The prevalence of concrete and cement-based materials in the civil infrastructure plus the risk of failure makes structural health monitoring an important issue in the understanding of the complete life cycle of civil structures. Correspondingly, the field of nondestructive evaluation (NDE) has been maturing and now concentrates on the detection of flaws and defects, as well as material damage in early stages of degradation. This defect detection is typically usually done by looking at the impulse response of the medium in question such as a cement-based material. The impulse response of a solid can be used to image a complex medium. Classically, the waveform is obtained by an active setup: an ultrasonic signal is generated at one location and recorded at another location. The waveform obtained from imaging can be used to quantitatively characterize the medium, for example by calculating the material's diffusivity coefficient or dissipation rate. In recent years, a different monitoring technique has been developed in seismology to measure the velocity of different kinds of waves, the Coda Wave Interferometry (CWI). In this CWI technique, the main focus is given to the late part of the recorded waveform, the coda. CWI is now successfully used in seismology and acoustics. In the current research, CWI is applied on concrete in different damage states to develop basic knowledge of the behavior of the wave velocity, and how it can be used to characterize cement-based materials. By comparing two impulse responses, the relative velocity change between the two impulse responses is used to characterize damage. Because of the stress-dependency of the velocity change, the calculations can also be used to directly calculate the Murnaghan's and Lam´e's coefficients. The newer technique of CWI is applied - the Stretching Technique (ST) [27]. The first goal of this research is to establish the viability of using CWI in cement-based materials. Next, we use the ST in the application of stress as we compress concrete samples for the detection of thermal damage, ASR-damage and mechanical softening.

A teoria da acustoelasticidade relaciona a variação de velocidade de propagação de ondas mecânicas à variação de tensão em um meio sólido. Em materiais frágeis como concreto, a danificação altera a velocidade de propagação paralelamente ao efeito acustoelástico. O objetivo deste trabalho é identificar e quantificar como a danificação e o efeito acustoelástico agem sobre a Velocidade de Pulso Ultrassônico (VPU) em corpos de prova de concreto submetidos a compressão uniaxial. Para tanto, foram realizadas três fases de ensaio. A primeira fase objetivou gerar dados para a análise da aplicação da interferometria de cauda de onda (Coda Wave Interferometry – CWI). Duas variações deste método foram estudadas e comparadas, com o propósito de determinar-se qual gera melhores resultados e quais parâmetros devem ser adotados para as análises. Para tal, um código computacional foi desenvolvido utilizando a linguagem Python 3.6.0. Foi constatado que a técnica do alongamento apresenta resultados melhores que a técnica tradicional da interferometria de cauda de onda. A segunda etapa foi dedicada ao estudo da variação de velocidade de propagação devido à recuperação de dano do corpo de prova. A terceira fase abordou a influência da geometria da amostra e da composição do concreto sobre a resposta do material à acustoelasticidade. Além disso, definiu-se um Índice de Dano (D) baseado na redução do módulo de elasticidade devido ao carregamento, a fim de isolar a variação de velocidade causada pelo efeito acustoelástico. Quanto ao estudo da recuperação de dano ao longo do tempo, a variação relativa de velocidade nas primeiras 24 horas após a retirada do carregamento se mostrou muito pequena em relação às variações geradas pelas condições de temperatura e umidade. Concluiu-se também que as amostras cilíndricas apresentaram respostas mais uniformes ao efeito acustoelástico que as amostras prismáticas. Por fim, o Índice de Dano se mostrou eficaz para isolar os efeitos da danificação e da acustoelasticidade sobre a VPU. / The acoustoelasticity theory relates the variation in propagation velocity of mechanical waves to the stress variation in a solid medium. In brittle materials such as concrete, damage affects the propagation velocity parallel to the acoustoelastic effect. This research aims to identify and quantify how damage and acoustoelastic effect act on Ultrasonic Pulse Velocity (UPV) in concrete samples subjected to uniaxial compression. In order to do so, three phases of testing were performed. The first one focused on generating data to analyze the application of the Coda Wave Interferometry (CWI). Two variations of this method were studied and compared, to the purpose of determining which variation shows better results and which parameters should be adopted in the analysis. To enable the analysis, a computational code using Python 3.6.0 language was developed. It was verified that the stretching technique shows better results than the traditional coda wave interferometry technique. The second phase was dedicated to study the variation in propagation velocity due to damage recovery in the sample. The third phase addressed the influence of the sample geometry and the concrete composition over the response from the material to the acoustoelasticity. Furthermore, a Damage Index (D) was defined based on the elastic modulus reduction due to loading, in order to isolate the variation of velocity due solely to the acoustoelastic effect. Regarding the study of damage recovery over time, the relative velocity variation in the first 24 hours following the withdrawal of the loading showed to be too little when compared to the variations caused by temperature and humidity conditions. It was also concluded that the cylindrical samples showed more uniform responses to the acoustoelastic effect than the prismatic samples. Finally, the Damage Index proved itself to be a reliable tool to isolate the effects of damage and acoustoelasticity over the UPV.

