Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The majorities of industrial systems may be considered as dynamic hybrid systems, that is, described by variables in the continuous and discrete domains. Therefore, due to the need to obtain high quality products with losses minimization, to have safer and environmental friendly process operation, is what justies the development of tools for evaluating fault propagation eects. In this work, some tools for the representation of dynamic hybrid systems were studied in order to address the abnormal behavior propagation. The rst tools are Petri nets that allow a qualitative mathematical representation, the analysis of the models and supply useful information on the structure and the dynamic behavior of the system. Another tool is the signed directed graph - SDG that includes the mechanisms of fault propagation and the development of a theory of modeling based on the knowledge of complicated systems. In this context, it was presented the qualitative modeling of dynamic systems by using Petri nets and SDG, with the occurrence of faults and the evaluation of their propagation during process operation. These techniques were applied to Chemical Engineering problems such as mixing tanks and batch reactors for Petri nets, tank systems and faulty continuous reactors with the use of SDG. In those processes, simulations were performed and the faults eects were evaluated for several operational scenarios. SDGs were also used to determine the fault propagation path, with the evaluation of each aected process variable of interest both in open and closed-loop operation. / A maioria dos sistemas industriais pode ser classificada como pertencente _a classe de sistemas híbridos dinâmicos, ou seja, sistemas compostos por variáveis no domínio contínuo e discreto. Assim, devido à necessidade de se obter produtos de alta qualidade com mínimo de perdas possíveis, operar com segurança respeitando o meio ambiente, gerenciar ou tolerar falhas é que se tem a necessidade do desenvolvimento de ferramentas de investigação dos efeitos de falhas em processos. Neste trabalho foram estudadas algumas ferramentas para a representação de sistemas híbridos dinâmicos com a finalidade de representar o comportamento anormal desses sistemas. Uma dessas ferramentas _e a rede de Petri que permite a representação matemática, análise dos modelos e fornece informações úteis sobre a estrutura e o comportamento dinâmico dos sistemas modelados. Outra ferramenta estudada são os grafos orientados - SDG (Signed Directed Graph) que incluem os mecanismos de propagação de falhas e o desenvolvimento de uma teoria de modelagem baseada no conhecimento de sistemas. Nesse contexto, foram apresentados nesta dissertação a modelagem qualitativa dos sistemas discretos utilizando redes de Petri e SDG, com a ocorrência de falhas e a avaliação da sua propagação nos processos. Essas técnicas foram aplicadas em problemas simples da Engenharia Química como tanques de mistura e reator em batelada para redes de Petri e sistemas com um e três tanques e CSTR para o SDG. Em cada um dos processos estudados foram feitas as simulações acopladas aos modelos qualitativos em diversos cenários de falhas. Com isso comparam-se os resultados para o sistema operando sob diversos cenários e avaliam-se as informações do modelo qualitativo. Para ilustrar a aplicação das técnicas investigadas, esse trabalho avaliou, com a utilização de SDG, os caminhos de propagação das falhas, obtendo as variáveis afetadas em cada cenário operacional investigado. / Mestre em Engenharia Química

