Otimização de planos de manutenções de componentes de sistemas de distribuição de energia elétrica centrados em confiabilidade

Pós-graduação em Engenharia Elétrica - FEIS

Índices de confiabilidade devido a vegetação e planejamento de podas de árvores em redes de distribuição

Pós-graduação em Engenharia Elétrica - FEIS

Non-Dominated Sorting Genetic Algorithm Based on Reinforcement Learning to Optimization of Broad-Band Reflector Antennas Satellite

This paper aims to provide an improved NSGA-II (Non-Dominated Sorting Genetic Algorithm-version II) which incorporates a parameter-free self-tuning approach by reinforcement learning technique, called Non-Dominated Sorting Genetic Algorithm Based on Reinforcement Learning (NSGA-RL). The proposed method is particularly compared with the classical NSGA-II when applied to a satellite coverage problem. Furthermore, not only the optimization results are compared with results obtained by other multiobjective optimization methods, but also guarantee the advantage of no time-spending and complex parameter tuning.

Multi-objective evolutionary algorithm for single and multiple fault service restoration in large-scale distribution systems

Network reconfiguration for service restoration (SR) in distribution systems is a complex optimization problem. For large-scale distribution systems, it is computationally hard to find adequate SR plans in real time since the problem is combinatorial and non-linear, involving several constraints and objectives. Two Multi-Objective Evolutionary Algorithms that use Node-Depth Encoding (NDE) have proved able to efficiently generate adequate SR plans for large distribution systems: (i) one of them is the hybridization of the Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) with NDE, named NSGA-N; (ii) the other is a Multi-Objective Evolutionary Algorithm based on subpopulation tables that uses NDE, named MEAN. Further challenges are faced now, i.e. the design of SR plans for larger systems as good as those for relatively smaller ones and for multiple faults as good as those for one fault (single fault). In order to tackle both challenges, this paper proposes a method that results from the combination of NSGA-N, MEAN and a new heuristic. Such a heuristic focuses on the application of NDE operators to alarming network zones according to technical constraints. The method generates similar quality SR plans in distribution systems of significantly different sizes (from 3860 to 30,880 buses). Moreover, the number of switching operations required to implement the SR plans generated by the proposed method increases in a moderate way with the number of faults.

Ottimizzazione delle reti di distribuzione idrica tramite algoritmi genetici multi-obiettivo

Water distribution networks optimization is a challenging problem due to the dimension and the complexity of these systems. Since the last half of the twentieth century this field has been investigated by many authors. Recently, to overcome discrete nature of variables and non linearity of equations, the research has been focused on the development of heuristic algorithms. This algorithms do not require continuity and linearity of the problem functions because they are linked to an external hydraulic simulator that solve equations of mass continuity and of energy conservation of the network. In this work, a NSGA-II (Non-dominating Sorting Genetic Algorithm) has been used. This is a heuristic multi-objective genetic algorithm based on the analogy of evolution in nature. Starting from an initial random set of solutions, called population, it evolves them towards a front of solutions that minimize, separately and contemporaneously, all the objectives. This can be very useful in practical problems where multiple and discordant goals are common. Usually, one of the main drawback of these algorithms is related to time consuming: being a stochastic research, a lot of solutions must be analized before good ones are found. Results of this thesis about the classical optimal design problem shows that is possible to improve results modifying the mathematical definition of objective functions and the survival criterion, inserting good solutions created by a Cellular Automata and using rules created by classifier algorithm (C4.5). This part has been tested using the version of NSGA-II supplied by Centre for Water Systems (University of Exeter, UK) in MATLAB® environment. Even if orientating the research can constrain the algorithm with the risk of not finding the optimal set of solutions, it can greatly improve the results. Subsequently, thanks to CINECA help, a version of NSGA-II has been implemented in C language and parallelized: results about the global parallelization show the speed up, while results about the island parallelization show that communication among islands can improve the optimization. Finally, some tests about the optimization of pump scheduling have been carried out. In this case, good results are found for a small network, while the solutions of a big problem are affected by the lack of constraints on the number of pump switches. Possible future research is about the insertion of further constraints and the evolution guide. In the end, the optimization of water distribution systems is still far from a definitive solution, but the improvement in this field can be very useful in reducing the solutions cost of practical problems, where the high number of variables makes their management very difficult from human point of view.

