Defected wheel are one the major reasons endangered state of railroad vehicles safety statue, due to vehicle derailment and worsen the quality of freight and passenger transportation. Therefore, timely defect detection for monitoring and detecting the state of defects is highly critical. This thesis presents a passive non-contact acoustic structural health monitoring approach using ultrasonic acoustic emissions (UAE) to detect certain defects on different structures, as well as, classifying the type of the defect on them. The acoustic emission signals used in this study are in the ultrasonic range (18-120 kHz), which is significantly higher than the majority of the research in this area thus far. For the proposed method, an impulse excitation, such as a hammer strike, is applied to the structure. In addition, ultrasound techniques have higher sensitivity to both surface and subsurface defects, which make the defect detection more accurate. Three structures considered for this study are: 1) a longitudinal beam, 2) a lifting weight, 3) an actual rail-wheel. A longitudinal beam was used at the first step for a better understanding of physics of the ultrasound propagation from the defect, as well, develop a method for extracting the signature response of the defect. Besides, the inherent directionality of the ultrasound microphone increases the signal to noise ratio (SNR) and could be useful in the noisy areas. Next, by considering the ultimate goal of the project, lifting weight was chosen, due to its similarity to the ultimate goal of this project that is a rail-wheel. A detection method and metric were developed by using the lifting weight and two type of synthetic defects were classified on this structure. Also, by using same extracted features, the same types of defects were detected and classified on an actual rail-wheel. / Master of Science

The largest and most difficult numerical approximation error to estimate is discretization error. Residual-based discretization error estimation methods are a category of error estimators that use an estimate of the source of discretization error and information about the specific application to estimate the discretization error using only one grid level. The higher-order terms are truncated from the discretized equations and are the local source of discretization error. The accuracy of the resulting discretization error estimate depends solely on the accuracy of the estimated truncation error. Residual-based methods require only one grid level compared to the more commonly used Richardson extrapolation which requires at least two. Reducing the required number of grid levels reduces computational expense and, since only one grid level is required, can be applied to unstructured grids where multiple quality grid levels are difficult to produce. The two residual-based discretization error estimators of interest are defect correction and error transport equations. The focus of this work is the development, improvement, and evaluation of various truncation error estimation methods considering the accuracy of the truncation error estimate and the resulting discretization error estimates. The minimum requirements for accurate truncation error estimation is specified along with proper treatment for several boundary conditions. The methods are evaluated using various Euler and Navier-Stokes applications. The discretization error estimates are compared to Richardson extrapolation. The most accurate truncation error estimation method was found to be the k-exact method where the fine grid with a correction factor was considerably reliable. The single grid methods including the k-exact require that the continuous operator be modified at the boundary to be consistent with the implemented boundary conditions. Defect correction showed to be more accurate for areas of larger discretization error; however, the cost was substantial (although cheaper than the primal problem) compared to the cost of solving the ETEs which was essential free due to the linearization. Both methods showed significantly more accurate estimates compared to Richardson extrapolation especially for smooth problems. Reduced accuracy was apparent with the presence of stronger shocks and some possible modifications to adapt to singularies are proposed for future work. / Ph. D.

Numerical solutions obtained through the use of Computational Fluid Dynamics (CFD) are subject to discretization error, which is locally generated by truncation error. The discretization error is extremely difficult to properly estimate and this in turn leads to uncertainty over the quality of the numerical solutions obtained via CFD methods and the engineering functionals computed using these solutions. Adjoint error estimation techniques specifically seek to estimate the error in functionals, but are dependent upon accurate truncation error estimates. This work examines the application of new, single-grid, truncation error estimation procedures to the problem of adjoint error estimation for both the quasi-1D and 2D Euler equations. The new truncation error estimation techniques are based on local reconstructions of the computed solutions and comparisons are made for the quasi-1D study in order to determine the most appropriate solution variables to reconstruct as well as the most appropriate reconstruction method. In addition, comparisons are made between the single-grid truncation error estimates and methods based on uniformally refining or coarsening the underlying numerical mesh on which the computed solutions are obtained. A method based on an refined grid error estimate is shown to work well for a non-isentropic flow for the quasi-1D Euler equations, but all truncation error estimations methods ultimately result in over prediction of functional discretization error in the presence of a shock in 2D. Alternatives to adjoint methods, which can only estimate the error in a single functional for each adjoint solution obtained, are examined for the 2D Euler equations. The defection correction method and error transport equations are capable of locally improving the entire computed solution, allowing for error estimates in multiple functionals. It is found that all three functional discretization error estimates perform similarly for the same truncation error estimate, although the defect correction method is the most costly from a computational viewpoint. Comparisons are made between truncation error and adjoint weighted truncation error based adaptive indicators. For the quasi-1D Euler equations it is found that both methods are competitive, however the truncation error based method is cheaper as a separate adjoint solve is avoided. For the 2D Euler equations, the truncation error estimates on the adapted meshes suffer due to a lack of smooth grid transformations which are used in reconstructing the computed solutions. In order to complete this work, a new CFD code incorporating a variety of best practices from the field of Computer Science is developed as well as a new method of performing code verification using the method of manufactured solutions which is significantly easier to implement than traditional manufactured solution techniques. / Ph. D.

