The research in this thesis is undertaken by observing that modern systems are becoming more and more complex and safety-critical due to the increasing requirements on system smartness and autonomy, and as a result health monitoring system needs to be developed to meet the requirements on system safety and reliability. The state-of-the-art approaches to monitoring system status are model based Fault Diagnosis (FD) systems, which can fuse the advantages of system physical modelling and sensors' characteristics. A number of model based FD approaches have been proposed. The conventional residual based approaches by monitoring system output estimation errors, however, may have certain limitations such as complex diagnosis logic for fault isolation, less sensitiveness to system faults and high computation load. More importantly, little attention has been paid to the problem of fault diagnosis system verification which answers the question that under what condition (i.e., level of uncertainties) a fault diagnosis system is valid. To this end, this thesis investigates the design and verification of fault diagnosis algorithms. It first highlights the differences between two popular FD approaches (i.e., residual based and fault estimation based) through a case study. On this basis, a set of uncertainty estimation algorithms are proposed to generate fault estimates according to different specifications after interpreting the FD problem as an uncertainty estimation problem. Then FD algorithm verification and threshold selection are investigated considering that there are always some mismatches between the real plant and the mathematical model used for FD observer design. Reachability analysis is drawn to evaluate the effect of uncertainties and faults such that it can be quantitatively verified under what condition a FD algorithm is valid. First the proposed fault estimation algorithms in this thesis, on the one hand, extend the existing approaches by pooling the available prior information such that performance can be enhanced, and on the other hand relax the existence condition and reduce the computation load by exploiting the reduced order observer structure. Second, the proposed framework for fault diagnosis system verification bridges the gap between academia and industry since on the one hand a given FD algorithm can be verified under what condition it is effective, and on the other hand different FD algorithms can be compared and selected for different application scenarios. It should be highlighted that although the algorithm design and verification are for fault diagnosis systems, they can also be applied for other systems such as disturbance rejection control system among many others.

Malgré les avantages de l'intégration 3D, le test, le rendement et la fiabilité des Through-Silicon-Vias (TSVs) restent parmi les plus grands défis pour les systèmes 3D à base de Réseaux-sur-Puce (Network-on-Chip - NoC). Dans cette thèse, une stratégie de test hors-ligne a été proposé pour les interconnections TSV des liens inter-die des NoCs 3D. Pour le TSV Interconnect Built-In Self-Test (TSV-IBIST) on propose une nouvelle stratégie pour générer des vecteurs de test qui permet la détection des fautes structuraux (open et short) et paramétriques (fautes de délaye). Des stratégies de correction des fautes transitoires et permanents sur les TSV sont aussi proposées aux plusieurs niveaux d'abstraction: data link et network. Au niveau data link, des techniques qui utilisent des codes de correction (ECC) et retransmission sont utilisées pour protégé les liens verticales. Des codes de correction sont aussi utilisés pour la protection au niveau network. Les défauts de fabrication ou vieillissement des TSVs sont réparé au niveau data link avec des stratégies à base de redondance et sérialisation. Dans le réseau, les liens inter-die défaillante ne sont pas utilisables et un algorithme de routage tolérant aux fautes est proposé. On peut implémenter des techniques de tolérance aux fautes sur plusieurs niveaux. Les résultats ont montré qu'une stratégie multi-level atteint des très hauts niveaux de fiabilité avec un cout plus bas. Malheureusement, il n'y as pas une solution unique et chaque stratégie a ses avantages et limitations. C'est très difficile d'évaluer tôt dans le design flow les couts et l'impact sur la performance. Donc, une méthodologie d'exploration de la résilience aux fautes est proposée pour les NoC 3D mesh. / 3D technology promises energy-efficient heterogeneous