Sediment yield reduction in an agricultural watershed

With proper application of Best Management Practices (BMPs), the impact from the sediment to the water bodies could be minimized. However, finding the optimal allocation of BMP can be difficult, since there are numerous possible options. Also, economics plays an important role in BMP affordability and, therefore, the number of BMPs able to be placed in a given budget year. In this study, two methodologies are presented to determine the optimal cost-effective BMP allocation, by coupling a watershed-level model, Soil and Water Assessment Tool (SWAT), with two different methods, targeting and a multi-objective genetic algorithm (Non-dominated Sorting Genetic Algorithm II, NSGA-II). For demonstration, these two methodologies were applied to an agriculture-dominant watershed located in Lower Michigan to find the optimal allocation of filter strips and grassed waterways. For targeting, three different criteria were investigated for sediment yield minimization, during the process of which it was found that the grassed waterways near the watershed outlet reduced the watershed outlet sediment yield the most under this study condition, and cost minimization was also included as a second objective during the cost-effective BMP allocation selection. NSGA-II was used to find the optimal BMP allocation for both sediment yield reduction and cost minimization. By comparing the results and computational time of both methodologies, targeting was determined to be a better method for finding optimal cost-effective BMP allocation under this study condition, since it provided more than 13 times the amount of solutions with better fitness for the objective functions while using less than one eighth of the SWAT computational time than the NSGA-II with 150 generations did.

Optimización de aperturas mediante técnicas de algoritmos evolutivos

Los sistemas de imagen por ultrasonidos son hoy una herramienta indispensable en aplicaciones de diagnóstico en medicina y son cada vez más utilizados en aplicaciones industriales en el área de ensayos no destructivos. El array es el elemento primario de estos sistemas y su diseño determina las características de los haces que se pueden construir (forma y tamaño del lóbulo principal, de los lóbulos secundarios y de rejilla, etc.), condicionando la calidad de las imágenes que pueden conseguirse. En arrays regulares la distancia máxima entre elementos se establece en media longitud de onda para evitar la formación de artefactos. Al mismo tiempo, la resolución en la imagen de los objetos presentes en la escena aumenta con el tamaño total de la apertura, por lo que una pequeña mejora en la calidad de la imagen se traduce en un aumento significativo del número de elementos del transductor. Esto tiene, entre otras, las siguientes consecuencias: Problemas de fabricación de los arrays por la gran densidad de conexiones (téngase en cuenta que en aplicaciones típicas de imagen médica, el valor de la longitud de onda es de décimas de milímetro) Baja relación señal/ruido y, en consecuencia, bajo rango dinámico de las señales por el reducido tamaño de los elementos. Complejidad de los equipos que deben manejar un elevado número de canales independientes. Por ejemplo, se necesitarían 10.000 elementos separados λ 2 para una apertura cuadrada de 50 λ. Una forma sencilla para resolver estos problemas existen alternativas que reducen el número de elementos activos de un array pleno, sacrificando hasta cierto punto la calidad de imagen, la energía emitida, el rango dinámico, el contraste, etc. Nosotros planteamos una estrategia diferente, y es desarrollar una metodología de optimización capaz de hallar de forma sistemática configuraciones de arrays de ultrasonido adaptados a aplicaciones específicas. Para realizar dicha labor proponemos el uso de los algoritmos evolutivos para buscar y seleccionar en el espacio de configuraciones de arrays aquellas que mejor se adaptan a los requisitos fijados por cada aplicación. En la memoria se trata el problema de la codificación de las configuraciones de arrays para que puedan ser utilizados como individuos de la población sobre la que van a actuar los algoritmos evolutivos. También se aborda la definición de funciones de idoneidad que permitan realizar comparaciones entre dichas configuraciones de acuerdo con los requisitos y restricciones de cada problema de diseño. Finalmente, se propone emplear el algoritmo multiobjetivo NSGA II como herramienta primaria de optimización y, a continuación, utilizar algoritmos mono-objetivo tipo Simulated Annealing para seleccionar y retinar las soluciones proporcionadas por el NSGA II. Muchas de las funciones de idoneidad que definen las características deseadas del array a diseñar se calculan partir de uno o más patrones de radiación generados por cada solución candidata. La obtención de estos patrones con los métodos habituales de simulación de campo acústico en banda ancha requiere tiempos de cálculo muy grandes que pueden hacer inviable el proceso de optimización con algoritmos evolutivos en la práctica. Como solución, se propone un método de cálculo en banda estrecha que reduce en, al menos, un orden de magnitud el tiempo de cálculo necesario Finalmente se presentan una serie de ejemplos, con arrays lineales y bidimensionales, para validar la metodología de diseño propuesta comparando experimentalmente las características reales de los diseños construidos con las predicciones del método de optimización. ABSTRACT Currently, the ultrasound imaging system is one of the powerful tools in medical diagnostic and non-destructive testing for industrial applications. Ultrasonic arrays design determines the beam characteristics (main and secondary lobes, beam pattern, etc...) which assist to enhance the image resolution. The maximum distance between the elements of the array should be the half of the wavelength to avoid the formation of grating lobes. At the same time, the image resolution of the target in the region of interest increases with the aperture size. Consequently, the larger number of elements in arrays assures the better image quality but this improvement contains the following drawbacks: Difficulties in the arrays manufacturing due to the large connection density. Low noise to signal ratio. Complexity of the ultrasonic system to handle large number of channels. The easiest way to resolve these issues is to reduce the number of active elements in full arrays, but on the other hand the image quality, dynamic range, contrast, etc, are compromised by this solutions In this thesis, an optimization methodology able to find ultrasound array configurations adapted for specific applications is presented. The evolutionary algorithms are used to obtain the ideal arrays among the existing configurations. This work addressed problems such as: the codification of ultrasound arrays to be interpreted as individuals in the evolutionary algorithm population and the fitness function and constraints, which will assess the behaviour of individuals. Therefore, it is proposed to use the multi-objective algorithm NSGA-II as a primary optimization tool, and then use the mono-objective Simulated Annealing algorithm to select and refine the solutions provided by the NSGA I I . The acoustic field is calculated many times for each individual and in every generation for every fitness functions. An acoustic narrow band field simulator, where the number of operations is reduced, this ensures a quick calculation of the acoustic field to reduce the expensive computing time required by these functions we have employed. Finally a set of examples are presented in order to validate our proposed design methodology, using linear and bidimensional arrays where the actual characteristics of the design are compared with the predictions of the optimization methodology.

