Structural health monitoring (SHM) is damage detection strategy for aerospace, civiland mechanical infrastructure. This project tries to show that Lamb waves, that are being generated and sensed with piezoelectric transducers, can be used for damage detection in a SHM system. For these piezoelectric transducers to work, filtering and amplification circuits needs to be connected to them. This report include the design,simulation, assembly and testing of these circuits. Due to lack of time, it was not possible to generate and sense actual Lamb waves. The result of the thesis is thatsimulations and tests show that it is possible to generate and sense Lamb waves for damage detection in a SHM system / Structural health monitoring (SHM) är en skadedetekteringsstrategi för flyg-,civil- och mekanisk infrastruktur. Detta projekt försöker visa att Lambvågor, som genereras och avkänns med piezoelektriska givare, kan användas för skadedetektering i ett SHM-system. För att dessa piezoelektriska givare ska fungera krävs att filtrerings- och förstärkningskretsar är anslutna till dem. Denna rapport inkluderar design, simulering, montering och testning av dessa kretsar. På grund av tidsbrist var det inte möjligt att generera eller avkänna Lambvågor. Resultatet av examensarbetet är att simuleringar och tester visar att det är möjligt att generera och avkänna Lambvågor för skadedetektering i ett SHM-system.

The steel through-truss Traffic Bridge, located in Saskatoon, Saskatchewan is over one hundred years old. The bridge has been subject to ongoing maintenance throughout its service life. However, inspection reports from 2005 and 2006 highlighted the severe deterioration experienced primarily by the steel members immediately above and below the deck surface. These reports prompted the City of Saskatoon (COS) to implement a rehabilitation project that involved the installation of a post-tensioning system to relieve the badly corroded bottom chord members of the axial loads due to the self-weight of the structure, in 2006. Due to the severe deterioration and the structural modifications that the Traffic Bridge has endured, a limited scope structural health monitoring (SHM) system, based on strain measurements, was implemented to reduce some of the uncertainty regarding the active load paths occurring at the deck level.<p> The objectives of the SHM study were to obtain more information regarding the actual load paths and ascertain possible types of structural redundancy, to determine how to best model this type of structure, and to find ways to track ongoing deterioration using instrumentation. The SHM study involved controlled truck loading scenarios to permit measurement of the load paths and provide data to compare the measured results to a finite element (FE) model of the instrumented span. In addition, random loading scenarios were used to capture the vertical dynamic response of the structure in order to further refine the FE model.<p> This study focused on the response of one-half of one interior span. A total of 72 strain gauges were installed. The downstream truss was highly instrumented at ten locations, three members of the upstream truss were instrumented to measure the distribution, and the floor joists in the downstream lane were instrumented to establish possible redundancy paths.<p> Using an FE model in combination with the measured strain data, it was found that redundant load paths only existed at the level of the deck. The bottom chord members experienced non-zero strains once the control vehicle was past the span, possibly indicating some level of redundancy. The members believed to relieve a portion of the bottom chord tensile forces included the car joists, edge joists, and the timber deck. The amount of force transferred from the bottom chord to the deck members was found by FE analysis to be highly related to the lateral stiffness of the floor beams.<p> The FE model was adjusted to match the measured results by modifying various modelling parameters. The most important features of the model were that all deck elements were modelled to be located at the elevation of the bottom chord, that the lateral stiffness of the floor beams was reduced by 50% to best represent the transfer of forces to deck elements, and that the stiffness of bottom chord members was reduced to 80% of their pristine values. In combination with calibrated modification factors applied to the measured values, this FE model is believed to be a useful tool to represent the behaviour of the structure to assist in detecting further damage by modelling the strain differential between members, and components of members.

