Objective of the present study is the development of design methods for the control of products’ architecture in order to obtain modular designs. Towards this target, an integrated approach is proposed, investigating the design architecture from two aspects: the -functions to parts- mapping as well as the point of view related to parts’ interactions. For the first aspect, an approach utilizing Axiomatic Design Theory is described in order to control the design architecture with regards to the -functions to parts- mapping. As far as the second aspect is concerned, two indexes are developed quantifying the design architecture in terms of the parts’ interactions perspective. Furthermore, an algorithm for clustering of product’s parts into clusters/modules is introduced. The algorithm utilizes Artificial Neural Networks (ANNs) and Design Structure Matrices (DSMs). The aforementioned developments were incorporated into a CAD based software tool, having as objective the support of modular design. Its main functions are: (a) DSM generation from product CAD model, (b) calculation of the aforementioned indexes, (c) facilitation of clustering and (d) representation of clustered DSM in CAD form. Application of the tool to real case studies from the automotive industry, provide an evaluation of the developed methods. The main outcome of the present work is the integrated approach that was proposed and realized through the software tool, which integrates methods for the handling of a product’s design architecture. This process assists in real time (during the design process) design engineers to the generation of modular designs. The evaluation of the case studies reveals the efficiency of the proposed approach to produce such designs and validates its applicability to industry. / Το αντικείμενο της παρούσας διατριβής είναι η ανάπτυξη μεθόδων για τον έλεγχο της σχεδιαστικής αρχιτεκτονικής των προϊόντων. Για το σκοπό αυτό προτείνεται μια ολοκληρωμένη προσέγγιση η οποία διερευνά τη σχεδιαστική αρχιτεκτονική και από τις δύο της διαστάσεις: την αντιστοίχιση των λειτουργιών του προϊόντος στα μέρη από τα οποία αποτελείται καθώς και την αλληλεπίδραση που έχουν τα μέρη αυτά μεταξύ τους. Για τη πρώτη διάσταση, προτείνεται ένας τρόπος χρησιμοποίησης της Θεωρίας του Αξιωματικού Σχεδιασμού (Axiomatic Design Theory) ώστε να γίνεται έλεγχος της σχεδιαστικής αρχιτεκτονικής ως προς την αντιστοίχιση των λειτουργιών στα μέρη του προϊόντος. Όσον αφορά τη δεύτερη διάσταση, αναπτύσσονται δύο δείκτες οι οποίοι ποσοτικοποιούν την σχεδιαστική αρχιτεκτονική που αφορά τη δομή των αλληλεπιδράσεων των τμημάτων του προϊόντος. Επίσης, εισάγεται ένας αλγόριθμος για την ομαδοποίηση (clustering) των μερών ενός προϊόντος. Ο αλγόριθμος αυτός χρησιμοποιεί Τεχνητά Νευρωνικά Δίκτυα και πίνακες DSM (Design Structure Matrix). Οι παραπάνω μέθοδοι ενσωματώθηκαν σε ένα λογισμικό εργαλείο που αναπτύχθηκε. Το εργαλείο αυτό συνεργάζεται με προγράμματα CAD και έχει ως στόχο την στήριξη του ομαδοποιημένου σχεδιασμού. Οι βασικές του λειτουργίες είναι η δημιουργία του πίνακα DSM ενός προϊόντος χρησιμοποιώντας το αντίστοιχο σχέδιο CAD, ο υπολογισμός των προαναφερθέντων δεικτών, η διευκόλυνση της διαδικασίας ομαδοποίησης καθώς και η αναπαράσταση σε CAD ενός ομαδοποιημένου πίνακα DSM. Μέσω της εφαρμογής του εργαλείου αυτού σε πραγματικές περιπτώσεις της αυτοκινητοβιομηχανίας, πραγματοποιήθηκε αξιολόγηση των μεθόδων που αναπτύχθηκαν. Το κυριότερο αποτέλεσμα της εργασίας είναι η ολοκληρωμένη λύση που προτάθηκε και υλοποιήθηκε μέσω ενός λογισμικού εργαλείου, η οποία ενσωματώνει μεθόδους ελέγχου της σχεδιαστικής αρχιτεκτονικής προϊόντων. Η λύση αυτή βοηθάει σε πραγματικό χρόνο (κατά τη διάρκεια της σχεδιαστικής διαδικασίας) τους σχεδιαστές μηχανικούς, στη δημιουργία καινοτόμων σχεδιασμών. Η αξιολόγηση των περιπτώσεων της αυτοκινητοβιομηχανίας έδειξε την δυνατότητα της προτεινόμενης λύσης να παράγει τέτοιους σχεδιασμούς και επικύρωσε την εφαρμοσιμότητά της σε βιομηχανικό περιβάλλον.

