Dans un contexte où la demande en électricité varie fréquemment et que l'utilisation de production d'électricité intermittente telle que l'éolien et le solaire est privilégiée, il est nécessaire d'avoir un moyen de production pour stabiliser le réseau. L'hydroélectricité est une option attirante puisqu'elle est renouvelable et a un temps de réponse au changement très rapide. Cette stabilisation entraîne cependant une augmentation drastique d'arrêts et départs des turbines hydrauliques, ce qui a un effet néfaste sur la durée de vie. En effet, d'importantes contraintes ont été observées lors du démarrage de certaines turbines, sans pour autant que soit comprise l'origine de cet effet. Afin d'améliorer notre compréhension sur la dynamique de la turbine en démarrage, des mesures expérimentales sur modèle réduit en laboratoire sont essentielles. Cette thèse présente l'évolution du comportement dynamique de l'écoulement lors de différentes séquences de démarrage sur le modèle réduit d'une turbine bulbe. Les bancs d'essai en circuit fermé comme celui installé au laboratoire de machines hydrauliques (LAMH) ne permettent pas en général de conserver constante la chute d'essai pour des phénomènes transitoires comme un démarrage. Cette recherche propose donc une méthode ingénieuse pour surmonter ce problème. Afin d'obtenir des conditions similaires aux turbines prototypes, les caractéristiques d'ouverture des directrices (angle finale, vitesse d'ouverture) ont du être adaptées. Basées sur une analyse dimensionnelle, elles sont déterminées pour évaluer l'effet de la vitesse d'ouverture des directrices sous une chute constante. L'ouverture finale des directrices est déterminée pour atteindre une vitesse unitaire choisie pour le modèle. Les trois vitesses d'ouverture sélectionnées représentent, toutes proportions gardées, ce qui peut être observé sur des prototypes. Cette étude propose également une technique innovante pour obtenir l'évolution du débit au cours du démarrage à l'aide d'un système de vélocimétrie par image de particules résolu dans le temps. L'évolution des quantités globales de la turbine, telles que le couple et le débit, est différente pour les basses et la haute vitesses d'ouverture. Un certain degré d'universalité est observé sur les nombres sans dimension pour les deux faibles vitesses d'ouverture. Pour la haute vitesse, l'écoulement semble avoir de la difficulté à suivre l'évolution du schéma d'ouverture des directrices. D'autre part, l'analyse du champ de pression sur les pales lors de la phase d'accélération de la roue montre que la roue opère comme une turbine à impulsion jusqu'à ce que le couple atteigne sa valeur maximale. Même si la machine réagit davantage comme une turbine à réaction par la suite, des phénomènes locaux apparaissent et entraînent la formation de structures importantes dans la roue, comme par exemple un écoulement de retour. Lorsque la turbine atteint le régime sans charge, à la fin de l'ouverture, la roue tourne à la vitesse unitaire choisie. Dans ces conditions, deux phénomènes sous-synchrones et des tourbillons inter-aubes sont présents. Ces phénomènes sous-synchrones sont constitués de une ou deux structures tournant autour de l'axe de la roue, en alternance apparemment aléatoire. Ce caractère bistable influence le champ de pression sur les pales dont l'évolution dépend des structures présentes. De plus, l'intensité des tourbillons inter-aubes qui apparaissent lorsque le couple atteint sa valeur maximale est également influencée par ces phénomènes. / In a context where the electricity consumption varies frequently and the use of intermittent power generation such as wind and solar is trendy, it is necessary to have a means of generation to stabilize the grid. Hydropower is an attractive option because it is renewable and has a very fast response time to load variation. However, the use of hydraulic turbines to stabilize the grid leads to a drastic increase in their stops and startups, which has a critical effect on the lifetime of the runner. Indeed, significant stresses have been observed during the start-up of some turbines, without understanding where this effect comes from. In order to improve our understanding of the dynamics of the turbine during start-up, experimental measurements on scale models in a laboratory are essential. This thesis presents the evolution of the dynamic behaviour of the flow during different startup sequences on a model scale bulb turbine. Closed-loop test benches such as the one at Laval University's LAMH are interesting because they allow experimental analyses to be undertaken on a wide range of turbines while being compact. However, they do not generally allow to keep the pressure head constant for transient phenomena such as start-up. This study proposes a first attempt to study the transient regime of start-up on this type of test bench. In order to obtain conditions similar to prototype turbines, the opening characteristics of the guide vanes (final angle, opening speed) had to be adapted. Based on a dimensional analysis, the characteristics of the experiment are determined in order to evaluate the effect of the opening speed of the guide vanes under a constant pressure head. The final opening of the guide vanes is determined in order to reach a unit speed chosen for the model. The three selected opening speeds represent, in all proportions, what can be observed on prototypes. This study also proposes an innovative technique to obtain the evolution of the flow rate during start-up using a time-resolved particle image velocimetry system. The turbine shows a different behavior between the two lowest opening speeds and the high opening speeds. The global turbine quantities such as torque and flow rate have different evolutions. A certain degree of universality is observed on the dimensionless numbers for the two lowest opening speeds. For the high speed, the flow seems to have difficulty to follow the evolution of the opening sequence of the guide vanes. In addition to the study on the global quantities of the turbine, the analysis of the pressure field on the blades during the acceleration phase of the runner presents some interesting characteristics. It shows that the runner operates like an impulse turbine until the torque reaches its maximum value. Even if the machine works more like a reaction turbine afterwards, local phenomena occur which lead to the formation of important structures in the runner, such as a return flow. When the turbine reaches speed no-load regime, at the end of the opening, the runner rotates at the selected unit speed. Under these conditions, two sub-synchronous phenomena and inter-blade vortices are present. These sub-synchronous phenomena are made up of one or two structures rotating around the axis of the runner, in apparently random alternation. This bistable character influences the pressure field on the blades whose evolution depends on the structures present. In addition, the intensity of the inter-blade vortices that appear when the torque reaches its maximum value is also influenced by these phenomena.