Proteger as linhas de transmissão é uma das tarefas mais importantes dentro dos sistemas elétricos de potência. Faltas em linhas de transmissão devem ser localizadas precisamente e extintas o mais rápido possível. Para tal, o esquema de proteção de linhas utiliza valores amostrados de correntes e tensões para a execução das tarefas de detecção, classificação e localização da falta. Neste esquema, grandezas trifásicas de corrente (IA, IB, IC) e tensão (VA, VB, VC) compõem as entradas do sistema. Após a detecção e classificação da falta, o relé efetua o cálculo da impedância aparente para a verificação da zona de proteção na qual a falta se insere (localização). Dentro deste contexto, a rápida detecção e a correta classificação da falta são passos fundamentais para a lógica de controle de um relé. Para a utilização de sistemas de proteção com alta velocidade de operação, o conjunto detector + classificador deve realizar uma decisão precisa do tipo de falta envolvida em menos de 10 ms após a ocorrência desta. Alguns métodos convencionais têm lidado com este problema, porém os tempos para estimação do tipo de falta são algumas vezes excessivamente longos. Este trabalho apresenta um novo sistema que provê uma rápida e confiável detecção e classificação de faltas através das medidas de valores de correntes trifásicas. O novo método utiliza-se da teoria de Redes Neurais Artificiais, baseada em dois diferentes tipos de redes (MLP e RBF), para a tarefa de detecção e classificação de faltas nos níveis de tempo requeridos para um moderno sistema de proteção. Um estudo comparativo em relação ao desempenho das redes mencionadas também foi realizado. Os testes efetuados para as redes dos tipos MLP e RBF mostraram que o sistema proposto foi capaz de detectar e classificar corretamente 100% dos casos estudados. Deve ainda ser ressaltado, que na maior parte dos casos (93% para a rede MLP e 84% para a rede RBF), o processo de detecção e classificação foi completado com no máximo 5 amostras de pós-falta (5ms). Isto demonstra a rapidez na tarefa de detecção e classificação embutida no método proposto, principalmente levando-se em consideração os tempos apresentados pelos métodos convencionais. / Transmission line protection is one of the major tasks for a power system. Transmission line faults must be located accurately and isolated as fast as possible. In order to perform this task, the power system protection system uses the three-phase currents (IA, IB, IC) and voltages (VA, VB, VC) to detect, classify and locate the fault. After detecting and classifying the fault, the relay calculates the apparent impedance to verify in which protection zone the fault is located. Taking this into account, precise and fast detection and classification methods are fundamental steps for the relay control algorithm. The combination detection + classification must carry out the correct response in less than 10 ms after the fault for a high-speed protection system. Some conventional methods have treated this problem but the time for a correct classification is sometimes excessively long. This work presents a fast and reliable new system for fault detection and classification using the three-phase current measurements. This new system is based on Artificial Neural Networks (RBF and MLP) for the detection and classification tasks. A comparative study involving both types of neural networks was done. Tests showed that the proposed system was able to correctly detect and classify 100% of the studied cases where the majority (93% of the cases for MLP net and 84% for RBF net) of them was done in up to 5 post-fault samples (5 ms). The afore-mentioned demonstrates the high speed of the new method for the detection and classification tasks when compared to the conventional ones.

Este trabalho apresenta o estado da arte das técnicas mais aplicadas para localização de faltas e modelagem de faltas de alta impedância e propõe a utilização de uma recente técnica baseada na decomposição dos sinais em componentes ortogonais. Este estudo avalia a aplicabilidade da técnica proposta utilizando dados reais de um sistema de distribuição de energia brasileiro, além de apresentar os conceitos teóricos sobre a decomposição em componentes ortogonais. / This work presents the state of the art of the most used techniques for locating and modelling high impedance faults and proposes the use of a recent technique based on the decomposition of the signals in orthogonal components. The objective of this study is to evaluate the application of the proposed technique using real data from a Brazilian distribution network, and presents the theory on orthogonal decomposition.

