des données extraites, ces dernières étant généralement quantitatives. De nombreux modèles de classification ont été proposés dans la littérature, le choix du modèle le plus adapté est souvent guidé par les performances en classification ainsi que la lisibilité du modèle. L'arbre de classification et le treillis de Galois sont deux modèles symboliques connus pour leur lisibilité. Dans sa thèse [Guillas 2007], Guillas a utilisé efficacement les treillis de Galois pour la classification d'images, et des liens structurels forts avec les arbres de classification ont été mis en évidence. Les travaux présentés dans ce manuscrit font suite à ces résultats, et ont pour but de définir un modèle hybride entre ces deux modèles, qui réunissent leurs avantages (leur lisibilité respective, la robustesse du treillis et le faible espace mémoire de l'arbre). A ces fins, l'étude des liens existants entre les deux modèles a permis de mettre en avant leurs différences. Tout d'abord, le type de discrétisation, les arbres utilisent généralement une discrétisation locale tandis que les treillis, initialement définis pour des données binaires, utilisent une discrétisation globale. A partir d'une étude des propriétés des treillis dichotomiques (treillis définis après une discrétisation), nous proposons une discrétisation locale pour les treillis permettant d'améliorer ses performances en classification et de diminuer sa complexité structurelle. Puis, le processus de post-élagage mis en œuvre dans la plupart des arbres a pour objectif de diminuer la complexité de ces derniers, mais aussi d'augmenter leurs performances en généralisation. Les simplifications de la structure de treillis (exponentielle en la taille de données dans les pires cas), quant à elles, sont motivées uniquement par une diminution de la complexité structurelle. En combinant ces deux simplifications, nous proposons une simplification de la structure du treillis obtenue après notre discrétisation locale et aboutissant à un modèle de classification hybride qui profite de la lisibilité des deux modèles tout en étant moins complexe que le treillis mais aussi performant que celui-ci. / Image classification is generally based on two steps namely the extraction of the image signature, followed by the extracted data analysis. Image signature is generally numerical. Many classification models have been proposed in the literature, among which most suitable choice is often guided by the classification performance and the model readability. Decision trees and Galois lattices are two symbolic models known for their readability. In her thesis {Guillas 2007}, Guillas efficiently used Galois lattices for image classification. Strong structural links between decision trees and Galois lattices have been highlighted. Accordingly, we are interested in comparing models in order to design a hybrid model between those two. The hybrid model will combine the advantages (robustness of the lattice, low memory space of the tree and readability of both). For this purpose, we study the links between the two models to highlight their differences. Firstly, the discretization type where decision trees generally use a local discretization while Galois lattices, originally defined for binary data, use a global discretization. From the study of the properties of dichotomic lattice (specific lattice defined after discretization), we propose a local discretization for lattice that allows us to improve its classification performances and reduces its structural complexity. Then, the process of post-pruning implemented in most of the decision trees aims to reduce the complexity of the latter, but also to improve their classification performances. Lattice filtering is solely motivated by a decrease in the structural complexity of the structures (exponential in the size of data in the worst case). By combining these two processes, we propose a simplification of the lattice structure constructed after our local discretization. This simplification leads to a hybrid classification model that takes advantage of both decision trees and Galois lattice. It is as readable as the last two, while being less complex than the lattice but also efficient.

abstract: Cigarette smoking remains a major global public health issue. This is partially due to the chronic and relapsing nature of tobacco use, which contributes to the approximately 90% quit attempt failure rate. The recent rise in mobile technologies has led to an increased ability to frequently measure smoking behaviors and related constructs over time, i.e., obtain intensive longitudinal data (ILD). Dynamical systems modeling and system identification methods from engineering offer a means to leverage ILD in order to better model dynamic smoking behaviors. In this dissertation, two sets of dynamical systems models are estimated using ILD from a smoking cessation clinical trial: one set describes cessation as a craving-mediated process; a second set was reverse-engineered and describes a psychological self-regulation process in which smoking activity regulates craving levels. The estimated expressions suggest that self-regulation more accurately describes cessation behavior change, and that the psychological self-regulator resembles a proportional-with-filter controller. In contrast to current clinical