Дисертації з теми "Proportional machines"

Ce travail de recherche est réalisé dans le cadre d'une collaboration internationale entre l'Université Technique de Cluj-Napoca en Roumanie et l'Ecole Centrale de Lille en France. Cette thèse a comme premier objectif, de réduire l’ondulation de couple d’une machine triphasée 6/8 à réluctance variable grâce aux paramètres de contrôle dans le mode de fonctionnement moteur et générateur. Par conséquent, un modèle analytique a été développé et mis en œuvre dans un environnement de calcul numérique. Un outil d'optimisation permettant de trouver les meilleurs paramètres de contrôle a été utilisé. Toutefois, le processus d’optimisation est très lent.En conséquence, une stratégie pour réduire le temps du processus d'optimisation sans abaisser la précision des résultats est proposée. Une stratégie de ‘mapping’ est utilisée entre un modèle grossier mais rapide et un modèle fin mais lent. Un travail est mené sur le choix du modèle grossier, dans le cadre d’un modèle fournissant des valeurs en fonction du temps. Deux stratégies de ‘mapping’ ont été utilisées dans cette thèse : l’‘Output Space Mapping Proportional’(OSMP) et le Manifold Mapping (MM). Ces deux stratégies sont utilisées pour déterminer les paramètres de contrôle optimaux. Le problème d’optimisation est ensuite complété en ajoutant des variables géométriques au processus d’optimisation du contrôle. Les outils développés sont ensuite utilisés dans le cadre du dimensionnement par optimisation d’un alterno-démarreur automobile. Finalement, des expériences et des essais sur un prototype de Machine à RV sont menés afin de valider les résultats du processus d'optimisation en mode moteur et générateur
The switched reluctance machine has a simple construction making it cheaper in execution but one of the drawbacks of this machine is the torque ripple. This thesis had as first target, the usage of an optimization tool to calculate the best control parameters to correct this major drawback in motor and generator operation modes. Hence, an analytical model that takes into account the machine geometry and that is able to be simulated in both operation modes by adjusting the commutation angles was provided and implemented in a calculation environment. The second target of this work is to investigate a method to reduce the optimization time without lowering the accuracy of the results. The strategy used in the optimization process is known in literature as the space mapping technique, more precisely for this thesis output space mapping proportional and manifold mapping were studied. After testing them on a mathematical model it was possible to continue the investigation on defining the optimal control parameters of a three-phases 6/8 SR machine prototype, being able that further to use this strategy in a much complicated process, i.e. defining the right geometry and control of a SR machine to be used in automotive integrated starter alternator systems. The final target of the thesis was to conduct experiments and tests on the existing prototype in order to partially validate the results of the optimization process

The organic materials used for stator winding insulation are subject to deterioration from thermal, electrical, and mechanical stresses. Stator winding insulation breakdown due to excessive thermal stress is one of the major causes of electric machine failures; therefore, prevention of such a failure is crucial for increasing machine reliability and minimizing financial loss due to motor failure. This work focuses on the development of an efficient and reliable stator winding temperature estimation scheme for small to medium size mains-fed induction machines. The motivation for the stator winding temperature estimation is to develop a sensorless temperature monitoring scheme and provide an accurate temperature estimate that is capable of responding to the changes in the motors cooling capability. A discussion on the two major types of temperature estimation techniques, thermal model-based and parameter-based temperature techniques, reveals that neither method can protect motors without sacrificing the estimation accuracy or motor performance. Based on the evaluation of the advantages and disadvantages of these two types of temperature estimation techniques, a new online stator winding temperature estimation scheme for small to medium size mains-fed induction machines is proposed in this work. The new stator winding temperature estimation scheme is based on a hybrid thermal model. By correlating the rotor temperature with the stator temperature, the hybrid thermal model unifies the thermal model-based and the parameter-based temperature estimation techniques. Experimental results validate the proposed scheme for stator winding temperature monitoring. The entire algorithm is fast, efficient and reliable, making it suitable for implementation in real time stator winding temperature monitoring.

