The concept of simulation-based engineering has been embraced by virtually every research and industry sector (Sinha, Liang et al. 2001; Mocko and Fenves 2003). Engineering and science communities have become increasingly aware that computer simulation is an indispensable tool for resolving a multitude of scientific and technological problems. It is clearly desirable to gain a reliable perspective on the behaviour of a system early in the design stage, long before building costly prototypes (Chul and Ro 2002; Letherwood, Gunter et al. 2004; Makarand Datar 2007; Ersal, Fathy et al. 2008; Mueller, Ferris et al. 2009). Simulation tools have become a critical part of the automotive industry due to their ability to reduce the time and money spent in the development process. Terrain is the principle source of vertical excitation to the vehicle and must be accurately represented in order to correctly predict the vehicle response in simulation. In this dissertation, non-deformable terrain surfaces are defined as a sequence of vectors, where each vector comprises terrain heights at locations oriented perpendicular to the direction of travel. The evolution and implications of terrain surface measurement techniques and existing methods for correcting INS drift are reviewed as a framework for a new compensation method for INS drift in terrain surface measurements. Each measurement is considered a combination of the true surface and the error surface, defined on a Hilbert vector space, in which the error is decomposed into drift (global error) and noise (local error). It is also desirable to develop a compact, path-specific, terrain surface representation that exploits the inherent anisotropicity in terrain over which vehicles traverse. In order to obtain this, a set of analytic basis vectors is formed from Gegenbauer polynomials, parameterized to approximate the empirical basis vectors of the true terrain surface. It is also desirable to evaluate vehicle models and tire models over a wide range of terrain types, but it is computationally impractical to store long distances of every terrain surface variation. This dissertation examines the terrain surface, rather than the terrain profile, to maximize the information available to the tire model (i.e. wheel path data). A method to decompose the terrain surface as a combination of deterministic and stochastic components is also developed. / Ph. D.

The double-sensor conductivity probe is one of the most commonly used techniques for obtaining local time-averaged parameters in two-phase flows. The uncertainty of this measurement technique has not been well understood in the past as it involves many different steps and influential factors in a typical measurement. This dissertation aims to address this gap by performing a systematic and comprehensive study on the measurement uncertainty of the probe. Three types of uncertainties are analyzed: that of measurands, of the model input parameters, and of the mathematical models. A Monte Carlo uncertainty evaluation framework closely simulating the actual measuring process is developed to link various uncertainty sources to the time-averaged two-phase flow quantities outputted by the probe. Based on the Monte Carlo uncertainty evaluation framework, an iteration method is developed to infer the true values of the quantities that are being measured. A better understanding of the uncertainty of the double-sensor conductivity probe is obtained. Multiple advanced techniques, such as high speed optical imaging and fast X-ray densitometry, recently become mature and easily accessible. To further improve the accuracy of local two-phase flow measurement, a method is developed to integrate these techniques with the double-sensor conductivity probe by considering the measuring principles and unique advantages of each technique. It has been demonstrated that after processing and synergizing the data from different techniques using the current integration method, the final results show improved accuracy for void fraction, gas velocity and superficial gas velocity, compared to the original probe measurements. High-resolution two-phase flow data is essential for the further development of various two-phase flow models and validation of two-phase CFD codes. Therefore, a comprehensive high-accuracy database of two-phase flows is acquired. The gas-phase information is obtained by the integration method developed in this dissertation, and the recently developed Particle Image Velocimetry and Planar Laser Induced Fluorescence (PIV-PLIF) technique is utilized to measure liquid-phase velocity and turbulence characteristics. Flow characteristics of bubbly flow, slug flow and churn-turbulent flow are investigated. The 1-D drift-flux model is re-evaluated by the newly obtained dataset. The distribution parameter model has been optimized based on a new void-profile classification method proposed in this study. The optimized drift-flux model has significant improvements in predicting both gas velocity and void fraction. / Doctor of Philosophy / The double-sensor conductivity probe is one widely used technique for measuring local time-averaged parameters in two-phase flows. Although a number of studies have been carried out in the past, a good understanding of the uncertainty of this technique is still lacking. This paper aims to address this gap by performing a systematic and comprehensive study on the measurement uncertainty of the probe. Three types of uncertainties are analyzed: that of measurands, of the model input parameters, and of the mathematical models. A better understanding of the uncertainty of the double-sensor conductivity probe has been obtained. Considering the unique measuring principles and advantages of multiple advanced techniques, a method is developed to integrate these techniques with the double-sensor conductivity probe to further improve the accuracy of local two-phase flow measurement. It has been demonstrated that the integration method significantly improves the accuracy of probe measurements. Realizing the needs of high-resolution two-phase flow data to the further development of various two-phase flow models and validation of two-phase CFD codes, a comprehensive database of two-phase flows is acquired. The gas-phase and liquid-phase information are acquired by the new integration method and the recently developed Particle Image Velocimetry and Planar Laser Induced Fluorescence (PIV-PLIF) technique, respectively. The classical 1-D drift-flux model is re-evaluated by the newly obtained dataset. The distribution parameter model has been optimized, resulting in significant improvements in predicting both gas velocity and void fraction.

