No description available.

The ability to rapidly assess the condition of a structure in a manner which enables the accurate prediction of its remaining capacity has long been viewed as a crucial step in allowing communities to make safe and efficient use of their public infrastructure. This objective has become even more relevant in recent years as both the interdependency and state of deterioration in infrastructure systems throughout the world have increased. Current practice for structural condition assessment emphasizes visual inspection, in which trained professionals will routinely survey a structure to estimate its remaining capacity. Though these methods have the ability to monitor gross structural changes, their ability to rapidly and cost-effectively assess the detailed condition of the structure with respect to its future behavior is limited.<div>Vibration-based monitoring techniques offer a promising alternative to this approach. As opposed to visually observing the surface of the structure, these methods judge its condition and infer its future performance by generating and updating models calibrated to its dynamic behavior. Bayesian inference approaches are particularly well suited to this model updating problem as they are able to identify the structure using sparse observations while simultaneously assessing the uncertainty in the identified parameters. However, a lack of consensus on efficient methods for their implementation to full-scale structural systems has led to a diverse set of Bayesian approaches, from which no clear method can be selected for full-scale implementation. The objective of this work is therefore to assess and enhance those techniques currently used for structural identification and make strides toward developing unified strategies for robustly implementing them on full-scale structures. This is accomplished by addressing several key research questions regarding the ability of these methods to overcome issues in identifiability, sensitivity to uncertain experimental conditions, and scalability. These questions are investigated by applying novel adaptations of several prominent Bayesian identification strategies to small-scale experimental systems equipped with nonlinear devices. Through these illustrative examples I explore the robustness and practicality of these algorithms, while also considering their extensibility to higher-dimensional systems. Addressing these core concerns underlying full-scale structural identification will enable the practical application of Bayesian inference techniques and thereby enhance the ability of communities to detect and respond to the condition of their infrastructure.<br></div>

Railways are a main element of a sustainable transport policy in several countries as they are considered a safe, efficient and green mode of transportation. Owing to these advantages, there is a cumulative request for the railway industry to increase the performance, the capacity and the availability in addition to safely transport goods and people at higher speeds. To meet the demand, large adjustment of the infrastructure and improvement of maintenance process are required.  Inspection activities are essential in establishing the required maintenance, and it is periodically required to reduce unexpected failures and to prevent dangerous consequences.  Maintenance of railway catenary systems is a critical task for warranting the safety of electrical railway operation.Usually, the catenary inspection is performed manually by trained personnel. However, as in all human-based inspections characterized by slowness and lack of objectivity, might have a number of crucial disadvantages and potentially lead to dangerous consequences. With the rapid progress of artificial intelligence, it is appropriate for computer vision detection approaches to replace the traditional manual methods during inspections.  In this thesis, a strategy for monitoring the health of catenary wires is developed, which include the various steps needed to detect anomalies in this component. Moreover, a solution for detecting different types of wires in the railway catenary system was implemented, in which a deep learning framework is developed by combining the Convolutional Neural Network (CNN) and the Region Proposal Network (RPN).

The Lamb wave based, non-contact damage detection techniques are developed using the Modified Time Reversal (MTR) method and the model based inverse problem approach. In the first part of this work, the Lamb wave-based MTR method along with the non-contacting sensors is used for structural damage detection. The use of non-contact measurements for MTR method is validated through experimental results and finite element simulations. A novel technique in frequency-time domain is developed to detect linear damages using the MTR method. The technique is highly suitable for the detection of damages in large metallic structures, even when the damage is superficial, and the severity is low. In this technique, no baseline data are used, and all the wave motion measurements are made remotely using a laser vibrometer. Additionally, this novel MTR based technique is not affected due to changes in the material properties of a structure, environmental conditions, or structural loading conditions. Further, the MTR method is improved for two-dimensional damage imaging. The damage imaging technique is successfully tested through experimental results and finite element simulations. In the second part of this work, an inverse problem approach is developed for the detection and estimation of major damage types experienced in adhesive joints. The inverse problem solution is obtained through an optimization algorithm wherein the objective function is formulated using the Lamb wave propagation data. The technique is successfully used for the detection/estimation of cohesive damages, micro-voids, debonds, and weak bonds. Further, the inverse problem solution is separately obtained through a fully connected artificial neural network. The neural network is trained using the Lamb wave propagation data generated from Wavelet Spectral Finite Element (WSFE) model which is computationally much faster than a conventional finite element model. This inverse problem approach technique requires a single point measurement for the inspection of the entire width of the adhesive joint. The proposed technique can be used as an automated quality assurance tool during the manufacturing process, and as an inspection tool during the operational life of adhesively bonded structures.