Магістерська дисертація: 101 с., 26 рис., 22 табл., 1 додаток, 55 джерел. Актуальність. Задачі теорії розкладів широко вивчаються в літературі останні 50 років, їх розгляду приділяється значна увага з боку багатьох вчених і фахівців-практиків по всьому світу. Складні реальні задачі, з якими стикаються різні типи компаній, стали викликом для пошуку шляхів їх розв’язання. Результати отримані дослідженнями різних науковців свідчать про актуальність проблеми та потребу в її подальшому дослідженні. Таким чином, можна з упевненістю сказати, що питання теорії розкладів є одним з найбільш поширених питань в галузі дослідження операцій і наук управління. Проблематика теорії розкладів охоплює дослідження обчислювальної складності задач, розробку точних, наближених та евристичних алгоритмів їх розв’язання. При цьому більшість праць присвячено розвитку комбінаторних підходів. Однак, як показує практика, можливості комбінаторних алгоритмів суттєво обмежені розмірністю розв’язуваних задач. У зв’язку з цим актуальною є розробка програмного продукту для складання розкладу виконання завдань паралельними пристроями, який допоможе мінімізувати сумарне відхилення від директивних строків. Зв'язок роботи з науковими програмами, планами, темами. Робота виконувалась на кафедрі автоматизованих систем обробки інформації та управління Національного технічного університету України «Київський політехнічний інститут ім. Ігоря Сікорського» в рамках теми «Ефективні методи розв’язання задач теорії розкладів» (№ 0117U000919). Мета і завдання дослідження – підвищення якості розв’язку задач календарного планування за рахунок побудови оптимального чи близького до оптимального розкладу, що дозволяє мінімізувати середній час відхилення від директивних строків. 4 Для досягнення мети необхідно виконати наступні завдання:  виконати аналітичний огляд відомих результатів, з розв’язання поставленої в рамках роботи задачі;  розробити алгоритм створення календарного плану виконання завдань паралельними пристроями, що мінімізує сумарне відхилення моментів завершення завдань від директивних строків;  розробити програмну реалізацію розробленого алгоритму;  виконати аналіз отриманих результатів. Об’єкт дослідження – оперативно-календарне планування дрібносерійного виробництва. Предмет дослідження – моделі та методи розв’язання задач календарного планування з метою мінімізації сумарного відхилення виконання завдань від директивних строків паралельними пристроями. Методи дослідження, застосовані в роботі, базуються на методах та алгоритмах теорії розкладів та дослідження операцій. Наукова новизна отриманих результатів. Розроблено евристичний алгоритм розв‘язання задачі мінімізації сумарного відхилення від спільного директивного строку при виконанні завдань паралельними пристроями. Публікації. Матеріали роботи опубліковано у Міжнародному науковому журналі «Науковий огляд» – №9(14). – с. 14–32 у 2017 році [1], також матеріали опубліковано в тезах 20-ї Міжнародної науково-технічної конференції SAIT 2018 [2], наукової конференції студентів, магістрантів та аспірантів «Інформатика та обчислювальна техніка» – ІОТ-2018 [3].
Master’s thesis: 101 pages, 26 figures, 22 tables, 1 appendix, 55 references. Relevance. Scheduling problems have been widely studying in the literature for 50 years, considerable attention from many scholars and practitioners around the world is given to their consideration. The complex real problems faced by different types of companies have become a challenge to find ways to solve them. The results obtained by researches of various scholars indicate the urgency of the problem and the need for its further research. Thus, one can confidently say that the question of the scheduling problems is one of the most common issues in the field of operations research and management sciences. The question of scheduling problems covers the studying the complexity of problems, the development of accurate, approximate and heuristic algorithms for their solution. At the same time, most works are devoted to the development of combinatorial approaches. However, as practice shows, the possibilities of combinatorial algorithms are essentially limited by the dimensionality of solvable problems. In this regard, it is relevant to develop a software product to scheduling tasks by parallel machines, which will help minimize the total deviation from a common due date. Relationship of work with scientific programs, plans, themes. The work was carried out at National Technical University of Ukraine «Igor Sikorsky Kyiv Polytechnic Institute» the department of Computer-Aided Management and Data Processing Systems within the theme “Effective methods for solving the problems of the theory of schedules” (state registration number 0117U000919). Purpose and objectives of the study. Improving the quality of solving problem scheduling tasks at the expense of construction an optimal or close to optimal schedule, which minimizes the average deviation time from due dates. The following tasks:  performing the known scheduling results analytical review;  developing an algorithm for minimizing the total deviation for common due date by parallel machine scheduling; 6  developing a software implementation of the algorithm in a form that can be used for schedule optimizing;  performing an analysis of the results. The object of study is operational planning of small-scale production. Subject of research: models and methods for solving scheduling problems in order to minimize the total deviation of completion times from a common due date by parallel machines. Scientific novelty of the research. A heuristic algorithm for solving the problem of minimizing the total deviation from the common policy term when performing tasks with parallel devices is developed. Publications. Materials of the work are published in the International Scientific Journal "Scientific Review" - № 9 (14). - p. 14-32 in 2017 [1], materials are also published in the theses of the 20th International Scientific and Technical Conference SAIT 2018 [2] and in the scientific conference of students, graduate students and postgraduates "Informatics and Computing" 2018 [3].