integrated systems, which may open the way to thousands cores chips. Silicon dies containing processing elements are stacked and connected by vertical wires called Through-Silicon-Vias. In 3D chips, interconnecting an increasing number of processing elements requires a scalable high-performance interconnect solution: the 3D Network-on-Chip. Despite the advantages of 3D integration, testing, reliability and yield remain the major challenges for 3D NoC-based systems. In this thesis, the TSV interconnect test issue is addressed by an off-line Interconnect Built-In Self-Test (IBIST) strategy that detects both structural (i.e. opens, shorts) and parametric faults (i.e. delays and delay due to crosstalk). The IBIST circuitry implements a novel algorithm based on the aggressor-victim scenario and alleviates limitations of existing strategies. The proposed Kth-aggressor fault (KAF) model assumes that the aggressors of a victim TSV are neighboring wires within a distance given by the aggressor order K. Using this model, TSV interconnect tests of inter-die 3D NoC links may be performed for different aggressor order, reducing test times and circuitry complexity. In 3D NoCs, TSV permanent and transient faults can be mitigated at different abstraction levels. In this thesis, several error resilience schemes are proposed at data link and network levels. For transient faults, 3D NoC links can be protected using error correction codes (ECC) and retransmission schemes using error detection (Automatic Retransmission Query) and correction codes (i.e. Hybrid error correction and retransmission).For transients along a source-destination path, ECC codes can be implemented at network level (i.e. Network-level Forward Error Correction). Data link solutions also include TSV repair schemes for faults due to fabrication processes (i.e. TSV-Spare-and-Replace and Configurable Serial Links) and aging (i.e. Interconnect Built-In Self-Repair and Adaptive Serialization) defects. At network-level, the faulty inter-die links of 3D mesh NoCs are repaired by implementing a TSV fault-tolerant routing algorithm. Although single-level solutions can achieve the desired yield / reliability targets, error mitigation can be realized by a combination of approaches at several abstraction levels. To this end, multi-level error resilience strategies have been proposed. Experimental results show that there are cases where this multi-layer strategy pays-off both in terms of cost and performance. Unfortunately, one-fits-all solution does not exist, as each strategy has its advantages and limitations. For system designers, it is very difficult to assess early in the design stages the costs and the impact on performance of error resilience. Therefore, an error resilience exploration (ERX) methodology is proposed for 3D NoCs.

Les véhicules modernes sont de plus en plus équipés de nouveaux organes visant à améliorer la sécurité des occupants. Ces nouveaux systèmes sont souvent des organes actifs utilisant des données de capteurs sur le véhicule. Cependant, en cas de mauvais fonctionnement d'un capteur, les conséquences pour le véhicule peuvent être dramatiques. Afin de garantir la sécurité dans le véhicule, des nouvelles méthodologies de détections de défauts adaptées pour les véhicules sont proposées. Les méthodologies présentées sont étendues de la méthode de l'espace de parité pour les systèmes à paramètres variant (LPV). En outre, la transformation du problème de détection de défauts pour la détection de situations critiques est également proposée. Des résultats applicatifs réalisés sur un véhicule réel dans le cadre du projet INOVE illustrent les performances des détections de défauts et la détection de perte de stabilité du véhicule. / Modern vehicles are increasingly equipped with new mechanisms to improve occupant safety. These new systems are often active parts using data from sensors on the vehicle. However, in case of malfunction of a sensor, the consequences for the vehicle can be dramatic. To ensure safety in the vehicle, new methodologies for detection of faults suitable for vehicles are proposed. The developed methodologies are extended from the method of parity space for linear parameter varying systems (LPV). In addition, the transformation of fault detection problem for the detection of critical situations is also available. Application of results achieved on a real vehicle within the INOVE project illustrate the performance of fault detection and detection of loss of stability of the vehicle.