Made available in DSpace on 2014-06-11T19:23:30Z (GMT). No. of bitstreams: 0 Previous issue date: 2005Bitstream added on 2014-06-13T18:09:43Z : No. of bitstreams: 1 cruz_nc_me_prud.pdf: 1504734 bytes, checksum: 24dad2fab48cdca8018cdd5f1df08e04 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Nos últimos anos, infestações de plantas aquáticas em reservatórios estão sendo estudadas como um efeito do desequilíbrio causado pela poluição e represamento dos rios. A quantidade excessiva de plantas, conseqüente desse desequilíbrio, dificulta tanto a navegação como a produção de energia elétrica. Esse tipo de ocorrência, assim como a presença de algumas substâncias na água, provocam mudanças na radiância da mesma, registradas por sensores orbitais. Nesse sentido, técnicas de processamento e análise de dados de sensoriamento remoto podem se constituir em uma fonte complementar de dados e fornecer informações relacionadas ao grau de infestação de reservatórios. Nesse contexto, o presente trabalho teve como objetivo verificar a influência da resolução espacial de imagens multiespectrais na detecção e mapeamento de áreas infestadas por plantas aquáticas emersas em um reservatório de pequeno porte, através de utilização de procedimentos de análise multiescala e classificação supervisionada usando redes neurais artificiais. Para isso foram utilizadas imagens IKONOS multiespectrais (4 metros de resolução espacial) do reservatório de Salto Grande localizado na cidade de Americana- SP. Assim, foram geradas imagens multiescala, resultando em imagens de 8, 16 e 32 metros de resolução espacial. Na classificação das imagens, utilizando Redes Neurais Artificiais, os dados de entrada constituíram-se de imagens multiespectrais IKONOS (bandas 1, 2, 3 e 4), imagem de textura (banda do IVP), e uma imagem de índice de vegetação (NDVI). O procedimento metodológico adotado mostrou-se adequado para o mapeamento das variações espectrais da água e detecção das infestações por plantas aquáticas, nos vários níveis de resolução da imagem. Os resultados obtidos mostraram que a classificação pela rede neural, com os parâmetros... / In past few years, great infestations of aquatic plants in reservoirs have been studied as an effect of the environmental unbalance caused by pollution and damming of rivers. The excessive amount of plants, deriving from this unbalance, makes navigation and the production of electricity difficult. This kind of occurrence, as well as the appearance of some substances in the water, cause changes in the water radiance detected by satellite sensors. Thus, processing techniques and data analysis may be used as a complementary data source to give information related to the degree of infestation of these plants in reservoirs. So, the present dissertation aimed at verifying the influence of the spatial resolution of multispectral images in the detection and mapping of areas infested by aquatic plants in a small reservoir , through the use of multiscale analysis procedures and supervised classification using artificial neural networks. Multiespectral imagens IKONOS (spatial resolution of 4 meters) of the reservoir of Salto Grande, in the city of Americana-SP were used. So, multiscale images were generated, resulting in images of 8, 16 and 32 meters of spatial resolution. In the classification of these images, using Artificial Neural Networks, the input data was constituted of multispectral images IKONOS (bands 1, 2, 3 and 4), image of texture (band of NIR), and one image of vegetation index (NDVI). The method used was adequate to map the spectral variation of the water and to detect infested areas of aquatic plants in the various levels of resolution of the image. The results obtained showed that the classification by the parameters defined for the original image and applied in the training of the scheme adopted for the different resolution levels was satisfactory. Furthermore, an analysis was made comparing multiscale images classified through crossed comparison, which permits comparing...(Complete abstract click electronic access below)