Le travail de mémoire porte sur la synthèse de commandes robustes pour machines synchrones à aimants permanents en tenant compte des amortisseurs et de variations de paramètres de la machine. Dans un premier temps, une commande basée sur le contrôle vectoriel et sur l'utilisation de correcteurs classiques de type proportionnel intégral a été élaborée. Le dimensionnement de ces correcteurs a été facilité par l'introduction d'un algorithme de découplage des courants statoriques direct et en quadrature. Dans un deuxième temps, nous nous sommes intéressé à la synthèse de commandes robustes et particulièrement, au contrôle H infini et à la commande par mode de glissement. Elles permettent la prise en compte des différentes perturbations et assurent en outre, un bon compromis robustesse/performances

Supplying haptic or force feedback to operators using hydraulic machinery such as excavators has the potential to increase operator capabilities. Haptic, robotic, human-machine interfaces enable several enhancing features including coordinated motion control and programmable haptic feedback. Coordinated or resolved motion control supplies a more intuitive means of specifying the equipment's motion. Haptic feedback is used to relay meaningful information back to the user in the form of force signals about digging force acting on the bucket, programmable virtual constraints and system limitations imposed by the mechanism, maximum pressure or maximum flow. In order to make this technology economically viable, the benefits must offset the additional cost associated with implementation. One way to minimize this cost is to not use high-end hydraulic components. For smaller backhoes and mini-excavators this means that the hydraulic systems are comprised of a constant displacement pump and proportional direction control valves. Hydraulic and haptic control techniques suitable for backhoes/excavators are developed and tested on a small backhoe test-bed. A virtual backhoe simulator is created for controller design and human evaluation. Not only is the virtual simulator modeled after the test-bed, but the control algorithm used in the simulator is the same as the actual backhoe test-bed. Data from human subject tests are presented that evaluate the control strategies on both the real and virtual backhoe. The end goal of this project is to incorporate coordinated haptic control algorithms that work with low-cost systems and maximize the enhancement of operator capabilities.

Le raisonnement par analogie est reconnu comme une des principales caractéristiques de l'intelligence humaine. En tant que tel, il a pendant longtemps été étudié par les philosophes et les psychologues, mais de récents travaux s'intéressent aussi à sa modélisation d'un point de vue formel à l'aide de proportions analogiques, permettant l'implémentation de programmes informatiques. Nous nous intéressons ici à l'utilisation des proportions analogiques à des fins prédictives, dans un contexte d'apprentissage artificiel. Dans de récents travaux, les classifieurs analogiques ont montré qu'ils sont capables d'obtenir d'excellentes performances sur certains problèmes artificiels, là où d'autres techniques traditionnelles d'apprentissage se montrent beaucoup moins efficaces. Partant de cette observation empirique, cette thèse s'intéresse à deux axes principaux de recherche. Le premier sera de confronter le raisonnement par proportion analogique à des applications pratiques, afin d'étudier la viabilité de l'approche analogique sur des problèmes concrets. Le second axe de recherche sera d'étudier les classifieurs analogiques d'un point de vue théorique, car jusqu'à présent ceux-ci n'étaient connus que grâce à leurs définitions algorithmiques. Les propriétés théoriques qui découleront nous permettront de comprendre plus précisément leurs forces, ainsi que leurs faiblesses. Comme domaine d'application, nous avons choisi celui des systèmes de recommandation. On reproche souvent à ces derniers de manquer de nouveauté ou de surprise dans les recommandations qui sont adressées aux utilisateurs. Le raisonnement par analogie, capable de mettre en relation des objets en apparence différents, nous est apparu comme un outil potentiel pour répondre à ce problème. Nos expériences montreront que les systèmes analogiques ont tendance à produire des recommandations d'une qualité comparable à celle des méthodes existantes, mais que leur complexité algorithmique cubique les pénalise trop fortement pour prétendre à des applications pratiques où le temps de calcul est une des contraintes principales. Du côté théorique, une contribution majeure de cette thèse est de proposer une définition fonctionnelle des classifieurs analogiques, qui a la particularité d'unifier les approches préexistantes. Cette définition fonctionnelle nous permettra de clairement identifier les liens sous-jacents entre l'approche analogique et l'approche par k plus-proches-voisins, tant au plan algorithmique de haut niveau qu'au plan des propriétés théoriques (taux d'erreur notamment). De plus, nous avons pu identifier un critère qui rend l'application de notre principe d'inférence analogique parfaitement certaine (c'est-à-dire sans erreur), exhibant ainsi les propriétés linéaires du raisonnement par analogie
Analogical reasoning is recognized as a core component of human intelligence. It has been extensively studied from philosophical and psychological viewpoints, but recent works also address the modeling of analogical reasoning for computational purposes, particularly focused on analogical proportions. We are interested here in the use of analogical proportions for making predictions, in a machine learning context. In recent works, analogy-based classifiers have achieved noteworthy performances, in particular by performing well on some artificial problems where other traditional methods tend to fail. Starting from this empirical observation, the goal of this thesis is twofold. The first topic of research is to assess the relevance of analogical learners on real-world, practical application problems. The second topic is to exhibit meaningful theoretical properties of analogical classifiers, which were yet only empirically studied. The field of application that was chosen for assessing the suitability of analogical classifiers in real-world setting is the topic of recommender systems. A common reproach addressed towards recommender systems is that they often lack of novelty and diversity in their recommendations. As a way of establishing links between seemingly unrelated objects, analogy was thought as a way to overcome this issue. Experiments here show that while offering sometimes similar accuracy performances to those of basic classical approaches, analogical classifiers still suffer from their algorithmic complexity. On the theoretical side, a key contribution of this thesis is to provide a functional definition of analogical classifiers, that unifies the various pre-existing approaches. So far, only algorithmic definitions were known, making it difficult to lead a thorough theoretical study. From this functional definition, we clearly identified the links between our approach and that of the nearest neighbors classifiers, in terms of process and in terms of accuracy. We were also able to identify a criterion that ensures a safe application of our analogical inference principle, which allows us to characterize analogical reasoning as some sort of linear process