Software testing is important, but judging whether a set of software tests is effective is difficult. This problem also appears in the classroom as educators more frequently include software testing activities in programming assignments. The most common measures used to assess student-written software tests are coverage criteria—tracking how much of the student’s code (in terms of statements, or branches) is exercised by the corresponding tests. However, coverage criteria have limitations and sometimes overestimate the true quality of the tests. This dissertation investigates alternative measures of test quality based on how many defects the tests can detect either from code written by other students—all-pairs execution—or from artificially injected changes—mutation analysis. We also investigate a new potential measure called checked code coverage that calculates coverage from the dynamic backward slices of test oracles, i.e. all statements that contribute to the checked result of any test. Adoption of these alternative approaches in automated classroom grading systems require overcoming a number of technical challenges. This research addresses these challenges and experimentally compares different methods in terms of how well they predict defect-detection capabilities of student-written tests when run against over 36,500 known, authentic, human-written errors. For data collection, we use CS2 assignments and evaluate students’ tests with 10 different measures—all-pairs execution, mutation testing with four different sets of mutation operators, checked code coverage, and four coverage criteria. Experimental results encompassing 1,971,073 test runs show that all-pairs execution is the most accurate predictor of the underlying defect-detection capability of a test suite. The second best predictor is mutation analysis with the statement deletion operator. Further, no strong correlation was found between defect-detection capability and coverage measures. / Ph. D.

Product defects concern various groups of people, such as customers, manufacturers, government officials, etc. Thus, defect-related knowledge and information are essential. In keeping with the growth of social media, online forums, and Internet commerce, people post a vast amount of feedback on products, which forms a good source for the automatic acquisition of knowledge about defects. However, considering the vast volume of online reviews, how to automatically identify critical product defects and summarize the related information from the huge number of user reviews is challenging, even when we target only the negative reviews. As a kind of opinion mining research, existing defect discovery methods mainly focus on how to classify the type of product issues, which is not enough for users. People expect to see defect information in multiple facets, such as product model, component, and symptom, which are necessary to understand the defects and quantify their influence. In addition, people are eager to seek problem resolutions once they spot defects. These challenges cannot be solved by existing aspect-oriented opinion mining models, which seldom consider the defect entities mentioned above. Furthermore, users also want to better capture the semantics of review text, and to summarize product defects more accurately in the form of natural language sentences. However, existing text summarization models including neural networks can hardly generalize to user review summarization due to the lack of labeled data. In this research, we explore topic models and neural network models for product defect discovery and summarization from user reviews. Firstly, a generative Probabilistic Defect Model (PDM) is proposed, which models the generation process of user reviews from key defect entities including product Model, Component, Symptom, and Incident Date. Using the joint topics in these aspects, which are produced by PDM, people can discover defects which are represented by those entities. Secondly, we devise a Product Defect Latent Dirichlet Allocation (PDLDA) model, which describes how negative reviews are generated from defect elements like Component, Symptom, and Resolution. The interdependency between these entities is modeled by PDLDA as well. PDLDA answers not only what the defects look like, but also how to address them using the crowd wisdom hidden in user reviews. Finally, the problem of how to summarize user reviews more accurately, and better capture the semantics in them, is studied using deep neural networks, especially Hierarchical Encoder-Decoder Models. For each of the research topics, comprehensive evaluations are conducted to justify the effectiveness and accuracy of the proposed models, on heterogeneous datasets. Further, on the theoretical side, this research contributes to the research stream on product defect discovery, opinion mining, probabilistic graphical models, and deep neural network models. Regarding impact, these techniques will benefit related users such as customers, manufacturers, and government officials. / Ph. D. / Product defects concern various groups of people, such as customers, manufacturers, and government officials. Thus, defect-related knowledge