Für den zielgerichteten Einsatz von präventiven und mitigativen Notfallmaßnahmen sowie zur Beurteilung ihrer Wirksamkeit während auslegungsüberschreitender Unfälle in Druckwasserreaktoren aber auch für Hinweise zum Störfallverlauf und für die Abschätzung der Auswirkungen auf die Anlagenumgebung müssen geeignete Störfallinstrumentierungen vorhanden sein. Insbesondere der Zustand des Reaktordruckbehälterinventars (RDB-Inventar) während der In-Vessel-Phase eines auslegungsüberschreitenden Unfalls lässt sich mit aktuellen Störfallinstrumentierungen nur stark eingeschränkt überwachen, wodurch die o. g. Forderungen nicht erfüllt werden können. Die vorliegende Arbeit beinhaltet detaillierte Untersuchungen für die Entwicklung einer Störfallinstrumentierung, welche eine durchgängige Zustandsüberwachung des RDB-Inventars während der In-Vessel-Phase eines auslegungsüberschreitenden Unfalls ermöglicht. Die Störfallinstrumentierung basiert auf der Messung und Klassifikation von unterschiedlichen Gammaflussverteilungen, welche während der In-Vessel-Phase außerhalb des Reaktordruckbehälters auftreten können. Ausgehend von der Analyse zum Stand von Wissenschaft und Technik wird der modell-basierte Ansatz für Structural Health Monitoring-Systeme genutzt, um eine grundlegende Vorgehensweise für die Entwicklung der Störfallinstrumentierung zu erarbeiten. Anschließend erfolgt eine detaillierte Analyse zu den Vorgängen während der In-Vessel-Phase und eine daraus abgeleitete Definition von Kernzuständen für einen generischen Kernschmelzunfall. Für die definierten Kernzustände werden mittels Simulationen (Monte-Carlo-Simulationen zum Gammastrahlungstransport in einem zu dieser Arbeit parallel laufenden Vorhaben) Gammaflüsse außerhalb des Reaktordruckbehälters berechnet. Die Simulationsergebnisse dienen dem Aufbau von Datenbasen für die Entwicklung und Analyse eines Modells zur Klassifikation von Gammaflussverteilungen. Für die Entwicklung des Klassifikationsmodells kommen drei diversitäre und unabhängig arbeitende Klassifikationsverfahren (Entscheidungsbaum, k-nächste-Nachbarn-Klassifikation, Multilayer Perzeptron) zur Anwendung, um die Identifikationsgenauigkeit und Robustheit der Störfallinstrumentierung zu erhöhen. Die abschließenden Betrachtungen umfassen die Validierung der Störfallinstrumentierung mittels eines Versuchstandes zur Erzeugung unterschiedlicher Gammaflussverteilungen. Im Ergebnis der Untersuchungen konnte die prinzipielle Funktionsweise der entwickelten Störfallinstrumentierung nachgewiesen werden. Unter der Voraussetzung, die Gültigkeit der definierten Kernzustände zu untermauern sowie weitere, nicht in dieser Arbeit betrachtete Kernschmelzszenarien mit in die Entwicklung der Störfallinstrumentierung einzubeziehen, steht somit insbesondere für zukünftige Kernkraftwerke mit Druckwasserreaktoren eine Möglichkeit für die messtechnische Überwachung des RDB-Inventars während auslegungsüberschreitender Unfälle bereit. Die Arbeit leistet einen wesentlichen Beitrag auf dem Gebiet der Reaktorsicherheitsforschung sowie für den sicheren Betrieb von kerntechnischen Anlagen.:1 Einleitung 2 Analyse zum Stand von Wissenschaft und Technik 2.1 Sicherheit in deutschen Kernkraftwerken mit Druckwasserreaktor 2.1.1 Mehrstufenkonzept 2.1.2 Störfallinstrumentierungen 2.2 Auslegungsüberschreitende Unfälle mit Kernschmelze in DWR 2.2.1 Auslösende Ereignisse 2.2.2 Grundlegender Ablauf eines auslegungsüberschreitenden Unfall mit Kernschmelze 2.3 Strahlungstechnik, Strahlungsmesstechnik 2.3.1 Grundlagen der Strahlungstechnik 2.3.2 Wechselwirkungen von Gammastrahlung mit Materie 2.3.3 Messung ionisierender Strahlung 2.4 Verfahren und Methoden der Zustandsüberwachung 2.4.1 Zustandsüberwachung 2.4.2 Structural Health Monitoring 2.4.3 Mustererkennung 2.4.4 Entscheidungsbäume 2.4.5 k-nächste-Nachbarn-Klassifikation 2.4.6 Künstliche neuronale Netze 2.5 Schlussfolgerungen aus der Analyse zum Stand von Wissenschaft und Technik 2.5.1 Zusammenfassung zum Kapitel 2 2.5.2 Zielstellung, Aufbau und Abgrenzung der Arbeit 3 Analyse der In-Vessel-Phase und Definition von Kernzuständen 3.1 Detaillierte Analyse der In-Vessel-Phase 3.1.1 Auftretende