Optimización multiobjetivo inteligente del dimensionamiento energético en sistemas ferroviarios señalizados y en plataformas software de tiempo real

La capacidad de transporte es uno de los baremos fundamentales para evaluar la progresión que puede llegar a tener un área económica y social. Es un sector de elevada importancia para la sociedad actual. Englobado en los distintos tipos de transporte, uno de los medios de transporte que se encuentra más en alza en la actualidad, es el ferroviario. Tanto para movilidad de pasajeros como para mercancías, el tren se ha convertido en un medio de transporte muy útil. Se encuentra dentro de las ciudades, entre ciudades con un radio pequeño entre ellas e incluso cada vez más, gracias a la alta velocidad, entre ciudades con gran distancia entre ellas. Esta Tesis pretende ayudar en el diseño de una de las etapas más importantes de los Proyectos de instalación de un sistema ferroviario: el sistema eléctrico de tracción. La fase de diseño de un sistema eléctrico de tracción ferroviaria se enfrenta a muchas dudas que deben ser resueltas con precisión. Del éxito de esta fase dependerá la capacidad de afrontar las demandas de energía de la explotación ferroviaria. También se debe atender a los costes de instalación y de operación, tanto costes directos como indirectos. Con la Metodología que se presenta en esta Tesis se ofrecerá al diseñador la opción de manejar un sistema experto que como soluciones le plantee un conjunto de escenarios de sistemas eléctricos correctos, comprobados por resolución de modelos de ecuaciones. Correctos desde el punto de vista de validez de distintos parámetros eléctrico, como de costes presupuestarios e impacto de costes indirectos. Por tanto, el diseñador al haber hecho uso de esta Metodología, tendría en un espacio de tiempo relativamente corto, un conjunto de soluciones factibles con las que poder elegir cuál convendría más según sus intereses finales. Esta Tesis se ha desarrollado en una vía de investigación integrada dentro del Centro de Investigaciones Ferroviarias CITEF-UPM. Entre otros proyectos y vías de investigación, en CITEF se ha venido trabajando en estudios de validación y dimensionamiento de sistemas eléctricos ferroviarios con diversos y variados clientes y sistemas ferroviarios. A lo largo de los proyectos realizados, el interés siempre ha girado mayoritariamente sobre los siguientes parámetros del sistema eléctrico: - Calcular número y posición de subestaciones de tracción. Potencia de cada subestación. - Tipo de catenaria a lo largo del recorrido. Conductores que componen la catenaria. Características. - Calcular número y posición de autotransformadores para sistemas funcionando en alterna bitensión o 2x25kV. - Posición Zonas Neutras. - Validación según normativa de: o Caídas de tensión en la línea o Tensiones máximas en el retorno de la línea o Sobrecalentamiento de conductores o Sobrecalentamiento de los transformadores de las subestaciones de tracción La idea es que las soluciones aportadas por la Metodología sugieran escenarios donde de estos parámetros estén dentro de los límites que marca la normativa. Tener la posibilidad de tener un repositorio de posibles escenarios donde los parámetros y elementos eléctricos estén calculados como correctos, aporta un avance en tiempos y en pruebas, que mejoraría ostensiblemente el proceso habitual de diseño para los sistemas eléctricos ferroviarios. Los costes directos referidos a elementos como subestaciones de tracción, autotransformadores, zonas neutras, ocupan un gran volumen dentro del presupuesto de un sistema ferroviario. En esta Tesis se ha querido profundizar también en el efecto de los costes indirectos provocados en la instalación y operación de sistemas eléctricos. Aquellos derivados del impacto medioambiental, los costes que se generan al mantener los equipos eléctricos y la instalación de la catenaria, los costes que implican la conexión entre las subestaciones de tracción con la red general o de distribución y por último, los costes de instalación propios de cada elemento compondrían los costes indirectos que, según experiencia, se han pensado relevantes para ejercer un cierto control sobre ellos. La Metodología cubrirá la posibilidad de que los diseños eléctricos propuestos tengan en cuenta variaciones de coste inasumibles o directamente, proponer en igualdad de condiciones de parámetros eléctricos, los más baratos en función de los costes comentados. Analizando los costes directos e indirectos, se ha pensado dividir su impacto entre los que se computan en la instalación y los que suceden posteriormente, durante la operación de la línea ferroviaria. Estos costes normalmente suelen ser contrapuestos, cuánto mejor es uno peor suele ser el otro y viceversa, por