Recent technological developments have implemented the use of proportional control in prosthetic hands, giving rise to the importance of appropriate myoelectric control. EMG models in the past have assumed a linear proportionality to simplify the EMG-force relationships. However, it has been shown that a non-linear EMG-force relationship may be a more effective model. This study focused on evaluating three different control algorithms for a novel myoelectric training device, consisting of a toy car controlled by EMG signals from the distal muscles in the arm. Sixteen healthy adult subjects (5 male and 11 female) with an average age of 23.6 years (SD = 2.7) were asked to drive the car through a slalom course. Completion times as well as number of errors (wall hits, cone hits, and reversals) were recorded to evaluate performance. The NASA TLX was administered to evaluate psychometrics such as mental demand, physical demand, frustration, and overall workload. The average total errors per trial on the final day of testing using the linear proportional algorithm was found to be statistically significantly (p < 0.05) lower than digital and non-linear proportional. The average course completion time per trial and overall workload using the non-linear proportional algorithm was found to be statistically significantly (p < 0.05) lower than digital and linear proportional. These results suggest that a non-linear algorithm would be most appropriate for myoelectric control in prosthetic hands.

L’insertion massive de sources interfacées par Électronique de Puissance (EP) dans le réseau va induire de grands défis techniques à surmonter pour garantir la stabilité du système électrique. L’une des principales préoccupations tient aux caractéristiques de ces sources, qui sont différentes de celles des Machines Synchrones (MS). Ce travail apporte un éclairage supplémentaire sur l’effet de l’insertion massive des sources interfacées par EP dans le système électrique, plus spécifiquement sur l’évolution du comportement de la fréquence de ces systèmes. Pour y répondre, un ensemble d’outils et de méthodes s’inspirant des fondamentaux des systèmes basés sur la production d’électricité avec les MSs a été développé. Des modèles d’ordre réduit des sources interfacées par EP se comportant comme sources de tension ou de courant sont proposés. Comme ces représentations partagent la même approche théorique que les modèles de MSs, il est ensuite possible de les fusionner en vue de l’analyse des systèmes électriques quelconques, indépendamment leur complexité. Les systèmes d’ordre réduit peuvent soit être modélisés en représentant chaque source ou bien en utilisant le concept des zones synchrones, ce qui offre une grande flexibilité d’analyse. Avec une approche progressive, le comportement de la fréquence des systèmes académiques relevé avec la méthodologie proposée est analysé en profondeur et comparé avec celui obtenu avec des modèles détaillés. L’approche développée dans ce travail peut être utilisée pour déterminer l’impact du ratio de pénétration des sources interfacées par EP sur le comportement global de la fréquence, mais aussi pour évaluer les oscillations intra et interzones des différents systèmes électriques
The massive insertion of Power Electronics (PE)-based sources into the grid is creating some technical challenges which must be overcome in order to guarantee the stability of the power system. One of the main concerns is due to the characteristics of these sources, which are different to those of Synchronous Machines (SMs). This research intends to provide a supplementary insight into the massive insertion of PE-based sources into the power system, more specifically the evolution of the frequency behaviour of these systems. For this purpose, a set of methods and tools already used on classical SM-based systems is proposed. Reduced order models of PE-based sources behaving as voltage or current sources are proposed, and, since the obtained representations share the same theoretical approach as that employed for SMs, they can be employed together to analyse any power system, regardless its complexity. The reduced order model of the systems can either be designed by representing each source of the original system or using the concept of synchronous areas, providing flexibility in the analysis. Using a step-by-step approach, the frequency behaviours observed with academic systems are deeply analysed with the proposed methodology, and validated with those obtained with detailed EMT models. The proposed approach developed in this work can be used to determine the impact of the penetration rate of PE-based sources on the overall frequency behaviour, and on the intra- and inter-area oscillations of different power systems

