Cette thèse porte sur le diagnostic et le pronostic pour l’aide à la maintenance de systèmes complexes. Elle présente deux approches de diagnostic/pronostic qui permettent de générer les indicateurs utiles pour l’optimisation de la stratégie de maintenance. Plus précisément, ces approches permettent d’évaluer l’état de santé et de prédire la durée de vie résiduelle du système. Les approches présentées visent en particulier à pallier le problème d’absence d’indicateurs de dégradation. Les développements sont fondés sur l’utilisation d’observateurs, de formalisme de Modèle de Markov Caché, des méthodes d’inférences statistiques et des méthodes de prédiction de séries temporelles à base d’apprentissage afin de caractériser et prédire les modes de fonctionnement du système. Les deux approches sont illustrées sur des exemples de dégradation d’un système de régulation de niveau d’eau, d’une machine asynchrone et d’une batterie Li-Ion. / The research presented in this thesis deals of diagnosis and prognosis of complex systems. It presents two approaches that generate useful indicators for optimizing maintenance strategies. Specifically, these approaches are used to assess the level of degradation and estimate the Remaining Useful Life of the system. The aim of these approaches is to overcome for the lack of degradation indicators. The developments are based on observers, Hidden Markov Model formalism, statistical inference methods and learning-based methods in order to characterize and predict the system operating modes. To illustrate the proposed failure diagnosis/prognosis approaches, a simulated tank level control system, an induction motor and a Li-Ion battery were used.

Made available in DSpace on 2014-12-17T14:54:55Z (GMT). No. of bitstreams: 1 MarconiCR_TESE.pdf: 2377871 bytes, checksum: c798a5eab76defef17ac0fe081e2453d (MD5) Previous issue date: 2010-03-16 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / This paper presents a new multi-model technique of dentification in ANFIS for nonlinear systems. In this technique, the structure used is of the fuzzy Takagi-Sugeno of which the consequences are local linear models that represent the system of different points of operation and the precursors are membership functions whose adjustments are realized by the learning phase of the neuro-fuzzy ANFIS technique. The models that represent the system at different points of the operation can be found with linearization techniques like, for example, the Least Squares method that is robust against sounds and of simple application. The fuzzy system is responsible for informing the proportion of each model that should be utilized, using the membership functions. The membership functions can be adjusted by ANFIS with the use of neural network algorithms, like the back propagation error type, in such a way that the models found for each area are correctly interpolated and define an action of each model for possible entries into the system. In multi-models, the definition of action of models is known as metrics and, since this paper is based on ANFIS, it shall be denominated in ANFIS metrics. This way, ANFIS metrics is utilized to interpolate various models, composing a system to be identified. Differing from the traditional ANFIS, the created technique necessarily represents the system in various well defined regions by unaltered models whose pondered activation as per the membership functions. The selection of regions for the application of the Least Squares method is realized manually from the graphic analysis of the system behavior or from the physical characteristics of the plant. This selection serves as a base to initiate the linear model defining technique and generating the initial configuration of the membership functions. The experiments are conducted in a teaching tank, with multiple sections, designed and created to show the characteristics of the technique. The results from this tank illustrate the performance reached by the technique in task of identifying, utilizing configurations of ANFIS, comparing the developed technique with various models of simple metrics and comparing with the NNARX technique, also adapted to identification / Este trabalho apresenta uma nova t?cnica de identifica??o multimodelos baseada em ANFIS para sistemas n?o lineares. Nesta t?cnica, a estrutura utilizada ? do tipo fuzzy Takagi-Sugeno cujos consequentes s?o modelos lineares locais que representam o sistema em diferentes pontos de opera??o e os antecedentes s?o fun??es de pertin?ncia cujos ajustes s?o realizados pela fase de aprendizagem da t?cnica neuro-fuzzy ANFIS. Modelos que representem o sistema em diferentes pontos de opera??o podem ser encontrados com t?cnicas de lineariza??o como, por exemplo, o m?todo dos M?nimos Quadrados que ? robusto a ru?dos e de simples aplica??o. Cabe ? fase de implica??o do sistema fuzzy informar a propor??o de cada modelo que deve ser empregada, utilizando, para isto, as fun??es de pertin?ncia. As fun??es de pertin?ncia podem ser ajustadas pelo ANFIS com o uso de algoritmos de redes neurais, como o de retropropaga??o do erro, de modo que os modelos encontrados para cada regi?o sejam devidamente interpolados e, assim, definam-se a atua??o de cada modelo para as poss?veis entradas do sistema. Em multimodelos a defini??o de atua??o de modelos ? conhecida por m?trica e, como neste trabalho ? realizada pelo ANFIS, ser? denominada de m?trica ANFIS. Desta forma, uma m?trica ANFIS ? utilizada para interpolar v?rios modelos, compondo o sistema a ser identificado. Diferentemente do ANFIS tradicional, a t?cnica desenvolvida necessariamente representa o sistema em v?rias regi?es bem definidas por modelos inalter?veis que, por sua vez, ter?o sua ativa??o ponderada a partir das fun??es de pertin?ncia. A sele??o de regi?es para a aplica??o do m?todo dos M?nimos Quadrados ? realizada manualmente a partir da an?lise gr?fica do comportamento do sistema ou a partir do conhecimento de caracter?sticas f?sicas da planta. Esta sele??o serve como base para iniciar a t?cnica definindo modelos lineares e gerando a configura??o inicial das fun??es de pertin?ncia. Experimentos s?o realizados em um tanque did?tico, com m?ltiplas se??es, projetado e desenvolvido com a finalidade de mostrar caracter?sticas da t?cnica. Os resultados neste tanque ilustram o bom desempenho alcan?ado pela t?cnica na tarefa de identifica??o, utilizando, para isto, v?rias configura??es do ANFIS, comparando a t?cnica desenvolvida com m?ltiplos modelos de m?trica simples e comparando com a t?cnica NNARX, tamb?m adaptada para identifica??o