Temperaturbereiche 3.1.2 Vorgänge während der frühen In-Vessel-Phase 3.1.3 Vorgänge während der späten In-Vessel-Phase 3.1.4 Spaltproduktfreisetzung 3.2 Definition von Kernzuständen für einen generischen Kernschmelzunfall 3.3 Zusammenfassung zum Kapitel 3 4 Datenbasen zur Entwicklung und Analyse des Klassifikationsmodells 4.1 Beschreibung der Monte-Carlo-Simulationsmodell 4.2 Beschreibung der Simulationsergebnisse und Merkmalsextraktion 4.3 Datenbasis zur Entwicklung 4.4 Datenbasen zur Analyse 4.5 Zusammenfassung zum Kapitel 4 5 Entwicklung und Analyse des Klassifikationsmodells 5.1 Beschreibung des Klassifikationsmodells 5.2 Teilmodell 1 - Entscheidungsbaum 5.2.1 Entwicklung 5.2.2 Analyse der Identifikationsgenauigkeit 5.3 Teilmodell 3 - k-nächste-Nachbarn-Klassifikation 5.3.1 Entwicklung 5.3.2 Analyse der Identifikationsgenauigkeit 5.4 Teilmodell 3 - Multilayer Perzeptron 5.4.1 Trainings- und Testdatenbasis 5.4.2 Entwicklung 5.4.3 Analyse der Identifikationsgenauigkeit 5.5 Teilmodell 4 - Vergleichsalgorithmus 5.5.1 Entwicklung 5.5.2 Analyse der Identifikationsgenauigkeit 5.6 Analysen zur Robustheit des Klassifikationsmodells 5.6.1 Ausfall einzelner Gammastrahlungsdetektoren 5.6.2 Gleichzeitiger Ausfall mehrerer Gammastrahlungsdetektoren 5.7 Zusammenfassung und Schlussfolgerungen für das Kapitel 5 6 Validierung der Kernzustandsüberwachungsverfahren 6.1 Zielstellung und Vorgehensweise 6.2 Versuchstand zur Validierung 6.2.1 Aufbau 6.2.2 Funktionsweise 6.3 Anpassung der Kernzustandsüberwachungsverfahren an den Versuchsstand 6.4 Validierungsexperimente 6.4.1 Experiment 1 - Füllstandsänderungen 6.4.2 Experiment 2 - Quellenbewegungen 6.4.3 Experiment 3 - Füllstandsänderungen, Quellenbewegungen und Änderung von Profilkonturen 6.5 Zusammenfassung und Schlussfolgerungen für das Kapitel 6 7 Zusammenfassung und Ausblick

Die verteilte faseroptische Sensorik (VFOS) auf Basis von Rayleighstreuung stellt ein besonderes und vielversprechendes Verfahren zur Dehnungsmessung im Betonbau und im Structural Health Monitoring (SHM) dar. Neben einer hohen Ortsauflösung und Messempfindlichkeit kann sie geringste Dehnungsänderungen an jedem Punkt der Messfaser erfassen. Für einen zuverlässigen Einsatz fehlen aktuell Material- und Handlungsempfehlungen; in der Literatur finden sich widersprüchliche Aussagen zu den Dehnungsübertragungsverlusten zwischen Substrat und Messfaser. Diesbezüglich beschäftigt sich die vorliegende Arbeit mit der Validierung des Messverfahrens für Szenarien im Betonbau. Neben der Applikation auf Stahl- und Betonoberflächen wurde die Integration in der Matrix untersucht. Im Zuge dessen sollten die Dehnungsübertragungsverluste verschiedener Fasercoatings bzw. Fasercoating-Klebstoff-Kombinationen überprüft werden. Darüber hinaus wurde ein Auswerteverfahren mithilfe der Programmiersprache Python entwickelt, das eine automatisierte Datenaufbereitung und Substituierung der Messabweichungen der enormen Datenmengen ermöglicht. Im Zuge der Validierung auf der Stahloberfläche wurden die Dehnungswerte der verschiedenen Coating-Klebstoff-Kombinationen an Präzisionsflachstählen im 4-Punkt-Biegeversuch mit einer photogrammetrischen Dehnungsmessung verglichen. Im Rahmen der Validierung auf der Betonoberfläche kamen Betondruckzylinder zum Einsatz, an denen zusätzlich zu den Coating-Klebstoff-Kombinationen der Einfluss einer Grundierung untersucht wurde. Induktive Wegaufnehmer dienten im Verlauf der Druckversuche als Vergleichsmesstechnik. Die Validierung verschiedener Coatingmaterialien in der Matrix fand anhand von Schwindversuchen an Betonprismen statt; als Vergleichsmessmethode dienten digitale Messuhren. Zur Bewertung der aus den Validierungsversuchen abgeleiteten Material- und Handlungsempfehlungen wurden Bauteilversuche an Betonbalken durchgeführt. Die Auswertung der Validierungsversuche zeigte, dass vor allem mit einem Ormocer-Coating gute Ergebnisse hinsichtlich der Dehnungsübertragung erzielt werden konnten. Im Falle einer Applikation empfiehlt sich die Verwendung eines Cyanacrylatklebstoffs -- besonders der M-Bond 200 überzeugte durch geringe Dehnungsverluste. Betonoberflächen sollten vorher geschliffen und mit Epoxidharz