and information are essential. In keeping with the growth of social media, online forums, and Internet commerce, people post a vast amount of feedback on products, which forms a good source for the automatic acquisition of knowledge about defects. However, considering the vast volume of online reviews, how to automatically identify critical product defects and summarize the related information from the huge number of user reviews is challenging, even when we target only the negative reviews. People expect to see defect information in multiple facets, such as product model, component, and symptom, which are necessary to understand the defects and quantify their influence. In addition, people are eager to seek problem resolutions once they spot defects. Furthermore, users also want to better summarize product defects more accurately in the form of natural language sentences. These requirements cannot be satisfied by existing methods, which seldom consider the defect entities mentioned above, or hardly generalize to user review summarization. In this research, we develop novel Machine Learning (ML) algorithms for product defect discovery and summarization. Firstly, we study how to identify product defects and their related attributes, such as Product Model, Component, Symptom, and Incident Date. Secondly, we devise a novel algorithm, which can discover product defects and the related Component, Symptom, and Resolution, from online user reviews. This method tells not only what the defects look like, but also how to address them using the crowd wisdom hidden in user reviews. Finally, we address the problem of how to summarize user reviews in the form of natural language sentences using a paraphrase-style method. On the theoretical side, this research contributes to multiple research areas in Natural Language Processing (NLP), Information Retrieval (IR), and Machine Learning. Regarding impact, these techniques will benefit related users such as customers, manufacturers, and government officials.

In this paper, a novel approach is utilized to investigate the deformation mechanisms at the microstructural level in 3D printed alloys. The complex in-situ heat treatments during 3D printing leaves a unique and complicated microstructure in the as-built 3D printed metals, particularly alloys. The microstructure is made of a hierarchical stacking of some interconnected geometrical shapes, namely meltpools, grains, and cells. These are connected to each other by boundaries that might have different element compositions, and consequently, material properties, compared to the interior region of each geometrical unit. Deformation mechanisms in this microstructure are still highly unexplored, mainly because of the challenges on the way of performing experiments at the micrometer length scale. In this work, we establish an image processing framework that directly converts the SEM images taken from the microstructure of 3D printed 316L stainless steel alloys into CAD models. The model of the complicated microstructure is then scaled up, and the scaled model is 3D printed using polymeric materials. For 3D printing these samples, two polymers with contrasting mechanical properties are used. Distribution of these two polymers mimics the arrangement of soft and stiff regions in the microstructure of 3D printed alloys. These representative samples are subjected to mechanical loads and digital image correlation is utilized to investigate the deformation mechanisms, particularly the delocalization of stress concentration and also the crack propagation, at the microstructural level of 3D printed metals. Besides experiments, computational modeling using finite element method is also performed to study the same deformation mechanisms at the microstructure of 3D printed 316L stainless steel. Our results show that the hierarchical arrangement of stiff and soft phases in 3D printed alloys delocalizes the stress concentration and has the potential to make microstructures with significantly improved damage tolerance capabilities. / Master of Science / Many researchers have studied the impacts of laser parameters on the bulk material properties of SLM printed parts; few if any have studied how these parts break at a microstructural level. In this work we show how SLM printed parts with complex microstructures including grains, meltpools, and cells, deform and break. The cellular network that occurs in some SLM printed parts leads to a multi-material hierarchical structure, with a stiff network of thin boundaries, and a bulk "matrix" of soft cell material. This leads to similar properties as some composites, whereby the stiff network of cell boundaries leads to increased damage tolerance. We show both computationally through finite element analysis, and experimentally through multi-material 3D fabrication, that the microstructure leads to increased crack length in failure, as well as lower toughness loss and strength loss in the event of a crack. Essentially, the complex nature of the formation of these parts (high heating and cooling rates from laser melting) leads to a beneficial microstructure for damage tolerance that has not been studied from this perspective before.