Modern semiconductor manufacturing is a complex process with a multitude of software applications. This application landscape has to be constantly monitored, since the communication and access patterns provide important insights. Because of the high event rates of the equipment log data stream in modern factories, big-data tools are required for scalable state and history analytics. The choice of suitable big-data solutions and their technical realization remains a challenging task. This thesis compares big-data architectures and discovers solutions for log-data ingest, enrichment, analytics and visualization. Based on the use cases and requirements of developers working in this field, a comparison of a custom assembled stack and a complete solution is made. Since the complete stack is a preferable solution, Datadog, Grafana Loki and the Elastic 8 Stack are selected for a more detailed study. These three systems are implemented and compared based on the requirements. All three systems are well suited for big-data logging and fulfill most of the requirements, but show different capabilities when implemented and used.:1 Introduction 1.1 Motivation 1.2 Structure 2 Fundamentals and Prerequisites 2.1 Logging 2.1.1 Log level 2.1.2 CSFW log 2.1.3 SECS log 2.2 Existing system and data 2.2.1 Production process 2.2.2 Log data in numbers 2.3 Requirements 2.3.1 Functional requirements 2.3.2 System requirements 2.3.3 Quality requirements 2.4 Use Cases 2.4.1 Finding specific communication sequence 2.4.2 Watching system changes 2.4.3 Comparison with expected production path 2.4.4 Enrichment with metadata 2.4.5 Decoupled log analysis 3 State of the Art and Potential Software Stacks 3.1 State of the art software stacks 3.1.1 IoT flow monitoring system 3.1.2 Big-Data IoT monitoring system 3.1.3 IoT Cloud Computing Stack 3.1.4 Big-Data Logging Architecture 3.1.5 IoT Energy Conservation System 3.1.6 Similarities of the architectures 3.2 Selection of software stack 3.2.1 Components for one layer 3.2.2 Software solutions for the stack 4 Analysis and Implementation 4.1 Full stack vs. a custom assembled stack 4.1.1 Drawbacks of a custom assembled stack 4.1.2 Advantages of a complete solution 4.1.3 Exclusion of a custom assembled stack 4.2 Selection of full stack solutions 4.2.1 Elastic vs. Amazon 4.2.2 Comparison of Cloud-Only-Solutions 4.2.3 Comparison of On-Premise-Solutions 4.3 Implementation of selected solutions 4.3.1 Datadog 4.3.2 Grafana Loki Stack 4.3.3 Elastic 8 Stack 5 Comparison 5.1 Comparison of components 5.1.1 Collection 5.1.2 Analysis 5.1.3 Visualization 5.2 Comparison of requirements 5.2.1 Functional requirements 5.2.2 System requirements 5.2.3 Quality requirements 5.3 Results 6 Conclusion and Future Work 6.1 Conclusion 6.2 Future Work / Die moderne Halbleiterfertigung ist ein komplexer Prozess mit einer Vielzahl von Softwareanwendungen. Diese Anwendungslandschaft muss ständig überwacht werden, da die Kommunikations- und Zugriffsmuster wichtige Erkenntnisse liefern. Aufgrund der hohen Ereignisraten des Logdatenstroms der Maschinen in modernen Fabriken werden Big-Data-Tools für skalierbare Zustands- und Verlaufsanalysen benötigt. Die Auswahl geeigneter Big-Data-Lösungen und deren technische Umsetzung ist eine anspruchsvolle Aufgabe. Diese Arbeit vergleicht Big-Data-Architekturen und untersucht Lösungen für das Sammeln, Anreicherung, Analyse und Visualisierung von Log-Daten. Basierend auf den Use Cases und den Anforderungen von Entwicklern, die in diesem Bereich arbeiten, wird ein Vergleich zwischen einem individuell zusammengestellten Stack und einer Komplettlösung vorgenommen. Da die Komplettlösung vorteilhafter ist, werden Datadog, Grafana Loki und der Elastic 8 Stack für eine genauere Untersuchung ausgewählt. Diese drei Systeme werden auf der Grundlage der Anforderungen implementiert und verglichen. Alle drei Systeme eignen sich gut für Big-Data-Logging und erfüllen die meisten Anforderungen, zeigen aber unterschiedliche Fähigkeiten bei der Implementierung und Nutzung.:1 Introduction 1.1 Motivation 1.2 Structure 2 Fundamentals and Prerequisites 2.1 Logging 2.1.1 Log level 2.1.2 CSFW log 2.1.3 SECS log 2.2 Existing system and data 2.2.1 Production process 2.2.2 Log data in numbers 2.3 Requirements 2.3.1 Functional requirements 