grundiert werden. Im Falle einer Integration in die Betonmatrix zeigten auch die Ormocer-Fasern minimale Unterschiede zur Vergleichsmessung. Die Ergebnisse der Bauteilversuche verifizieren die Handlungs- und Materialempfehlungen: Die Dehnungswerte der Fasern decken sich mit denen der analytischen Bemessung der Betonbalken. Lediglich an den Lasteinleitungsstellen konnten Oszillationen des Dehnungsverlaufs durch Gefügestörungen festgestellt werden. Forschungsbedarf besteht v.a. hinsichtlich der Validierung weiterer Komponenten (Klebstoff, Coating, Grundierungsmittel) und deren Langzeitstabilität, insbesondere bei chemischen und dynamischen Beanspruchungen. Im Rahmen dieser Arbeit konnte ein Überblick über verschiedene Materialien geschaffen werden, jedoch ist das Repertoire an verfügbaren Komponenten immens, gerade bei den Klebstoffen. Bei der Validierung auf der Betonoberfläche und in der Matrix kam eine Feinkornbetonmischung zum Einsatz. Diesbezüglich sollten in weiteren Forschungen unterschiedliche Matrices und Korngrößen Untersuchungsgegenstand sein. / Distributed fiber optic sensor (DFOS) technology based on Rayleigh scattering is a unique and promising method for strain measurement in concrete structures as well as structural health monitoring (SHM). It can detect the smallest strain changes at any point in the measuring fiber with a high spatial resolution and sensitivity. Currently, there exist no material and handling recommendations for a reliable application, and the literature contains contradictory statements on strain transfer losses between substrate and fiber. The present study deals with the validation of this measuring method for scenarios in concrete structures. Besides applications on steel and concrete surfaces, the integration in a concrete matrix was investigated. The validation yields results for strain transfer losses for different fiber coatings or fiber coating/adhesive combinations. Furthermore, the development of an evaluation method using the computer language Python provides automated data preparation and measurement error substitution of the enormous data volumes. For the validation on steel surfaces, the strain values of different coating-adhesive combinations on precision flat steels were compared in a 4-point bending test with a photogrammetric strain measurement. For the validation on concrete surfaces, concrete pressure cylinders were used to investigate the influence of a primer and different coating-adhesive combinations. Inductive displacement transducers served as a comparative measuring technique during the compression tests. Shrinkage tests allowed the validation of different coating materials in a matrix on concrete prisms. Digital dial gauges were used as a comparative measuring method. For evaluation purposes, the material and handling recommendations derived from the validation were tested on concrete beams. The results of the validation tests indicate good results regarding strain transfer with an Ormocer coating. For application as tested , the use of a cyanoacrylate adhesive is recommended -- especially M-Bond 200 was convincing due to its low strain losses. Concrete surfaces should be sanded and primed with epoxy resin. If integrated into a concrete matrix, the Ormocer fibers also showed minor differences from the comparative measurement. The results of the evaluation tests verify the handling and material recommendations: the strain values of the fibers correspond to the values obtained in concrete-beam analysis. Oscillations of the strain profile due to microstructural disturbances could only be detected at the load application points. Further research regarding the validation of additional components (adhesives, coating, primer) and their long-term stability, especially chemical and dynamic loads, is necessary. This study provides an overview of different coating and adhesives. However, the repertoire of available components is immense - especially for adhesives. During the validation on concrete surfaces and in matrix, a fine-grained concrete mixture was used. In this respect, different matrices and grain sizes should be the subject of further research.