<p>This dissertation discusses theoretical methods for describing photoionization in different systems in the context of time-dependent and time-independent non-relativistic and time-independent relativistic systems. We introduce a multichannel quantum defect theory (MQDT) model for describing photoionization in the context of pump-probe experiments. The basics of MQDT are introduced and specialized to the argon atom. Two energy regimes are studied in detail and compared to the experiment: (i) a perturbative calculation describing the dynamics of an autoionizing wave packet, (ii) a time-resolved calculation describing the two-photon ionization of a deeply bound wave packet. In both cases, the model accurately describes the relative ionization with respect to the two spin-orbit split thresholds of the ion and the oscillations shown in the delay between the pump and probe. We finalize with a brief presentation, which is primarily pedagogical, of how to use MQDT inside a finite box.</p> <p>Next, we use MQDT to describe the ytterbium atom in different energy regimes and varying degrees of approximation. The motivation behind this lies in the context of quantum information science, but our study is only concerned with calculating atomic properties. We start with a minimal MQDT model to describe the data observed in the experiment, followed by the presentation of an ab initio two-electron model. Both models compare very well to the experiment, and the ab initio method compares favorably with older spectroscopic results. In addition, we show unpublished results that incorporate the hyper-fine effects into the approximate model.</p> <p>Finally, we present an implementation of the two-electron variational R-matrix method for the Dirac equation, including the complete derivation of the solution of the Dirac equation in a central potential. We provide explicit analytic forms for the solutions of the Coulomb potential and use them to derive the generalized quantum defect parameters. A discussion of the variational R-matrix method for the Dirac equation in single and multichannel contexts is presented, with sample calculations for the beryllium and radium atoms. A chapter that summarizes and points to future work for each one of the projects concludes the work.</p>

Over the past few decades, additive manufacturing (AM) or 3D printing (3DP) technologies witnessed revolutionary growth in the manufacturing sector. Parts produced with metal AM techniques, especially Laser Powder Bed Fusion (LPBF), are often prone to porosity issues. The presence of pores leads to harmful effects such as crack formation and, eventually, premature failure of the component. Consequently, research in defect detection and pore prediction attracted substantial attention. Utilizing image-based porosity detection in preexisting systems is a simple, effective, and cost-efficient approach for final part inspection. This thesis investigates the possibility of predicting porosity using U-Net and its novel network architectures named RU-Net and RAU-Net, on an X-ray computed tomography (XCT) image dataset. Later, the performance of these models is analyzed and compared using precision, recall, F1 score, mAP, IoU metrics, and their hybrid losses combining BCG and Dice loss. RAU-Net outperforms RU-Net and U-Net in all these metrics by detecting more than 90% of actual pores while retaining 95% precision. While RU-Net and U-Net required additional training, RAU-Net achieved high performance in only 50 epochs, demonstrating its data efficiency and convergence. Due to its shorter training period, also leading to lower computational overhead, RAU-Net is suited for practical high throughput and low latency applications. Particularly in time-sensitive applications, RAU-Net can enable more widespread adoption of dense prediction networks. A custom script is developed for estimating the porosity percentage level in 3D printed metal components precisely, further enhancing final product inspection procedures. As a result, the entire quality control process is simplified, which allows for the quicker inspection of final components to deliver, by ensuring they meet required quality and reliability standards.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第25106号 / 理博第5013号 / 新制||理||1715(附属図書館) / 京都大学大学院理学研究科物理学・宇宙物理学専攻 / (主査)教授 杉本 茂樹, 教授 青木 慎也, 教授 橋本 幸士 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DFAM