L'activité de "pronostic de défaillances" est aujourd'hui considérée comme un processus clef dans les stratégies de maintenance industrielle. Cependant, dans la pratique, les outils de pronostic sont encore rares. Les approches aujourd'hui stabilisées reposent sur un historique des incidents assez conséquent pour être représentatif des événements potentiellement prévisibles. L'objet de cette thèse est de proposer un "outil" permettant de prédire la dégradation d'un équipement sans connaissance a priori sur son comportement, et de générer les indicateurs de pronostic permettant d'optimiser les stratégies de maintenance. Dans cet objectif, notre contribution se décline en trois aspects complémentaires.<br>- Un premier volet de travail traite de la formalisation du processus de pronostic. Le concept de pronostic est défini et positionné par rapport aux stratégies de maintenance. Différents mesures typées pronostic sont proposées et les outils utilisables dans ce contexte sont étudiés (nature, applicabilité, guide de choix).<br>- Le coeur du travail porte ensuite sur la spécification d'un système neuro-flou permettant de reproduire l'évolution des propriétés d'un équipement, et de prédire un état de dégradation au cours du temps. Plus précisément les développements visent la proposition d'une architecture neuro-floue permettant de satisfaire un objectif de contrôle de l'erreur de prédiction, quel que soit l'horizon de prédiction.<br>- Nous développons finalement une approche floue/possibiliste d'adaptation des processus classiques d'évaluation prévisionnelle des grandeurs de sûreté de fonctionnement au cas prédictif (fiabilité, MTTF). Ces indicateurs doivent permettre in fine d'optimiser les stratégies de maintenance en tenant compte de l'incertitude inhérente à l'étape de prédiction des dégradations.