Applications of vibration structural health monitoring (VHM) techniques are increasing rapidly. This is because of the advances in sensors and instrumentation during the last decades. VHM uses the vibration properties to evaluate many civil structures during the design steps, building steps and service life. The stiffness and frequencies of tensegrity structures are primarily related to the level of pre-stress. The present work investigates the possibilities to use this relation in designing, constructing and evaluating the tensegrity structures. The first part of the  present work studies the improvement of current models for resonance frequency simulation of tensegrities by introducing the bending behaviour of all components, and by a one-way coupling between the axial force and the stiffness. From this, both local and global vibration modes are obtained. The resonance frequencies are seen as non-linearly dependent on the pre-stress level in the structure, thereby giving a basis for diagnosis of structural conditions from measured frequencies. The new aspects of tensegrity simulations are shown for simple, plane structures but the basic methods are easily used also for more complex structures. In the second part, the environmental temperature effects on vibration properties of tensegrity structures have been investigated, considering primarily seasonal temperature differences (uniform temperature differences). Changes in dynamic characteristics due to temperature variations were compared with the changes due to decreasing pre-tension in one of the cables. In general, it is shown that the change in structural frequencies made by temperature changes could be equivalent to the change made by damage (slacking). Different combinations of materials used and boundary conditions are also investigated. These are shown to have a significant impact on the pre-stress level and the natural frequencies of the tensegrity structures when the environment temperature is changed. / Användandet av vibrationsbaserade hälsokontrollsmetoder (VHM) för strukturer ökar snabbt.Detta har möjliggjorts av utvecklingen inom  mätmetoder och mätutrus- tning under de senaste decennierna.Dessa metoder använder sig av de uppmätta eller simulerade vibrationsegenskaperna  underdesign-, uppbyggnads- och  nyttjandestadierna hos många slag av byggnadsverk. Styvheten och resonansfrekvenserna hos tensegritets-strukturer är i hög grad beroendepå den aktuella förspänningsnivån. Föreliggande arbete undersöker möjlig- heterna attanvända detta beroende i konstruktion, byggande och utvärdering av sådana strukturer. Den första delen av  föreliggande arbete studerar förbättringar av de vanligen användamodellerna för simulering av resonansfrekvenser hos tensegritetsstrukturergenom att införa de ingående komponenternas böjningsegenskaper, och genom att i enriktning koppla normalkraften till böjstyvheten. Genom detta kan såväl lokala som globalavibrationsmoder hittas. Resonansfrekvenserna ses därmed som icke-linjärt beroende avförspänningsnivån i strukturen. Detta ger därmed möjligheter att diagnosticera strukturenskondition från uppmätta frekvenser. De nya simuleringsmöjligheternademonstreras för enkla, plana strukturer, men de utvecklade metoderna kan också lättanpassas till mera komplexa fall. Den andra delen av arbetet undersöker hur strukturernas vibrationsegenskaper ärberoende på temperatureffekter i omgivningen. I första hand beaktas säsongsvisa (likformiga)temperaturvariationer. Förändringar i de dynamiska egenskaperna beroende påtemperaturförändringar jämfördes med dem som beror på en minskande förspänning hos någonav de ingående kablarna. I allmänhet gäller att förändringarna i resonansfrekvenser kanvara av samma storleksordning som de som beror på skador (minskad förspänning).Olika kombinationer av material, och olika upplagsförhållanden undersöktes.Dessa egenskaper visades ha en betydande effekt på förspänningsnivån, och därmed ocksåpå resonansfrekvenserna, hos tensegritets-strukturerna som utsätts för temperaturvariationer. / <p>QC 20140514</p>

Sverige har som många länder runt om i världen en åldrande infrastruktur och behovet av underhåll stiger. I en artikel i Svenska Dagbladet från 21/9–2018 redovisar analys- och teknikkonsultföretaget WSP en grov uppskattning att 300 miljarder kronor behöver investeras för att rusta upp existerande infrastruktur. Efter upprustningen skulle det krävas ca 25 miljarder kronor årligen för att utföra löpande tillståndsbedömning och underhåll av Sveriges väg- och järnvägsnät. Idag används inspektioner för att bedöma broars tillstånd. Det finns tre olika inspektionstyper; huvudinspektion, allmäninspektion och särskild inspektion. Structural Health Monitoring (SHM) är en teknik som globalt används mer och mer som en metod vid tillståndsbedömning av broar. SHM siktar mot att i realtid utföra automatisk bedömning av hela konstruktionens och de enskilda skadornas tillstånd. SHM använder sig av sensorer placerade på kritiska positioner för att samla in mätdata som jämförs med i förtid definierade gränsvärden. I Sverige används SHM sparsamt idag där kunskapen ligger på akademisk nivå och det råder en osäkerhet om hur SHM ska implementeras i praktiken. Genom att implementera SHM, kan skador upptäckas i ett tidigt skede och minimeras genom att snabbt utföra reparationer på konstruktionen innan skadorna blivit kritiska. Studien syftar till att producera en rapport som kan användas som en guide för hur SHM kan implementeras samt visa på hur SHM-systemen har implementerats