The steel through-truss Traffic Bridge, located in Saskatoon, Saskatchewan is over one hundred years old. The bridge has been subject to ongoing maintenance throughout its service life. However, inspection reports from 2005 and 2006 highlighted the severe deterioration experienced primarily by the steel members immediately above and below the deck surface. These reports prompted the City of Saskatoon (COS) to implement a rehabilitation project that involved the installation of a post-tensioning system to relieve the badly corroded bottom chord members of the axial loads due to the self-weight of the structure, in 2006. Due to the severe deterioration and the structural modifications that the Traffic Bridge has endured, a limited scope structural health monitoring (SHM) system, based on strain measurements, was implemented to reduce some of the uncertainty regarding the active load paths occurring at the deck level.<p> The objectives of the SHM study were to obtain more information regarding the actual load paths and ascertain possible types of structural redundancy, to determine how to best model this type of structure, and to find ways to track ongoing deterioration using instrumentation. The SHM study involved controlled truck loading scenarios to permit measurement of the load paths and provide data to compare the measured results to a finite element (FE) model of the instrumented span. In addition, random loading scenarios were used to capture the vertical dynamic response of the structure in order to further refine the FE model.<p> This study focused on the response of one-half of one interior span. A total of 72 strain gauges were installed. The downstream truss was highly instrumented at ten locations, three members of the upstream truss were instrumented to measure the distribution, and the floor joists in the downstream lane were instrumented to establish possible redundancy paths.<p> Using an FE model in combination with the measured strain data, it was found that redundant load paths only existed at the level of the deck. The bottom chord members experienced non-zero strains once the control vehicle was past the span, possibly indicating some level of redundancy. The members believed to relieve a portion of the bottom chord tensile forces included the car joists, edge joists, and the timber deck. The amount of force transferred from the bottom chord to the deck members was found by FE analysis to be highly related to the lateral stiffness of the floor beams.<p> The FE model was adjusted to match the measured results by modifying various modelling parameters. The most important features of the model were that all deck elements were modelled to be located at the elevation of the bottom chord, that the lateral stiffness of the floor beams was reduced by 50% to best represent the transfer of forces to deck elements, and that the stiffness of bottom chord members was reduced to 80% of their pristine values. In combination with calibrated modification factors applied to the measured values, this FE model is believed to be a useful tool to represent the behaviour of the structure to assist in detecting further damage by modelling the strain differential between members, and components of members.

L'objectif de cette thèse est de réaliser un système de Contrôle Santé Intégré des structures aéronautiques (CSI ou SHM) autonome et à double-fonctionnalité. Ce système doit être en mesure d'assurer son autonomie énergétique tout en réalisant les tâches de détection et de localisation des endommagements. Latechnique retenue pour alimenter ce système est basée sur la récupération d'énergie vibratoire par transducteurs piézoélectriques SHM collés. Durant ces travaux, un modèle analytique complet de la chaîne de récupération d'énergie vibratoire a d'abord été créé. Ce modèle, validé par la Méthode des ÉlémentsFinis (MEF), permet d'améliorer le rendement du système en déterminant les dimensions, les locali-sations et le type de matériau piézoélectrique idéals des transducteurs. Ce modèle a ensuite été étendu à une configuration plus représentative des conditions de vibrations d'une structure en vol. Une bonne corrélation entre les résultats provenant du modèle prédictif et les essais sur un banc de mesures a étémise en évidence. Une puissance de 1.67mW a été récupérée et la capacité large bande des transducteurs a été vérifiée. L'application de la récupération d'énergie au contrôle de structures composites en cours d'assemblage sur les lignes de production a également été étudiée. Dans ce cas, un transducteur stratégiquement localisé et alimenté par une source de tension disponible génère des ondes de Lambdans la structure afin de pallier l'absence de vibrations naturelles. Un réseau de transducteurs secondaires disséminés sur cette structure récupère et convertit cette énergie vibratoire en énergie électrique. Une puissance de 7.36 mW a été récoltée et ce système a été en mesure de détecter une chute d'outil sur le composite et d'éclairer de façon autonome une diode électroluminescente (DEL) simulant ici la consommation de la transmission sans fil de l'information.