Made available in DSpace on 2014-06-11T19:22:34Z (GMT). No. of bitstreams: 0 Previous issue date: 2010-07-29Bitstream added on 2014-06-13T19:08:09Z : No. of bitstreams: 1 spadotto_mm_me_bauru.pdf: 1459647 bytes, checksum: c67d870286e648ad917f7e25b8b18d56 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / A necessidade de aplicação de novos equipamentos em ambientes cada vez mais agressivos demandou a busca por novos produtos capazes de suportar altas temperaturas, inertes às corroções químicas e com alta rigidez mecânica. O avanço tecnógico na produção de materiais cerâmicos tornou possível o emprego de processos de fabricação que antes eram somente empregados em metais. Dentre os processos de usinagem de cerâmicas avançadas, a retificação é o mais utilizado devido às maiores taxas de remoção diferentemente do brunimento e das limitações geométricas do processo de lapidação. A rugosidade é um do parâmetros de saída do processo de retificação que influi, dentre outros fatores, na qualidade do deslizamento entre estruturas, podendo gerar aquecimento. Além disso, o desgaste da ferramenta de corte gerado durante o processo está associado aos custos fixos e a problemas relacionados com o acabamento superficial bem como a danos estruturais. Essas duas variáveis, rugosidade e desgaste, são objetos de estudos de muitos pesquisadores. Entretanto, o controle automático tem sido uma difícil tarefa de ser realizada devido às variações de parâmetros ocorridas no processo. Dessa maneira, o presente trabalho tem por objetivo aplicar a lógica ANFIS (Adaptive Neuro-Fuzzy Inference System) na estimação da rogosidade e do desgaste da ferramenta de corte no processo de retificação plana de cerâmicas avançadas. A ferramenta de corte aplicada para retificar os corpos-de-prova de alumina (96%) foi um rebolo diamantado. A partir do processamento digital dos sinais de emissão acústica e potência média de corte foram calculadas as estatísticas: média, desvio padrão, potência máxima, DPO e DPKS. As estatísticas foram aplicadas com entradas de duas redes ANFIS, uma estimando valores de rugosidade e outra estimando valores de desgaste... / The need for implementation of new equipaments in an increasingly agressive environmentl demanded a search for new products capable of withstanding high temperatures, inert to chemical corrosion and high mechanical stiffeness. Technological advances in the production of ceramic materials have become possible with the employment of manufacturing processes that previously were only employed in metals. Among the advanced ceramics machining processes, the grinding process is the most used, because of higher removal rates in constrast with the honing process and geometric limitations of lapping process. The surface reoughness is one of the output parameters of grinding process that affects, among other factors, the quality of sliding between structures that may generate heat. Moreover, the wear of the cutting tool generated during the process is associated with fixed costs and problems related to suface finishing as well as structural damages. These two variables, surface roughness and wear, have been studied by many researchers; however, the automatic control has been a difficult task to be carry out due to parameters variations occurring in the process. Hence, this work aims to apply logic ANFIS (Adaptive Neuro-Fuzzy Inference System) in the estimation of surface roughness and wear of the cutting tool in the tangential griding process of advanced ceramics. The cutting tools used to grind workpieces of alumina (96%) was a diamond grinding wheel. From the digital processing of acoustic emission and average cutting power signals some statistics were calculated: mean, standard deviation, maximum power, DPO and DPKS. The statistics were applied as inputs of two ANFIS networks estimating surface roughess and wear values. The results had demonstrated that the statistics associated with the ANFIS network can be used in the estimation of surface roughness and wear. However, the wear ANFIS network... (Complete abstract click electronic access below)

Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2017-01-27T12:26:54Z No. of bitstreams: 1 JoseSoaresBatistaLopes_TESE.pdf: 3701361 bytes, checksum: 945caee9725d682534c235543f919e4b (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2017-01-30T13:07:29Z (GMT) No. of bitstreams: 1 JoseSoaresBatistaLopes_TESE.pdf: 3701361 bytes, checksum: 945caee9725d682534c235543f919e4b (MD5) / Made available in DSpace on 2017-01-30T13:07:29Z (GMT). No. of bitstreams: 1 JoseSoaresBatistaLopes_TESE.pdf: 3701361 bytes, checksum: 945caee9725d682534c235543f919e4b (MD5) Previous issue date: 2016-06-17 / Este trabalho descreve o estudo e a implementa??o digital embarcado em um DSP TMS 3208F28335 para o controle vetorial de velocidade do motor-mancal com bobinado dividido de 4 p?los com 250W de pot?ncia. As t?cnicas inteligentes: ANFIS e as Redes Neurais foram investigadas e implementadas computacionalmente para a avalia??o do desempenho do motor-mancal nas seguintes condi??es: operando como estimador de par?metros incertos, e como controlador de velocidade, respectivamente. Para isso, utilizou-se o programa MATLAB? e seu toolbox para as simula??es e os ajustes dos par?metros envolvendo a estrutura ANFIS, e tamb?m para as simula??es com a Rede Neural. Os resultados simulados mostraram um bom desempenho para as duas t?cnicas aplicadas, de forma diferente: como estimador, e como controlador de velocidade utilizando ambas um modelo do motor de indu??o operando como um motor-mancal. A parte experimental para o controle vetorial de velocidade utiliza tr?s malhas de controles: corrente, posi??o radial e velocidade, onde foram investigados a configura??o dos perif?ricos, as interfaces ou drivers para o acionamento do motor-mancal. Detalhes de configura??o dos perif?ricos do DSP TMS 3208F335 s?o descritas neste trabalho, assim como, as interfaces respons?veis pela aquisi??o da corrente, posi??o radial e velocidade do rotor. Por ?ltimo, s?o mostrados os resultados experimentas do motor-mancal comparando o funcionamento do controle vetorial cl?ssico com o controle neural.