During an earthquake, columns in a one-story building must support vertical gravity loads while undergoing large lateral drifts associated with deflections of the vertical seismic force resisting system and deflections of the flexible roof diaphragm. Analyzing the behavior of these gravity columns is complex since not only is there an interaction between compression and bending, but also the boundary conditions are not perfectly pinned or fixed. In this research, the behavior of steel columns that are square hollow structural sections (HSS) is investigated for stability using three design methods: elastic design, plastic hinge design, and pinned base design. First, for elastic design, the compression and flexural strength of the HSS columns are calculated according to the AISC specifications, and the story drift ratio that causes the interaction equation to be violated for varying axial force demands is examined. Then, a simplified design procedure is proposed; this procedure includes a modified interaction equation applicable to HSS column design based on a parameter, Pnh/Mn, and a set of design charts are provided. Second, a plastic hinge design is grounded in the concept that a stable plastic hinge makes the column continue to resist the gravity load while undergoing large drifts. Based on the available test data and the analytical results from finite element models, three limits on the width to thickness ratios are developed for steel square HSS columns. Lastly, for pinned base design, the detailing of a column base connection is schematically described. Using FE modeling, it is shown that it is possible to create rotational stiffness below a limit such that negligible moment develops at the column base. All the design methods are demonstrated with a design example / Master of Science / One-story buildings are one of the most economical types of structures built for industrial, commercial, or recreational use. During an earthquake, columns in a one-story building must support vertical gravity loads while undergoing large lateral displacements, referred to as story drift. Vertical loads cause compression forces, and lateral drifts produce bending moments. The interaction between these forces makes it more complex to analyze the behavior of these gravity columns. Moreover, since the column base is not perfectly fixed to the ground, there are many boundary conditions applicable to the column base depending on the fixity condition. For these reasons, the design for columns subjected to lateral drifts while supporting axial compressive forces has been a growing interest of researchers in the field. However, many researchers have focused more on wide-flange section (I-shape) steel columns rather than on tube section columns, known as hollow structural section (HSS) steel columns. In this research, the behavior of steel square tube section columns is investigated for stability using three design methods: elastic design, plastic hinge design, and pinned base design. First, for elastic design, the compression and flexural strength of the HSS columns are calculated according to current code equations, and the story drift that causes failure for varying axial force demands is examined. Then, a simplified design procedure is proposed including design charts. Second, a plastic hinge design is grounded in the concept that controlled yielding at the column base makes the column continue to resist the gravity load while undergoing large drifts. Based on the available test data and results from computational models, three limits on the width to thickness ratios of the tubes are developed. Lastly, for pinned base design, concepts for detailing a column base connection with negligible bending resistance is schematically described. Using a computational model, it is shown that the column base can be detailed to be sufficiently flexible to allow rotation. All the design methods are demonstrated with a design example.