L’insertion massive de sources interfacées par Électronique de Puissance (EP) dans le réseau va induire de grands défis techniques à surmonter pour garantir la stabilité du système électrique. L’une des principales préoccupations tient aux caractéristiques de ces sources, qui sont différentes de celles des Machines Synchrones (MS). Ce travail apporte un éclairage supplémentaire sur l’effet de l’insertion massive des sources interfacées par EP dans le système électrique, plus spécifiquement sur l’évolution du comportement de la fréquence de ces systèmes. Pour y répondre, un ensemble d’outils et de méthodes s’inspirant des fondamentaux des systèmes basés sur la production d’électricité avec les MSs a été développé. Des modèles d’ordre réduit des sources interfacées par EP se comportant comme sources de tension ou de courant sont proposés. Comme ces représentations partagent la même approche théorique que les modèles de MSs, il est ensuite possible de les fusionner en vue de l’analyse des systèmes électriques quelconques, indépendamment leur complexité. Les systèmes d’ordre réduit peuvent soit être modélisés en représentant chaque source ou bien en utilisant le concept des zones synchrones, ce qui offre une grande flexibilité d’analyse. Avec une approche progressive, le comportement de la fréquence des systèmes académiques relevé avec la méthodologie proposée est analysé en profondeur et comparé avec celui obtenu avec des modèles détaillés. L’approche développée dans ce travail peut être utilisée pour déterminer l’impact du ratio de pénétration des sources interfacées par EP sur le comportement global de la fréquence, mais aussi pour évaluer les oscillations intra et interzones des différents systèmes électriques
The massive insertion of Power Electronics (PE)-based sources into the grid is creating some technical challenges which must be overcome in order to guarantee the stability of the power system. One of the main concerns is due to the characteristics of these sources, which are different to those of Synchronous Machines (SMs). This research intends to provide a supplementary insight into the massive insertion of PE-based sources into the power system, more specifically the evolution of the frequency behaviour of these systems. For this purpose, a set of methods and tools already used on classical SM-based systems is proposed. Reduced order models of PE-based sources behaving as voltage or current sources are proposed, and, since the obtained representations share the same theoretical approach as that employed for SMs, they can be employed together to analyse any power system, regardless its complexity. The reduced order model of the systems can either be designed by representing each source of the original system or using the concept of synchronous areas, providing flexibility in the analysis. Using a step-by-step approach, the frequency behaviours observed with academic systems are deeply analysed with the proposed methodology, and validated with those obtained with detailed EMT models. The proposed approach developed in this work can be used to determine the impact of the penetration rate of PE-based sources on the overall frequency behaviour, and on the intra- and inter-area oscillations of different power systems

Dans l'industrie, et notamment au niveau des fabricants de composants hydrauliques et pneumatiques, il est de plus en plus important de maîtriser la fiabilité des produits fabriqués. La détermination de cette fiabilité exige le développement de méthodes de réduction de la durée et du coût des essais. Les techniques bayésiennes sont une première étape vers la réduction des durées d'essais nécessaires à l'estimation de la fiabilité. Une seconde consiste à modéliser l'influence de l'environnement sur le comportement de composants afin de pouvoir estimer les performances de ces composants dans des conditions données sans aucun essai. La méthode que nous avons développée permet de calculer à la fois les facteurs environnementaux permettant de passer d'un essai à un autre et la loi de fiabilité de base d'un composant. Cet outil fournit un modèle complet de la fiabilité d'un composant ; on détermine alors la fiabilité à partir des valeurs des contraintes extérieures appliquées (pression, température, fréquence).

Master thesis describes a single-purpose machine design, that is used for testing a basic features of proportional solenoid. The design is created in cooperation with company NUVIA a.s. and contains mechanical solution that is processed in respect with ergonomic requirements, safety and overall costs. Machine concept is affected by customer requirements, whose name, as well as real parameters of the product will not be published, because of non-disclosure agreement.

A temperature sensor is a device to measure the temperature through an electrical signal. It requires a thermocouple or resistance temperature detectors (RTD) and will interface with a programmable logic controller (PLC). A temperature transmitter is a device that connects to a temperature sensor to transmit the signal elsewhere for monitoring and control purposes that its role is to convert the temperature sensor's signal to a 0-10V DC voltage in the PLC. The PLC voltage signal setting is compared to the temperature deviation after the Proportional Integral Derivative (PID) operation. Then, the system will issue a temperature control signal to achieve the cooling system voltage control so that it implements continuous monitoring and temperature control. The temperature control system in the industrial field has a certain value that it is an important factor for controlling temperature. A certain value of the temperature of the desired environment has to be set with a minimum and maximum set-point temperature which is suitable for that environment and the set-point value can be variable and can be changed with Human Machine Interface (HMI) easily.

Thesis (M. S.)--Mechanical Engineering, Georgia Institute of Technology, 2010.
Committee Chair: Book, Wayne; Committee Member: Glezer, Ari; Committee Member: Sadegh, Nader. Part of the SMARTech Electronic Thesis and Dissertation Collection.