lo que hace falta un sistema que trate ambos objetivos por separado. Para conseguir los objetivos comentados, se ha construido la Metodología sobre tres pilares básicos: - Simulador ferroviario Hamlet: Este simulador integra módulos para construir esquemas de vías ferroviarios completos; módulo de simulación mecánica y de la tracción de material rodante; módulo de señalización ferroviaria; módulo de sistema eléctrico. Software realizado en C++ y Matlab. - Análisis y estudio de cómo focalizar los distintos posibles escenarios eléctricos, para que puedan ser examinados rápidamente. Pico de demanda máxima de potencia por el tráfico ferroviario. - Algoritmos de optimización: A partir de un estudio de los posibles algoritmos adaptables a un sistema tan complejo como el que se plantea, se decidió que los algoritmos genéticos serían los elegidos. Se han escogido 3 algoritmos genéticos, permitiendo recabar información acerca del comportamiento y resultados de cada uno de ellos. Los elegidos por motivos de tiempos de respuesta, multiobjetividad, facilidad de adaptación y buena y amplia aplicación en proyectos de ingeniería fueron: NSGA-II, AMGA-II y ɛ-MOEA. - Diseño de funciones y modelo preparado para trabajar con los costes directos e indirectos y las restricciones básicas que los escenarios eléctricos no deberían violar. Estas restricciones vigilan el comportamiento eléctrico y la estabilidad presupuestaria. Las pruebas realizadas utilizando el sistema han tratado o bien de copiar situaciones que se puedan dar en la realidad o directamente sistemas y problemas reales. Esto ha proporcionado además de la posibilidad de validar la Metodología, también se ha posibilitado la comparación entre los algoritmos genéticos, comparar sistemas eléctricos escogidos con los reales y llegar a conclusiones muy satisfactorias. La Metodología sugiere una vía de trabajo muy interesante, tanto por los resultados ya obtenidos como por las oportunidades que puede llegar a crear con la evolución de la misma. Esta Tesis se ha desarrollado con esta idea, por lo que se espera pueda servir como otro factor para trabajar con la validación y diseño de sistemas eléctricos ferroviarios. ABSTRACT Transport capacity is one of the critical points to evaluate the progress than a specific social and economical area is able to reach. This is a sector of high significance for the actual society. Included inside the most common types of transport, one of the means of transport which is elevating its use nowadays is the railway. Such as for passenger transport of weight movements, the train is being consolidated like a very useful mean of transport. Railways are installed in many geography areas. Everyone know train in cities, or connecting cities inside a surrounding area or even more often, taking into account the high-speed, there are railways infrastructure between cities separated with a long distance. This Ph.D work aims to help in the process to design one of the most essential steps in Installation Projects belonging to a railway system: Power Supply System. Design step of the railway power supply, usually confronts to several doubts and uncertainties, which must be solved with high accuracy. Capacity to supply power to the railway traffic depends on the success of this step. On the other hand is very important to manage the direct and indirect costs derived from Installation and Operation. With the Methodology is presented in this Thesis, it will be offered to the designer the possibility to handle an expert system that finally will fill a set of possible solutions. These solutions must be ready to work properly in the railway system, and they were tested using complex equation models. This Thesis has been developed through a research way, integrated inside Citef (Railway Research Centre of Technical University of Madrid). Among other projects and research ways, in Citef has been working in several validation studies and dimensioning of railway power supplies. It is been working by a large range of clients and railways systems. Along the accomplished Projects, the main goal has been rounded mostly about the next list of parameters of the electrical system: - Calculating number and location of traction substations. Power of each substation. - Type of Overhead contact line or catenary through the railway line. The wires which set up the catenary. Main Characteristics. - Calculating number and position of autotransformers for systems working in alternating current bi-voltage of called 2x25 kV. - Location of Neutral Zones. - Validating upon regulation of: o Drop voltages along the line o Maximum return voltages in the line o Overheating/overcurrent of the wires of the catenary o Avoiding overheating in the transformers of the traction