Les convertisseurs DC-DC suscitent un intérêt considérable en raison de leur puissance élevée et de leurs bonnes performances. Ils sont particulièrement utiles dans les systèmes multisources de production d'énergie électrique. Toutefois, en raison du grand nombre de composants sensibles utilisés dans ces circuits et comprenant des semi-conducteurs de puissance, des bobines et des condensateurs, une probabilité non négligeable de défaillance des composants doit être prise en compte. Cette thèse considère l'un des convertisseurs DC-DC les plus prometteurs - le convertisseur ZVS à pont isolé de type Buck. Une approche en deux étapes est présentée pour détecter et isoler les défauts en circuit ouvert dans les semi-conducteurs de puissance des convertisseurs DC-DC. La première étape concerne la détection et la localisation des défauts dans un convertisseur donne. La seconde étape concerne sur les systèmes munis de plusieurs convertisseurs DC-DC. Les méthodes proposées sont basées sur les réseaux Bayesiens (BBN). Les signaux utilisés dans ces méthodes sont ceux des entrées de mesure du système de commande et aucune mesure supplémentaire n'est requise. Un convertisseur expérimental ZVS à pont isolé de type Buck a été conçu et construit pour valider la détection et la localisation des défauts Sur un seul convertisseur. Ces méthodes peuvent être étendues à d'autres types de convertisseurs DC-DC. / DC-DC converters have received significant interest recently as a result of their high power capabilities and good power quality. They are of particular interest in systems with multiple sources of energy. However due to the large number of sensitive components including power semiconductor devices, coils, and capacitors used in such circuits there is a high likelihood of component failure. This thesis considers one of the most promising DC-DC converters—the ZVS full bridge isolated Buck converter. An approach with two stages is presented to detect and isolate opencircuit faults in the power semiconductor devices in systems with DC-DC converters. The first stage is the fault detection and isolation for a single DC-DC converter, while the second stage works on a system with multiple DC-DC converters. The proposed methods are based on Bayesian Belief Network (BBN). The signals used in the proposed methods are already available as measurement inputs to control system and no additional measurements are required. An experimental ZVS full bridge isolated Buck converter has been designed and built to validate the fault detection and isolation method on a single converter. The methods can be used with other DC-DC converter typologies employing similar analysis and principals.

This work takes place in the context of distribution power system protection and tries to improve the detection and location of earth faults. The protection problem is vast and many ideas emerge every year to enhance the reliability of the grid. The author has focused his energy into the compensated and isolated network protection in the specific case of single phase earth fault. This PhD thesis is divided in two main parts that might be considered as independent. The first part studies the detection of single phase earth fault and the second analyzes the fault location of such fault.<p><p>Pragmatism was asked during these three years because a product development was necessary especially regarding the fault detection problem. The first part of the thesis took 18 months of research and development to obtain a prototype of transient protection able to detect single phase earth fault in compensated and isolated network. The sensitivity of the algorithm has been emphasized regarding the fault impedance and to detect earth fault up to 5 kOhm depending on the network characteristic. The fault location problem has been much more theoretical although the problem links to the accuracy of the algorithm and its robustness regarding wrong fault location indication has been strongly considered.<p><p>Compensated networks and in some conditions isolated networks are distribution from 12 kV up to 110 kV mostly used in East and North Europe but also in China. Others areas also work with such networks but they also have others systems and they do not use them on all the territory. These networks have the particularity to obtain very small fault current in case of single phase earth fault. Low current means the difference between a faulty and a sound feeder is not significant. Therefore classic overcurrent protection is completely useless to protect the network, forcing the development of more complex algorithm. A possibility to overcome the problem of the small fault current is to develop a transient protection. The transient occurring at the beginning of the fault has strong information to distinguish a faulty from a sound feeder. In this work I have chosen to use not only the transient but also the steady state to get the best sensitivity.<p><p>Then the fault location has been investigated but the small information coming from the faulty feeder is not sufficient to have a precise enough position of the fault. Therefore, active system has been suggested to be implemented in the grid to increase the faulty current and have enough power for a precise location. Different existing algorithms based on the steady state at the nominal frequency are compared using a tool developed during this work. Recommendations are then made depending on the topology, the network parameters, the measurements precision, etc. Due to the complexities of the problem, a simulator has been coded in Matlab .The user of a possible fault location must then use this tool to understand and see the future fault location precision that he could obtain from different algorithm on his network. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Abstract<p><p>This thesis deals with the operation of and winding short-circuit fault detection in a Doubly-Fed Induction Generator (DFIG) based Wind Turbine Generator System (WTGS). Both the faulted and faultless condition of operation has been studied, where the focus is on the electrical part of the system. The modelled electrical system is first simulated and the developed control system is then validated on a test bench. The test-bench component dimensioning is also discussed.<p><p>The faultless condition deals with the start-up and power production mode of operation. Control design based on the Proportional Integral (PI) control technique has been compared for power and torque control strategies against the Linear Quadratic Gaussian (LQG) control technique, at different operating points through the variable-speed region of WTGS operation following the maximum power curve of the system. It was found that the torque control strategy offered less degradation in performance for both the control techniques at operating points different for the one for which the control system was tuned. The start-up procedure of the DFIG based WTGS has been clarified and simplified. The phase difference between the stator and the grid voltage, which occurs due to the arbitrary rotor position when the rotor current control is activated, is minimized by using a sample-and-hold technique which eliminates the requirement of designing an additional controller. This method has been validated both in simulation and experiments.<p><p>The faulted condition of operation deals with the turn-turn short-circuit fault in the phase winding of the generator. The model of the generator, implemented using the winding-function approach, allows the fault to be created online both in a stator and a rotor phase. It has been demonstrated that the magnitude of the current harmonics, used extensively in literature for the Machine Current Signature Analysis (MCSA) technique for winding short-circuit fault detection, is very different when the location of the fault is changed to another coil within the phase winding. This makes the decision on the threshold selection for alarm generation difficult. Furthermore, the control system attenuates the current harmonics by an order of magnitude. This attenuation property is also demonstrated through experiments. The attention is then shifted to the negative-sequence current component, resulting from the winding unbalance, as a possible fault residual. Its suitability is tested in the presence of noise for scenarios with different fault locations, fault severity in terms of the number of shorted-turns and grid voltage unbalance. It is found that due to the presence of a control system the magnitude of the negative-sequence current, resulting from the fault, remains almost the same for all fault locations and fault severity. Thus, it was deemed more suitable as a fault residual. In order to obtain a fast detection method, the Cumulative Sum (CUSUM) algorithm was used. The test function is compared against a threshold, determined on the basis of expected residual magnitude and the time selected for detection, to generate an alarm. The validation is carried out with noise characteristics different from the ones used during the design and it is shown that the voltage unbalance alone is not able to trigger a false alarm. In all the scenarios considered, the detection was achieved within 40 ms despite the presence of measurement filters. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

In this dissertation, we study two new approaches to fault detection for autonomous robots. The first approach involves the synthesis of software components that give a robot the capacity to detect faults which occur in itself. Our hypothesis is that hardware faults change the flow of sensory data and the actions performed by the control program. By detecting these changes, the presence of faults can be inferred. In order to test our hypothesis, we collect data in three different tasks performed by real robots. During a number of training runs, we record sensory data from the robots both while they are operating normally and after a fault has been injected. We use back-propagation neural networks to synthesize fault detection components based on the data collected in the training runs. We evaluate the performance of the trained fault detectors in terms of the number of false positives and the time it takes to detect a fault.<p>The results show that good fault detectors can be obtained. We extend the set of possible faults and go on to show that a single fault detector can be trained to detect several faults in both a robot's sensors and actuators. We show that fault detectors can be synthesized that are robust to variations in the task. Finally, we show how a fault detector can be trained to allow one robot to detect faults that occur in another robot.<p><p>The second approach involves the use of firefly-inspired synchronization to allow the presence of faulty robots to be determined by other non-faulty robots in a swarm robotic system. We take inspiration from the synchronized flashing behavior observed in some species of fireflies. Each robot flashes by lighting up its on-board red LEDs and neighboring robots are driven to flash in synchrony. The robots always interpret the absence of flashing by a particular robot as an indication that the robot has a fault. A faulty robot can stop flashing periodically for one of two reasons. The fault itself can render the robot unable to flash periodically.<p>Alternatively, the faulty robot might be able to detect the fault itself using endogenous fault detection and decide to stop flashing.<p>Thus, catastrophic faults in a robot can be directly detected by its peers, while the presence of less serious faults can be detected by the faulty robot itself, and actively communicated to neighboring robots. We explore the performance of the proposed algorithm both on a real world swarm robotic system and in simulation. We show that failed robots are detected correctly and in a timely manner, and we show that a system composed of robots with simulated self-repair capabilities can survive relatively high failure rates.<p><p>We conclude that i) fault injection and learning can give robots the capacity to detect faults that occur in themselves, and that ii) firefly-inspired synchronization can enable robots in a swarm robotic system to detect and communicate faults.<p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