On s'intéresse à la construction et l'estimation - à partir d'observations incomplètes - de modèles de variables aléatoires à valeurs réelles sur un graphe. Ces modèles doivent être adaptés à un problème de régression non standard où l'identité des variables observées (et donc celle des variables à prédire) varie d'une instance à l'autre. La nature du problème et des données disponibles nous conduit à modéliser le réseau sous la forme d'un champ markovien aléatoire, choix justifié par le principe de maximisation d'entropie de Jaynes. L'outil de prédiction choisi dans ces travaux est l'algorithme Belief Propagation - dans sa version classique ou gaussienne - dont la simplicité et l'efficacité permettent son utilisation sur des réseaux de grande taille. Après avoir fourni un nouveau résultat sur la stabilité locale des points fixes de l'algorithme, on étudie une approche fondée sur un modèle d'Ising latent où les dépendances entre variables réelles sont encodées à travers un réseau de variables binaires. Pour cela, on propose une définition de ces variables basée sur les fonctions de répartition des variables réelles associées. Pour l'étape de prédiction, il est nécessaire de modifier l'algorithme Belief Propagation pour imposer des contraintes de type bayésiennes sur les distributions marginales des variables binaires. L'estimation des paramètres du modèle peut aisément se faire à partir d'observations de paires. Cette approche est en fait une manière de résoudre le problème de régression en travaillant sur les quantiles. D'autre part, on propose un algorithme glouton d'estimation de la structure et des paramètres d'un champ markovien gaussien, basé sur l'algorithme Iterative Proportional Scaling. Cet algorithme produit à chaque itération un nouveau modèle dont la vraisemblance, ou une approximation de celle-ci dans le cas d'observations incomplètes, est supérieure à celle du modèle précédent. Cet algorithme fonctionnant par perturbation locale, il est possible d'imposer des contraintes spectrales assurant une meilleure compatibilité des modèles obtenus avec la version gaussienne de Belief Propagation. Les performances des différentes approches sont illustrées par des expérimentations numériques sur des données synthétiques.

Orientador: Prof. Dr. José Luis Azcue Puma
Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia Elétrica, 2016.
Este trabalho apresenta o controle direto de potencia do gerador de relutancia variavel de 12/8 utilizando o controlador fuzzy PI auto-ajustavel e o controle vetorial do inversor conectado à rede eletrica. Inicialmente são estudados os principios de operação da maquina e a topologia do conversor eletronico de potencia que sera utilizado para aciona-lo. Usando o software de simulaçãoMatlab/Simulink é implementado o modelo da maquina de relutancia variavel utilizando as curvas caracteristicas, de corrente e torque, que foram extraídas utilizando dados experimentais. Utilizando o modelo da maquina é projetado o controlador de velocidade para sua operação no modo motor. Tambem sera projetado o controlador de potência PI utilizando o modelo estimado do sistema de conversão de energia eólica. A partir dos ganhos do controlador PI é projetado o controlador fuzzy PI autoajustavel com o objetivo de melhorar a resposta em regime permanente da potencia controlada. A potencia gerada pelo gerador de relutancia variavel é injetada à rede eletrica, atraves do controle independente das potencias ativa e reativa, para este proposito é utilizado um conversor trifasico de dois níveis. Na parte experimental deste trabalho foi implementado o controle de corrente por histerese para maquina operando no modo motor, posteriormente foi implementado o controle de corrente e controle da tens~ao no barramento CC para a máquina operando no modo gerador. Os resultados de simulação e experimentais se mostraram coerentes validando os controladores propostos.
This work presents the direct power control of the switched reluctance generator of 12/8 using the self-tuning fuzzy PI controller and the vector control for inverter connected to the grid. Initially, the machine's operating principles and the topology of the power electronics converter used to drive this machine are studied. The model of switched reluctance machine is implemented in Matlab/Simulink simulation software using the characteristic curves of current and torque extracted using experimental data. Using the machine's model, It is designed the speed controller for its operation in motor mode. Also, the PI controller is designed for the power control loop based on the estimation of the wind energy conversion system. Using the gains of the PI controller, It is designed the self-tuning fuzzy PI controller, in order to improve the steady state response of the power control loop. The power generated by the switched reluctance generator is injected into the power grid through the independent control of active and reactive power, for this purpose is used a three-phase two level converter. In the experimental part of this work was implemented the hysteresis current control for machine operating in motor mode, also, It was implemented the current and DC bus voltage control for the machine operating in generator mode. The simulation and experimental results were proved consistent and these results validate the proposed controllers.

碩士
國立雲林科技大學
機械工程技術研究所
88
In this paper, a pneumatic proportional pressure control valve (PPPCV) is developed and constructed. This PPPCV is intended to be installed in a low-pressure diecasting machine and serves primarily as the pressure control element. An efficient dynamic model for this PPCV is also proposed. To calculate the nonlinear mass flow-rate of air through the valve orifice, we propose that the two important parameters of the valve orifice, the conductive value C and the critical pressure ratio b, are a linear function with exponential decay and a constant respectively. In addition, a Computational Fluid Dynamics (CFD) analytical software is used to find the functions of the parameters. Experimental results prove that the presented analytical model can well simulate the dynamic behavior of the developed PPPCV.