published_or_final_version / Mechanical Engineering / Master / Master of Philosophy

This dissertation considers certain issues in support vector machines (SVMs), including a description of their construction, aspects of certain exponential kernels used in some SVMs, and a presentation of an algorithm that computes the necessary elements of their operation with proof of convergence. In its first section, this dissertation provides a reasonably complete description of SVMs and their theoretical basis, along with a few motivating examples and counterexamples. This section may be used as an accessible, stand-alone introduction to the subject of SVMs for the advanced undergraduate. Its second section provides a proof of the positive-definiteness of a certain useful function here called E and dened as follows: Let V be a complex inner product space. Let N be a function that maps a vector from V to its norm. Let p be a real number between 0 and 2 inclusive and for any in V , let ( be N() raised to the p-th power. Finally, let a be a positive real number. Then E() is exp(()). Although the result is not new (other proofs are known but involve deep properties of stochastic processes) this proof is accessible to advanced undergraduates with a decent grasp of linear algebra. Its final section presents an algorithm by Dr. Kallman (preprint), based on earlier Russian work by B.F. Mitchell, V.F Demyanov, and V.N. Malozemov, and proves its convergence. The section also discusses briefly architectural features of the algorithm expected to result in practical speed increases.

This thesis proposes a novel experimental environment (non-linear stochastic gradient descent, NL-SGD), as well as a novel online learning algorithm (OL SVM), for solving a classic nonlinear Soft Margin L1 Support Vector Machine (SVM) problem using a Stochastic Gradient Descent (SGD) algorithm. The NL-SGD implementation has a unique method of random sampling and alpha calculations. The developed code produces a competitive accuracy and speed in comparison with the solutions of the Direct L2 SVM obtained by software for Minimal Norm SVM (MN-SVM) and Non-Negative Iterative Single Data Algorithm (NN-ISDA). The latter two algorithms have shown excellent performances on large datasets; which is why we chose to compare NL-SGD and OL SVM to them. All experiments have been done under strict double (nested) cross-validation, and the results are reported in terms of accuracy and CPU times. OL SVM has been implemented within MATLAB and is compared to the classic Sequential Minimal Optimization (SMO) algorithm implemented within MATLAB's solver, fitcsvm. The experiments with OL SVM have been done using k-fold cross-validation and the results reported in % error and % speedup of CPU Time.