O objetivo desta tese consiste na implementação de sistemas inteligentes para monitoramento e diagnósticos de falhas ocorrentes em motores de indução trifásicos. Para tanto, desenvolveu-se uma bancada de experimentos que visa ensaios de falhas relacionados a curto-circuito entre as bobinas do enrolamento de estator, quebras nas barras da gaiola de esquilo do rotor e, finalmente, rolamentos defeituosos. Mais especificamente, o enfoque principal consiste na proposição de uma abordagem neural de detecção de quebras nas barras de rotores de motores de indução trifásicos mediante a análise do espectro de frequência e aplicação de técnicas de análise das componentes principais. Considerou-se o acionamento do motor de indução tanto pela tensão de alimentação da rede quanto por inversor trifásico em diferentes frequências, operando sob diversas condições de torque de carga para a avaliação da metodologia. / The objective of this thesis consists of the implementation of intelligent systems for three-phase induction motors fault diagnosis and condition monitoring. Therefore, an experimental test stand for stator winding inter-turn short circuit faults, broken rotor bar in squirrel cage and, finally, defective wheel bearing has been designed. The main focus is to propose a neural network approach, which uses spectral frequency analysis and principal component analysis techniques to detect broken rotor bar in squirrel cage induction motor. Induction motor operating at different load torque conditions and supplied with sinusoidal voltage supply and three-phase inverter at different frequency was considered in the experiment for methodology evaluation.

Made available in DSpace on 2014-06-11T19:30:50Z (GMT). No. of bitstreams: 0 Previous issue date: 2009-08-21Bitstream added on 2014-06-13T19:19:29Z : No. of bitstreams: 1 pereira_als_dr_ilha.pdf: 1505539 bytes, checksum: 8b3fa09211b5f80a63a93c6fd21675aa (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Este trabalho apresenta estudos referentes à inclusão do dispositivo FACTS STATCOM e a utilização de Redes Neurais Artificiais para o ajuste dos parâmetros de sinais adicionais estabilizantes (PSS’s e POD’s) no sistema de potência multimáquinas. O objetivo é a melhoria da estabilidade frente às pequenas perturbações do sistema de energia elétrica. O modelo matemático utilizado para o estudo das oscilações eletromecânicas de baixa freqüência em sistemas de energia elétrica foi o Modelo de Sensibilidade de Potência (MSP), modificado para permitir a inclusão do dispositivo STATCOM. Este modelo baseia-se no princípio de que o balanço de potência ativa e reativa deve ser satisfeito continuamente em qualquer barra do sistema durante um processo dinâmico. Prosseguindo na realização do trabalho foram desenvolvidos os modelos matemáticos para a inclusão dos PSS’s e POD’s no sistema elétrico, bem como foi realizada uma discussão a respeito da escolha do local de instalação destes controladores e técnicas clássicas para o ajuste de seus parâmetros. A partir disto foram utilizadas redes neurais artificiais (RNA’s) com o objetivo de ajustar os parâmetros dos controladores. A rede utilizada para este fim foi a do tipo Perceptron Multicamadas, sendo utilizado para o seu treinamento o algoritmo Backpropagation. Simulações foram realizadas e os resultados obtidos para os sistemas testes Sul Brasileiro, Simétrico de Duas Áreas e New England são comentados / This work presents studies about the inclusion of FACTS STATCOM devices using neural networks tune stability additional signal parameters (PSS’s and POD’s) on a multi-machine power system. The objective is to improve the stability to small perturbations in electric power systems. The mathematical model used for studying the lower frequency electromechanical oscillations is the Power Sensitivity Model (PSM), modified to allow the inclusion of the STATCOM device. This model is based on the principle that the active and reactive power balance must be continuously satisfied in every bus of the system during the dynamical process. Mathematical models were developed to include the PSS’s and POD’s on the electrical system, as well as the local to install these control devices and the classical techniques to adjust these parameters. Afterwards, the neural networks were used to adjust the parameters of the controllers. The neural network used is a Perceptron Multi Layer, with the training by backpropagation. Simulations were effectuated for the South Brazilian Power System, the Two Areas Symmetrical Power System and the New England Power System