Orientador: Gilberto Pechoto de Melo / Banca: Aparecido Carlos Gonçalves / Banca: Cleudmar Amaral de Araújo / Resumo: Neste trabalho foi desenvolvida uma metodologia de detecção e localização de falhas, para sistemas mecânicos que utilizam absorvedores dinâmicos de vibrações tipo lâmina vibrante (ADVLV). Através de modificações em alguns de seus parâmetros estruturais, o ADVLV desenvolvido absorve toda ou parte da vibração do sistema mecânico onde se encontra acoplado. Durante o trabalho foi utilizada a teoria de algoritmos genéticos baseada na evolução das espécies como ferramenta de otimização e em seguida, aplicou-se a metodologia de diagnóstico de falhas via observadores de estado para detecção e identificação de possíveis irregularidades no sistema. Na seqüência, a fim de validar a metodologia desenvolvida foram apresentados resultados obtidos através de simulações computacionais e experimentais, realizados com a construção de um ADVLV em uma estrutura primária pertencente ao laboratório de vibrações mecânicas do Departamento de Engenharia Mecânica da UNESP, Campus de Ilha Solteira. / Abstract: In this work it was developed a technique of fault detection and location, to mechanical systems using dynamic vibration absorbers type blade vibrant (DVABV). Through changes in some of its structural parameters, the DVABV developed absorbs all or part of the mechanical vibration of the system which it is attached. During the work, it was used the theory of genetic algorithms based on the evolution of species as a tool for optimization, then applied to the methodology for the diagnosis of faults by state observers in the detection and identification of possible flaws in the system. Following, in order to validate the methodology developed present themselves results through computer simulations and experimental made with the construction of a DVABV and their use in a structure primary, in the Mechanical Vibrations Laboratory, at Mechanical Engineering Department of UNESP, Ilha Solteira. / Mestre

Orientador: Suely Cunha Amaro Mantovani / Banca: Nobuo Oki / Banca: Luis Gustavo Wesz da Silva / Resumo: Atualmente as empresas do setor elétrico deparam-se cada vez mais com as exigências do mercado energético sendo obrigadas a assegurarem aos seus clientes bons níveis de continuidade e confiabilidade no serviço de fornecimento da energia elétrica e também atender os índices de continuidade do serviço estabelecidos pela agência reguladora do setor elétrico (ANEEL - Agência Nacional de Energia Elétrica). Para alcançar estes objetivos além de investir na otimização dos seus sistemas de transmissão e distribuição, as empresas responsáveis têm investido na automação de suas operações, buscando alternativas que reduzam os tempos de interrupção por faltas permanentes nos sistemas de potência. Através de informações disponíveis em uma subestação, é possível estabelecer um procedimento para determinar e classificar condições de faltas, localizando o elemento de proteção acionado, e assim fornecer o apoio à tomada de decisão no ambiente de subestações de sistemas de distribuição de energia elétrica. Neste trabalho é proposta uma metodologia que fornece respostas rápidas (controle on line), para detecção e classificação de faltas em sistemas de distribuição de energia elétrica através de informações analógicas disponíveis em uma subestação, tais como amostras de sinais de tensões e correntes na saída dos alimentadores, com uma arquitetura reconfigurável paralela que usa dispositivos lógicos programáveis (Programables Logics Devices - PLDs) -FPGAs e a linguagem de descrição de hardware - VDHL (Very High Speed Integraded Circuit - VHSIC). Para validar o sistema proposto, foram gerados dados de forma aleatória, compatíveis com informações fornecidas em tempo real pelo sistema SCADA (supervisory control and data-acquisition) de uma subestação real. Os resultados obtidos com as simulações realizadas, mostram que a... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Currently companies of the energy industry is facing increasingly with the requirements of the energy market are obliged to ensure their customers good levels of continuity in service and reliability of supply of electric energy and also meet the rates of continuity of service established by the agency regulator of the energy industry (ANEEL - National Electric Energy Agency). To achieve these goals than to invest in optimization of its transmission and distribution systems, the companies responsible have invested in automation of its operations, seeking alternatives that reduce the time of interruption by failures in the systems of permanent power. Through information available in a substation, it is possible to establish a procedure for identifying and classifying conditions of absence, finding the element of protection driven, and thus provide support for decision-making within the environment of substations to distribution systems for power. This work is proposed a methodology that provides quick answers (control online), for detection and