Detecção de dano em estruturas de engenharia de grandes dimensões através da análise de suas características dinâmicas envolve diversos campos de estudo. O primeiro deles trata da identificação dos parâmetros modais da estrutura, uma vez que executar testes de vibração livre em tais estruturas não é uma tarefa simples, necessita-se de um método robusto que seja capaz de identificar os parâmetros modais dessa estrutura a ações ambientais, campo esse chamado de análise modal operacional. Este trabalho trata do problema de detecção de dano em estruturas que possam ser representadas através de modelos em pórticos planos e vigas e que estejam submetidos à ação de vibrações ambientais. A localização do dano é determinada através de um algoritmo de otimização conhecido como Backtracking Search Algorithm (BSA) fazendo uso de uma função objetivo que utiliza as frequências naturais e modos de vibração identificados da estrutura. Simulações e testes são feitos a fim de verificar a concordância da metodologia para ambos os casos. Para as simulações, são utilizados casos mais gerais de carregamentos dinâmicos, e dois níveis de ruído (3% e 5%) são adicionados ao sinal de respostas para que esses ensaios se assemelhem aos ensaios experimentais, onde o ruído é inerente do processo. Já nos ensaios experimentais, apenas testes de vibração livre são executados. Diversos cenários de dano são propostos para as estruturas analisadas a fim de se verificar a robustez da rotina de detecção de dano. Os resultados mostram que a etapa de identificação modal estocástica através do método de identificação estocástica de subespaço (SSI) teve ótimos resultados, possibilitando, assim, a localização da região danificada da estrutura em todos os casos analisados. / Damage detection in large dimensions engineering structures through the analysis of their dynamic characteristics involves several fields. The first one deals with the structure modal identification parameter, since running free vibration tests in such structures is not a simple task, robust methods are needed in order to identify the modal parameters of this structure under ambient vibrations, this field is known as operational modal analysis. This work deals with the problem of damage detection in structures under ambient vibrations that can be represented by FEM using frame and beam elements. The damage location is determined through an optimization algorithm know as Backtracking Search Algorithm (BSA). It uses as objective function the identified natural frequencies and modes of vibration of the structure. Numerical and experimental tests are performed to assess the agreement of the methodology for both cases. For the numerical tests, more general cases of dynamic loads are used, and two noise levels (3% and 5%) are added to the response signal to assessing the robustness of the methodology close to the field conditions, in which noise is inherent of the process. In the experimental tests, only free vibration tests are performed. Several damage scenarios are proposed for the analyzed structures to check the robustness of the damage detection routine. The results show that the stochastic modal identification using the stochastic subspace identification (SSI) method had excellent results, thus allowing the location of the damaged region of the structure in all analyzed cases.

Submitted by DANIELA CABRAL DE OLIVEIRA null (danielacaboliveira@gmail.com) on 2017-07-31T18:25:34Z No. of bitstreams: 1 Dissertacao.pdf: 4071736 bytes, checksum: 8334dda6779551cc88a5687ed7937bb3 (MD5) / Approved for entry into archive by Luiz Galeffi (luizgaleffi@gmail.com) on 2017-08-03T16:52:18Z (GMT) No. of bitstreams: 1 oliveira_dc_me_ilha.pdf: 4071736 bytes, checksum: 8334dda6779551cc88a5687ed7937bb3 (MD5) / Made available in DSpace on 2017-08-03T16:52:18Z (GMT). No. of bitstreams: 1 oliveira_dc_me_ilha.pdf: 4071736 bytes, checksum: 8334dda6779551cc88a5687ed7937bb3 (MD5) Previous issue date: 2017-06-28 / Este trabalho introduz uma nova metodologia de Monitoramento da Integridade de Estruturas (SHM, do inglês Structural Health Monitoring) utilizando algoritmos de aprendizado de máquina não-supervisionado para localização e detecção de dano. A abordagem foi testada em material isotrópico (placa de alumínio). Os dados experimentais foram cedidos por Rosa (2016). O banco de dados disponibilizado é abrangente e inclui medidas em diversas situações. Os transdutores piezelétricos foram colados na placa de alumínio com dimensões de 500 x 500 x 2mm, que atuam como sensores e atuadores ao mesmo tempo. Para manipulação dos dados foram analisados os sinais definindo o primeiro pacote do sinal (first packet), considerando apenas o intervalo de tempo igual ao tempo da força de excitação. Neste caso, na há interferência dos sinais refletidos nas bordas da estrutura. Os sinais são obtidos na situação sem dano (baseline) e, posteriormente nas diversas situações de dano. Como método de avaliação do quanto o dano interfere em cada caminho, foram implementadas as seguintes métricas: pico máximo, valor médio quadrático (RMSD), correlação entre os sinais, normas H2 e H∞ entre os sinais baseline e sinais com dano. Logo após o cálculo das métricas para as diversas situações de dano, foi implementado o algoritmo de aprendizado de máquina não-supervisionado K-Means no matlab e também testado no toolbox Weka. No algoritmo K-Means há a necessidade da pré-determinação do número de clusters