Terrain surface measurement is an important tool in vehicle design work as well as pavement classification and health monitoring. �Non-deformable terrains are the primary excitation to vehicles traveling over it, and therefore it is important to be able to quantify these terrain surfaces. Knowledge of the terrain can be used in combination with vehicle models in order to predict force loads the vehicles would experience while driving over the terrain surface. �This is useful in vehicle design, as it can speed the design process through the use of simulation as opposed to prototype construction and durability testing. �Additionally, accurate terrain maps can be used by highway engineers and maintenance personnel to identify deterioration in road surface conditions for immediate correction. �Repeated measurements of terrain surfaces over an extended length of time can also allow for long term pavement health monitoring. Many systems have been designed to measure terrain surfaces, most of them historically single line profiles, with more modern equipment capable of capturing three dimensional measurements of the terrain surface. �These more modern systems are often constructed using a combination of various sensors which allow the system to measure the relative height of the terrain with respect to the terrain measurement system. �Additionally, these terrain measurement systems are also equipped with sensors which allow the system to be located in some global coordinate space, as well as the angular attitude of that system to be estimated. �Since all sensors return estimated values, with some uncertainty, the combination of a group of sensors serves to also combine their uncertainties, resulting in a system which is less precise than any of its individual components. �In order to predict the precision of the system, the individual probability densities of the components must be quantified, in some cases transformed, and finally combined in order to predict the system precision. �This thesis provides a proof-of-concept as to how such an evaluation of final precision can be performed. / Master of Science

The premise for AI systems like personal assistants to provide guidance and suggestions to an end-user is to understand, at any moment in time, the personal context that the user is in. The context – where the user is, what she is doing and with whom – allows the machine to represent the world in user’s terms. The context must be inferred from a stream of sensor readings generated by smart wearables such as smartphones and smartwatches, and the labels are acquired from the user directly. To perform robust context prediction in this real-world scenario, the machine must handle the egocentric nature of the context, adapt to the changing world and user, and maintain a bidirectional interaction with the user to ensure the user-machine alignment of world representations. To this end, the machine must learn incrementally on the input stream of sensor readings and user supervision. In this work, we: (i) introduce interactive classification in the wild and present knowledge drift (KD), a special form of concept drift, occurring due to world and user changes; (ii) develop simple and robust ML methods to tackle these scenarios; (iii) showcase the advantages of each of these methods in empirical evaluations on controlled synthetic and real-world data sets; (iv) design a flexible and modular architecture that combines the methods above to support context recognition in the wild; (v) present an evaluation with real users in a concrete social science use case.

Supporting fish populations by restocking with artificially bred young fish is a common practice worldwide, which may sometimes have a negative effect on natural genetic diversity. In this study, the currently ongoing restocking program of brown trout (Salmo trutta Linnaeus, 1758) in the subarctic Pasvik river in Norway was investigated. Previous investigations of this population showed a loss of genetic diversity and both genetic and morphological differentiation between historical breeding stocks and wild trout populations. However, the genetic impact of the breeding program used in this river today has never been investigated. To accomplish this, I compared the parent fish, the offspring, and fish caught in the river using 16 short tandem repeat loci and estimated pairwise relatedness, inbreeding, genetic distances, and differentiation. The results showed, by estimating the effective population size, that the number of parents used in the breeding program was generally insufficient to maintain a genetically sustainable river population. Further, genetic diversity in stocked fish was typically not maximized due to lack of performing all possible crossings between males and females. In combination, this induced genetic drift in the offspring generations and thus genetic differentiation away from the river populations. In summary, the results from this study show that the current stocking practises used in the Pasvik river are not sustainable and that measures need to be taken to prevent continued loss of genetic diversity in this large-growing trout population.

