Automatic multi-resolution shape modeling of multi-organ structures.

Point Distribution Models (PDM) are among the most popular shape description techniques and their usefulness has been demonstrated in a wide variety of medical imaging applications. However, to adequately characterize the underlying modeled population it is essential to have a representative number of training samples, which is not always possible. This problem is especially relevant as the complexity of the modeled structure increases, being the modeling of ensembles of multiple 3D organs one of the most challenging cases. In this paper, we introduce a new GEneralized Multi-resolution PDM (GEM-PDM) in the context of multi-organ analysis able to efficiently characterize the different inter-object relations, as well as the particular locality of each object separately. Importantly, unlike previous approaches, the configuration of the algorithm is automated thanks to a new agglomerative landmark clustering method proposed here, which equally allows us to identify smaller anatomically significant regions within organs. The significant advantage of the GEM-PDM method over two previous approaches (PDM and hierarchical PDM) in terms of shape modeling accuracy and robustness to noise, has been successfully verified for two different databases of sets of multiple organs: six subcortical brain structures, and seven abdominal organs. Finally, we propose the integration of the new shape modeling framework into an active shape-model-based segmentation algorithm. The resulting algorithm, named GEMA, provides a better overall performance than the two classical approaches tested, ASM, and hierarchical ASM, when applied to the segmentation of 3D brain MRI.

Assessment of CTF boiling transition and critical heat flux modeling capabilities using the OECD/NRC BFBT and PSBT benchmark databases

The need to refine models for best-estimate calculations, based on good-quality experimental data, has been expressed in many recent meetings in the field of nuclear applications. The modeling needs arising in this respect should not be limited to the currently available macroscopic methods but should be extended to next-generation analysis techniques that focus on more microscopic processes. One of the most valuable databases identified for the thermalhydraulics modeling was developed by the Nuclear Power Engineering Corporation (NUPEC), Japan. From 1987 to 1995, NUPEC performed steady-state and transient critical power and departure from nucleate boiling (DNB) test series based on the equivalent full-size mock-ups. Considering the reliability not only of the measured data, but also other relevant parameters such as the system pressure, inlet sub-cooling and rod surface temperature, these test series supplied the first substantial database for the development of truly mechanistic and consistent models for boiling transition and critical heat flux. Over the last few years the Pennsylvania State University (PSU) under the sponsorship of the U.S. Nuclear Regulatory Commission (NRC) has prepared, organized, conducted and summarized the OECD/NRC Full-size Fine-mesh Bundle Tests (BFBT) Benchmark. The international benchmark activities have been conducted in cooperation with the Nuclear Energy Agency/Organization for Economic Co-operation and Development (NEA/OECD) and Japan Nuclear Energy Safety (JNES) organization, Japan. Consequently, the JNES has made available the Boiling Water Reactor (BWR) NUPEC database for the purposes of the benchmark. Based on the success of the OECD/NRC BFBT benchmark the JNES has decided to release also the data based on the NUPEC Pressurized Water Reactor (PWR) subchannel and bundle tests for another follow-up international benchmark entitled OECD/NRC PWR Subchannel and Bundle Tests (PSBT) benchmark. This paper presents an application of the joint Penn State University/Technical University of Madrid (UPM) version of the well-known subchannel code COBRA-TF, namely CTF, to the critical power and departure from nucleate boiling (DNB) exercises of the OECD/NRC BFBT and PSBT benchmarks

A region based approach to background modeling in a wavelet multi-resolution framework

In the field of detection and monitoring of dynamic objects in quasi-static scenes, background subtraction techniques where background is modeled at pixel-level, although showing very significant limitations, are extensively used. In this work we propose a novel approach to background modeling that operates at region-level in a wavelet based multi-resolution framework. Based on a segmentation of the background, characterization is made for each region independently as a mixture of K Gaussian modes, considering the model of the approximation and detail coefficients at the different wavelet decomposition levels. Background region characterization is updated along time, and the detection of elements of interest is carried out computing the distance between background region models and those of each incoming image in the sequence. The inclusion of the context in the modeling scheme through each region characterization makes the model robust, being able to support not only gradual illumination and long-term changes, but also sudden illumination changes and the presence of strong shadows in the scene