[pt] Mesmo sistemas de software desenvolvidos com um controle de qualidade rigoroso podem apresentar falhas durante seu ciclo de vida. Quando uma falha é observada no ambiente de produção, mantenedores são responsáveis por produzir o diagnóstico e remover o seu defeito correspondente. No entanto, em um serviço crítico este tempo pode ser muito longo, logo, se for possível, a assinatura da falha deve ser utilizada para gerar um mecanismo de recuperação automático capaz de detectar e tratar futuras ocorrências similares, até que o defeito possa ser removido. Nesta tese, a atividade de recuperação consiste em restaurar o sistema para um estado correto, que permita continuar a execução com segurança, ainda que com limitações em suas funcionalidades. Para serem eficazes, as tarefas de diagnóstico e recuperação requerem informações detalhadas sobre a execução que falhou. Falhas que ocorrem durante a fase de testes em um ambiente controlado podem ser depuradas através da inserção de nova instrumentação e re-execução da rotina que contem o defeito, tornando mais fácil o estudo de comportamentos inesperados. No entanto, falhas que ocorrem no ambiente de produção apresentam informações limitadas à situação específica em que ocorrem, além de serem imprevisíveis. Para mitigar essa adversidade, informações devem ser coletadas sistematicamente com o intuito de detectar, diagnosticar para recuperar e, eventualmente, diagnosticar para remover a circunstância geradora da falha. Além disso, há um balanceamento entre a informação inserida como instrumentação e a performance do sistema: técnicas de logging geralmente apresentam baixo impacto no desempenho, porém não provêm informação suficiente sobre a execução; por outro lado, as técnicas de tracing podem registrar informações precisas e detalhadas, todavia são impraticáveis para um ambiente de produção. Esta tese propõe uma abordagem hibrida para gravação e extração de informações durante a execução do sistema. A solução proposta se baseia no registro de eventos, onde estes são enriquecidos com propriedades contextuais sobre o estado atual da execução no momento em que o evento é gravado. Através deste registro de eventos com informações de contexto, uma técnica de diagnóstico e uma ferramenta foram desenvolvidas para permitir que eventos pudessem ser filtrados com base na perspectiva de interesse do mantenedor. Além disso, também foi desenvolvida uma abordagem que utiliza estes eventos enriquecidos para detectar falhas automaticamente visando recuperação. As soluções propostas foram avaliadas através de medições e estudos conduzidos em sistemas implantados, baseando-se nas falhas que de fato ocorreram enquanto se utilizava o software em um contexto de produção. / [en] Even software systems developed with strict quality control may expect failures during their lifetime. When a failure is observed in a production environment the maintainer is responsible for diagnosing the cause and eventually removing it. However, considering a critical service this might demand too long a time to complete, hence, if possible, the failure signature should be identified in order to generate a recovery mechanism to automatically detect and handle future occurrences until a proper correction can be made. In this thesis, recovery consists of restoring a correct context allowing dependable execution, even if the causing fault is still unknown. To be effective, the tasks of diagnosing and recovery implementation require detailed information about the failed execution. Failures that occur during the test phase run in a controlled environment, allow adding specific code instrumentation and usually can be replicated, making it easier to study the unexpected behavior. However, failures that occur in the production environment are limited to the information present in the first occurrence of the failure. But run time failures are obviously unexpected, hence run time data must be gathered systematically to allow detecting, diagnosing with the purpose of recovering, and eventually diagnosing with the purpose of removing the causing fault. Thus there is a balance between the detail of information inserted as instrumentation and the system performance: standard logging techniques usually present low impact on performance, but carry insufficient information about the execution; while tracing techniques can record precise and detailed information, however are impracticable for a production environment. This thesis proposes a novel hybrid approach for recording and extracting system s runtime information. The solution is based on event logs, where events are enriched with contextual properties about the current state of the execution at the moment the event is recorded. Using these enriched log events a diagnosis technique and a tool have been developed to allow event filtering based on the maintainer s perspective of interest. Furthermore, an approach using these enriched events has been developed that allows detecting and diagnosing failures aiming at recovery. The proposed solutions were evaluated through measurements and studies conducted using deployed systems, based on failures that actually occurred while using the software in a production context.