La minéralisation aurifère de la propriété Amaruq (Nunavut, CA) est recouverte par divers sédiments glaciaires ayant formé des traînées de débris minéralisés sur des surfaces importantes. Ce projet a pour objectif de caractériser la signature minéralogique et géochimique du gisement aurifère dans le till de surface afin de développer des méthodes de prospection pour les travaux d’exploration. 61 échantillons de till ont été récoltés le long de quatre transects pour déterminer les variations de composition dans la direction principale de l’écoulement glaciaire (NNO). Quatre profils verticaux d’ostioles, un par transect, ont été échantillonnés pour documenter la variation de la signature géochimique avec la profondeur. La signature chimique de la scheelite dans le gisement est comparée à celle retrouvée dans le till et permet de tester son utilisation comme outil de prospection glacio-sédimentaire. Trois des quatre profils d’ostioles présentent des concentrations stables des éléments traces avec la profondeur, indiquant une bonne homogénéisation du till par les processus cryogéniques. Les ostioles sont donc un milieu d’échantillonnage pertinent en région de pergélisol. Les analyses en composantes principales (ACP) de la géochimie des échantillons d’une campagne de till fournie par Mines Agnico Eagle ltée (AEM), et les données des transects, décrivent la variance de la roche encaissante mafique/ultramafique et de la minéralisation aurifére (CP1 et CP2 respectivement). Le krigeage des scores de la CP1 et CP2 indiquent des trains de dispersion vers le NNO, en aval glaciaire des zones minéralisées. Les comptages des minéraux indicateurs (MI : or et scheelite) augmentent de façon importante à environ 1.4 km en aval de la minéralisation dans le secteur est de la propriété, dans un till distal formant des formes drumlinoides. Dans le secteur ouest, recouvert d’un till proximal formant des crêtes morainiques, les comptages des MI et les scores de la CP1 augmentent directement en aval de la minéralisation. Les différences de distance de transport entre ces deux secteurs reflètent un changement de la dynamique glaciaire et démontrent l’importance du contexte local dans les travaux de prospection glaciaire. La comparaison de la signature géochimique de la scheelite entre le gisement et et des échantillons de till du transect de Whale Tail, grâce à des diagrammes discriminants et spectres des terres rares, montrent d’importantes similitudes, démontrant sa potentielle utilisation en prospection glacio-sédimentaire pour des gisements de type orogéniques. / On the Amaruq property (Nunavut, CA), the gold mineralization is covered by a patchwork of glacial sediments, including trains of mineralized debris dispersed over large surfaces. This project aims to define the deposit mineralogical and multi-elemental signatures in surface sediments to guide exploration in the area. A total of 61 samples were collected from frost boils along four NNW transects, parallel to the major ice flow direction, in order to assess the spatial variability of the deposit signature using till matrix geochemistry and indicator minerals. Four profiles in frost boils were sampled, one in each transect, in order to document trace elements depth variation. The scheelite geochemical signature from the deposit is compared to that in till from the Whale Tail transect to test its applicability in drift prospecting. Three out of four frost boil profiles exposed constant trace elements concentrations at depth, reflecting a good till homogenization by cryoturbation. Therefore, frost boils are a suitable medium of sampling for exploration purpose in permafrost terrain. Principal component analysis (PCA) of a previous till survey by Agnico Eagle Mines ldt (AEM) and transect samples define the ultramafic/mafic host rock of the mineralization and a pathfinder element suite characteristic of lode gold deposits (PC1 and PC2, respectively). Kriging of PC1 and PC2 defines a dispersal train oriented NNW, down ice of known mineralization. In the eastern part of the property, covered mostly by a distal till forming drumlinoid ridges, indicator mineral counts (IM: gold and scheelite) increase to a maximum approximately 1.4 km down ice from the outcropping zone. Western transects, largely covered by a proximal till forming morainal ridges, show a sharp increase of counts directly down-ice of mineralization. The difference in glacial transport between these two sectors represents a local change of ice dynamic and shows the importance of understanding the local context in exploration when interpreting glacial dispersal patterns. The comparison of the trace element signature between the deposit and the till scheelite using discriminant diagrams and REE abundance shows similarities, supporting its usefullness in drift prospecting for orogenic gold deposits.