classification of faults in distribution systems of electric energy through analog information available on a substation, such as samples for signs of tensions and currents in the output of feeders, with an architecture that uses parallel reconfigurable programmable logic devices (Programables Logics Devices - PLDs)-FPGAs and the language of description of hardware - VDHL (Very High Speed Circuit Integraded - VHSIC). To validate the proposed system, data were generated at random, consistent with information provided by the system in real time SCADA (supervisory control and data-acquisition) of a real substation. The results obtained with the simulations conducted, show that the proposed methodology, presents satisfactory results, and times of reasonable answers. / Mestre

Made available in DSpace on 2014-06-11T19:27:14Z (GMT). No. of bitstreams: 0 Previous issue date: 2008-03-19Bitstream added on 2014-06-13T19:35:08Z : No. of bitstreams: 1 fernandes_fv_me_ilha.pdf: 2370293 bytes, checksum: dd12c9f677a5b0c0516c7c8361bbee5e (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Neste trabalho foi desenvolvida uma metodologia de detecção e localização de falhas, para sistemas mecânicos que utilizam absorvedores dinâmicos de vibrações tipo lâmina vibrante (ADVLV). Através de modificações em alguns de seus parâmetros estruturais, o ADVLV desenvolvido absorve toda ou parte da vibração do sistema mecânico onde se encontra acoplado. Durante o trabalho foi utilizada a teoria de algoritmos genéticos baseada na evolução das espécies como ferramenta de otimização e em seguida, aplicou-se a metodologia de diagnóstico de falhas via observadores de estado para detecção e identificação de possíveis irregularidades no sistema. Na seqüência, a fim de validar a metodologia desenvolvida foram apresentados resultados obtidos através de simulações computacionais e experimentais, realizados com a construção de um ADVLV em uma estrutura primária pertencente ao laboratório de vibrações mecânicas do Departamento de Engenharia Mecânica da UNESP, Campus de Ilha Solteira. / In this work it was developed a technique of fault detection and location, to mechanical systems using dynamic vibration absorbers type blade vibrant (DVABV). Through changes in some of its structural parameters, the DVABV developed absorbs all or part of the mechanical vibration of the system which it is attached. During the work, it was used the theory of genetic algorithms based on the evolution of species as a tool for optimization, then applied to the methodology for the diagnosis of faults by state observers in the detection and identification of possible flaws in the system. Following, in order to validate the methodology developed present themselves results through computer simulations and experimental made with the construction of a DVABV and their use in a structure primary, in the Mechanical Vibrations Laboratory, at Mechanical Engineering Department of UNESP, Ilha Solteira.

Made available in DSpace on 2014-06-11T19:27:14Z (GMT). No. of bitstreams: 0 Previous issue date: 2006-08-03Bitstream added on 2014-06-13T20:27:41Z : No. of bitstreams: 1 morais_ts_me_ilha.pdf: 1779620 bytes, checksum: b9fc2164652299690dfd01e30502deab (MD5) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / Neste trabalho desenvolveram-se metodologias de diagnóstico de falhas utilizando observadores de estado do tipo Filtro de Kalman, nas quais, as entradas para os observadores são identificadas utilizando as funções ortogonais de Fourier, Legendre e Chebyshev. Um tipo de observador denominado Proporcional-Integral é apresentado para a identificação de entradas desconhecidas. Este observador consegue estimar, sob certas condições, as entradas e ou distúrbios presentes no sistema e essas entradas são utilizadas para a diagnose de falha utilizando um observador do tipo Filtro de Kalman. Também é apresentado o desenvolvimento de uma metodologia de identificação de parâmetros bem como das forças de excitação, através das funções ortogonais, utilizando somente a resposta. Apresentam-se resultados obtidos por meio de simulações computacionais e realizados experimentalmente numa bancada de teste pertencente ao laboratório de vibrações mecânicas do Departamento de Engenharia Mecânica de Ilha Solteira. / In this work a methodology for fault diagnosis of mechanical systems was developed by using Kalman Filter state observes, in which the input of the observers are identified by using Fourier, Legendre and Chebyshev orthogonal functions. A proportional-integral observer is presented to the unknown input identification. This observer is able to find the unknown inputs of the system and these inputs are used to fault detection purposes by using a Kalman Filter Observer. The methodology for the identification of system parameters and excitation forces by using only the response of the system, through orthogonal functions. The methodology developed is applied to a mechanical structure containing vibrating tables, in the Mechanical Vibrations Laboratory, at Unesp, Ilha Solteira.