2.3.2 System requirements 2.3.3 Quality requirements 2.4 Use Cases 2.4.1 Finding specific communication sequence 2.4.2 Watching system changes 2.4.3 Comparison with expected production path 2.4.4 Enrichment with metadata 2.4.5 Decoupled log analysis 3 State of the Art and Potential Software Stacks 3.1 State of the art software stacks 3.1.1 IoT flow monitoring system 3.1.2 Big-Data IoT monitoring system 3.1.3 IoT Cloud Computing Stack 3.1.4 Big-Data Logging Architecture 3.1.5 IoT Energy Conservation System 3.1.6 Similarities of the architectures 3.2 Selection of software stack 3.2.1 Components for one layer 3.2.2 Software solutions for the stack 4 Analysis and Implementation 4.1 Full stack vs. a custom assembled stack 4.1.1 Drawbacks of a custom assembled stack 4.1.2 Advantages of a complete solution 4.1.3 Exclusion of a custom assembled stack 4.2 Selection of full stack solutions 4.2.1 Elastic vs. Amazon 4.2.2 Comparison of Cloud-Only-Solutions 4.2.3 Comparison of On-Premise-Solutions 4.3 Implementation of selected solutions 4.3.1 Datadog 4.3.2 Grafana Loki Stack 4.3.3 Elastic 8 Stack 5 Comparison 5.1 Comparison of components 5.1.1 Collection 5.1.2 Analysis 5.1.3 Visualization 5.2 Comparison of requirements 5.2.1 Functional requirements 5.2.2 System requirements 5.2.3 Quality requirements 5.3 Results 6 Conclusion and Future Work 6.1 Conclusion 6.2 Future Work

[ES] La industria de la construcción es un amplio sector industrial que abarca desde el diseño y la gestión de grandes infraestructuras como puentes hasta la construcción de viviendas civiles. Es mundialmente reconocido como un sector impulsor fundamental del Producto Interno Bruto, pero también se encuentra entre los de menor rendimiento y retraso en la adopción y explotación de mejoras tecnológicas. Estas limitaciones están induciendo a las partes interesadas a tomar prestadas e integrar muchas mejoras de otros campos industriales en el sector. Esta tendencia de digitalización se está extendiendo a lo largo de todo el ciclo de vida del proceso de construcción e identifica un enfoque desafiante debido al cambio de paradigma necesario de los sistemas físicos a los ciberfísicos. El concepto Industria 4.0 impulsó esta tendencia por lo que tanto en la academia como en la industria de la construcción se ha concretado como Construcción 4.0. Toma prestada de la Industria 4.0 la adopción de muchas tecnologías habilitadoras clave como Internet de las Cosas, Inteligencia Artificial y Fabricación Aditiva. Esta tesis investiga specíficamente esta integración tecnológica, centrándose en la aplicación de tales tecnologías habilitadoras en el campo de la construcción y considerando diferentes etapas en el ciclo de vida en diferentes tipologías de infraestructura. A partir de una investigación bibliográfica sobre sistemas inteligentes "holísticos" en la construcción de Edificios Inteligentes, a la manera de Gemelos Digitales, se estudia la influencia y la aplicación de tecnologías habilitadoras y herramientas TIC operativas relacionadas, como Internet de las Cosas y Big Data, desde una perspectiva de todo el ciclo de vida de las construcciones. Se estudia la fase de mantenimiento de grandes infraestructuras en materia de seguridad estructural y detección de fallos, mediante el desarrollo de un método de detección de daños en puentes ferroviarios de celosía metálica mediante inteligencia artificial. Luego se presenta una innovadora tecnología de fabricación aditiva para construcciones de gran altura. Consiste en una mejora de la tecnología de las grúas torre estándar con una extrusora personalizada, mientras que todo el sistema está controlado por un agente de inteligencia artificial. Concluimos que la Construcción 4.0 aún se encuentra en su etapa embrionaria. Se pueden obtener resultados más avanzados en la implantación tecnológica sobre infraestructuras existentes para su gestión de operación y mantenimiento debido al enfoque relacionado principalmente con la sensorización y análisis de datos. La innovación en la fase integrada de diseño/construcción sigue siendo más desafiante, debido a la necesidad de un paradigma completamente nuevo e innovaciones industriales en muchos campos diferentes. / [CA] La indústria de la construcció és un ampli sector industrial que abasta des del disseny