substations. Main objective is that the solutions given by the Methodology, could be suggest scenarios where all of these parameters from above, would be between the limits established in the regulation. Having the choice to achieve a repository of possible good scenarios, where the parameters and electrical elements will be assigned like ready to work, that gives a great advance in terms of times and avoiding several tests. All of this would improve evidently the regular railway electrical systems process design. Direct costs referred to elements like traction substations, autotransformers, neutral zones, usually take up a great volume inside the general budget in railway systems. In this Thesis has been thought to bear in mind another kind of costs related to railway systems, also called indirect costs. These could be enveloped by those enmarked during installation and operation of electrical systems. Those derived from environmental impact; costs generated during the maintenance of the electrical elements and catenary; costs involved in the connection between traction substations and general electric grid; finally costs linked with the own installation of the whole electrical elements needed for the correct performance of the railway system. These are integrated inside the set has been collected taking into account own experience and research works. They are relevant to be controlled for our Methodology, just in case for the designers of this type of systems. The Methodology will cover the possibility that the final proposed power supply systems will be hold non-acceptable variations of costs, comparing with initial expected budgets, or directly assuming a threshold of budget for electrical elements in actual scenario, and achieving the cheapest in terms of commented costs from above. Analyzing direct and indirect costs, has been thought to divide their impact between two main categories. First one will be inside the Installation and the other category will comply with the costs often happens during Railway Operation time. These costs normally are opposed, that means when one is better the other turn into worse, in costs meaning. For this reason is necessary treating both objectives separately, in order to evaluate correctly the impact of each one into the final system. The objectives detailed before build the Methodology under three basic pillars: - Railway simulator Hamlet: This software has modules to configure many railway type of lines; mechanical and traction module to simulate the movement of rolling stock; signaling module; power supply module. This software has been developed using C++ and Matlab R13a - Previously has been mandatory to study how would be possible to work properly with a great number of feasible electrical systems. The target comprised the quick examination of these set of scenarios in terms of time. This point is talking about Maximum power demand peaks by railway operation plans. - Optimization algorithms. A railway infrastructure is a very complex system. At the beginning it was necessary to search about techniques and optimization algorithms, which could be adaptable to this complex system. Finally three genetic multiobjective algorithms were the chosen. Final decision was taken attending to reasons such as time complexity, able to multiobjective, easy to integrate in our problem and with a large application in engineering tasks. They are: NSGA-II, AMGA-II and ɛ-MOEA. - Designing objectives functions and equation model ready to work with the direct and indirect costs. The basic restrictions are not able to avoid, like budgetary or electrical, connected hardly with the recommended performance of elements, catenary and safety in a electrical railway systems. The battery of tests launched to the Methodology has been designed to be as real as possible. In fact, due to our work in Citef and with real Projects, has been integrated and configured three real railway lines, in order to evaluate correctly the final results collected by the Methodology. Another topic of our tests has been the comparison between the performances of the three algorithms chosen. Final step has been the comparison again with different possible good solutions, it means power supply system designs, provided by the Methodology, testing the validity of them. Once this work has been finished, the conclusions have been very satisfactory. Therefore this Thesis suggest a very interesting way of research and work, in terms of the results obtained and for the future opportunities can be created with the evolution of this. This Thesis has been developed with this idea in mind, so is expected this work could adhere another factor to work in the difficult task of validation and design of railway power supply systems.