Made available in DSpace on 2014-12-17T14:55:03Z (GMT). No. of bitstreams: 1 RaphaelaGF.pdf: 1672960 bytes, checksum: 5b6b120f4026f9849183e5f96e363672 (MD5) Previous issue date: 2007-02-08 / ?This master dissertation presents the development of a fault detection and isolation system based in neural network. The system is composed of two parts: an identification subsystem and a classification subsystem. Both of the subsystems use neural network techniques with multilayer perceptron training algorithm. Two approaches for identifica-tion stage were analyzed. The fault classifier uses only residue signals from the identification subsystem. To validate the proposal we have done simulation and real experiments in a level system with two water reservoirs. Several faults were generated above this plant and the proposed fault detection system presented very acceptable behavior. In the end of this work we highlight the main difficulties found in real tests that do not exist when it works only with simulation environments / ?Esta disserta??o de mestrado apresenta o desenvolvimento de um sistema de detec??o e isolamento de falhas (DIF) baseado em redes neurais. O sistema ? dividido em duas etapas: uma de identifica??o neural do sistema e outra de detec??o e classifica??o de falhas. Ambos subsistemas usam t?cnicas de redes neurais com o algoritmoBackpropa- gation para redes Perceptronde M?ltiplas Camadas. Duas abordagens para identifica??o neural foram testadas e uma delas selecionada para fazer parte do sistema DIF. Oclassifi-cador de falhas utiliza apenas valores residuais para a classifica??o das mesmas. Todos os testes foram realizados tanto em ambiente simulado quanto em ambiente real, no intuito de comprovar dificuldades encontradas em testes reais n?o existentes quando se trabalha apenas com simula??es

Made available in DSpace on 2014-12-17T14:55:21Z (GMT). No. of bitstreams: 1 MarceloRBGV_TESE.pdf: 4018928 bytes, checksum: 7940c018115fd94c74a5dbbd7d3f7fb6 (MD5) Previous issue date: 2014-06-27 / The industries are getting more and more rigorous, when security is in question, no matter is to avoid financial damages due to accidents and low productivity, or when it s related to the environment protection. It was thinking about great world accidents around the world involving aircrafts and industrial process (nuclear, petrochemical and so on) that we decided to invest in systems that could detect fault and diagnosis (FDD) them. The FDD systems can avoid eventual fault helping man on the maintenance and exchange of defective equipments. Nowadays, the issues that involve detection, isolation, diagnose and the controlling of tolerance fault are gathering strength in the academic and industrial environment. It is based on this fact, in this work, we discuss the importance of techniques that can assist in the development of systems for Fault Detection and Diagnosis (FDD) and propose a hybrid method for FDD in dynamic systems. We present a brief history to contextualize the techniques used in working environments. The detection of fault in the proposed system is based on state observers in conjunction with other statistical techniques. The principal idea is to use the observer himself, in addition to serving as an analytical redundancy, in allowing the creation of a residue. This residue is used in FDD. A signature database assists in the identification of system faults, which based on the signatures derived from trend analysis of the residue signal and its difference, performs the classification of the faults based purely on a decision tree. This FDD system is tested and validated in two plants: a simulated plant with coupled tanks and didactic plant with industrial instrumentation. All collected results of those tests will be discussed / As ind?strias est?o cada vez mais rigorosas quando o assunto ? seguran?a, tanto para evitar preju?zos financeiros com acidentes e baixa produtividade, quanto para preservar o meio ambiente. Diante dos grandes acidentes em todo o mundo envolvendo aeronaves e processos industriais (nucleares, petroqu?micos etc) procurou-se