på tidigare projekt. Rapporten skall ge läsaren en helhetsbild över hur tillståndsbedömning går till idag, vilka skador som är vanligt förekommande samt hur SHM kan användas som ett verktyg vid tillståndsbedömning. Rapporten syftar också till att ge en övergripande förklaring på svenska av SHM-tekniken och vad som behövs för implementering. Öppna ostrukturerade intervjuer genomfördes med forskare inom SHM och skador på betongkonstruktioner samt sakkunniga inom tillståndsbedömning. Intervjuerna användes som utgångspunkt för vidare studier av vanligt förekommande skador och SHM-tekniker. I litteraturstudien användes olika vetenskapliga databaser såsom Diva och ScienceDirect, samt KTH:s bibliotek för att få fram tidigare material om betongskador på broar, deras skademekanismer och om SHM-tekniken. Vidare hämtades information från ett antal doktorsavhandlingar, artiklar och tidigare examensarbeten. Genom litteraturstudie och intervjuer med sakkunniga kom studien fram till att armeringskorrosion och sprickor i betongen är skador som är vanligt förekommande i betongbroar. Skadornas skademekanismer beskrivs i rapporten. Studien identifierade tre olika system som kan användas för detektering och övervakning av armeringskorrosion och sprickor, de systemen är akustisk emission, SOFO-system (SOFO är en förkortning för Surveillance d'Ouvrage par Fibers Optics) och MuST-system (MuST är ett kommersiellt namn). Rapporten ger exempel på hur systemen kan användas för detektering och övervakning av skadorna genom att presentera fyra fallstudier där systemen har använts. Två av fallstudierna presenterar hur akustisk emission har implementeras och två av fallstudierna presenterar hur SOFO-systemet har implementerats. Studien kunde inte identifiera någon fallstudie för MuST-systemet och hur systemet har implementerats. Slutsatsen av studien är att SHM kan användas som ett komplement vid tillståndsbedömningar men man kan inte utföra automatiserade tillståndsbedömningar idag. / Like many countries around the world, Sweden has an aging infrastructure and the need for maintenance is increasing. In an article in Svenska Dagbladet from 21 / 9–2018, the analysis and engineering consulting company WSP reports a rough estimate that SEK 300 billion needs to be invested to upgrade existing infrastructure. After the upgrade, it would require approximately SEK 25 billion annually to carry out ongoing condition assessment and maintenance of Sweden's road and railway networks. Today, inspections are used to assess the condition of bridges. There are three types of inspection; main inspection, general inspection and special inspection. Structural Health Monitoring (SHM) is a technology that is globally more used as a method for condition assessment of bridges. SHM aims to perform automatic assessment of the state of the entire construction and individual damages in real time. SHM uses sensors placed at critical positions to collect measurement data compared to pre-defined limit values. In Sweden, SHM is used sparingly today, where knowledge is at an academic level and there is uncertainty about how SHM should be implemented in practice. By implementing SHM, damage can be detected at an early stage and minimized by quickly performing repairs on the design before the damage becomes critical. The study aims to produce a report that can be used as a guide on how SHM can be implemented and show how the SHM systems have been implemented on previous projects. The report should give the reader an overall picture of how condition assessment is performed today, which damages are common and how SHM can be used as a tool when assessing the condition. The report also aims to provide an overall explanation in Swedish of the SHM technology and what is needed for implementation. Open unstructured interviews were conducted with researchers within SHM and damages to concrete structures as well as experts in condition assessment. The interviews were used as a starting point for further studies of commonly occurring damages and SHM techniques. In the literature study, various scientific databases were used, such as Diva and ScienceDirect, as well as KTH's library to obtain earlier material on concrete damage to bridges, its damage mechanisms and about SHM technology. Furthermore, information was obtained from a number of doctoral dissertations, articles and previous degree projects. Through literature study and interviews with experts, the study concluded that reinforcement corrosion and cracks in the concrete are damages that are commonly found in concrete bridges. The damage mechanisms are described in the report. The study identified three different systems that can be used for detection and monitoring of reinforcement corrosion and cracks, those systems are acoustic emission, SOFO system (SOFO is an abbreviation for Surveillance d'Ouvrage pair of Fiber's Optics) and MuST system (MuST is a commercial name). The report gives examples of how the systems can be used for the detection and monitoring of the damages by presenting four case studies where the systems have been used. Two of the case studies present how acoustic emission has been implemented and two of the case studies present how the SOFO-system has been implemented. The study could not identify any case study for the MuST-system and how the system was implemented. The conclusion of the study is that SHM can be used as a supplement to condition assessments but cannot for the time being used for performing automated condition assessments today.