Estimación de la resistencia de estructuras de hormigón armado reforzadas a cortante mediante FRP

El empleo de refuerzos de FRP en vigas de hormigón armado es cada vez más frecuente por sus numerosas ventajas frente a otros métodos más tradicionales. Durante los últimos años, la técnica FRP-NSM, consistente en introducir barras de FRP sobre el recubrimiento de una viga de hormigón, se ha posicionado como uno de los mejores métodos de refuerzo y rehabilitación de estructuras de hormigón armado, tanto por su facilidad de montaje y mantenimiento, como por su rendimiento para aumentar la capacidad resistente de dichas estructuras. Si bien el refuerzo a flexión ha sido ampliamente desarrollado y estudiado hasta la fecha, no sucede lo mismo con el refuerzo a cortante, debido principalmente a su gran complejidad. Sin embargo, se debería dedicar más estudio a este tipo de refuerzo si se pretenden conservar los criterios de diseño en estructuras de hormigón armado, los cuales están basados en evitar el fallo a cortante por sus consecuencias catastróficas Esta ausencia de información y de normativa es la que justifica esta tesis doctoral. En este pro-yecto se van a desarrollar dos metodologías alternativas, que permiten estimar la capacidad resistente de vigas de hormigón armado, reforzadas a cortante mediante la técnica FRP-NSM. El primer método aplicado consiste en la implementación de una red neuronal artificial capaz de predecir adecuadamente la resistencia a cortante de vigas reforzadas con este método a partir de experimentos anteriores. Asimismo, a partir de la red se han llevado a cabo algunos estudios a fin de comprender mejor la influencia real de algunos parámetros de la viga y del refuerzo sobre la resistencia a cortante con el propósito de lograr diseños más seguros de este tipo de refuerzo. Una configuración óptima de la red requiere discriminar adecuadamente de entre los numerosos parámetros (geométricos y de material) que pueden influir en el compor-tamiento resistente de la viga, para lo cual se han llevado a cabo diversos estudios y pruebas. Mediante el segundo método, se desarrolla una ecuación de proyecto que permite, de forma sencilla, estimar la capacidad de vigas reforzadas a cortante con FRP-NSM, la cual podría ser propuesta para las principales guías de diseño. Para alcanzar este objetivo, se plantea un pro-blema de optimización multiobjetivo a partir de resultados de ensayos experimentales llevados a cabo sobre vigas de hormigón armado con y sin refuerzo de FRP. El problema multiobjetivo se resuelve mediante algoritmos genéticos, en concreto el algoritmo NSGA-II, por ser más apropiado para problemas con varias funciones objetivo que los métodos de optimización clásicos. Mediante una comparativa de las predicciones realizadas con ambos métodos y de los resulta-dos de ensayos experimentales se podrán establecer las ventajas e inconvenientes derivadas de la aplicación de cada una de las dos metodologías. Asimismo, se llevará a cabo un análisis paramétrico con ambos enfoques a fin de intentar determinar la sensibilidad de aquellos pa-rámetros más sensibles a este tipo de refuerzo. Finalmente, se realizará un análisis estadístico de la fiabilidad de las ecuaciones de diseño deri-vadas de la optimización multiobjetivo. Con dicho análisis se puede estimar la capacidad resis-tente de una viga reforzada a cortante con FRP-NSM dentro de un margen de seguridad espe-cificado a priori. ABSTRACT The use of externally bonded (EB) fibre-reinforced polymer (FRP) composites has gained acceptance during the last two decades in the construction engineering community, particularly in the rehabilitation of reinforced concrete (RC) structures. Currently, to increase the shear resistance of RC beams, FRP sheets are externally bonded (EB-FRP) and applied on the external side surface of the beams to be strengthened with different configurations. Of more recent application, the near-surface mounted FRP bar (NSM-FRP) method is another technique successfully used to increase the shear resistance of RC beams. In the NSM method, FRP rods are embedded into grooves intentionally prepared in the concrete cover of the side faces of RC beams. While flexural strengthening has been widely developed and studied so far, the same doesn´t occur to shearing strength mainly due to its great complexity. Nevertheless, if design criteria are to be preserved more research should be done to this sort of strength, which are based on avoiding shear failure and its catastrophic consequences. However, in spite of this, accurately calculating the shear capacity of FRP shear strengthened RC beams remains a complex challenge that has not yet been fully resolved due to the numerous variables involved in the procedure. The objective of this Thesis is to develop methodologies to evaluate the capacity of FRP shear strengthened RC beams by dealing with the problem from a different point of view to the numerical modeling approach by using artificial intelligence techniques. With this purpose two different approaches have been developed: one concerned with the use of artificial neural networks and the other based on the implementation of an optimization approach developed jointly with the use of artificial neural networks (ANNs) and solved with genetic algorithms (GAs). With these approaches some of the difficulties concerned regarding the numerical modeling can be overcome. As an alternative tool to conventional numerical techniques, neural networks do not provide closed form solutions for modeling problems but do, however, offer a complex and accurate solution based on a representative set of historical examples of the relationship. Furthermore, they can adapt solutions over time to include new data. On the other hand, as a second proposal, an optimization approach has also been developed to implement simple yet accurate shear design equations for this kind of strengthening. This approach is developed in a multi-objective framework by considering experimental results of RC beams with and without NSM-FRP. Furthermore, the results obtained with the previous scheme based on ANNs are also used as a filter to choose the parameters to include in the design equations. Genetic algorithms are used to solve the optimization problem since they are especially suitable for solving multi-objective problems when compared to standard optimization methods. The key features of the two proposed procedures are outlined and their performance in predicting the capacity of NSM-FRP shear strengthened RC beams is evaluated by comparison with results from experimental tests and with predictions obtained using a simplified numerical model. A sensitivity study of the predictions of both models for the input parameters is also carried out.