La minimisation des coûts et la maximisation de la satisfaction du client sont les facteurs majeurs dans le développement des activités d'une entreprise. Ceci pourrait être assuré par les bonnes pratiques de gestion et de prise de décision à travers un processus de planification hiérarchique bien déterminé: Stratégique, tactique et opérationnel. Dans la plupart des cas, les industries manufacturières ont tendance à planifier la production de façon à minimiser les coûts et remédier à la pénurie des produits. Ceci permet à l’entreprise d'avoir une vision stratégique permettant d'optimiser ses activités sur plusieurs plans (maintenance, approvisionnement, qualité...). Mais pour assurer une meilleure optimisation et assurer la continuité de l'entreprise, les managers doivent joindre la maintenance à la production et les considérer comme des systèmes interdépendants. Dans ce contexte, plusieurs travaux ont traité le cas de maintenance intégrée à la production et l'ont considéré comme outil d'optimisation afin de minimiser les coûts et assurer une certaine compétitivité. Des travaux ont aussi intégré la dégradation des machines pour planifier les actions de maintenance. Cependant, la production et la maintenance sont liées à diverses ressources physiques (ouvriers, machines etc..) que l'entreprise doit prendre en considération. Autrefois, les industries ont toujours tendance à se procurer toutes sortes de ressources physiques afin de satisfaire ses demandes. Ces ressources sont alors toujours disponibles, même si elles sont inutilisées pendant certaines périodes de production. Ayant conscience des pertes que peut causer ces ressources en cas de non utilisation, certaines entreprises choisissent de sous-traiter ou de louer quelques ressources pendant une période donné afin de satisfaire une demande bien déterminée. La particularité de notre travail consiste à trouver la quantité des ressources à louer afin de satisfaire la demande. En effet, en s'inspirant du modèle HMMS, qui étudie la planification de la production et des stocks en prenant en considération la variation de nombre de ressources humaines travaillant, on a eu l'idée de transformer le problème et de chercher alors le nombre optimal de machines à louer pendant chaque période. Ayant conscience aussi de l'importance de la planification de maintenance, celle-ci a été intégrée au plan de production afin de trouver la meilleure solution. Le nombre de machines à louer étant variable, on a considéré qu'il varie selon la demande des clients. Des stratégies de maintenance améliorées ainsi que de nouvelles contraintes ont été prises en considération lors la planification des plans de production et de maintenance du système étudié / Ameliorating the situation of an industry requires certainly reducing costs and maximizing the customer satisfaction. These two goals can't be achieved without a good management and a good knowledge of making decision tools. These decisions are generally associated, at least, with three levels of the hierarchical planning process: strategic, tactical and operational levels. Generally, manufacturing industries aims at determining the most adequate production planning which helps them to minimize costs and hedge against capacity shortages. Releasing such an efficient planning urges firms to have a global vision on their production process which may be looked upon as an inter-dependant set of sub systems performing various functions including ordering raw materials, assembling pieces, controlling quality, repairing machines, storage, etc. But, it is worth mentioning that to ensure the continuous productivity of a firm, managers have to focus especially on both production and maintenance sub-systems and consider them as strongly dependent systems. In this context, many researches were carried out in industrial field dealing with joint production and maintenance planning optimization as a solution to minimize costs and guard against important competitiveness. Such an optimization may be more efficient when considering the degradation of workstations over time while scheduling maintenance actions. Moreover, production and maintenance activities deal with diverse physical resources including mainly workforce and workstations. So, it is worth noting that, while making decisions and proposing planning, a special attention should be made on the manner of managing these resources which remain indispensable for the firm's sustainability. Previously, firms used to buy (and recruit) all physical resources required to fulfill orders. So, these resources are always available at the industry even if they are sometimes more and less needless and useless during some production periods. This may take place especially when the firm has to fulfill fluctuating demand causing consequently drastic costs. Aware about the superfluous costs of possessing, for a long period, resources which can be sometimes pointless, subcontracting and leasing have become very important for many manufacturing firms because of the advantage that these solutions can bring. So, nowadays, these industrial solutions are becoming more and more in vogue concerning mainly subcontracting workforce to perform some tasks (maintenance, supervision, audit,...) or leasing workstations in order to produce the required quantities. In the present work, we deal with a new concept, that no previous work has dealt with, consisting on leasing machines depending on the ordered quantities. In fact, inspired mainly from the HMMS model, carried out by a group of economists who studied the production, inventories and workforce planning considering variable number of human resources, we had the idea to transform the problem from a workforce planning problem to a workstations planning one. Also, aware about the drastic importance of managing the production and the maintenance planning simultaneously, we aim in the present work at suggesting a joint production and maintenance planning optimization under leasing constraint. Otherwise, in our work, dealing with many leased machines, we consider that the number of these used machines does not remain constant and may vary from a production period to another depending on a fluctuating demand over a finite time horizon. Moreover, in the present work we aim at performing a disturbance analysis and to study its contribution to the development of conjugated production and maintenance planning considering leasing constraint and taking into account used machines degradation. Improved maintenance strategy and related constraints are also studied in this work in order to obtain optimal production and maintenance schedules

Thesis. 1976. M.S.--Massachusetts Institute of Technology. Dept. of Mechanical Engineering. / Microfiche copy available in Archives and Engineering. / Includes bibliographical references. / M.S.

O presente trabalho apresenta o desenvolvimento de um protótipo de motor de indução linear tubular para aplicação na extração de petróleo. A principal função do motor é acionar diretamente uma bomba de sucção instalada no fundo de um poço em terra. São estabelecidas as diretrizes de projeto elétrico do motor, em conjunto com uma análise preliminar baseada na teoria convencional das máquinas elétricas. Propõe-se a aplicação do método das matrizes de transmissão para obter um modelo analítico mais preciso do motor de indução linear tubular. Também é apresentado o modelo obtido pela aplicação do Método dos Elementos Finitos. Os resultados dos modelos teóricos convencional e matricial são confrontados com os resultados do modelo por Elementos Finitos e de testes, mostrando boa concordância. De maneira geral, os resultados indicam que o motor tubular pode substituir com vantagens o sistema mecânico de superfície e o conjunto de hastes utilizados na extração de petróleo. / This report presents the development of a tubular linear induction motor for oil well applications. The aim of the motor is to directly drive an on-shore downhole piston pump with no rods. Guidance to the electric and magnetic design of the motor is established. Two forms of analysis are proposed: the rotating electrical machines (conventional) based and the transmission matrix based. A model using the Finite Element Method is also described. Conventional and matricial solutions are compared to the Finite Element and tests results, showing good agreement. In a general sense, the work results show that the tubular motor may substitute the mechanical surface system and the rod string with advantage.