Much of North America’s reinforced concrete infrastructure is reaching the end of its service life and careful inspection and assessment is required to ensure the appropriate capacity is maintained in these structures. The research conducted herein seeks to further the development of two new sensor technologies: fibre optic strain sensors and digital image correlation, which have the potential to provide comprehensive performance data for structures to a level of accuracy previously not possible. The research involves determining the accuracy of these sensor systems to monitor both strain and crack widths in reinforced concrete compared to conventional techniques, such as electrical resistance strain gauges. Preliminary work was also undertaken on correcting the sensor results for temperature. It was determined that temperature variations in the range of +21 °C to 20 °C, result in significant strain errors for both sensor systems. Once the results obtained from the sensors systems are corrected for temperature, crack widths are monitored in four small-scale reinforced concrete tension specimens, and strain and crack width behaviour is monitored in four full-scale beams under four point bending. One of the major problems faced when using the digital image correlation technique is out of plane movement which results in significant error. Techniques to lower this error are addressed. In addition, obtaining a more robust understanding of the effects of temperature on crack widths, stiffness, strength and short term creep behaviour of reinforced concrete elements is explored to improve structural monitoring and numerical models used for analysis. Four full-scale beams, two at room temperature and two at 20 °C, were loaded to failure under four point bending. A comparison of the room temperature and low temperature test results show that the cracks tend to close up at lower temperatures in members that are free to expand and contract. This behaviour results in a potential increase in shear capacity for beams at lower temperatures. The low temperature beams also saw a minor increase in strength, but saw no noticeable increase in stiffness. Lastly, short term creep behaviour was reduced in the low temperature beams once the formation of ice occurred. / Thesis (Master, Civil Engineering) -- Queen's University, 2012-10-31 11:08:32.631

The power plant and aerospace industries have been facing a huge loss, due to structural failure. The structural failure occurs due to the presence of the crack in it. Hence, it becomes necessary to monitor the structural health under operating condition. Most of the techniques, for structural health monitoring are used for a specific purpose. Some of these techniques require structure dismantling, which is very much expensive and time consuming. So the vibration based structural health monitoring is advantageous, compared to other techniques. Most of the vibration based Structural Health Monitoring (SHM) approaches, use linear vibration theory. But, these linear vibration based procedures, have inherently low sensitivity to crack. Since crack introduces nonlinearities in the system, their merits in damage detection need to be investigated for SHM. In this thesis, the problem is focused on studying nonlinear dynamics of cracked structures for Structural Health Monitoring. For this, simulations and experiments are performed. The new procedure for the simulation is developed using Matlab-Simulink. It uses the numerical approximation for dynamic compliance operators and a nonlinear model of cracks contact faces interaction to study the dynamic behaviour of the cracked bar. Furthermore, the finite element model of the cracked cantilever bar with crack- tip plasticity is developed and the dynamic behaviour of the elasto-plastic bar is studied. Additionally, numerous experiments are performed to study the dynamics of cantilever bar with the fatigue crack in it. The results from Matlab-Simulink simulation shows the distribution of higher harmonics generated along the bar length, as a function of distance from the crack. In finite element simulation, comparison is made between the resonance frequency of cracked cantilever bar with and without crack-tip plasticity. It is found that, there is decrease in resonance frequency of the cracked bar with cracked tip plasticity, when compared with the resonance frequency of cracked bar without crack-tip plasticity. This reduction in resonance frequency is due to the crack-induced plasticity near the crack tip which affects the overall stiffness of bar. In experiments, the response is measured at four different points on the cracked cantilever bar at a given resonant frequency of excitation at lower and higher vibration amplitude. For lower vibration amplitude, it is found that the response obtained near the vicinity of the crack shows the presence of higher harmonics of resonant frequency, which disappears in the response obtained far away from the crack. For higher vibration amplitude, it is found that the response obtained near the vicinity of the crack shows the presence of higher harmonics along with the low frequency component. This low frequency component causes modulation, which leads to the generation of side band frequencies near the resonant frequency. The occurrence of low frequency component and side band frequencies is due to the vibro-impact behaviour of crack. The amplitude of these side band frequencies and higher harmonics are reduced in the response obtained far away from the crack. This indicates that crack-induced nonlinearity has a localized effect on the dynamics of bar. It is also observed that the magnitude of low frequency component is proportional to the magnitude of resonant frequency of excitation. This indicates that crack behaves like a signal modulator, detector of low frequency component and amplifier as the magnitude of low frequency component is proportional to the magnitude of resonant frequency excitation. From the Matlab-Simulink simulation and experimental results, it is concluded that crackinduced nonlinearity affects the dynamic behaviour of the cracked bar significantly, which will be effective in structural health monitoring. Keywords: vibro-impact, crack, dynamic compliance, harmonics, modulator, detector, amplifier, crack-tip plasticity, resonance frequency, structural health monitoring.