i la gestió de grans infraestructures com a ponts fins a la construcció d'habitatges civils. És mundialment reconegut com un sector impulsor fonamental del Producte Intern Brut, però també es troba entre els de menor rendiment i retard en l'adopció i explotació de millores tecnològiques. Aquestes limitacions estan induint a les parts interessades a amprar i integrar moltes millores d'altres camps industrials en el sector. Aquesta tendència de digitalització s'està estenent al llarg de tot el cicle de vida del procés de construcció i identifica un enfocament desafiador a causa del canvi de paradigma necessari dels sistemes físics als ciberfísics. El concepte Indústria 4.0 va impulsar aquesta tendència pel que tant en l'acadèmia com en la indústria de la construcció s'ha concretat com a Construcció 4.0. Ampra de la Indústria 4.0 l'adopció de moltes tecnologies habilitants clau com a Internet de les Coses, Intel·ligència Artificial i Fabricació Additiva. Aquesta tesi investiga específicament aquesta integració tecnològica, centrant-se en l'aplicació de tals tecnologies habili- tants en el camp de la construcció i considerant diferents etapes en el cicle de vida en diferents tipologies d'infraestructura. A partir d'una investigació bibliogràfica sobre sistemes intel·ligents "holístics" en la construcció d'Edificis Intel·ligents, a la manera de Bessons Digitals, s'estudia la influència i l'aplicació de tecnologies habilitants i eines TIC operatives relacionades, com a Internet de les coses i Big Data, des d'una perspectiva de tot el cicle de vida de les construccions. S'estudia la fase de manteniment de grans infraestructures en matèria de seguretat estructural i detecció de fallades, mitjançant el desenvolupament d'un mètode de detecció de danys en ponts ferroviaris de gelosia metàl·lica mitjançant intel·ligència artificial. Després es presenta una innovadora tecnologia de fabricació additiva per a construccions de gran altura. Consisteix en una millora de la tecnologia de les grues torre estàndard amb una extrusora personalitzada, mentre que tot el sistema està controlat per un agent d'intel·ligència artificial. Concloem que la Construcció 4.0 encara es troba en la seua etapa embrionària. Es poden obtindre resultats més avançats en la implantació tecnològica sobre infraestructures existents per a la seua gestió d'operació i manteniment degut a l'enfocament relacionat principalment amb la sensorització i anàlisi de dades. La innovació en la fase integrada de disseny/construcció continua sent més desafiadora, a causa de la necessitat d'un paradigma completament nou i innovacions industrials en molts camps diferents. / [EN] The construction industry is a wide industrial sector ranging from the design and management of major infrastructures, such as bridges, to civil dwelling construction. It is worldwide acknowledged as a fundamental driving sector for the Gross Domestic Product, but it is also among the less performing and delayed ones in the adoption and exploitation of technological improvements. These limitations are inducing stakeholders to borrow and integrate many enhancements from other industrial fields into the sector. This digitalization trend is spreading through the entire life cycle of the construction process and identifying a challenging approach because of the paradigm shift needed from physical to cyber-physical systems. The Industry 4.0 concept boosted this trend so that both in the academy and in the construction industry it has been specified as Construction 4.0. It borrows from the Industry 4.0 the adoption of many key enabling technologies such as Internet of Things, Artificial Intelligence and Additive Manufacturing. This thesis investigates specifically this technological integration, focusing on the application of such enabling technologies in the construction field and considering different stages in the life cycle in varying infrastructure typologies. Starting from a literature investigation on "holistic" intelligent systems in Intelligent Buildings construction, in a Digital Twin fashion, the influence and the application of enabling technologies and related operative ICT tools such as Internet of Things and Big Data are studied, from a perspective of the whole constructions' life cycle. The maintenance