A teoria da acustoelasticidade relaciona a variação de velocidade de propagação de ondas mecânicas à variação de tensão em um meio sólido. Em materiais frágeis como concreto, a danificação altera a velocidade de propagação paralelamente ao efeito acustoelástico. O objetivo deste trabalho é identificar e quantificar como a danificação e o efeito acustoelástico agem sobre a Velocidade de Pulso Ultrassônico (VPU) em corpos de prova de concreto submetidos a compressão uniaxial. Para tanto, foram realizadas três fases de ensaio. A primeira fase objetivou gerar dados para a análise da aplicação da interferometria de cauda de onda (Coda Wave Interferometry – CWI). Duas variações deste método foram estudadas e comparadas, com o propósito de determinar-se qual gera melhores resultados e quais parâmetros devem ser adotados para as análises. Para tal, um código computacional foi desenvolvido utilizando a linguagem Python 3.6.0. Foi constatado que a técnica do alongamento apresenta resultados melhores que a técnica tradicional da interferometria de cauda de onda. A segunda etapa foi dedicada ao estudo da variação de velocidade de propagação devido à recuperação de dano do corpo de prova. A terceira fase abordou a influência da geometria da amostra e da composição do concreto sobre a resposta do material à acustoelasticidade. Além disso, definiu-se um Índice de Dano (D) baseado na redução do módulo de elasticidade devido ao carregamento, a fim de isolar a variação de velocidade causada pelo efeito acustoelástico. Quanto ao estudo da recuperação de dano ao longo do tempo, a variação relativa de velocidade nas primeiras 24 horas após a retirada do carregamento se mostrou muito pequena em relação às variações geradas pelas condições de temperatura e umidade. Concluiu-se também que as amostras cilíndricas apresentaram respostas mais uniformes ao efeito acustoelástico que as amostras prismáticas. Por fim, o Índice de Dano se mostrou eficaz para isolar os efeitos da danificação e da acustoelasticidade sobre a VPU. / The acoustoelasticity theory relates the variation in propagation velocity of mechanical waves to the stress variation in a solid medium. In brittle materials such as concrete, damage affects the propagation velocity parallel to the acoustoelastic effect. This research aims to identify and quantify how damage and acoustoelastic effect act on Ultrasonic Pulse Velocity (UPV) in concrete samples subjected to uniaxial compression. In order to do so, three phases of testing were performed. The first one focused on generating data to analyze the application of the Coda Wave Interferometry (CWI). Two variations of this method were studied and compared, to the purpose of determining which variation shows better results and which parameters should be adopted in the analysis. To enable the analysis, a computational code using Python 3.6.0 language was developed. It was verified that the stretching technique shows better results than the traditional coda wave interferometry technique. The second phase was dedicated to study the variation in propagation velocity due to damage recovery in the sample. The third phase addressed the influence of the sample geometry and the concrete composition over the response from the material to the acoustoelasticity. Furthermore, a Damage Index (D) was defined based on the elastic modulus reduction due to loading, in order to isolate the variation of velocity due solely to the acoustoelastic effect. Regarding the study of damage recovery over time, the relative velocity variation in the first 24 hours following the withdrawal of the loading showed to be too little when compared to the variations caused by temperature and humidity conditions. It was also concluded that the cylindrical samples showed more uniform responses to the acoustoelastic effect than the prismatic samples. Finally, the Damage Index proved itself to be a reliable tool to isolate the effects of damage and acoustoelasticity over the UPV.