Automated start–up and shutdown procedures increase the profitability and safety of a process, but are difficult to implement due to the complex nature of the concepts that must be incorporated. Generic components used specifically for the implementation of automated startup and shutdown procedures were defined to streamline the implementation process. The generic components developed are based on Sequential Function Charts and were applied to the startup of a fixed–bed gasification unit, for which a dynamic simulation model was developed. The application showed that the automated startup can be defined by a few generic components and that the flexibility of the startup procedure is increased through the incorporation of a fault accommodation module. The use of a visual–based definition of sequential processes increases the understanding of the complex scheduling procedures as well as the efficiency of the development of these automated procedures. In addition, iterative learning was incorporated into the generic definition to optimise controller performance during the non–linear phases of operation. / Dissertation (MEng (Control Engineering))--University of Pretoria, 2006. / Chemical Engineering / MEng / Unrestricted

Orientador: Robson Pederiva / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-07-23T22:28:49Z (GMT). No. of bitstreams: 1 Chiarello_AndreGarcia_D.pdf: 7903318 bytes, checksum: cd3f513a26e60d59ef55edbb6453d942 (MD5) Previous issue date: 1998 / Resumo: Este trabalho aborda o problema de detecção e localização de falhas em sistemas mecânicos através de modelos matemáticos. Considera-se que o sistema mecânico seja representado por um modelo dinâmico na forma de variáveis de estado, com parâmetros constantes e entradas estacionárias. Definindo-se, apropriadamente, funções de correlação e explorando as propriedades inerentes aos sistemas estacionários, consegue-se obter funções analíticas envolvendo parâmetros físicos do sistema. Posteriormente, demonstra-se que estas funções analíticas podem ser utilizadas para monitorar parâmetros físicos do modelo, ou ainda, identificar parâmetros físicos relacionados a uma falha no sistema. Duas abordagens diferentes para o problema de localização de falhas são desenvolvidas: uma utiliza a estimação de parâmetros e outra utiliza funções de resíduo. Visando uma aplicação em sistemas reais, foram simulados sistemas mecânicos em diversas condições de operação. Os resultados numéricos comprovaram que a metodologia proposta pode ser utilizada no monitoramento de sistemas mecânicos / Abstract: This work deals with the problem of fault detection and location in mechanical systems through mathematical models. It is considered that the mechanical system is represented in a state space form, with constant parameters and stationary inputs. Some correlation functions are appropriately defined involving physical parameters of the stationary system. It is demonstrated that these correlation functions can be transformed to appropriate analytical form (redundancy equations) for fault detection and isolation purposes. Two different approaches are considered: one based on physical parameters estimation and the other based on residuaIs generation. Three different mechanical systems were numerically simulated with conditions (or restrictions) usually found in practical situations. The results showed that the proposed approach is a promising methodology for monitoring stationary mechanical systems / Doutorado / Projeto Mecanico / Doutor em Engenharia Mecânica