practice, adaptive smoking cessation interventions seek to personalize cessation treatment over time. An intervention of this nature generally reflects a control system with feedback and feedforward components, suggesting its design could benefit from a control systems engineering perspective. An adaptive intervention is designed in this dissertation in the form of a Hybrid Model Predictive Control (HMPC) decision algorithm. This algorithm assigns counseling, bupropion, and nicotine lozenges each day to promote tracking of target smoking and craving levels. Demonstrated through a diverse series of simulations, this HMPC-based intervention can aid a successful cessation attempt. Objective function weights and three-degree-of-freedom tuning parameters can be sensibly selected to achieve intervention performance goals despite strict clinical and operational constraints. Such tuning largely affects the rate at which peak bupropion and lozenge dosages are assigned; total post-quit smoking levels, craving offset, and other performance metrics are consequently affected. Overall, the interconnected nature of the smoking and craving controlled variables facilitate the controller's robust decision-making capabilities, even despite the presence of noise or plant-model mismatch. Altogether, this dissertation lays the conceptual and computational groundwork for future efforts to utilize engineering concepts to further study smoking behaviors and to optimize smoking cessation interventions. / Dissertation/Thesis / Doctoral Dissertation Bioengineering 2014

L'intégration de la production des énergies renouvelables intermittentes dans le mix énergétique est aujourd'hui limitée à un seuil de 30 % de la puissance totale produite. Cette mesure vise à assurer la sécurité de l'alimentation électrique des réseaux insulaires en France. La levée de ce verrou technique ne pourra se faire qu'en apportant des solutions au caractère intermittent des sources d'énergies éolienne et photovoltaïque. Les difficultés énergétiques auxquelles sont confrontés aujourd'hui les milieux insulaires préfigurent celles que rencontreront la planète à plus ou moins long terme. Ces territoires sont des laboratoires uniques pour éprouver les nouvelles technologies de stockage, de gestion et de prévision de l'énergie. La contribution de ce travail de thèse se focalise sur la prévision du rayonnement solaire global à différents horizons de temps car la puissance photovoltaïque produite découle directement de l'intensité du rayonnement solaire global. Dans un premier temps, l'étude bibliographique a permis de classer les modèles de prévision numériques et les modèles de prévision statistiques en fonction de la résolution spatiale et temporelle. Par ailleurs, elle montre que les meilleurs performances sont obtenues avec les modèles hybrides. Dans un deuxième temps, un modèle de prévisions à court terme (J+1) est proposé avec le modèle Weather Research and Forecasting (WRF) et un réseau de neurone bayésien. L'hybridation de ces deux méthodes améliore les performances de prévisions à J+1. Dans un troisième temps, un modèle de prévision à très court terme (t+h) est proposé avec le modèle hybride de Kalman. Cette méthode produit d'une part une prévision énergétique et d'autre part une prévision multi-horizon. La comparaison de la performance de ces modèles avec la méthode de référence dite de persistance montre une amélioration de la qualité de la prévision. Enfin, la combinaison du filtre de Kalman avec le modèle numérique WRF permet une mise en œuvre opérationnelle de la prévision. / The integration of intermittent renewable energy in the energy mix is currently limited to a threshold of 30% of the total power being produced. This restriction aims at ensuring the safety of the power input. The elimination of this technical obstacle will be possible with solutions to energy intermittence of wind and solar energy. The energy issues which islands are facing today prefigure global problems in a more or less long term. These territories constitute unique laboratories for testing new technologies of storage, management and forecasting of energy. The contribution of this thesis focuses on the forecasting of global horizontal irradiance at different time horizons. Indeed, the generated PV power stems directly from the intensity of the global horizontal irradiance. First, the review of solar irradiance forecasting methods allows to classify numerical weather models and statistical forecasting methods depending on spatial and temporal resolution. Moreover, it shows that best performance is obtained with hybrid models. Second, a short-term forecast model (day ahead forecast) is developed with the Weather Research and Forecasting model (WRF) and a Bayesian neural network. The hybridization of these methods improves the day ahead forecast performance. Third, a model for forecasting the very short term is developped with the Kalman hybrid model. This method offers on the one hand an energy forecasting and on the other hand a multi-horizon forecast. Comparing the performance of the aforesaid with the reference method, namely the persistence method, shows an improvement of the quality of the forecasts. Combining the Kalman filter with the WRF numerical model allows an operational implementation of the forecast.