Ce travail de thèse est une contribution au développement des méthodes d’acoustique non linéaire pour le contrôle non destructif et l’imagerie de défauts de type contact dans les solides.Dans ce travail, des modifications sont proposées pour deux méthodes récentes de contrôle non destructif par acoustique non linéaire : l’interférométrie de coda couplée au retournement temporel, et l’imagerie non linéaire par ultrasons aériens. Le principal avantage de la première méthode est sa sensibilité extrême liée à l’accumulation des effets induits par des changements, même faibles, des propriétés de l’échantillon durant la formation de la coda. La deuxième méthode apporte une approche complémentaire en permettant de réaliser un contrôle sans contact. Les techniques développées ont été testées sur des échantillons présentant des défauts artificiels à des emplacements connus. Les performances de chacune des méthodes ont été étudiées.La deuxième partie de ce travail porte sur la description théorique des non-linéarités acoustiques de contact et leur utilisation pour le développement d’une boite à outils numériques permettant la simulation d’ondes acoustiques dans des structures complexes contenant des contacts internes. Un model physique décrivant le décalage tangentiel de deux corps en contact en présence de friction est proposé. Il aboutit à une solution analytique pour la relation présentant une hystérésis entre les déplacements de contact normal et tangentiel et les chargements. Ce modèle est utilisé comme condition aux frontières pour les surfaces de contact internes (défauts) dans un modèle de propagation d’ondes acoustiques utilisant un logiciel d’éléments finis commercial / This PhD thesis work contributes to the development of nonlinear elastic methods for non-destructive testing and imaging of contact-type defects in solids.In this work, two modifications of recent nonlinear nondestructive testing methods are suggested: the coda wave interferometry combined with the nonlinear time reversal principle and air-coupled nonlinear ultrasonic imaging. The principal advantage of former technique is in its extremely high sensitivity owing to the fact that weak changes in sample's parameters are accumulated and finally greatly amplified during the formation of the coda wave. The other technique has a complimentary strength and offers a possibility of a remote detection. The developed techniques are tested on samples with artificially fabricated defects at known locations. The performance of each method is accessed and the potential for obtaining robust nonlinear images is demonstrated.The second part of the work is concerned with a theoretical description of contact acoustical nonlinearity and its use for creating of a numerical toolbox capable of simulating wave propagation in complex structures containing internal contacts. A physical model describing the tangential shift of two contacting bodies in the presence of friction has been proposed. Its result is an analytical computer-assisted solution for hysteretic relationships between normal and tangential contact displacements and loads. The contact model and derived load-displacement relationships are used as boundary conditions posed at the internal boundaries (contact surfaces) in a finite element wave propagation model programmed via commercial software

Nous nous intéressons, dans cette étude, à la caractérisation d'un défaut en développe une méthode ultrasonore basée sur la diffusion multiple des ondes. Pour cela, nous avons utilisé un béton base polymère endommagé par un essai de flexion en trois points. La technique d’'Interférométrie par Onde de Coda (CWI) prend en compte les ondes issues de la diffusion multiple et qui parcourent de ce fait une distance très grande devant celle séparant la source et du récepteur. Cette technique montre la sensibilité de la coda quand le matériau est soumis en résonance non linéaire où l’ influence des conditions de l’ environnement est considéré comme limité par l’ utilisation d’ un signal différent de référence. Les résultats expérimentaux ont montré que l’ efficacité de la méthode dépend du plan de vibrations considérées. Dans le but d'étudier l’ anisotropie d’ un défaut créé dans un béton base polymère nous avons utilisé les données de l’'Emission Acoustique (EA) récoltées au cours d’ essais acoustiques en Résonance Non Linéaire pour les deux plans. L’ existence d'une différence entre les signaux enregistrés au cours de mesures de l’ EA montrent que nous produisons différents micro-mécanismes dont la présence et / ou l'absence ont un impact important important sur l'interaction entre l'onde ultrasonore et le défaut. / Nonlinear Resonance (NR) and Coda Wave Interferometry (CWI) have proved to be efficient to detect and follow the evolution of micro-cracks within a strongly scattering media (concrete, rocks, etc.). Nevertheless, the localization of the cracks using the same techniques is not straight forward. In order to avoid the conditioning and its subsequent relaxation effect related to NR, CWI is simultaneously applied when concrete samples are vibrating in the linear regime. Based on a comparative study of the coda signals contents (non ballistic part) in the absence and under the weak linear vibration, the localization of the mechanically induced scatterers was possible depending on the scatterers' main direction with respect to the vibration plane. The latter point raises the issue of the generated types of vibration at the scatterers. Therefore, investigations were performed using the Acoustic Emission (AE) technique, which has served to verify that the acoustic activity during the linear vibrations does change depending on the considered experimental configuration. The latter, has also a direct effect on the frequency content of the recorded AE hits showing the potential link existing between the quantitative analysis of AE hits and the generated vibration mechanisms of the existing micro-cracks.

