The thesis addresses the problem of structural damage detection using inverse sensitivity based methods. The focus here is on characterization with regard to identification, location, and, quantification of structural damage in linear time invariant (LTI) systems, using vibration data. The study encompasses both analytical and experimental methods. A suite of five algorithms for damage detection, namely, inverse eigensensitivity method that is refined to account for cross orthogonality between distinct modes, damping dependent eigensolutions, and sensitivity with respect to points of antiresonance and minima, inverse FRF method that includes refinements in terms of inclusion of second order sensitivity, response function method (RFM) based on first order Taylor’s expansion, a newly proposed inverse sensitivity method based on singular values of FRF matrix, and method based on response time histories, are presented. The scope of these methods vis-à-vis the need for model reduction, ability to deal with incomplete data, ill-posedness of governing equations and the need for regularization, sensitivity with respect to measurement noise, ability to identify damping characteristics, the highest and lowest magnitudes of changes in structural properties, and the ability to characterize systems with closely spaced natural frequencies that the methods can detect are discussed. The performance of proposed procedures is illustrated by considering a five degrees-of-freedom (dof) mass-spring-dashpot system and subsequently applied on three archetypal structural systems using analytical and experimental methods. In the examples presented, factors, such as, completeness of measured data in time and frequency, nature (proportional/non-proportional) and magnitude of damping, levels of changes in structural properties, modal truncations, number of governing equations for system parameters, and efficacy of regularization techniques are investigated. The study also highlights the difficulties in implementing the damage detection algorithm based on real life noisy vibration data. A comparative study on the suitability of each of these methods in locating and quantifying of different damage scenarios has been reported. A critical review of performance of the various methods is presented. The thesis concludes with a summary on the contributions made and also deliberates on future avenues for research and development in this area of research.

The research work presented in this thesis concern to the stability analysis of linear time-delay systems with low-order controllers. This thesis is divided into three parts.The first part of the thesis focus on the study of linear SISO (single-input/single-output) systems with input/output delays, where the feedback loop is closed with a controller of PID-type. Inspired by the geometrical approach developed by Gu et al. we propose an analytical method to find the stability regions of all stabilizing controllers of PID-type for the time-delay system. Based on this same approach, we propose an algorithm to calculate the degree of fragility of a given controller of PID- type (PI, PD and PID).The second part of the thesis focuses on the stability analysis of linear systems under an NCS (Networked System Control) based approach. More precisely, we first focus in the stabilization problem by taking into account the induced network delays and the effects induced by the sampling period. To carry out such an analysis we have adopted an eigenvalue perturbation-based approach.Finally, in the third part of the thesis we tackle certain problems concerning to the behavior of the zeros of a certain class of sampled-data SISO systems. More precisely, given a continuous-time system, we obtain the sampling intervals guaranteeing the invariance of the number of unstable zeros in each interval. To perform such an analysis, we adopt an eigenvalue perturbation-based approach.

This Thesis is concerned with the approximation of high order controllers or the controller reduction problem. We firstly consider approximating high-order controllers by low order controllers based on the closed-loop system approximation. By approximating the closed-loop system transfer function, we derive a new parametrized double-sided frequency weighted model reduction problem. The formulas for the input and output weights are derived using three closed-loop system configurations: (i) by placing a controller in cascade with the plant, (ii) by placing a controller in the feedback path, and (iii) by using the linear fractional transformation (LFT) representation. One of the weights will be a function of a free parameter which can be varied in the resultant frequency weighted model reduction problem. We show that by using standard frequency weighted model reduction techniques, the approximation error can be easily reduced by varying the free parameter to give more accurate low order controllers. A method for choosing the free parameter to get optimal results is being suggested. A number of practical examples are used to show the effectiveness of the proposed controller reduction method. We have then considered the relationships between the closed-loop system con gurations which can be expressed using a classical control block diagram or a modern control block diagram (LFT). Formulas are derived to convert a closed-loop system represented by a classical control block diagram to a closed-loop system represented by a modern control block diagram and vice versa.