phase of major infrastructures is studied concerning structural safety and fault detection, by developing a method to detect damages in railway steel truss bridges via artificial intelligence. An innovative additive manufacturing technology for high-rise constructions is then presented. It consists of an improvement with a custom extruder of standard tower crane technology, while the whole system is driven by an artificial intelligence agent. We conclude that Construction 4.0 is still at its embryonic stage. More advanced results are obtainable for the operation and maintenance management of existing infrastructures because of the already mature approach related to sensorization and data analysis. Innovation in the design/construction phase remains more challenging,because of the need for a completely new paradigm and industrial innovations in many different fields. / Parisi, F. (2023). Automation and Information Approaches to Support Maintenance and Production Management in the Construction Industry [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/192826

[pt] Na engenharia estrutural, um dos critérios de projeto consiste na necessidade de que as estruturas mantenham suas condições de segurança e funcionalidade ao longo da vida útil. Desta forma, diversas metodologias de avaliação da integridade estrutural (SHM) vêm sendo desenvolvidas. Esta avaliação ocorre através do sensoriamento contínuo de estruturas de modo a validar seu comportamento no tempo. A resposta dinâmica pode ser considerada uma metodologia gradualmente explorada como SHM. Estruturas em serviço tendem a sofrer vibrações que geram uma resposta dinâmica. A existência de danos proporciona a degradação de propriedades como massa e rigidez, resultando em variação nos parâmetros modais, como frequências naturais, modos de vibração e amortecimento. Na presente pesquisa, com o auxílio de um modelo de elementos finitos devidamente validado, propõem-se uma metodologia de avaliação de danos através de parâmetros modais experimentais. A proposta será avaliada através da comparação entre respostas numéricas e resultados de ensaios dinâmicos em laje danificada. Objetiva-se estimar a variação de rigidez da estrutura, que corrobora para alteração dos parâmetros dinâmicos. Serão avaliados ainda fenômenos como breathing cracks, fissuras que se encontram abertas ou fechadas em instantes distintos durante a vibração. Além disso, será ponderado o efeito de enrijecimento do concreto na região fissurada. Estes processos afetam consideravelmente a rigidez do elemento fissurado. Desta forma, objetiva-se um modelo robusto de análise de lajes de concreto armado fissuradas a partir da análise do comportamento dinâmico. / [en] In structural engineering, one of the design criteria is the need to maintain structures safety and functionality conditions throughout their useful life. This way, several structural health monitoring methodologies (SHM) have been developed. This evaluation occurs through the continuous sensing of structures in order to validate their behavior over time. Dynamic response can be considered a gradually explored methodology as SHM. In-service structures tend to suffer with vibrations that produce a dynamic response. The existence of structural damage produces a degradation of properties such as mass and stiffness, resulting in variation of modal properties such as natural frequencies, vibration modes and damping. In the present research, with the aid of a duly validated finite element model, it is proposed a damages diagnosis methodology through experimental modal parameters. The proposal will be through the comparison of numerical responses and dynamic tests results made with a damaged slab. The analysis aim to estimate the structure s stiffness variation, which corroborates the alteration of the dynamic parameters. Phenomena such as breathing cracks, where the fissure is open or closed at different times during the vibration, will also be evaluated. In addition, the effect of concrete stiffening around the cracked region will be considered. These processes considerably affect the stiffness of the cracked element. In this way, the objective is a robust model for the analysis of cracked reinforced concrete slabs from the analysis of their dynamic behavior.