Heart signals allow for a comprehensive analysis of the heart. Electrocardiography (ECG or EKG) uses electrodes to measure the electrical activity of the heart. Extracting ECG signals is a non-invasive process that opens the door to new possibilities for the application of advanced signal processing and data analysis techniques in the diagnosis of heart diseases. With the help of today’s large database of ECG signals, a computationally intelligent system can learn and take the place of a cardiologist. Detection of various abnormalities in the patient’s heart to identify various heart diseases can be made through an Adaptive Neuro-Fuzzy Inference System (ANFIS) preprocessed by subtractive clustering. Six types of heartbeats are classified: normal sinus rhythm, premature ventricular contraction (PVC), atrial premature contraction (APC), left bundle branch block (LBBB), right bundle branch block (RBBB), and paced beats. The goal is to detect important characteristics of an ECG signal to determine if the patient’s heartbeat is normal or irregular. The results from three trials indicate an average accuracy of 98.10%, average sensitivity of 94.99%, and average specificity of 98.87%. These results are comparable to two artificial neural network (ANN) algorithms: gradient descent and Levenberg Marquardt, as well as the ANFIS preprocessed by grid partitioning.

No description available.

Submitted by Luciclea Silva (luci@ufpa.br) on 2018-10-01T14:39:49Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertacao_Previsaodistorcaoharmonica.pdf: 4236129 bytes, checksum: bb47a1edb3151361639a5867d6c2c545 (MD5) / Approved for entry into archive by Luciclea Silva (luci@ufpa.br) on 2018-10-01T14:40:33Z (GMT) No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertacao_Previsaodistorcaoharmonica.pdf: 4236129 bytes, checksum: bb47a1edb3151361639a5867d6c2c545 (MD5) / Made available in DSpace on 2018-10-01T14:40:33Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertacao_Previsaodistorcaoharmonica.pdf: 4236129 bytes, checksum: bb47a1edb3151361639a5867d6c2c545 (MD5) Previous issue date: 2018-07-11 / Este trabalho apresenta uma modelagem para DHTv%, DHTi% e harmônicos individuais utilizando previsões de um sistema ANFIS que aprende com dados medidos e prevê o comportamento da rede para valores futuros. Estas previsões podem ajudar a atender as normas nacionais de DHTv% estipuladas pela Agência Nacional de Energia Elétrica (ANEEL) através dos Procedimentos de Distribuição (PRODIST), como as normas internacionais de DHTi%., desta forma se antecipando às normas que atualmente são recomendativas, mas em um futuro próximo serão punitivas. A modelagem é realizada por meio de um sistema Neuro-Fuzzy denominado ANFIS, o qual utiliza rede neural para aprender o comportamento do sistema e ajuste dos parâmetros e regra Fuzzy para a determinação dos valores de saída do sistema levando em consideração o aprendizado da rede Neural. A grande vantagem desta ferramenta é o poder de se modelar padrões utilizando uma previsão de estado harmônico das cargas conectadas na baixa tensão, o que ajuda na criação de pseudomedidas para as redes de distribuição, onde é difícil e oneroso a obtenção de medições reais. Entre as aplicações práticas para esta ferramenta pode-se destacar a utilização dos valores previstos em substituição a valores anômalos medidos, a utilização em medidores de energia para prever e evitar a ultrapassagem dos valores de Distorção Harmônico estipulados em norma e a utilização como base para a previsão de harmônicas individuais, que podem ser utilizadas em estudos de fluxo de carga harmônicos. / This work presents a modeling for THDv%, THDi% and individual harmonics using predictions from an ANFIS system that learns with measured data and predicts the behavior of the network for future values. These forecasts can help meet national THDv% standards stipulated by the Agência Nacional de Energia Elétrica (ANEEL) through Distribution Procedures (PRODIST), such as THDi% international standards, thus anticipating the currently recommended standards, but in the near future will be punitive. The modeling is performed by means of a Neuro-Fuzzy system called ANFIS, which uses neural network to learn the behavior of the system and adjustment of the parameters and Fuzzy rule for the determination of the system output values taking into account the learning of the Neural network. The great advantage of this tool is the power of modeling standards using a prediction of the harmonic state of the connected loads in the low voltage, which helps in the creation of pseudomedidas for the distribution networks, where it is difficult and costly to obtain real measurements. Among the practical applications for this tool is the use of the predicted values instead of measured anomalous values, the use in energy meters to predict and avoid exceeding the values of Harmonic Distortion stipulated in standard and the use as a basis for the prediction of individual harmonics that can be used in harmonic load flow studies.