Orientador: Pedro Luis Dias Peres / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-13T19:35:14Z (GMT). No. of bitstreams: 1 Borges_RenatoAlves_D.pdf: 1110575 bytes, checksum: 60707903258309c73eb16c6b81e8e7ba (MD5) Previous issue date: 2009 / Resumo: A principal contribuição desta tese é a proposta de condições de sintese de filtros e controladores lineares, tanto robustos quanto dependentes de parametros, para sistemas discretos variantes no tempo.Os controladores, ou filtros, são obtidos solucionando problemas de otimização formulados em termos de desigualdades matriciais bililineares, por meio de um metodo que se baseia na alternancia de problemas convexos descritos por desigualdades matriciais lineares. Para obtenção das condiçoes de sintese foram utilizadas tanto funções de Lyapunov afins nos parâmetros quanto ametros, alem de variáveis multi-afins em diferentes instantes de tempo dos parâmetros, alem de variaveis extras introduzidas pelo lema de Finsler. Nesse contexto, sao tratados problemas de sintese com custo garantido H, assegurando robustez em relação a incertezas não estruturadas. Simulaçoes numéricas ilustram a eficiencia dos metodos propostos em termos de desempenho H quando comparados com outros metodos da literatura / Abstract: The main contribution of this dissertation is to propose conditions for linear filter and controller design, considering both robust and parameter dependent structures, for discrete time-varying systems. The controllers, or filters, are obtained through the solution of optimization problems, formulated in terms of bilinear matrix inequalities, using a method that alternates convex optimization problems described in terms of linear matrix inequalities. Both affine and multi-affine in different instants of time (path dependent)Lyapunov functions were usedto obtain the design conditions, as wellas extra variables introduced bythe Finsler's lemma.Design problems that take into account an H guaranteed cost were investigated, providing robustness with respect to unstructured uncertainties. Numerical simulations show the effciency of the proposed methods in terms of H performance when compared with other strategies from the literature / Doutorado / Automação / Doutor em Engenharia Elétrica

Orientadores: Pedro Luis Dias Peres, Ricardo Coração de Leão Fontoura de Oliveira / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-25T15:07:40Z (GMT). No. of bitstreams: 1 Lacerda_MarcioJunior_D.pdf: 1206868 bytes, checksum: c5abb83323581461eb40c95f27c95097 (MD5) Previous issue date: 2014 / Resumo: Este trabalho apresenta novas condições na forma de desigualdades matriciais lineares para o projeto de filtros H-infinito de ordem completa em três diferentes contextos: i) sistemas lineares incertos discretos com um atraso variante no tempo afetando os estados; ii) sistemas lineares com parâmetros variantes no tempo, contínuos e discretos, sujeitos a incertezas nas medições dos parâmetros; iii) sistemas não lineares quadráticos contínuos e discretos no tempo. Para cada contexto, o objetivo é projetar filtros: i) com termos atrasados nos estados; ii) dependentes dos parâmetros incertos medidos; iii) com termos quadráticos. Em cada um dos casos, o ponto de partida é a existência de uma função de Lyapunov que assegure estabilidade e um limitante para a norma H-infinito do sistema aumentado, ou seja, o sistema original conectado com o filtro de ordem completa. As condições de projeto são obtidas impondo-se uma determinada estrutura para as variáveis de folga, resultando em desigualdades matriciais com parâmetros escalares. A eficácia das condições apresentadas é ilustrada por meio de comparações numéricas utilizando exemplos da literatura / Abstract: This work presents new conditions in the form of linear matrix inequalities for full order H-infinity filter design in three different contexts: i) uncertain linear discrete-time systems with a time-varying delay affecting the states ii) linear parameter-varying systems, continuous and discrete-time, subject to inexactly measured parameters; iii) continuous and discrete-time nonlinear quadratic systems. For each context, the aim is to design filters: i) with state-delayed terms; ii) dependent upon the inexactly measured parameters; iii) with quadratic terms. In each case, the starting point is the existence of a Lyapunov function that assures stability and a bound to the H-infinity norm of the augmented system, that is, the original system conected with the full order filter. The design conditions are obtained by imposing a given structure to the slack variables, resulting in matrix inequalities with scalar parameters. The effectiveness of the proposed conditions is illustrated by means of numerical comparisons and benchmark examples from the literature / Doutorado / Automação / Doutor em Engenharia Elétrica