Made available in DSpace on 2015-03-26T13:28:40Z (GMT). No. of bitstreams: 1 texto completo.pdf: 1954884 bytes, checksum: e05012da0f1e291b4123e380e1682d25 (MD5) Previous issue date: 2007-03-06 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Remote sensing images have spatial, spectral, radiometric and temporal characteristics, which become an important tool for agricultural applications, in several aspects. This thesis describes a methodology for classification of extensive agricultural areas, in regional scale, using MODIS (Moderate Resolution Imaging Spectro-radiometer) images. It was used a spectral-temporal surface response, where each pixel of the image is represented in a three-dimensional space and the axes are: time, wavelength and reflectance, respectively. The methodology consists of an interpolation of analytical surfaces, passing through control points, using two types of interpolators (Collocation and Polynomial Trend Surfaces). Through these interpolation methods, the surface coefficients were generated, which describe the distribution of the surface in the three-dimensional space. These coefficients were then used into the classification process. Two classification algorithms were used, the maxima likelihood and artificial neural network classifiers. One of the drawbacks, in supervised classification process, is the acquisition of the reference image. For this work were used three distinct methodologies for its attainment: thematic image sampling from the spatial resolution of 30m to 250m; digitalization of homogeneous polygons on the screen; and neighborhood operation, which consists of the elimination of pixels with neighborhood with high variation on the digital level. Statistical analyses were used in order to validate the results. The results show that the classification using neural networks had the best resulted, even with small training sample size. The results also show the importance of high quality reference image generations. / Imagens do sensoriamento remoto possuem características espaciais, espectrais, radiométricas e temporais, tornando-se assim uma importante ferramenta para aplicações agrícolas nos mais diversos aspectos. Neste sentido, esta dissertação descreve uma metodologia para a classificação de extensas áreas agrícolas, em escala regional, utilizando imagens MODIS (Moderate Resolution Imaging Spectroradiometer). Fez-se uso de superfícies de resposta espectral-temporal, onde cada pixel da imagem é representado em um espaço tridimensional, onde os eixos são respectivamente: o tempo, o comprimento de onda e a refletância. A metodologia consiste na interpolação de uma superfície analítica passando por pontos de controle, usando dois tipos de interpoladores (Collocation e Análise de Tendência Polinomial). Através dessa interpolação obtiveram-se os coeficientes que descrevem a distribuição da superfície no espaço tridimensional, e os mesmos foram utilizados para a classificação das imagens digitais. Para a classificação foram utilizados dois algoritmos, o classificador Gaussiano da Máxima verossimilhança e as Redes Neurais Artificiais. Uma das limitações, no processo de classificação supervisionada, é a aquisição da imagem de referência, assim para este trabalho foram usadas três metodologias distintas para sua obtenção: Reamostragem das imagens temáticas com resolução espacial de 30 para a resolução de 250 metros; Digitalização de polígonos homogêneos em tela; e Operação de vizinhança, que consiste na eliminação de pixels com vizinhança com alta variação no nível digital. Para a avaliação dos resultados obtidos foram utilizados testes e análises estatísticas. Os resultados mostram que as classificações pelas redes neurais apresentam os melhores resultados, até mesmo com poucas amostras de treinamento. Os resultados também mostram importância da alta qualidade na geração da imagem de referência.