In recent years, response analysis of complex structures under impact loads has attracted a great deal of attention. For example, a collision or an accident that produces impact loads that exceed the design load can cause severe damage on the structural components. Although the AASHTO specification is used for impact-resistant bridge design, it has many limitations. The AASHTO specification does not incorporate complex and uncertain factors. Thus, a well-designed structure that can survive a collision under specific conditions in one region may be severely damaged if it were impacted by a different vessel, or if it were located elsewhere with different in-situ conditions. With these limitations in mind, we propose different solutions that use smart control technology to mitigate impact hazard on structures. However, it is challenging to develop an accurate mathematical model of the integrated structure-smart control systems. The reason is due to the complicated nonlinear behavior of the integrated nonlinear systems and uncertainties of high impact forces. In this context, novel algorithms are developed for identification, control and monitoring of nonlinear responses of smart structures under high impact forces. To evaluate the proposed approaches, a smart aluminum and two smart reinforced concrete beam structures were designed, manufactured, and tested in the High Impact Engineering Laboratory of Civil and Environmental Engineering at WPI. High-speed impact force and structural responses such as strain, deflection and acceleration were measured in the experimental tests. It has been demonstrated from the analytical and experimental study that: 1) the proposed system identification model predicts nonlinear behavior of smart structures under a variety of high impact forces, 2) the developed structural health monitoring algorithm is effective in identifying damage in time-varying nonlinear dynamic systems under ambient excitations, and 3) the proposed controller is effective in mitigating high impact responses of the smart structures.

Les contraintes des marchés et les attentes de la société vis-à-vis des systèmes industriels en termes économique, sécuritaire, environnementaux... requièrent de considérer les performances de ces derniers de façon globale sur l'ensemble de leur cycle de vie. Cela nécessite de mettre en synergie, par exemple avec des ingénieries couplées dès la conception, le système principal et ses systèmes contributeurs, et notamment celui de soutien avec son processus pivot de maintenance. Cette focalisation intégrative sur la maintenance a conduit à évoluer d'anciennes pratiques de maintenance vers de nouvelles plus proactives faisant émerger des stratégies prévisionnelles dont le processus clé est le pronostic. Cependant ce processus fait l'objet d'un réel manque de formalisation et les travaux existants restent principalement centrés sur les composants, sans prendre en compte les performances des systèmes. Ainsi notre contribution porte sur la proposition d'architectures génériques de pronostic système permettant d'obtenir les évolutions futures des dégradations/défaillances des composants et des performances de niveaux système/sous-systèmes/composants : soit directement par un pronostic adapté, soit par modélisation de la causalité dysfonctionnelle sous forme de relations logiques supportées par un réseau de neurones flou ANFIS. Une méthodologie est associée pour définir les indicateurs de dégradation et de performance, aboutissant à la réalisation des architectures. Enfin la faisabilité de cette approche est démontrée sur un système de déroulage/pressage de la plateforme TELMA.