Le cœur des vertébrés est un organe modulaire qui requiert le " patterning " complexe des champs morphogénétiques cardiogènes et la convergence coordonnée des diverses sous-populations de progéniteurs cardiogéniques. Au moins 7 facteurs de transcription de la famille T-box coopèrent au sein de ces nombreuses sous-populations de progéniteurs cardiogéniques afin de réguler la morphogenèse et l’agencement de multiples structures le long de l’ébauche cardiaque, ce qui explique que les mutations humaines de ces gènes engendrent diverses malformations congénitales cardiaques (MCCs). L’un de ces gènes T-box, Tbx5, dont l’haploinsuffisance génère le syndrome de Holt-Oram (SHO), intervient dans une grande variété de réseaux de régulation géniques (RRGs) qui orchestrent la morphogenèse des oreillettes, du ventricule gauche, de la valve mitrale, des septums inter-auriculaire et inter-ventriculaire, ainsi que du système de conduction cardiaque. La diversité des RRGs impliqués dans la formation de ces structures cardiaques suggère que Tbx5 détient une profusion de fonctions qui ne seront identifiables qu’en répertoriant ses activités moléculaires dans chaque lignée cardiaque examinée isolément. Afin d’aborder cette problématique, une ablation génétique de Tbx5 dans l’endocarde a été réalisée. Cette expérience a démontré le rôle crucial de Tbx5 dans la survie des cellules endocardiques bordant le septum primum et des cardiomyocytes au sein de cette structure embryonnaire qui contribuera à la morphogenèse du septum inter-auriculaire. En outre, cette étude a révélé l’existence d’une communication croisée entre la sous-population de cellules endocardiques Tbx5+ et le myocarde au niveau du septum primum, afin d’assurer la survie des cardiomyocytes, et ultimement de garantir la maturation du septum inter-auriculaire. Nos résultats confirment aussi l’importance de l’interdépendance génétique (Tbx5 et Gata4 ainsi que Tbx5 et Nos3) entre différents loci dans la morphogenèse de la cloison inter-auriculaire, et particulièrement de l’influence que peut avoir l’environnement sur la pénétrance et l’expressivité des communications inter-auriculaires (CIAs) dans le SHO. En outre, puisque les fonctions d’un gène dépendent ordinairement des différents isoformes qu’il peut générer, une deuxième étude a focalisé davantage sur l’aspect transcriptionnel de Tbx5. Cette approche a mené à la découverte de 6 transcrits alternatifs exhibant des fonctions à la fois communes et divergentes. La caractérisation de 2 de ces isoformes a révélé le rôle de l’isoforme long (Tbx5_v1) dans la régulation de la croissance des cardiomyocytes durant la cardiogénèse, tandis que l’isoforme court (Tbx5_v2), préférentiellement exprimé dans le cœur mature, réprime la croissance cellulaire. Il est donc entièrement concevable que les mutations de TBX5 entraînant une troncation de la région C-terminale accroissent la concentration d’une protéine mutée qui, à l’instar de Tbx5_v2, interfère avec la croissance de certaines structures cardiaques. En revanche, la divergence de fonctions de ces isoformes, caractérisée par les disparités de localisation subcellulaire et de d’interaction avec d’autres cofacteurs cardiaques, suggère que les mutations affectant davantage un isoforme favoriseraient l’émergence d’un type particulier de MCC. Finalement, un dernier objectif était d’identifier le ou les mécanisme(s) moléculaire(s) par le(s)quel(s) Tbx5 régule son principal gène cible, Nppa, et d’en extraire les indices qui éclairciraient sa fonction transcriptionnelle. Cet objectif nécessitait dans un premier lieu d’identifier les différents modules cis-régulateurs (MCRs) coordonnant la régulation transcriptionnelle de Nppa et Nppb, deux gènes natriurétiques dont l’organisation en tandem et le profil d’expression durant la cardiogénèse sont conservés dans la majorité des vertébrés. L’approche d’empreinte phylogénétique employée pour scanner le locus Nppb/Nppa a permis d’identifier trois MCRs conservés entre diverses espèces de mammifères, dont un (US3) est spécifique aux euthériens. Cette étude a corroboré que la régulation de l’expression du tandem génique Nppb/Nppa requérait l’activité transcriptionnelle d’enhancers en complément aux promoteurs de Nppa et Nppb. La concordance quasiment parfaite entre les profils d’expression de Tbx5 et de ces deux gènes natriurétiques chez les mammifères, suggère que le gradient d’expression ventriculaire de Tbx5 est interprété par le recrutement de ce facteur au niveau des différents enhancers identifiés. En somme, les études présentées dans cette thèse ont permis de clarifier la profusion de fonctions cardiaques que possède Tbx5. Certaines de ces fonctions émanent de l’épissage alternatif de Tbx5, qui favorise la synthèse d’isoformes dotés de propriétés spécifiques. Les diverses interactions combinatoires entre ces isoformes et d’autres facteurs cardiaques au sein des diverses sous-populations de progéniteurs cardiogènes