Orientador: Gilmar Barreto / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-26T06:42:54Z (GMT). No. of bitstreams: 1 Battaglin_PauloDavid_D.pdf: 3180685 bytes, checksum: 5e1e9893bb97a4df42116a4c0d8b10d6 (MD5) Previous issue date: 2014 / Resumo: Este trabalho mostrará a aplicação do filtro de Kalman a um processo computacional discreto, o qual será representado por um modelo matemático que é um sistema de equações lineares, multivariáveis, discretas, estocásticas e variantes no tempo. As contribuições desta pesquisa evidenciam a construção de um modelo matemático apropriado de observabilidade instantânea para representar sistemas que variam rapidamente no tempo; a construção dos fundamentos teóricos do filtro de Kalman a ser aplicado em sistemas lineares, multivariáveis, discretos, estocásticos e variantes no tempo; bem como a construção deste filtro neste contexto e sua aplicação a um processo computacional discreto. Neste trabalho propomos um método para determinar: a matriz de observabilidade instantânea, o vetor de estimação de estado interno, a matriz de covariâncias de erros de estimação de estado interno e a latência de um processo computacional discreto, quando as medidas na saída do computador são conhecidas. Aqui mostramos que quando a propriedade observabilidade instantânea do sistema é verificada, a latência de um processo computacional pode ser estimada. Esta é uma vantagem comparada com os métodos de observabilidade usual, os quais são baseados em cenários estáticos. A aplicação potencial dos resultados deste trabalho é na predição de congestionamentos em processos que variam no tempo e acontecem em computadores digitais. Em uma perspectiva mais ampla, o método da observabilidade instantânea pode ser aplicado na identificação de patologias, na previsão de tempo, em navegação e rastreamento no solo, na água e no ar; no mercado de ações e em muitas outras áreas / Abstract: This work will show the application of the Kalman filter to a discrete computational process, which will be represented by a mathematical model: a system of linear, multivariable, discrete, stochastic and time-varying equations. The contributions of this research show the construction of an appropriate mathematical model of instantaneous observability to represent systems that vary quickly in time; the construction of the theoretical foundations of the Kalman filter to be applied to a linear, multivariable, discrete, stochastic and time-varying system; the construction of this filter in this context and its application to a discrete computational process. In this research we propose a method to determine: the instantaneous observability matrix, the internal state vector estimation, Covariance matrix of internal state estimation error and the latency of a digital computational process, when the measures on the computer output are known. Here we show that when the instantaneous observability property of the system comes true, a computing process latency can be estimated. This is an advantage compared to usual observability methods, which are based on static scenarios. The potential application of the results of this work is to predict bottlenecks in time-varying processes which happen inside the discrete computers. In a broader perspective, the instantaneous observability method can be applied on identification of a pathology, weather forecast, navigation and tracking on ground, in the water and in the air; in stock market prediction and many other areas / Doutorado / Automação / Doutor em Engenharia Elétrica

Orientador: Juan Francisco Camino dos Santos / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-22T01:28:47Z (GMT). No. of bitstreams: 1 Rodrigues_LuisAntonio_M.pdf: 1882720 bytes, checksum: efb662d3473168965c88de41357ef029 (MD5) Previous issue date: 2012 / Resumo: O presente trabalho investiga os problemas de estabilidade assintótica e desempenho H'INFINITO' de sistemas lineares a parâmetros variantes discretos no tempo. São fornecidas condições suficientes para análise de estabilidade, análise de desempenho H'INFINITO' e síntese de controladores estáticos de realimentação de saída robustos e por ganho escalonado. Além disso, é proposto um método de parametrização polinomial homogênea de sistemas LPV afins. Assume-se que a matriz do sistema tem dependência polinomial homogênea de grau arbitrário sobre os parâmetros que variam dentro de um politopo com conhecidos limitantes sobre suas taxas de variação. As propriedades geométricas do domínio politópico são exploradas para se obter um conjunto finito de desigualdades matriciais lineares que levam em consideração os limitantes sobre as taxas de variação dos parâmetros. As condições LMIs são obtidas usando uma função de Lyapunov quadrática nos estados com dependência polinomial homogênea dos parâmetros variantes em instantes sucessivos de tempo. As condições fornecidas são aplicadas no modelo LPV de um sistema vibroacústico. Comparações com resultados numéricos encontrados na literatura mostram os benefícios das técnicas propostas / Abstract: This work investigates stability and H'INFINITE' performance of discrete-time linear parameter varying systems. Sufficient conditions for stability analysis, H'INFINITE' performance analysis and synthesis of both robust and gain-scheduled static output feedback controller are provided. It is assumed that the system matrices have a homogeneous polynomial dependence of arbitrary degree on the time-varying parameters. Thus, a homogeneous-polynomially parametrization method for affine LPV systems is proposed. The parameters are assumed to vary inside a polytope and to have known bounds on their rates of variation. The geometric properties of the polytopic domain are exploited to derive a finite set of LMIs that take into account the bounds on the rates of variation of the scheduling parameters. The LMI conditions are obtained using a quadratic in the state Lyapunov function with a homogeneous polynomial dependence on the scheduling parameters at successive instants of time. The proposed techniques are applied to an LPV model of a vibroacoustic setup. Comparisons with numerical results found in literature show the benefits of the proposed approach / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica

Cette thèse porte sur l’analyse de graphes dynamiques, définis en temps discret ou continu. Nous introduisons une nouvelle extension dynamique du modèle a blocs stochastiques (SBM), appelée dSBM, qui utilise des processus de Poisson non homogènes pour modéliser les interactions parmi les paires de nœuds d’un graphe dynamique. Les fonctions d’intensité des processus ne dépendent que des classes des nœuds comme dans SBM. De plus, ces fonctions d’intensité ont des propriétés de régularité sur des intervalles temporels qui sont à estimer, et à l’intérieur desquels les processus de Poisson redeviennent homogènes. Un récent algorithme d’estimation pour SBM, qui repose sur la maximisation d’un critère exact (ICL exacte) est ici adopté pour estimer les paramètres de dSBM et sélectionner simultanément le modèle optimal. Ensuite, un algorithme exact pour la détection de rupture dans les séries temporelles, la méthode «pruned exact linear time» (PELT), est étendu pour faire de la détection de rupture dans des données de graphe dynamique selon le modèle dSBM. Enfin, le modèle dSBM est étendu ultérieurement pour faire de l’analyse de réseau textuel dynamique. Les réseaux sociaux sont un exemple de réseaux textuels: les acteurs s’échangent des documents (posts, tweets, etc.) dont le contenu textuel peut être utilisé pour faire de la classification et détecter la structure temporelle du graphe dynamique. Le modèle que nous introduisons est appelé «dynamic stochastic topic block model» (dSTBM). / This thesis focuses on the statistical analysis of dynamic graphs, both defined in discrete or continuous time. We introduce a new extension of the stochastic block model (SBM) for dynamic graphs. The proposed approach, called dSBM, adopts non homogeneous Poisson processes to model the interaction times between pairs of nodes in dynamic graphs, either in discrete or continuous time. The intensity functions of the processes only depend on the node clusters, in a block modelling perspective. Moreover, all the intensity functions share some regularity properties on hidden time intervals that need to be estimated. A recent estimation algorithm for SBM, based on the greedy maximization of an exact criterion (exact ICL) is adopted for inference and model selection in dSBM. Moreover, an exact algorithm for change point detection in time series, the "pruned exact linear time" (PELT) method is extended to deal with dynamic graph data modelled via dSBM. The approach we propose can be used for change point analysis in graph data. Finally, a further extension of dSBM is developed to analyse dynamic net- works with textual edges (like social networks, for instance). In this context, the graph edges are associated with documents exchanged between the corresponding vertices. The textual content of the documents can provide additional information about the dynamic graph topological structure. The new model we propose is called "dynamic stochastic topic block model" (dSTBM).Graphs are mathematical structures very suitable to model interactions between objects or actors of interest. Several real networks such as communication networks, financial transaction networks, mobile telephone networks and social networks (Facebook, Linkedin, etc.) can be modelled via graphs. When observing a network, the time variable comes into play in two different ways: we can study the time dates at which the interactions occur and/or the interaction time spans. This thesis only focuses on the first time dimension and each interaction is assumed to be instantaneous, for simplicity. Hence, the network evolution is given by the interaction time dates only. In this framework, graphs can be used in two different ways to model networks. Discrete time […] Continuous time […]. In this thesis both these perspectives are adopted, alternatively. We consider new unsupervised methods to cluster the vertices of a graph into groups of homogeneous connection profiles. In this manuscript, the node groups are assumed to be time invariant to avoid possible identifiability issues. Moreover, the approaches that we propose aim to detect structural changes in the way the node clusters interact with each other. The building block of this thesis is the stochastic block model (SBM), a probabilistic approach initially used in social sciences. The standard SBM assumes that the nodes of a graph belong to hidden (disjoint) clusters and that the probability of observing an edge between two nodes only depends on their clusters. Since no further assumption is made on the connection probabilities, SBM is a very flexible model able to detect different network topologies (hubs, stars, communities, etc.).