Graphene nanoplatelets have been introduced into the interphase between electrically insulating glass fibre and polymer matrix to functionalize the traditional composite. Owing to the distribution of network structure of GNPs, the interphase can transfer the signals about various internal change of material. Consequently, due to the novel bio-inspired overlapping structure, our GNPs-glass fibre shows a unique opportunity as a micro-scale multifunctional sensor. The following conclusions can be drawn from present research: • We prepared GNPs solution via a scalable and highly effective liquid-phase exfoliation method. This method produces high-quality, unoxidized graphene flakes from flake graphite. We control the thickness and size of GNPs by varying the centrifugation rate. • A simple fibre oriented capillary flow which can suppress ‘coffee ring’ effect to deposit GNPs onto the curved glass fibre surface. The GNPs form continuous fish scales like overlapping structure. • The electrical conductivity of our GNPs-glass fibre shows semiconductive property. The electrical resistance value scattering and the advancing contact angle value scattering indicate a uniform deposit structure. The uniform overlapping structure is a key factor for higher electrical conductivity compared with our previous work with CNTs. • The contact angles of our GNPs-glass fibre with water indicate that the GNPs are almost unoxidized, so the inert GNPs coating decreases the interfacial shears strength. • A micro scale GNPs-glass fibre sensor for gas sensing is achieved by deposit GNPs onto glass fibre surface. This sensor can be used to detect solvents vapours, such as water, ethanol and acetone. All these vapours work as electron acceptor when reacting with GNPs. The acetone shows the highest sensitivity (45000%) compared with water and ethanol. • The doping-dedoping of GNPs-glass fibres during adsorption-desorption cycles of acetone result in the efficient “break-junction” (GNPs lost electron carrier concentration) mechanism, which provides the possibility to fabricate the electrochemical “switch” in a simple and unique way. • The resistance of our GNPs-glass fibre shows exponential relationship with RH. This is attributed to two points. Firstly, the water vapours show similar exponential adsorption on carbon surface; secondly, the bandgap of GNPs increases with the increase of adsorbed water vapour concentration. • Due to the weak van der Waals interaction when water molecules are adsorbed on GNPs surface, our GNPs-glass fibre shows extreme fast response and recovery time with RH. It is potential for our GNPs-glass fibre being used to monitor the breath frequency. • Utilizing the negative temperature coefficient of GNPs, our GNPs-glass fibre can be used as temperature sensor with a sensing region of -150 to 30 °C. • Through the observed abnormal resistance change at a temperature of about – 18 °C, we discovered a phase change of the trance confined water in graphene layers. Based on the resistance change, we can study the interaction of water and carbon nanoparticles. • The bio-inspired novel overlapped multilayer structure of GNPs coating shows structural colours. Even more, our GNPs-glass fibre can be used to monitor the loading force in the interphase when it is embedded into epoxy resin. • Our GNPs-glass fibre shows an excellent piezoresistive property, the single GNPs-glass fibre shows a larger gauge factor than the commercial strains sensor. • The semiconductive interphase was formed when the GNPs-glass fibre was embedded in polymer matrix. This semiconductive interphase is very sensitive to the deformation of material, therefore, an in-situ strain sensor was manufactured to real-time monitor the microcracks in a composite instead of external sensors. The area of resistance ‘jump’ increase can be seen as the feature area for damage’s early warning. • Monitoring the resistance variation of the single fibre composite was conducted under cyclic loading with progressively increasing the strain peaks in order to further investigate the response of in-situ sensor to the interphase damage process. The deviation of resistance/strain when the stress is larger than 2 % highlights the accumulation of damage, which gives insight into the mechanism of resistance change.