contribuent à l’émergence de RRGs cardiaques divergents. / The vertebrate heart is a modular organ, which requires the complex patterning of the morphogenetic heart fields and the coordinated convergence of the diverse subpopulations of cardiogenic progenitors. At least 7 transcription factors of the T-box family cooperate within these numerous subpopulations of cardiogenic progenitors to regulate the morphogenesis and the layout of multiple structures along the primordial heart tube, which explains that the human mutations of these genes induce various congenital heart defects (CHDs). One of these T-box genes, Tbx5, whose haploinsufficiency generates the Holt-Oram syndrome (HOS), intervenes in a wide variety of gene regulatory networks (GRNs) that orchestrate the morphogenesis of the atria, the left ventricle, the mitral valve, the inter-atrial and inter-ventricular septa, as well as the cardiac conduction system. The diversity of GRNs involved in the formation of these cardiac structures suggests that Tbx5 holds a profusion of functions which will be identifiable only by indexing its molecular activities in each separately examined cardiac lineage. To address this problem, a conditional knockout of Tbx5 in the endocardium was generated. This experiment demonstrated a crucial role of Tbx5 in the survival of the endocardial cells lining the septum primum and the cardiomyocytes within this embryonic structure, which will contribute to the morphogenesis of the inter-atrial septum. Moreover, this study revealed a crosstalk between the Tbx5-positive endocardial cells subpopulation and the myocardium at the level of the septum primum to ensure the survival of cardiomyocytes, and ultimately to guarantee the maturation of the inter-atrial septum. Our results also confirmed the importance of genetic interdependence (Tbx5 and Gata4 as well as Tbx5 and Nos3) between different loci in the morphogenesis of the inter-atrial septum, and particularly the influence that the environment can have on the penetrance and the expressivity of atrial septal defects (ASDs) in the HOS. Besides, since the functions of a gene usually depend on the different isoforms it can generate, a second study focused more on the transcriptional aspect of Tbx5. This approach led to the discovery of 6 alternative transcripts exhibiting both common and specific functions. The characterization of 2 of these isoforms revealed the role of the long isoform (Tbx5_v1) in the regulation of cardiomyocytes growth during cardiogenesis, whereas the short isoform (Tbx5_v2), preferentially expressed in the mature heart, represses cell growth. It is thus entirely conceivable that TBX5 mutations leading to a C-terminal truncation increase the concentration of a mutated protein, which, like Tbx5_v2, interferes with the growth of certain cardiac structures. On the other hand, the divergence of functions of these isoforms, characterized by the disparities of subcellular localization and interaction with other cardiac cofactors, suggests that mutations affecting more one isoform would favor the emergence of a particular type of CHD. Finally, a last objective was to identify one or several molecular mechanism(s) by which Tbx5 regulates its main target gene, Nppa, and to extract clues that might clarify its transcriptional function. This objective required in a first place to identify the various cis-regulatory modules (CRMs) coordinating the transcriptional regulation of Nppa and Nppb, two natriuretic genes whose tandem organization and expression pattern during cardiogenesis are preserved in most vertebrates. The phylogenetic footprint approach employed to scan the Nppb/Nppa locus allowed the identification of three CRMs evolutionary conserved between different mammals species, one of which (US3) is specific to eutherians. This study confirmed that the regulation of the tandem genes Nppb/Nppa required the transcriptional activity of enhancers in complement to Nppa and Nppb promoters. The almost perfect concordance between the expression profiles of Tbx5 and these two natriuretic genes in mammals, suggests that the ventricular expression gradient of Tbx5 is interpreted by the recruitment of this factor to the identified enhancers. Altogether, the studies presented in this thesis allowed clarifying the profusion of Tbx5 cardiac functions. Some of these functions emanate from the alternative splicing of Tbx5, which favors the synthesis of isoforms endowed with specific properties. The diverse combinatorial interactions between these isoforms and other cardiac factors within the various cardiogenic progenitor subpopulations contribute to the emergence of distinct cardiac RRGs.