e isto pode dificultar sua utilização nas situações reais. Então, fez se necessário a implementação de um algoritmo de aprendizado de máquina não-supervisionado que utiliza propagação de afinidades, onde a determinação do número de clusters é definida pela matriz de similaridades. O algoritmo de propagação de afinidades foi desenvolvido para todas as métricas separadamente para cada dano. / This paper introduces a new Structural Health Monitoring (SHM) methodology using unsupervised machine learning algorithms for locating and detecting damage. The approach was tested with isotropic material in an aluminum plate. Experimental data were provided by Rosa (2016). This provided database is open and includes measures in a variety of situations. The piezoelectric transducers were bonded to the aluminum plate with dimensions 500 x 500 x 2mm, and act as sensors and actuators simultaneously. In order to manipulate the data, signals defining the first packet were analyzed. It considers strictly the time interval equal to excitation force length. In this case, there is no interference of reflected signals in the structure boundaries. Signals are gathered at undamaged situation (baseline) and at several damage situations. As an evaluating method of how damage interferes in each path, it was implemented the following metrics: maximum peak, root-mean-square deviation (RMSD), correlation between signals, H2 and H∞ norms regarding baseline and damaged signals. The metrics were computed for numerous damage situations. The data were evaluated in an unsupervised K-Means machine learning algorithm implemented in matlab and also tested in Weka toolbox. However, the K-Means algorithm requires the specification of the number of clusters and it is a problem for practical applications. Therefore, an implementation of an unsupervised machine learning algorithm, which uses affinity propagation was made. In this case, the determination of the number of clusters is defined by the data similarity matrix. The affinity propagation algorithm was developed for all metrics separately for each damage.

Detecção de dano em estruturas de engenharia de grandes dimensões através da análise de suas características dinâmicas envolve diversos campos de estudo. O primeiro deles trata da identificação dos parâmetros modais da estrutura, uma vez que executar testes de vibração livre em tais estruturas não é uma tarefa simples, necessita-se de um método robusto que seja capaz de identificar os parâmetros modais dessa estrutura a ações ambientais, campo esse chamado de análise modal operacional. Este trabalho trata do problema de detecção de dano em estruturas que possam ser representadas através de modelos em pórticos planos e vigas e que estejam submetidos à ação de vibrações ambientais. A localização do dano é determinada através de um algoritmo de otimização conhecido como Backtracking Search Algorithm (BSA) fazendo uso de uma função objetivo que utiliza as frequências naturais e modos de vibração identificados da estrutura. Simulações e testes são feitos a fim de verificar a concordância da metodologia para ambos os casos. Para as simulações, são utilizados casos mais gerais de carregamentos dinâmicos, e dois níveis de ruído (3% e 5%) são adicionados ao sinal de respostas para que esses ensaios se assemelhem aos ensaios experimentais, onde o ruído é inerente do processo. Já nos ensaios experimentais, apenas testes de vibração livre são executados. Diversos cenários de dano são propostos para as estruturas analisadas a fim de se verificar a robustez da rotina de detecção de dano. Os resultados mostram que a etapa de identificação modal estocástica através do método de identificação estocástica de subespaço (SSI) teve ótimos resultados, possibilitando, assim, a localização da região danificada da estrutura em todos os casos analisados. / Damage detection in large dimensions engineering structures through the analysis of their dynamic characteristics involves several fields. The first one deals with the structure modal identification parameter, since running free vibration tests in such structures is not a simple task, robust methods are needed in order to identify the modal parameters of this structure under ambient vibrations, this field is known as operational modal analysis. This work deals with the problem of damage detection in structures under ambient vibrations that can be represented by FEM using frame and beam elements. The damage location is determined through an optimization algorithm know as Backtracking Search Algorithm (BSA). It uses as objective function the identified natural frequencies and modes of vibration of the structure. Numerical and experimental tests are performed to assess the agreement of the methodology for both cases. For the numerical tests, more general cases of dynamic loads are used, and two noise levels (3% and 5%) are added to the response signal to assessing the robustness of the methodology close to the field conditions, in which noise is inherent of the process. In the experimental tests, only free vibration tests are performed. Several damage scenarios are proposed for the analyzed structures to check the robustness of the damage detection routine. The results show that the stochastic modal identification using the stochastic subspace identification (SSI) method had excellent results, thus allowing the location of the damaged region of the structure in all analyzed cases.

Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2018-07-27T14:09:42Z No. of bitstreams: 2 Dissertação - Michael dos Santos Durval - 2018.pdf: 12689287 bytes, checksum: a4aa8729f88220ca5e557b61addf7b0a (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-07-27T14:35:49Z (GMT) No. of bitstreams: 2 Dissertação - Michael dos Santos Durval - 2018.pdf: 12689287 bytes, checksum: a4aa8729f88220ca5e557b61addf7b0a (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-07-27T14:35:49Z (GMT). No. of bitstreams: 2 Dissertação - Michael dos Santos Durval - 2018.pdf: 12689287 bytes, checksum: a4aa8729f88220ca5e557b61addf7b0a (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-07-04 / Structural Health Monitoring (SHM) is a very cost-effective technique to reduce costs, increase life-cycle, and improve the performance of engineering structures. The impedancebased methodology uses the electromechanical behavior of piezoelectric materials (PZTs) to detect structural anomalies and damages. This technique uses high frequencies and excites the local modes, thus providing the monitoring of any change of the structural mechanical impedance in the region of influence of PZT. From the variation of the impedance signals, it can be concluded whether or not there is a damage. Artificial neural networks (RNA) are part of a broad concept called artificial systems. The foundation of neural networks is associated with the functioning of the human brain, which after training has the ability to perform associations. This science has great applicability in the solution of artificial intelligence problems, through the modeling of systems that use connections that make it possible to simulate the human nervous system. This work uses Kohonen’s self-organizing maps (SOM) associated to SHM based on electromechanical impedance for the detection and classification of damages in an aluminum beam. Based on the system under analysis, the network was trained to five different failure and severity positions. Through the neural network model of self-organizing maps, the network provided 30 maps as answers to the training and learning process. With this, it was realized qualitatively based on the concentration of energy of the maps that the grouping and classification of the different conditions of damages in which the engineering structure was submitted, happened with success. In order to establish a quantitative analysis proving the potential of the SOM network, the Hamming distance formula was applied, in which the results confirmed its accuracy. / O Monitoramento de Integridade Estrutural (SHM – Structural Health Monitoring) é uma técnica bem viável para se reduzir custos, aumentar a vida útil e melhorar o desempenho de estruturas de engenharia. A metodologia baseada em impedância usa o comportamento eletromecânico de materiais piezelétricos (PZT) para detectar anomalias e danos estruturais. Esta técnica utiliza altas frequências para excitar os modos locais, proporcionando, assim, o monitoramento de qualquer mudança da impedância mecânica estrutural na região de influência do PZT. A partir da variação dos sinais de impedância pode-se concluir pela existência ou não de um dano. Redes neurais artificiais (RNA) fazem parte de um amplo conceito chamado sistemas artificiais. O fundamento de redes neurais está associado ao funcionamento do cérebro humano, que após treinamento detém a capacidade de realizar associações. Esta ciência tem grande aplicabilidade na solução de problemas de inteligência artificial, através da modelagem de sistemas que usam conexões que possibilitam simular o sistema nervoso humano. Este trabalho utilizou a técnica dos mapas auto-organizáveis (SOMs – Self-Organizing Maps) de Kohonen associado ao SHM baseado na impedância eletromecânica para a detecção e a classificação de danos em uma viga de alumínio. Com base no sistema em análise, treinou-se a rede para cinco posições de falhas e severidades distintas. Por meio do modelo de rede neural dos mapas auto-organizáveis, a rede forneceu 30 mapas como respostas ao processo de treinamento e aprendizagem. Com isto, percebeu-se qualitativamente com base na concentração de energia dos mapas que o agrupamento e classificação das diferentes condições de danos em que a estrutura de engenharia foi submetida, ocorreu com sucesso. Para se estabelecer uma análise quantitativa que comprovasse o potencial da rede SOM, aplicou-se a fórmula da distância de Hamming, nos quais os resultados confirmaram sua precisão.