During vibration of engineering structures, fatigue cracks may exhibit repetitive crack open-close breathing like phenomenon. In this thesis, the concept of entropy is employed to quantify this bi-linearity/irregularity of the vibration response so as to evaluate crack severity. To increase the sensitivity of the entropy calculation to detect the damage severity, entropy is merged with wavelet transformation (WT). A cantilever beam with a breathing crack is studied to asses proposed crack identification method under two vibration conditions: sinusoidal and random excitations. Through numerical simulations and experimental testing, the breathing crack identification under sinusoidal excitation is studied first and proven to be effective. Then, the crack identification sensitivity under lower excitation frequencies is further improved by parametric optimization of sample entropy and WT. Finally, breathing crack identification under general random excitations are experimentally studied and realized using frequency response functions (FRFs) as an add-in tool with the proposed crack identification technique. / October 2016

Orientador: Fábio Roberto Chavarette / Resumo: Redes Neurais Artificiais (RNAs) são algoritmos de aprendizado, geralmente estruturados em torno de categorização de dados de entrada e/ou seu agrupamento por similaridade. Tendo em vista características desejáveis como aprendizado rápido e estabilidade frente a vetores de entrada altamente mutáveis, adotou-se uma RNA do tipo Fuzzy ARTMAP como mecanismo central de um método de monitoramento de saúde estrutural para detectar e categorizar falhas em dados experimentais provenientes de um sistema mecânico similar a um pequeno prédio de dois andares. Mais especificamente, com o objetivo de detectar alterações das frequências naturais da estrutura, fenômeno ligado à deterioração da mesma, e determinar qual(is) andar(es) está(ão) ligado(s) ao comportamento anômalo, se detectado. A acurácia da rede foi avaliada, sendo realizado um estudo da quantidade de dados necessárias para o desempenho satisfatório da rede. Observou-se desempenho satisfatório, a acurácia do método tendendo a aproximadamente 94% a partir de certas quantidades de dados. / Abstract: Artificial Neural Networks (ANNs) are learning algorithms, largely revolving around categorizing data sets based on measures of similarity between its members. Due to desirable characteristics such as fast learning and stability when dealing with highly mutable input vectors, a Fuzzy ARTMAP ANN was selected as the core mechanism of a structural health monitoring method. Its goal was to detect and categorize faults in experimental data collected from a mechanical system akin to a small two-story building. More specifically, to detect disturbances on the structure's natural frequencies, phenomenon linked to its deterioration, and to determine which story or stories are linked to anomalous behavior, if any. The accuracy of the method was evaluated, and the amount of data needed for optimal operation was determined. Satisfactory performance was observed; the method's accuracy tended towards 94% with enough training samples. / Mestre

La détection d'impacts est une préoccupation majeure pour suivre l'intégrité d'une structure évoluant dans un environnement comportant des risques de collision. Ce travail consiste à développer une méthode d’identification d'impact applicable à une structure composite de grandes dimensions qui soit à la fois robuste, peu coûteuse en temps de calcul, et qui nécessite peu de capteurs. Dans un premier temps, la démarche a été de décrire l'image du point d'impact dans la réponse vibratoire par un vecteur de participations modales. L'idée a ensuite été d'introduire l'hypothèse d'existence d'une famille de modes discriminants permettant d'établir le lien bijectif entre les participations modales et le point d'impact. Une procédure d'estimation de ces participations modales à partir d'un unique point de mesure a été proposée. Dans un second temps, la démarche a été d'étendre la procédure à l'identification d'une loi décrivant les paramètres principaux d'une force d'impact. Des conditions portant sur les paramètres de mesure et les propriétés modales de la structure sont déterminées pour garantir la précision et la robustesse de l'identification. Ce travail a permis de développer une approche en rupture par rapport à l'état de l'art, en ce sens qu'elle ne nécessite, en théorie, qu'un seul point de mesures vibratoires pour identifier un impact. Des essais expérimentaux sur un pavillon d'A350 ont permis de valider cette approche sur une structure composite de grandes dimensions. / Many industrial structures operate in an environment with a high risk of collision. The detection of impacts and the assessment of their severity is a major preoccupation in Structural Health Monitoring. This work deals with the development of an impact identification technique that is applicable to a large composite structure, numerically robust and time efficient, and that requires a low number of sensors. The first step was to describe the image of the impact point in the vibration response as a modal participation vector. The idea was to introduce the existence of a discriminating modes familly to make a bijective link between the modal participations and the impact point. A least-squares procedure is developed to estimate those modal participations by measuring the vibration response of a single point on the structure. The second step was to extend the procedure to the identification of a parametric law representing the impact load history. Some conditions on measurement parameters and modal properties of the structure are identified to guarantee both the accuracy and the robustness of the procedure. The approach developed in this work is new regarding the state of the art: only one measurement point is considered for identifying an impact event. An experimental validation on an A350 crown panel indicates that the methodology is valid for large composite structures.