investir em sistemas que pudessem detectar e diagnosticar falhas (FDD-Fault Detection and Diagnosis). Os sistemas FDD podem evitar eventuais falhas auxiliando o homem na manuten??o e troca de equipamentos defeituosos. Nos dias de hoje os assuntos que envolvem detec??o, isolamento, identifica??o e diagn?stico de falhas est?o ganhando for?a no meio acad?mico e industrial. Diante deste impulso, neste trabalho ser? discutido a import?ncia do estudo de t?cnicas que possam auxiliar o desenvolvimento de sistemas de detec??o e diagn?stico de falhas e proposto um m?todo h?brido para a detec??o e diagn?stico de falhas em sistemas din?micos. Um breve hist?rico ? apresentado de forma a contextualizar as t?cnicas utilizadas no trabalho. A detec??o de falhas pelo sistema proposto ? baseada em observadores de estado juntamente com outras t?cnicas estat?sticas. A ideia principal ? utilizar o pr?prio observador, para al?m de servir como redund?ncia anal?tica, permitir a cria??o de um res?duo que ser? utilizado na detec??o da falha e tamb?m no seu diagn?stico. Um banco de assinaturas auxiliar? o sistema de identifica??o de falhas, que, baseado nas assinaturas oriundas das an?lises de tend?ncia do sinal do res?duo e sua derivada, ir? realizar a classifica??o das falhas baseada em uma ?rvore de decis?o. Este sistema FDD ser? submetido a alguns testes e valida??es em duas plantas: uma planta simulada de tanques acoplados e em uma planta did?tica com instrumenta??o industrial. Os resultados colhidos desses ensaios se mostraram satisfat?rios para um grupo de falhas testadas e ser?o discutidos no decorrer do trabalho

Made available in DSpace on 2014-12-17T14:55:50Z (GMT). No. of bitstreams: 1 DiogoLR_DISSERT.pdf: 2166475 bytes, checksum: 3d70392d30b6b18a8ab0f79aa206e435 (MD5) Previous issue date: 2011-06-21 / In a real process, all used resources, whether physical or developed in software, are subject to interruptions or operational commitments. However, in situations in which operate critical systems, any kind of problem may bring big consequences. Knowing this, this paper aims to develop a system capable to detect the presence and indicate the types of failures that may occur in a process. For implementing and testing the proposed methodology, a coupled tank system was used as a study model case. The system should be developed to generate a set of signals that notify the process operator and that may be post-processed, enabling changes in control strategy or control parameters. Due to the damage risks involved with sensors, actuators and amplifiers of the real plant, the data set of the faults will be computationally generated and the results collected from numerical simulations of the process model. The system will be composed by structures with Artificial Neural Networks, trained in offline mode using Matlab? / Em um processo real, todos os recursos utilizados, sejam f?sicos ou desenvolvidos em software, est?o sujeitos a interrup??es ou a comprometimentos operacionais. Contudo, nas situa??es em que operam os sistemas cr?ticos, qualquer tipo de problema pode vir a trazer grandes consequ?ncias. Sabendo disso, este trabalho se prop?e a desenvolver um sistema capaz de detectar a presen?a e indicar os tipos de falhas que venham a ocorrer em um determinado processo. Para implementa??o e testes da metodologia proposta, um sistema de tanques acoplados foi escolhido como modelo de estudo de caso. O sistema desenvolvido dever? gerar um conjunto de sinais que notifiquem o operador do processo e que possam vir a ser p?s-processados, possibilitando que sejam feitas altera??es nas estrat?gias ou nos par?metros dos controladores. Em virtude dos riscos envolvidos com rela??o ? queima dos sensores, atuadores e amplificadores existentes na planta real, o conjunto de dados das falhas ser? gerado computacionalmente e os resultados coletados a partir de simula??es num?ricas do modelo do processo, n?o havendo risco de dano aos equipamentos. O sistema ser? composto por estruturas que fazem uso de Redes Neurais Artificiais, treinadas em modo offline pelo software matem?tico Matlab?