Les propriétés hydrauliques d'un till du Nord du Québec en condition saturée et non saturée ont été étudiées en laboratoire à l'aide de l'essai de conductivité hydraulique et l'essai de succion qui permet d'obtenir la courbe de rétention d'eau. Les essais ont été réalisés avec des échantillons compactés sous différentes conditions de teneur en eau et de masse volumique sèche, de manière à évaluer l'influence de ces conditions sur les propriétés hydrauliques. Les études précédentes ont montré qu'une structure agrégée se développe dans le till lorsqu'il est compacté de côté sec de l'optimum Proctor, c'est-à-dire à un degré de saturation de compactage inférieur au degré de saturation optimum (Sr0 < Sr-opt). Le till étudié dans le cadre de cette étude diffère de ceux utilisés dans les études précédentes par son très faible pourcentage de particules argileuses (< 2 μm) de 0,7%. L'objectif principal de cette étude est d'évaluer l'influence des conditions de compactage sur la conductivité hydraulique et la courbe de rétention d'eau et de déterminer si une structure agrégée se développe lorsque le till est compacté du côté sec de l'optimum. Les résultats expérimentaux typiques de la conductivité hydraulique saturée obtenus pour les échantillons compactés du côté humide de l'optimum (Sr0 > Sr-opt) avec un indice des vides de 0,328 varient de 2,6 x 10-7 à 8,6 x 10-7 m/s. Les conductivités hydrauliques typiques obtenues pour les échantillons compactés du côté sec de l'optimum (Sr0 < Sr-opt) avec le même indice des vides étaient très similaires, variant entre 2,4 x 10-7 et 9,1 x 10-7 m/s. Les courbes de rétention d'eau mesurées en laboratoire sont très semblables entre elles, présentant toutes une structure de sol homogène, peu importe si l'échantillon a été compacté du côté sec ou du côté humide. Il apparait que la conductivité hydraulique et la courbe de rétention d'eau de ce till sont très peu influencées par le degré de saturation de compactage. Ceci peut être expliqué par le très faible pourcentage de particules argileuses, qui prévient la formation d'agrégations de particules dans la structure du till lorsqu'il est compacté du côté sec de l'optimum Proctor. Plusieurs modèles d'estimation de la conductivité hydraulique sont utilisés. Cette étude propose trois nouveaux modèles d'estimation de la conductivité hydraulique saturée pour les tills du Nord du Québec. Le premier modèle met en relation la conductivité hydraulique saturée avec le diamètre des grains de la fraction fine (d50 FF) et le deuxième modèle est une équation modifiée de Kozeny-Carman et Hazen basée sur la porosité et une taille de particule effective (d10). Finalement, un modèle permettant d'évaluer l'effet de l'agrégation avec l'équation de Kozeny-Carman modifiée dans les tills compactés est proposé.

This study examined the Voice onset time (VOT) of Mandarin and English initial position stops as produced by native American-English speaking learners of Mandarin and also native speakers. The study primarily investigated the role of second language experience on native-like production of L2 VOT for the Mandarin stops /p, t, k, ph, th, kh/ and additionally examined any evidence for phonetic drift in the L1 English VOT as a result of L2 Mandarin learning and acquisition. To achieve this, the production of L2 Mandarin and L1 English by native AmericanEnglish speaking learners of Mandarin of two different levels of experience (beginner and advanced learners) were compared to that of monolingual Native speakers of Mandarin andEnglish. In total, thirty native American-English speaking learners of Mandarin, eight monolingual native Mandarin speakers, and eight monolingual American-English speakers participated in the study. Data was elicited via participants reading aloud prepared lists of Mandarin and/or English words. Statistical testing indicates that as a group beginner learners produced significantly different mean VOT than native Mandarin speakers for all six stops. In contrast, the advanced learner group's mean VOT only significantly differed with native speakers for /t/, indicating native-like VOT for the remaining five stops. Individual analysis of VOT data for L2 Mandarin long-lag stops /ph, th, kh/ showed that advanced (7 of 15) and even some beginner (3 of 15) participants were capable of producing native-like VOT for the whole set of long-lag stops and also that a number of participants were able to produce native-like VOT for some but not all of the long-lag stops. Logistic regression analysis examining the effect and influence of L2 experience and demographic related variables indicates only gender and Mandarin-speaking church missionary service to have a significant effect on native-like VOT probability with females and returned missionaries significantly more likely to achieve native-like VOT. Analysis of learner participants' L1 English VOT showed that for all three of the English long-lag stops (/p, t, k/), advanced learners of Mandarin produced significantly longer VOT than monolingual English speakers, indicating L1 phonetic drift in the form of assimilation towards the longer length of Mandarin long-lag stops. Significant difference was only found between beginner learner English VOT and monolingual VOT for /p/, suggesting instances of significant L1 drift increase when more L2 experience has been gained.