[pt] O monitoramento de integridade estrutural tem sido o foco de desenvolvimentos recentes no campo da avaliação baseada em vibração e, mais recentemente, no escopo da internet das coisas à medida que medição e computação se tornam distribuídas. Os dados se tornaram abundantes, embora a transmissão nem sempre seja viável em frequências mais altas especialmente em aplicações remotas. Portanto, é importante conceber fluxos de trabalho de modelo orientados por dados que garantam a melhor relação entre a precisão do modelo para avaliação de condição e os recursos computacionais necessários para soluções incorporadas, tópico que não tem sido amplamente utilizado no contexto de medições baseadas em vibração. Neste contexto, a presente pesquisa propõe abordagens para duas aplicações: na primeira foi proposto um fluxo de trabalho de modelagem capaz de reduzir a dimensão dos parâmetros de modelos autorregressivos usando análise de componentes principais e classificar esses dados usando algumas técnicas de aprendizado de máquina como regressão logística, máquina de vetor de suporte, árvores de decisão, k-vizinhos próximos e floresta aleatória. O exemplo do prédio de três andares foi usado para demonstrar a eficácia do método. No segundo caso, é utilizado um equipamento de teste composto por inércias rotativas onde a solução de monitoramento foi testada em uma plataforma baseada em GPU embarcada. Os modelos implementados para distinguir eficazmente os diferentes estados de atrito foram análise de componentes principais, deep autoencoders e redes neurais artificiais. Modelos rasos têm melhor desempenho em tempo de execução e precisão na detecção de condições de falha. / [en] Structural health monitoring has been the focus of recent developments in vibration-based assessment and, more recently, in the scope of the internet of things as measurement and computation become distributed. Data has become abundant even though the transmission is not always feasible, especially in remote applications. It is thus essential to devise data-driven model workflows that ensure the best compromise between model accuracy for condition assessment and the computational resources needed for embedded solutions. This topic has not been widely used in the context of vibration-based measurements. In this context, the present research proposes two approaches for two applications, a static and a rotating one. In case one, a modeling workflow capable of reducing the dimension of autoregressive model features using principal component analysis and classifying this data using some of the main machine learning techniques such as logistic regression, support vector machines, decision tree classifier, k-nearest neighborhood and random forest classifier was proposed. The three-story building example was used to demonstrate the method s effectiveness, together with ways to assess the best compromise between accuracy and model size. In case two, a test rig composed of rotating inertias and slender connecting rods is used, and the monitoring solution was tested in an embedded GPU-based platform. The models implemented to effectively distinguish between different friction states were principal component analysis, deep autoencoder and artificial neural networks. Shallow models perform better concerning running time and accuracy in detecting faulty conditions.