The focus of this thesis was the development of failure prognosis methods (prognostics) in rotating machinery with use of multiple sensors digital signal processing and machine learning techniques. The motivation stems from the void in literature concerning prognostics in meshing gearboxes. Moreover, there are several but inconclusive works regarding bearing prognosis. Few research groups have studied multi-hour fatigue gear experiments and this was one of the contributions of this thesis. Moreover, the study expanded beyond the sheer application of vibration monitoring with the addition of an Oil Debris Monitoring probe (ODM) as well as Acoustic emission (AE). The method of AE monitoring is, once again, proposed as a robust technique for failure prognosis being better correlated with gear pitting level compared to the classic vibration monitoring technique. Moreover, judging from ODM recordings the gear pitting comprises of two phases i) a linear phase, with an almost constant pitting rate and ii) a very short non linear phase where the pitting rate increases exponentially, an explicit indication of a critical failure. Multi-hour gear experiments that are close to real scale applications are very demanding in time as well as in invested capital. To bypass this shortfall a gear failure like simulation is proposed. The simulation framework is based on real life experiments and is applied to assess a number of data-driven Remaining Useful Life (RUL) estimation techniques namely i) Proportional Hazards Μodel (PHM), ii) ε- Support Vector Regression ε-SVR and iii) Exponential extrapolation based on bootstrap sampling. In the current thesis a feature extraction scheme for prognosis is proposed and assessed based on time domain, frequency domain statistical features and Wavelet Packet (WP) energy derived from AE and vibration recordings. ICA is proposed as a preferable fusion technique for gear failure prognostics. Application of ICA for feature fusion provided a clear improvement regarding the earlier presented bootstrap extrapolation technique. Bearings are also taken into account since they are closely connected to gearboxes. In the current thesis a wavelet denoising method is proposed for bearing vibration recordings aiming to the improvement of the diagnostic and prognostic potential of vibration. Finally the importance of data fusion is highlighted in the case of bearings. It is observed that a feature extraction scheme can generalize the application of prognostics, even in cases where RMS may yield no important degradation trend.
Η παρούσα εργασία εστιάζεται στην ανάπτυξη μεθοδολογιών πρόβλεψης τελικής αστοχίας σε περιστρεφόμενα συστήματα με χρήση πολλαπλών αισθητήρων και μεθόδων μηχανικής μάθησης και επεξεργασίας σήματος. Το κίνητρο προήλθε από το κενό που υπάρχει στη βιβλιογραφία όσον αφορά την προγνωστική σε κιβώτια ταχυτήτων. Η προγνωστική σε έδρανα έχει μεν μελετηθεί αλλά σε μικρό βαθμό και η παρούσα εργασία έρχεται να συμβάλλει και σε αυτό τον τομέα. Στα πλαίσια αυτής της εργασίας εκπονήθηκε ένας αριθμός πειραμάτων κόπωσης κιβωτίων ταχυτήτων. Η μελέτη επεκτάθηκε πέραν της παρακολούθησης κατάστασης με τη μέθοδο των κραδασμών και συγκεκριμένα μελετήθηκαν καταγραφές σωματιδίων σιδήρου στο λιπαντικό (ODM) καθώς και Ακουστική Εκπομπής (AE). Η μέθοδος ΑΕ ευρέθη πιο στενά συσχετισμένη με τη σταδιακή υποβάθμιση της ακεραιότητας του κιβωτίου ταχυτήτων σε σχέση με τις καταγραφές κραδασμών. Επίσης με βάση τις καταγραφές του αισθητήρα σωματιδίων σιδήρου διακρίθηκαν δύο στάδια  υποβάθμισης i) μια γραμμική περιοχή με σχεδόν σταθερό ρυθμό απελευθέρωσης υλικού από την επιφάνεια των δοντιών και ii) μια σύντομη αλλά έντονα μη γραμμική αύξηση στο ρυθμό αυτό πολύ κοντά στο τέλος της λειτουργίας του κιβωτίου. Tα πολύωρα πειράματα κόπωσης σε γρανάζια είναι πολύ απαιτητικά. Για να παρακαμφθεί αυτή η δυσκολία αναπτύχθηκε ένα φαινομενολογικό μοντέλο για αναπαραγωγή χρονοσειρών που ομοιάζουν σε καταγραφές γραναζιών σε κόπωση. Το μοντέλο αυτό στηρίχθηκε σε πραγματικά πειράματα κόπωσης. Έτσι έγινε δυνατό να εξεταστούν και να συγκριθούν ένας αριθμός μεθοδολογιών εκτίμησης εναπομένουσας ζωής και συγκεκριμένα i) Proportional Hazards Model (PHM), ii) ε- Support Vector Regression ε-SVR και iii) Exponential extrapolation βασισμένο σε μια διαδικασία bootstrap sampling. Στην παρούσα μελέτη προτείνεται ένα σύνολο παραμέτρων προερχόμενο από το πεδίο της συχνότητας, του χρόνου και κυματοπακέτων. Αυτό, συνδυαζόμενο με μια διαδικασία σύμπτυξης δεδομένων (ανάλυση σε πρωταρχικές και ανεξάρτητες συνιστώσες) αξιοποιείται για πρόγνωση σε γρανάζια σε κόπωση. Η τεχνική ανεξάρτητων συνιστωσών προτείνεται σαν προτιμότερη από τη σκοπιά της προγνωστικής καθώς βελτιώνει την εκτίμηση της εναπομένουσας ζωής. Η εργασία επεκτάθηκε και σε έδρανα κύλισης. Προτάθηκε μια διαδικασία wavelet denoising η οποία ενισχύει τόσο τη διαγνωστική όσο και την προγνωστική δυνατότητα του αισθητήρα κραδασμών. Τέλος, η σημασία της εξαγωγής παραμέτρων υπογραμμίζεται και στην περίπτωση της προγνωστικής σε έδρανα. Συνδυάζοντας πολλαπλές παραμέτρους και αισθητήρες κραδασμών μαζί με ένα μοντέλο ε-SVR παρέχεται ένα ολοκληρωμένο μοντέλο πιθανοτικής εκτίμησης εναπομένουσας ζωής σε έδρανα κύλισης ακόμα και σε περιπτώσεις που η τιμή RMS των κραδασμών δεν παρέχει πληροφορία.