Les tills compactés sont largement utilisés comme noyau des barrages en remblai dans le Nord québécois. La mise en place en différentes couches successives laisse présager un rapport d’anisotropie (rK) de la perméabilité différent de l’unité. Différents montages expérimentaux sont disponibles dans la littérature pour évaluer l’anisotropie de la conductivité hydraulique sur plusieurs types de sols. Toutefois, ces méthodes présentent deux principaux désavantages : i) le rapport d’anisotropie est évalué sur deux échantillons distincts et ii) la procédure expérimentale peut potentiellement remanier l’échantillon par des manipulations hasardeuses. Dans l’objectif d’améliorer la caractérisation du rapport d’anisotropie des tills compactés, un nouveau perméamètre cubique a été conçu. Cette nouvelle cellule permet de quantifier le paramètre rK dans le domaine saturé ou quasi-saturé, et ce sans démanteler la cellule. Successivement, une percolation verticale et une horizontale sont effectuées sur un même échantillon. Une procédure spéciale a été développée afin de limiter l’écoulement dans les pierres poreuses inactives, i.e. parallèles à l’écoulement. Une solution de gélatine bovine qui est solide à une température inférieure à 30°C est utilisée afin de sceller ces pierres inactives. De l’eau chaude à 60°C permet de chasser cette gélatine sans démanteler la cellule. Un nouvel essai de perméabilité peut ainsi être effectué sur le même échantillon, mais selon une autre direction. Le nouveau perméamètre et la procédure ont permis d’évaluer l’effet de la mise en place et de la teneur en pierre sur rK. Pour une analyse dans le domaine quasi-saturé, le rapport d’anisotropie varie de 3 à 7 pour une mise en place du côté sec de la courbe Proctor. Pour le côté humide, le rapport d’anisotropie est près de l’unité. Cependant, la procédure développée pour des conditions saturées devrait être utilisée pour confirmer ces résultats. En conclusion, un outil intéressant a été conçu pour quantifier le rapport d’anisotropie des tills compactés, une problématique importante pour l’analyse des barrages en remblai. Mots clés: perméabilité, anisotropie, perméamètre cubique, tills compactés, barrage en remblai / Compacted tills are widely use as core of embankment dams in northern Quebec. It is expected that placement method in successive layers may result in a permeability anisotropy different from unity. Various experimental techniques are available to assess the permeability anisotropy for different types of materials. However, these methods have two principal disadvantages: i) the anisotropy rapport is evaluated on separate samples and ii) the experimental procedure could potentially disturb the sample by inappropriate manipulations. In order to improve the characterization of the anisotropy rapport of compacted tills, a new square cell was built. This new cell is able to quantify the anisotropy rapport in both states, saturated and almost saturated conditions. The permeability under vertical and horizontal flow is assessed on a single sample and a new experimental procedure has been developed to minimize preferential seepage paths along the porous stones. Injection of bovine gelatin is used to temporarily seal the porous stones perpendicular to the flow path (unused). Continuous hot water flow is used to extract the gelatin once the permeability testing is done. This new device and procedure have been successfully tested to evaluate the effect of placing method and stone content on the permeability anisotropy of compacted tills. For an analysis in quasi-saturated state, the anisotropy rapport is greater than 3 or 7 for a dry compaction. On the wet side, the anisotropy rapport is close to one. However, the procedure developed in the saturated condition should be used to confirm these results. In conclusion, this project has helped to build a useful tool to quantify the anisotropy rapport of tills, an important issue in the evaluation of embankment dams. Key words: permeability, anisotropy, square permeability cell, compacted tills, embankment dam.