An extensive amount of natural gas trapped in the subsurface is found as methane hydrate. A fundamental understanding of natural hydrate-bearing sediments is required to engineer production strategies and to assess the risks hydrates pose to global climate change and large-scale seafloor destabilization. This thesis reports fundamental studies on hydrate nucleation, morphology and the evolution of unsaturation during dissociation, followed by additional studies on sampling and pressure core testing. Hydrate nucleation is favored on mineral surfaces and it is often triggered by mechanical vibration. Continued hydrate crystal growth within sediments is governed by capillary and skeletal forces; hence, the characteristic particle size d10 and the sediment burial depth determine hydrate morphologies in natural sediments. In aged hydrate-bearing sand, Ostwald ripening leads to patchy hydrate formation; the stiffness approaches to the lower bound at low hydrate saturation and the upper bound at high hydrate saturation. Hydrate saturation and pore habit alter the pore size variability and interconnectivity, and change the water retention curve in hydrate-bearing sediments. The physical properties of hydrate-bearing sediments are determined by the state of stress, porosity, and hydrate saturation. Furthermore, hydrate stability requires sampling, handling, and testing under in situ pressure, temperature, and stress conditions. Therefore, the laboratory characterization of natural hydrate-bearing sediments faces inherent sampling disturbances caused by changes in stress and strain as well as transient pressure and temperature changes that affect hydrate stability. While pressure core technology offers unprecedented opportunities for the study of hydrate-bearing sediments, careful data interpretation must recognize its inherent limitations.

Lorsqu'elle est chauffée, la roche peut subir une microfissuration thermique, qui influence ses propriétés physiques, mécaniques, thermiques, et de transport. La surveillance de la microfissuration thermique en laboratoire a été principalement réalisée pendant le chauffage, et rarement lors du refroidissement ou du chauffage cyclique que la roche subit dans les volcans et les réservoirs géothermiques. Un nouvel appareil a été élaboré pour surveiller les émissions acoustiques et mesurer les vitesses des ondes élastiques à haute température. L'état de fissuration a été évalué grâce à un nouvel algorithme d'analyse microstructurale, et l'influence des microfissures sur les propriétés des roches a été mesurée et modélisée. Selon la microstructure, la microfissuration peut avoir lieu pendant le chauffage ou le refroidissement, et les microfissures existantes peuvent s’ouvrir et se fermer de façon réversible avec des changements de température, et influencer les propriétés de la roche. / Rock may undergo thermal microcracking when heated, affecting its physical, mechanical, thermal, and transport properties. Thermal microcrack monitoring in the laboratory has mainly been performed during heating, and rarely during the cyclic heating and cooling relevant for volcanoes and geothermal reservoirs. For this, a new dedicated apparatus for the acoustic emission monitoring and wave velocity measurement at high temperatures was elaborated, building on previous designs. Microcrack damage was assessed with a new algorithm for quantitative microstructural analysis, and the influence of thermal microcracks on rock properties was measured and modelled. Depending on the rock type and initial microcrack content, microcracking occurred during either heating, cooling, or neither, and existing microcracks reversibly opened or closed with increasing temperature. In Earth's crust, the evolution of rock properties with temperature may be significant and is determined by the microstructure.