Avaliação de incertezas na calibração automática do modelo SWMM

Este trabalho é referente ao desenvolvimento de um calibrador multiobjetivo automático do modelo SWMM (Storm Water Management Model), e avaliação de algumas fontes de incertezas presentes no processo de calibração, visando à representação satisfatória da transformação chuva-vazão. O código foi escrito em linguagem C, e aplica os conceitos do método de otimização multiobjetivo NSGAII (Non Dominated Sorting Genetic Algorithm) com elitismo controlado, além de utilizar o código fonte do modelo SWMM para a determinação das vazões simuladas. Paralelamente, também foi criada uma interface visual, para melhorar a facilidade de utilização do calibrador. Os testes do calibrador foram aplicados a três sistemas diferentes: um sistema hipotético disponibilizado no pacote de instalação do SWMM; um sistema real de pequenas dimensões, denominado La Terraza, localizado no município de Sierra Vista, Arizona (EUA); e um sistema de maiores dimensões, a bacia hidrográfica do Córrego do Gregório, localizada no município de São Carlos (SP). Os resultados indicam que o calibrador construído apresenta, em geral, eficiência satisfatória, porém é bastante dependente da qualidade dos dados observados em campo e dos parâmetros de entrada escolhidos pelo usuário. Foi demonstrada a importância da escolha dos eventos utilizados na calibração, do estabelecimento de limites adequados nos valores das variáveis de decisão, da escolha das funções objetivo e, principalmente, da qualidade e representatividade dos dados de monitoramento pluvio e fluviométrico. Conclui-se que estes testes desenvolvidos contribuem para o entendimento mais aprofundado dos processos envolvidos na modelagem e calibração, possibilitando avanços na confiabilidade dos resultados da modelagem.

Intensificação da transferência de calor e otimização de trocadores de calor compactos tipo venezianas com geradores de vórtices tipo delta-winglets.

Esta pesquisa visa a análise da contribuição de cinco variáveis de entrada e a otimização do desempenho termo-hidráulico de trocadores de calor com venezianas combinados com geradores de vórtices delta-winglets. O desempenho termohidráulico de duas geometrias distintas, aqui nomeadas por GEO1 e GEO2, foram avaliadas. Smoothing Spline ANOVA foi usado para avaliar a contribuição dos parâmetros de entrada na transferência de calor e perda de carga. Considerando aplicação automotiva, foram investigados números de Reynolds iguais a 120 e 240, baseados no diâmetro hidráulico. Os resultados indicaram que o ângulo de venezianas é o maior contribuidor para o aumento do fator de atrito para GEO1 e GEO2, para ambos os números de Reynolds. Para o número de Reynolds menor, o parâmetro mais importante em termos de transferência de calor foi o ângulo das venezianas para ambas as geometrias. Para o número de Reynolds maior, o ângulo de ataque dos geradores de vórtices posicionados na primeira fileira é o maior contribuidor para a tranfesferência de calor, no caso da geometria GEO1, enquanto que o ângulo de ataque dos geradores de vórtices na primeira fileira foi tão importante quanto os ângulos das venezianas para a geometria GEO2. Embora as geometrias analisadas possam ser consideradas como técnicas compostas de intensificação da transferência de calor, não foram observadas interações relevantes entre ângulo de venezianas e parâmetros dos geradores de vórtices. O processo de otimização usa NSGA-II (Non-Dominated Sorting Genetic Algorithm) combinado com redes neurais artificiais. Os resultados mostraram que a adição dos geradores de vórtices em GEO1 aumentaram a transferência de calor em 21% e 23% com aumentos na perda de carga iguais a 24,66% e 36,67% para o menor e maior números de Reynolds, respectivamente. Para GEO2, a transferência de calor aumentou 13% e 15% com aumento na perda de carga de 20,33% e 23,70%, para o menor e maior número de Reynolds, respectivamente. As soluções otimizadas para o fator de Colburn mostraram que a transferência de calor atrás da primeira e da segunda fileiras de geradores de vórtices tem a mesma ordem de magnitude para ambos os números de Reynolds. Os padrões de escoamento e as características de transferência de calor das soluções otimizadas apresentaram comportamentos vi particulares, diferentemente daqueles encontrados quando as duas técnicas de intensificação de transferência de calor são aplicadas separadamente.