Detecção de dano em estruturas de engenharia de grandes dimensões através da análise de suas características dinâmicas envolve diversos campos de estudo. O primeiro deles trata da identificação dos parâmetros modais da estrutura, uma vez que executar testes de vibração livre em tais estruturas não é uma tarefa simples, necessita-se de um método robusto que seja capaz de identificar os parâmetros modais dessa estrutura a ações ambientais, campo esse chamado de análise modal operacional. Este trabalho trata do problema de detecção de dano em estruturas que possam ser representadas através de modelos em pórticos planos e vigas e que estejam submetidos à ação de vibrações ambientais. A localização do dano é determinada através de um algoritmo de otimização conhecido como Backtracking Search Algorithm (BSA) fazendo uso de uma função objetivo que utiliza as frequências naturais e modos de vibração identificados da estrutura. Simulações e testes são feitos a fim de verificar a concordância da metodologia para ambos os casos. Para as simulações, são utilizados casos mais gerais de carregamentos dinâmicos, e dois níveis de ruído (3% e 5%) são adicionados ao sinal de respostas para que esses ensaios se assemelhem aos ensaios experimentais, onde o ruído é inerente do processo. Já nos ensaios experimentais, apenas testes de vibração livre são executados. Diversos cenários de dano são propostos para as estruturas analisadas a fim de se verificar a robustez da rotina de detecção de dano. Os resultados mostram que a etapa de identificação modal estocástica através do método de identificação estocástica de subespaço (SSI) teve ótimos resultados, possibilitando, assim, a localização da região danificada da estrutura em todos os casos analisados. / Damage detection in large dimensions engineering structures through the analysis of their dynamic characteristics involves several fields. The first one deals with the structure modal identification parameter, since running free vibration tests in such structures is not a simple task, robust methods are needed in order to identify the modal parameters of this structure under ambient vibrations, this field is known as operational modal analysis. This work deals with the problem of damage detection in structures under ambient vibrations that can be represented by FEM using frame and beam elements. The damage location is determined through an optimization algorithm know as Backtracking Search Algorithm (BSA). It uses as objective function the identified natural frequencies and modes of vibration of the structure. Numerical and experimental tests are performed to assess the agreement of the methodology for both cases. For the numerical tests, more general cases of dynamic loads are used, and two noise levels (3% and 5%) are added to the response signal to assessing the robustness of the methodology close to the field conditions, in which noise is inherent of the process. In the experimental tests, only free vibration tests are performed. Several damage scenarios are proposed for the analyzed structures to check the robustness of the damage detection routine. The results show that the stochastic modal identification using the stochastic subspace identification (SSI) method had excellent results, thus allowing the location of the damaged region of the structure in all analyzed cases.

<div>This dissertation broadly focuses on autonomous condition assessment of civil infrastructures using vision-based methods, which present a plausible alternative to existing manual techniques. A region-based convolutional neural network (Faster R-CNN) is exploited for the detection of various earthquake-induced damages in reinforced concrete buildings. Four different damage categories are considered such as surface crack, spalling, spalling with exposed rebars, and severely buckled rebars. The performance of the model is evaluated on image data collected from buildings damaged under several past earthquakes taking place in different parts of the world. The proposed algorithm can be integrated with inspection drones or mobile robotic platforms for quick assessment of damaged buildings leading to expeditious planning of retrofit operations, minimization of damage cost, and timely restoration of essential services. </div><div><br></div><div> </div><div> Besides, a computer vision-based approach is presented to track the evolution of a damage over time by analysing historical visual inspection data. Once a defect is detected in a recent inspection data set, its spatial correspondences in the data collected during previous rounds of inspection are identified leveraging popular computer vision-based techniques. A single reconstructed view is then generated for each inspection round by synthesizing the candidate corresponding images. The chronology of damage thus established facilitates time-based quantification and lucid visual interpretation. This study is likely to enhance the efficiency structural inspection by introducing the time dimension into the autonomous condition assessment pipeline.</div><div><br></div><div> </div><div> Additionally, this dissertation incorporates depth fusion into a CNN-based semantic segmentation model. A 3D animation and visual effect software is exploited to generate a synthetic database of spatially aligned RGB and depth image pairs representing various damage categories which are commonly observed in reinforced concrete buildings. A number of encoding techniques are explored for representing the depth data. Besides, various schemes for fusion of RGB and depth data are investigated to identify the best fusion strategy. It was observed that depth fusion enhances the performance of deep learning-based damage segmentation algorithms significantly. Furthermore, strategies are proposed to manufacture depth information from corresponding RGB frame, which eliminates the need of depth sensing at the time of deployment without compromising on segmentation performance. Overall, the scientific research presented in this dissertation will be a stepping stone towards realizing a fully autonomous structural condition assessment pipeline.</div>