[pt] Recentemente, ondas guiadas por ultrassom têm mostrado grande potencial para ensaios não destrutivos e monitoramento de integridade estrutural (SHM) em um cenário de avaliação de danos. As medições obtidas por meio de ondas elásticas são particularmente úteis devido a sua capacidade de se propagarem em diferentes materiais, como meios sólidos e fluidos e, também, a capacidade de abrangerem áreas amplas. Ao possuir suficientes medições oriundas de ondas guiadas, técnicas avançadas baseadas em dados, como aprendizado de máquina, podem ser aplicadas ao problema, tornando o procedimento de avaliação de danos ainda mais poderoso e robusto. Com base nessas circunstâncias, o presente trabalho trata da aplicação de modelos de aprendizado de máquina para fornecer inferências de avaliação de falhas baseadas em informações de ondas guiadas por ultrassom. Dois principais estudos de caso são abordados. Primeiramente, uma placa de polímero reforçado com fibra de carbono (PRFC) é avaliada, utilizando dados da literatura de sinais de onda guiada do tipo Lamb na detecção de defeitos pontuais. Os resultados demonstraram que uma abordagem que utiliza um sinal de referência foi capaz de obter excelentes acurácias ao usar a extração de características baseadas em técnicas de identificação de sistemas. Em um segundo momento, defeitos semelhantes à corrosão em uma placa de alumínio são classificados de acordo com sua gravidade. A metodologia é auxiliada por um esquema de separação de modos em sinais de ondas guiadas do tipo SH pré-adquiridos. Os resultados obtidos mostraram que a adoção da separação de modos pode, de fato, melhorar os resultados do aprendizado de máquina. / [en] Recently ultrasonic guided waves have shown great potential for nondestructive testing and structural health monitoring (SHM) in a damage evaluation scenario. Measurements utilizing elastic waves are particularly useful due to their capability to propagate in different materials such as solid and fluid bounded media, and, also, the ability to cover broad areas. When enough guided waves measurements are available and advanced data-driven techniques such as machine learning can be applied to the problem, the damage evaluation procedure becomes then even more powerful and robust. Based on these circumstances, the present work deals with the application of machine learning models to provide fault evaluation inferences based on ultrasonic guided waves information. Two main case studies are tackled in the mentioned subject. Firstly, a carbon fiber reinforced polymer (CFRP) plate is assessed using open data of Lamb guided wave signals in the detection of dot type defects. Results demonstrated that a baseline dependent approach can obtain excellent results when using system identification feature extraction. Secondly, corrosion-like defects in an aluminium plate are classified according to their severity. The methodology is assisted by a mode separation scheme of SH guided waves signals of pre-acquired data. Results have shown that the adoption of mode separation can in fact improve the machine learning results.

This is probably the most appropriate time for the development of robust and reliable structural damage detection systems as aging civil engineering structures, such as bridges, are being used past their life expectancy and beyond their original design loads. Often, when a significant damage to the structure is discovered, the deterioration has already progressed far and required repair is substantial. This is both expensive and has negative impact on the environment and traffic during replacement. For the exposed reasons the demand for efficient Structural Health Monitoring techniques is currently extremely high. This licentiate thesis presents a two-stage model-free damage detection approach based on Machine Learning. The method is applied to data gathered in a numerical experiment using a three-dimensional finite element model of a railway bridge. The initial step in this study consists in collecting the structural dynamic response that is simulated during the passage of a train, considering the bridge in both healthy and damaged conditions. The first stage of the proposed algorithm consists in the design and unsupervised training of Artificial Neural Networks that, provided with input composed of measured accelerations in previous instants, are capable of predicting future output acceleration. In the second stage the prediction errors are used to fit a Gaussian Process that enables to perform a statistical analysis of the distribution of errors. Subsequently, the concept of Damage Index is introduced and the probabilities associated with false diagnosis are studied. Following the former steps Receiver Operating Characteristic curves are generated and the threshold of the detection system can be adjusted according to the trade-off between errors. Lastly, using the Bayes’ Theorem, a simplified method for the calculation of the expected cost of the strategy is proposed and exemplified. / <p>QC 20170420</p>