Health is an invaluable asset that prevails in any society throughout human history. As a priceless good, humans are willing to make all sacrifices to ensure this precious good. The techniques and technologies used over time, their massification and constant evolution, end up raising a highly competitive and increasingly sophisticated industry in a constant search directed to the population's most pressing needs. The digitalization of clinical processes and the immensity data acquired is currently an enormous knowledge capital. Cancer is the second leading cause of death worldwide, and its increasing number is a concern for world health organisations that are trying to reverse this trend. In this context, there is an urgent need for projects and research such as this dissertation that aims to achieve early diagnosis, predict and prevent disease repercussions, improve quality of life, and increase survivorship rates in cancer patients. This work focuses explicitly on lung cancer patients. Lung cancer is one of the most typical cancers worldwide, being the most common global cause of cancer death in men and the third most common in women. Non-small cell lung cancer (NSCLC) accounts for approximately 80% of all lung malignancies. In addition, the incidence of lung cancer has been gradually increasing over the last 50 years, becoming a worldwide public health issue. The dataset analysed is specific for non-small cell lung cancer patients, and all its attributes were reviewed from a clinical perspective. After understanding the dataset's content and the pre-processing data phase followed a descriptive analysis of each attribute, and the use of the Kaplan-Meier method. Finally, this work proposes the use of Cox's Multivariate Proportional Hazard Model. Additionally, this dissertation reviews the applications domain, including the Healthcare industry structure and Information Systems and the technical knowledge necessary to implement Machine Learning algorithms. This work was supported by Holos S.A. and engaged with the CLARIFY project (European Union Horizon 2020- under grant agreement nº 875160).
A Saúde constitui um bem inestimável que prevalece em qualquer sociedade ao longo da história do ser humano. Sendo um bem inestimável os humanos estão dispostos a todos os sacrifícios de modo a assegurarem este precioso bem. As técnicas e tecnologias usadas ao longo do tempo, a sua massificação e constante evolução acabam por elevar esta indústria extremamente competitiva e cada vez mais sofisticada numa procura constante e direcionada para a necessidades mais urgentes da população humana. A digitalização dos processos clínicos e a imensidão de dados adquiridos constitui atualmente um enorme capital de conhecimento. O cancro é a segunda maior causa de morte no mundo, com tendência crescente, e consequentemente uma preocupação para a organização mundial de saúde que está a tentar reverter esta tendência. Neste contexto, há uma necessidade urgente de projetos e investigações, como esta dissertação, que visam antecipar o diagnóstico, prevenir as repercussões da doença, melhorar estilos de vida e aumentar as taxas de sobrevivência em pacientes com cancro. Este trabalho concentra-se explicitamente em pacientes com cancro do pulmão. O cancro de pulmão é um dos cancros mais comuns em todo o mundo, sendo a causa global mais comum de morte por cancro em homens e a terceira mais comum em mulheres. O cancro do pulmão de células não pequenas é responsável por aproximadamente 80% de todas as doenças malignas do pulmão. Além disso, a incidência de cancro do pulmão tem aumentando gradualmente nos últimos 50 anos, tornando-se um problema de saúde pública mundial. O conjunto de dados analisado é específico para pacientes com cancro de pulmão de células não pequenas e todos os seus atributos foram revistos de uma perspectiva clínica. Após a compreensão do conteúdo do conjunto de dados, seguiu-se a fase de pré-processamento dos dados, uma análise descritiva de cada atributo e a utilização do método de Kaplan-Meier. Finalmente, este trabalho propõe o uso do modelo de risco proporcional multivariado de Cox. Além disso, esta dissertação revê o domínio de aplicações, incluindo a estrutura da indústria de Saúde e Sistemas de Informação, tal como o conhecimento técnico necessário para implementar algoritmos de aprendizagem automática. Esta dissertação é apoiada pela Holos S.A. e envolvida no projeto CLARIFY (European Union Horizon 2020- ao abrigo do acordo da bolsa nº 875160).