Dans le contexte des séries temporelles linéaires, on étudie les processus strictement stationnaires dits anticipatifs dépendant potentiellement de tous les termes d'une suite d'erreurs alpha-stables indépendantes et identiquement distribuées.On considère en premier lieu les processus autoregressifs (AR) et l'on montre que des moments conditionnels d'ordres plus élevés que les moments marginaux existent dès lors que le polynôme caractéristique admet au moins une racine à l'intérieur du cercle unité.Des formules fermées sont obtenues pour les moments d'ordre un et deux dans des cas particuliers.On montre que la méthode des moindres carrés permet d'estimer une représentation all-pass causale du processus dont la validité peut être vérifiée par un test de type portmanteau, et l'on propose une méthode fondée sur des propriétés d'extreme clustering pour retrouver la représentation AR originale.L'AR(1) stable anticipatif est étudié en détails dans le cadre des vecteurs stables bivariés et des formes fonctionnelles pour les quatre premiers moments conditionnels sont obtenues pour toute paramétrisation admissible.Lors des évènements extrêmes, il est montré que ces moments deviennent équivalents à ceux d'une distribution de Bernoulli chargeant deux évolutions futures opposées: accroissement exponentiel ou retour aux valeurs centrales.Des résultats parallèles sont obtenus pour l'analogue de l'AR(1) en temps continu, le processus d'Ornstein-Uhlenbeck stable anticipatif.Pour des moyennes mobiles alpha-stables infinies, la distribution conditionnelle des chemins futurs sachant la trajectoire passée est obtenue lors des évènements extrêmes par le biais d'une nouvelle représentation des vecteurs stables multivariés sur des cylindres unités relatifs à des semi-normes.Contrairement aux normes, ce type de représentation donne lieu à une propriété de variations régulières des queues de distribution utilisable dans un contexte de prévision, mais tout vecteur stable n'admet pas une telle représentation. Une caractérisation est donnée et l'on montre qu'un chemin fini de moyenne mobile alpha-stable sera représentable pourvu que le processus soit "suffisamment anticipatif".L'approche s'étend aux processus résultant de la combinaison linéaire de moyennes mobiles alpha-stables, et la distribution conditionnelle des chemins futurs s'interprète naturellement en termes de reconnaissance de formes. / In the framework of linear time series analysis, we study a class of so-called anticipative strictly stationary processes potentially depending on all the terms of an independent and identically distributed alpha-stable errors sequence.Focusing first on autoregressive (AR) processes, it is shown that higher order conditional moments than marginal ones exist provided the characteristic polynomials admits at least one root inside the unit circle. The forms of the first and second order moments are obtained in special cases.The least squares method is shown to provide a consistent estimator of an all-pass causal representation of the process, the validity of which can be tested by a portmanteau-type test. A method based on extreme residuals clustering is proposed to determine the original AR representation.The anticipative stable AR(1) is studied in details in the framework of bivariate alpha-stable random vectors and the functional forms of its first four conditional moments are obtained under any admissible parameterisation.It is shown that during extreme events, these moments become equivalent to those of a two-point distribution charging two polarly-opposite future paths: exponential growth or collapse.Parallel results are obtained for the continuous time counterpart of the AR(1), the anticipative stable Ornstein-Uhlenbeck process.For infinite alpha-stable moving averages, the conditional distribution of future paths given the observed past trajectory during extreme events is derived on the basis of a new representation of stable random vectors on unit cylinders relative to semi-norms.Contrary to the case of norms, such representation yield a multivariate regularly varying tails property appropriate for prediction purposes, but not all stable vectors admit such a representation.A characterisation is provided and it is shown that finite length paths of a stable moving average admit such representation provided the process is "anticipative enough".Processes resulting from the linear combination of stable moving averages are encompassed, and the conditional distribution has a natural interpretation in terms of pattern identification.

No description available.