Made available in DSpace on 2014-12-17T14:08:33Z (GMT). No. of bitstreams: 1 HeitorPG.pdf: 2527761 bytes, checksum: 19f772cea6bcb648ba6562ac9a4eeffd (MD5) Previous issue date: 2009-06-26 / The artificial lifting of oil is needed when the pressure of the reservoir is not high enough so that the fluid contained in it can reach the surface spontaneously. Thus the increase in energy supplies artificial or additional fluid integral to the well to come to the surface. The rod pump is the artificial lift method most used in the world and the dynamometer card (surface and down-hole) is the best tool for the analysis of a well equipped with such method. A computational method using Artificial Neural Networks MLP was and developed using pre-established patterns, based on its geometry, the downhole card are used for training the network and then the network provides the knowledge for classification of new cards, allows the fails diagnose in the system and operation conditions of the lifting system. These routines could be integrated to a supervisory system that collects the cards to be analyzed / A necessidade da eleva??o artificial de petr?leo se d? quando a press?o do reservat?rio n?o ? suficientemente elevada para que os fluidos nele contidos possam alcan?ar a superf?cie espontaneamente. Assim a eleva??o artificial fornece energia de forma suplementar ou integral aos fluidos do po?o para que cheguem ? superf?cie. O Bombeio mec?nico com hastes ? o m?todo de eleva??o artificial mais utilizado no mundo e a Carta Dinamom?trica (de superf?cie e de fundo) ? a melhor ferramenta de an?lise de um po?o equipado com esse tipo de m?todo de eleva??o.Um Sistema que utiliza Redes Neurais Artificiais MLP foi desenvolvido e usando padr?es pr?-estabelecidos, baseadas em sua geometria, as cartas de fundo dos po?os s?o utilizadas para o treinamento da rede e posteriormente essa rede disponibiliza o conhecimento adquirido para a classifica??o de novas cartas, permitindo o diagn?stico de poss?veis falhas no sistema de bombeio e das condi??es de funcionamento desse sistema de eleva??o. Essas rotinas podem ser integradas a um Sistema Supervis?rio que coleta as cartas a serem analisadas. Palavras-chave: Automa??o industrial, Eleva??o artificial, Bombeio mec?nico, Redes Neurais

Made available in DSpace on 2014-12-17T14:08:36Z (GMT). No. of bitstreams: 1 AldayrDAJ.pdf: 1169839 bytes, checksum: a47b70e79b9bb61b42503d47bffbccd3 (MD5) Previous issue date: 2010-01-27 / One of the main activities in the petroleum engineering is to estimate the oil production in the existing oil reserves. The calculation of these reserves is crucial to determine the economical feasibility of your explotation. Currently, the petroleum industry is facing problems to analyze production due to the exponentially increasing amount of data provided by the production facilities. Conventional reservoir modeling techniques like numerical reservoir simulation and visualization were well developed and are available. This work proposes intelligent methods, like artificial neural networks, to predict the oil production and compare the results with the ones obtained by the numerical simulation, method quite a lot used in the practice to realization of the oil production prediction behavior. The artificial neural networks will be used due your learning, adaptation and interpolation capabilities / Uma das atividades essenciais na engenharia de petroleo e a estimativa de producao de oleo existente nas reservas petroliferas. O calculo dessas reservas e crucial para a determina??o da viabilidade economica de sua explotacao. Atualmente, a industria do petroleo tem se deparado com problemas para analisar a producao enquanto facilidades operacionais disponibilizam um volume de informacoes que crescem exponencialmente. Tecnicas convencionais de modelagem de reservatorios como simulacao matematica e visualizacao estao bem desenvolvidas e disponiveis. A proposta deste trabalho e o uso de tecnicas inteligentes, como as redes neurais artificiais, para a predicao de producao de petroleo e comparar seus resultados com os obtidos pela simulacao numerica, metodo bastante utilizado na pratica para a realizacao de predicao do comportamento da producao de petroleo. As redes neurais artificiais serao usadas devido a sua capacidade de aprendizado, adaptacao e interpolacao