Orientadores: Rubens Maciel Filho, Aline Carvalho da Costa / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Quimica / Made available in DSpace on 2018-08-07T06:46:12Z (GMT). No. of bitstreams: 1 Mantovaneli_IvanaCristinaCordeiro_M.pdf: 4744367 bytes, checksum: 648b4b28d0ad3fff1bfeda9851005fc1 (MD5) Previous issue date: 2005 / Resumo: A utilização do etanol como combustível tem muitas vantagens; ele tem competido economicamente com a gasolina e diversos outros combustíveis, substituindo-os em várias utilidades. Desta forma, existe um grande interesse em se otimizar todos os passos da produção de etanol. Apenas um conhecimento profundo da dinâmica do processo gera uma operação ótima, e este pode ser conseguido através de simulações realizadas usando um modelo preciso. Muitos modelos fenomenológicos foram desenvolvidos considerando condições industriais, mas estes só são válidos para condições específicas nas quais foram determinados, invalidando a predição do modelo em outras condições. Mudanças acontecem normalmente em uma unidade industrial e a re-estimação freqüente dos parâmetros do modelo é usualmente difícil e demorada. O objetivo deste trabalho é desenvolver um modelo híbrido neuronal para o processo de fermentação alcoólica usando balanços de massa combinados com redes neuronais do tipo Functional Link. Será implementado um esquema para atualização dos pesos da rede sempre que esta não descrever o comportamento dinâmico da planta. O modelo desenvolvido será usado para descrever um processo real no lugar dos modelos fenomenológicos existentes, já que estes têm sido capazes de descrever o processo apenas por curtos espaços de tempo / Abstract: The use of ethanol as a fuel has many advantages; it has economically competed with gasoline and others fuels, substituting them in various uses. Thus, there is a great interest in optimizing all the steps of ethanol production. Only a detailed knowledge of the process dynamics can lead to optimal operation and this can be achieved through simulation using an accurate model. Many phenomenological models were developed considering industrial fermentations, but they are only valid for specific conditions. Changes occur frequently in an industrial unity and frequent reestimation of model parameters is usually expensive and time consuming. The objective of this work is to develop a hybrid neural model for the alcoholic fermentation process using mass balances combined with Functional Link Neural Networks. A scheme to update network weights always that it does not describe plant behavior accurately is implemented. The developed model is used to describe an industrial process substituting the existing phenomenological models, since they have been able to describe the process only for short periods / Mestrado / Desenvolvimento de Processos Biotecnologicos / Mestre em Engenharia Química

Le contrôle non destructif par ondes guidées générées et détectées par des transducteurs ultrasonores à couplage par air, présente deux avantages majeurs. Le premier réside dans la capacité des ondes guidées à transporter l’information sur la qualité du milieu sur une grande distance. De plus, l’absence d’un milieu de couplage liquide entre les capteurs et le milieu à tester, rend le contrôle plus commode. Ce travail consiste d’abord à développer un procédé de simulation numérique qui prend en considération de nombreux paramètres du système de contrôle. Dans une optique de réduire le nombre de degrés de liberté, un modèle hybride a été développé qui consiste en une combinaison entre un modèle analytique basé sur l’intégrale de Kirchhoff pour la propagation des ultrasons dans l’air et un modèle éléments finis de la propagation des ondes guidées dans le matériau. La mesure des caractéristiques du transducteur à couplage par air (efficacité de l’émetteur et sensibilité du récepteur) permet, d’une part, de calculer la valeur exacte de la pression dans l’air et les valeurs exactes des champs de contraintes et de déplacements dans la structure, pour une tension et une fréquence d’excitation, et d’autre part, de remonter à la tension électrique aux bornes de ce récepteur pour une pression rayonnée par le matériau. Par suite, cette caractérisation rend possible la comparaison entre les prédictions numériques de la réponse (en niveau de tension) du système et les mesures expérimentales correspondantes. A la lumière du modèle numérique développé, une optimisation des paramètres du système de contrôle (angle, fréquence,diamètre, direction de propagation, champ proche et champ lointain) a été effectuée pour améliorer la pureté des modes guidés par le matériau. Une manipulation expérimentale, basée sur un transducteur à couplage par air pour l’émission et une sonde laser pour la réception, a été alors mise en place pour valider quelques prédictions numériques. Ensuite, on a étudié l’interaction des ondes guidées ultrasonores avec des défauts de type délaminage enfouis dans une plaque composite à symétrie quadratique. Pour cela, on a analysé la sensibilité des deux modes fondamentaux A0 et S0 au délaminage en terme de détectabilité. En parallèle, on a traité un problème inverse qui consiste à dimensionner un délaminage par le calcul du spectre fréquentiel du coefficient de réflexion. Enfin, on a mis en évidence le potentiel des transducteurs à couplage par air à ausculter des pièces aéronautiques impactées. / Non-destructive testing (NDT) using guided waves generated and detected by air-coupled ultrasonic transducers have two main advantages. First, this non-contact technique without coupled medium allows obvious convenience of use. Moreover, the ability of guided waves to carry information about medium quality over long distance. In this context, a numerical model has been developed, which takes into account many parameters of the control system. In order to reduce the number of degrees of freedom, a hybrid model has been developed which consists of a combination between an analytical model, based on the Kirchhoff integral for the propagation of ultrasound in air and a finite element model for the propagation of guided waves in the material. The measured characteristics (efficiency and sensitivity) of two air-coupled transducers allow the prediction of the accurate values of the pressure of bulk waves generated in air and the measurement of the pressure of the radiated field in air by guided waves propagating in a structure. This process enables the comparison between predicted and measured guided waves modes. Based on the hybrid model, an optimization of the parameters of the control system (angle, frequency, diameter, direction of propagation, near and far field) was performed to improve the purity of guided modes along the material plate. To validate some numerical predictions, an aircoupled ultrasonic transducer is used and oriented at a specific angle chosen for generating one specific Lambmode guided along a composite plate sample, and a laser probe measures the normal velocity at different locations on the surface of the plate. Then, the interaction of ultrasonic guided waves with delamination in acomposite plate was studied. In particular, the sensitivity of the two fundamental modes A0 and S0 was analyzed in order to predict the detectability of the defect. In parallel, the inverse problem is solved and the defect size is quantified by calculating the spectrum of the reflection coefficient. Finally, the potential of air-coupled transducers to examine an aircraft structure, has been demonstrated.