Modelo de otimização multiobjetivo aplicado ao projeto de concepção de submarinos convencionais.

Este trabalho apresenta um modelo de otimização multiobjetivo aplicado ao projeto de concepção de submarinos convencionais (i.e. de propulsão dieselelétrica). Um modelo de síntese que permite a estimativa de pesos, volume, velocidade, carga elétrica e outras características de interesse para a o projeto de concepção é formulado. O modelo de síntese é integrado a um modelo de otimização multiobjetivo baseado em algoritmos genéticos (especificamente, o algoritmo NSGA II). A otimização multiobjetivo consiste na maximização da efetividade militar do submarino e na minimização de seu custo. A efetividade militar do submarino é representada por uma Medida Geral de Efetividade (OMOE) estabelecida por meio do Processo Analítico Hierárquico (AHP). O Custo Básico de Construção (BCC) do submarino é estimado a partir dos seus grupos de peso. Ao fim do processo de otimização, é estabelecida uma Fronteira de Pareto composta por soluções não dominadas. Uma dessas soluções é selecionada para refinamento preliminar e os resultados são discutidos. Subsidiariamente, esta dissertação apresenta discussão sucinta sobre aspectos históricos e operativos relacionados a submarinos, bem como sobre sua metodologia de projeto. Alguns conceitos de Arquitetura Naval, aplicada ao projeto dessas embarcações, são também abordados.

Multi-objective adaptive evolutionary strategy for tuning compilations

Tuning compilations is the process of adjusting the values of a compiler options to improve some features of the final application. In this paper, a strategy based on the use of a genetic algorithm and a multi-objective scheme is proposed to deal with this task. Unlike previous works, we try to take advantage of the knowledge of this domain to provide a problem-specific genetic operation that improves both the speed of convergence and the quality of the results. The evaluation of the strategy is carried out by means of a case of study aimed to improve the performance of the well-known web server Apache. Experimental results show that a 7.5% of overall improvement can be achieved. Furthermore, the adaptive approach has shown an ability to markedly speed-up the convergence of the original strategy.

Fault tolerant embedded systems design by multi-objective optimization

The design of fault tolerant systems is gaining importance in large domains of embedded applications where design constrains are as important as reliability. New software techniques, based on selective application of redundancy, have shown remarkable fault coverage with reduced costs and overheads. However, the large number of different solutions provided by these techniques, and the costly process to assess their reliability, make the design space exploration a very difficult and time-consuming task. This paper proposes the integration of a multi-objective optimization tool with a software hardening environment to perform an automatic design space exploration in the search for the best trade-offs between reliability, cost, and performance. The first tool is commanded by a genetic algorithm which can simultaneously fulfill many design goals thanks to the use of the NSGA-II multi-objective algorithm. The second is a compiler-based infrastructure that automatically produces selective protected (hardened) versions of the software and generates accurate overhead reports and fault coverage estimations. The advantages of our proposal are illustrated by means of a complex and detailed case study involving a typical embedded application, the AES (Advanced Encryption Standard).

Tuning compilations by multi-objective optimization: Application to Apache web server

Modern compilers present a great and ever increasing number of options which can modify the features and behavior of a compiled program. Many of these options are often wasted due to the required comprehensive knowledge about both the underlying architecture and the internal processes of the compiler. In this context, it is usual, not having a single design goal but a more complex set of objectives. In addition, the dependencies between different goals are difficult to be a priori inferred. This paper proposes a strategy for tuning the compilation of any given application. This is accomplished by using an automatic variation of the compilation options by means of multi-objective optimization and evolutionary computation commanded by the NSGA-II algorithm. This allows finding compilation options that simultaneously optimize different objectives. The advantages of our proposal are illustrated by means of a case study based on the well-known Apache web server. Our strategy has demonstrated an ability to find improvements up to 7.5% and up to 27% in context switches and L2 cache misses, respectively, and also discovers the most important bottlenecks involved in the application performance.