Orientador: Percival Bueno de Araujo / Banca: Carlos Roberto Minussi / Banca: Laurence Duarte Colvara / Banca: Roberto Apolonio / Banca: Vivaldo Fernando da Costa / Resumo: Este trabalho apresenta estudos referentes à inclusão do dispositivo FACTS STATCOM e a utilização de Redes Neurais Artificiais para o ajuste dos parâmetros de sinais adicionais estabilizantes (PSS's e POD's) no sistema de potência multimáquinas. O objetivo é a melhoria da estabilidade frente às pequenas perturbações do sistema de energia elétrica. O modelo matemático utilizado para o estudo das oscilações eletromecânicas de baixa freqüência em sistemas de energia elétrica foi o Modelo de Sensibilidade de Potência (MSP), modificado para permitir a inclusão do dispositivo STATCOM. Este modelo baseia-se no princípio de que o balanço de potência ativa e reativa deve ser satisfeito continuamente em qualquer barra do sistema durante um processo dinâmico. Prosseguindo na realização do trabalho foram desenvolvidos os modelos matemáticos para a inclusão dos PSS's e POD's no sistema elétrico, bem como foi realizada uma discussão a respeito da escolha do local de instalação destes controladores e técnicas clássicas para o ajuste de seus parâmetros. A partir disto foram utilizadas redes neurais artificiais (RNA's) com o objetivo de ajustar os parâmetros dos controladores. A rede utilizada para este fim foi a do tipo Perceptron Multicamadas, sendo utilizado para o seu treinamento o algoritmo Backpropagation. Simulações foram realizadas e os resultados obtidos para os sistemas testes Sul Brasileiro, Simétrico de Duas Áreas e New England são comentados / Abstract: This work presents studies about the inclusion of FACTS STATCOM devices using neural networks tune stability additional signal parameters (PSS's and POD's) on a multi-machine power system. The objective is to improve the stability to small perturbations in electric power systems. The mathematical model used for studying the lower frequency electromechanical oscillations is the Power Sensitivity Model (PSM), modified to allow the inclusion of the STATCOM device. This model is based on the principle that the active and reactive power balance must be continuously satisfied in every bus of the system during the dynamical process. Mathematical models were developed to include the PSS's and POD's on the electrical system, as well as the local to install these control devices and the classical techniques to adjust these parameters. Afterwards, the neural networks were used to adjust the parameters of the controllers. The neural network used is a Perceptron Multi Layer, with the training by backpropagation. Simulations were effectuated for the South Brazilian Power System, the Two Areas Symmetrical Power System and the New England Power System / Doutor

No description available.

In this thesis, the evolution of adaptive behaviour in artificial agents is studied. More specifically, two types of adaptive behaviours are studied: articulated and cognitive ones. Chapter 1 presents a general introduction together with a brief presentation of the research area of this thesis, its main goals and a brief overview of the experimental studies done, the results and conclusions obtained. On chapter 2, I briefly present some promising methods that automatically generate robot controllers and/or body plans and potentially could help in the development of adaptive robots. Among these methods I present in details evolutionary robotics, a method inspired on natural evolution, and the biological background regarding adaptive behaviours in biological organisms, which provided inspiration for the studies presented in this thesis. On chapter 3, I present a detailed study regarding the evolution of articulated behaviours, i.e., behaviours that are organized in functional sub-parts, and that are combined and used in a sequential and context-dependent way, regardless if there is a structural division in the robot controller or not. The experiments performed with a single goal task, a cleaning task, showed that it is possible to evolve articulated behaviours even in this condition and without structural division of the robot controller. Also the analysis of the results showed that this type of integrated modular behaviours brought performance advantages compared to structural divided controllers. Analysis of robots' behaviours helped to clarify that the evolution of this type of behaviour depended on the characteristics of the neural network controllers and the robot's sensorimotor capacities, that in turn defined the capacity of the robot to generate opportunity for actions, which in psychological literature is often called affordances. In chapter 4, a study seeking to understand the role of reactive strategies in the evolution of cognitive solutions, i.e. those capable of integrating information over time encoding it on internal states that will regulate the robot's behaviour in the future, is presented. More specifically I tried to understand whether the existence of sub-optimal reactive strategies prevent the development of cognitive solutions, or they can promote the evolution of solutions capable of combining reactive strategies and the use of internal information for solving a response delayed task, the double t-maze. The results obtained showed that reactive strategies capable of offloading cognitive work to the agent/environmental relation can promote, rather than prevent the evolution of solutions relying on internal information. The analysis of these results clarified how these two mechanisms interact producing a hybrid superior and robust solution for the delayed response task.