The work described in this thesis is inspired by natural-occurring molecules that are used throughout biology to perform specific, highly-selective tasks. This thesis illustrates the design, synthesis and investigation of novel molecular devices based on acyl hydrazones for the synthesis of a small molecule walker and light-driven molecular shuttles. Chapter One outlines a general overview of the design and synthesis of molecular devices, including molecular walkers and molecular shuttles. Some of the most important examples of walking molecules (both natural and synthetic) are described in detail, along with a comprehensive introduction of molecular shuttles and their synthetic mimics. Examples of stimuli-responsive molecular shuttles that have been developed are highlighted throughout the chapter. Chapter Two describes the design and synthetic progress towards a molecular walker, as well as detailing the optimisation of the synthetic steps achieved thus far. In this chapter, most of the work presented is based on the design and optimisation of the synthesis of an acyl hydrazone-based molecular walker, which will be able to walk directionally and repetitively along its conjugate track when the conditions are changed. A novel acyl hydrazone pyridine moiety is introduced to the system to achieve a high directional bias during the walking process. First, the concept and basis of the design is explained and further, the synthesis of the walker system is discussed in detail. Chapter Three illustrates the synthesis and operation of 1- and 2- station [2]-rotaxanes which exhibit all the requirements for a light-driven molecular shuttle. The effect of a new photo switchable binding station, an acyl pyridyl hydrazone, on the shuttling process is investigated by comparing the positional distribution of the macrocycle between the acyl hydrazone station and the succinamide-ester station, while the acyl hydrazone undergoes photo- and thermal isomerisation. The successful synthesis of this molecular architecture is described along with its operation, demonstrating high positional integrity and efficiency during the shuttling process.

O presente trabalho apresenta o desenvolvimento de um protótipo de motor de indução linear tubular para aplicação na extração de petróleo. A principal função do motor é acionar diretamente uma bomba de sucção instalada no fundo de um poço em terra. São estabelecidas as diretrizes de projeto elétrico do motor, em conjunto com uma análise preliminar baseada na teoria convencional das máquinas elétricas. Propõe-se a aplicação do método das matrizes de transmissão para obter um modelo analítico mais preciso do motor de indução linear tubular. Também é apresentado o modelo obtido pela aplicação do Método dos Elementos Finitos. Os resultados dos modelos teóricos convencional e matricial são confrontados com os resultados do modelo por Elementos Finitos e de testes, mostrando boa concordância. De maneira geral, os resultados indicam que o motor tubular pode substituir com vantagens o sistema mecânico de superfície e o conjunto de hastes utilizados na extração de petróleo. / This report presents the development of a tubular linear induction motor for oil well applications. The aim of the motor is to directly drive an on-shore downhole piston pump with no rods. Guidance to the electric and magnetic design of the motor is established. Two forms of analysis are proposed: the rotating electrical machines (conventional) based and the transmission matrix based. A model using the Finite Element Method is also described. Conventional and matricial solutions are compared to the Finite Element and tests results, showing good agreement. In a general sense, the work results show that the tubular motor may substitute the mechanical surface system and the rod string with advantage.

L'objet de cette thèse est la conception de systèmes informatiques bases sur des unités spécialisées dans des fonctions ou des types de traitement particulier. Nous nous intéressons plus particulièrement a la programmation en logique comme type de langage d'un tel système. L'intérêt des machines spécialisées et les problèmes de communication et de coopération dans un système a unités spécialisées (distribution fonctionnelle) sont d'abord discutes. Nous nous intéressons ensuite aux machines bases de données, en présentant le projet Opale, qui vise à la conception d'une machine bases de données adaptée au traitement des connaissances. Les deux points essentiels de ce projet sont la définition d'une stratégie de recherche orientée vers le traitement d'ensembles de solutions, et un algorithme, implementable en matériel, permettant d'unifier des ensembles de buts avec des termes lus sur disque, ceci en suivant le débit de transfert du disque. Enfin, nous abordons le problème général de la conception de machines pour la programmation en logique. Diverses approches sont présentées et discutées à partir de mesures et évaluations sur l'activité d'un interpréteur Prolog