An Interval-based Multiobjective Approach to Feature Subset Selection Using Joint Modeling of Objectives and Variables

This paper studies feature subset selection in classification using a multiobjective estimation of distribution algorithm. We consider six functions, namely area under ROC curve, sensitivity, specificity, precision, F1 measure and Brier score, for evaluation of feature subsets and as the objectives of the problem. One of the characteristics of these objective functions is the existence of noise in their values that should be appropriately handled during optimization. Our proposed algorithm consists of two major techniques which are specially designed for the feature subset selection problem. The first one is a solution ranking method based on interval values to handle the noise in the objectives of this problem. The second one is a model estimation method for learning a joint probabilistic model of objectives and variables which is used to generate new solutions and advance through the search space. To simplify model estimation, l1 regularized regression is used to select a subset of problem variables before model learning. The proposed algorithm is compared with a well-known ranking method for interval-valued objectives and a standard multiobjective genetic algorithm. Particularly, the effects of the two new techniques are experimentally investigated. The experimental results show that the proposed algorithm is able to obtain comparable or better performance on the tested datasets.

Efficient numerical methods for global linear instability. Analysis and modeling of non-orthogonal swept attachment-line boundary layer flow

The aim of this thesis is to study the mechanisms of instability that occur in swept wings when the angle of attack increases. For this, a simplified model for the a simplified model for the non-orthogonal swept leading edge boundary layer has been used as well as different numerical techniques in order to solve the linear stability problem that describes the behavior of perturbations superposed upon this base flow. Two different approaches, matrix-free and matrix forming methods, have been validated using direct numerical simulations with spectral resolution. In this way, flow instability in the non-orthogonal swept attachment-line boundary layer is addressed in a linear analysis framework via the solution of the pertinent global (Bi-Global) PDE-based eigenvalue problem. Subsequently, a simple extension of the extended G¨ortler-H¨ammerlin ODEbased polynomial model proposed by Theofilis, Fedorov, Obrist & Dallmann (2003) for orthogonal flow, which includes previous models as particular cases and recovers global instability analysis results, is presented for non-orthogonal flow. Direct numerical simulations have been used to verify the stability results and unravel the limits of validity of the basic flow model analyzed. The effect of the angle of attack, AoA, on the critical conditions of the non-orthogonal problem has been documented; an increase of the angle of attack, from AoA = 0 (orthogonal flow) up to values close to _/2 which make the assumptions under which the basic flow is derived questionable, is found to systematically destabilize the flow. The critical conditions of non-orthogonal flows at 0 _ AoA _ _/2 are shown to be recoverable from those of orthogonal flow, via a simple analytical transformation involving AoA. These results can help to understand the mechanisms of destabilization that occurs in the attachment line of wings at finite angles of attack. Studies taking into account variations of the pressure field in the basic flow or the extension to compressible flows are issues that remain open. El objetivo de esta tesis es estudiar los mecanismos de la inestabilidad que se producen en ciertos dispositivos aerodinámicos cuando se aumenta el ángulo de ataque. Para ello se ha utilizado un modelo simplificado del flujo de base, así como diferentes técnicas numéricas, con el fin de resolver el problema de estabilidad lineal asociado que describe el comportamiento de las perturbaciones. Estos métodos; sin y con formación de matriz, se han validado utilizando simulaciones numéricas directas con resolución espectral. De esta manera, la inestabilidad del flujo de capa límite laminar oblicuo entorno a la línea de estancamiento se aborda en un marco de análisis lineal por medio del método Bi-Global de resolución del problema de valores propios en derivadas parciales. Posteriormente se propone una extensión simple para el flujo no-ortogonal del modelo polinomial de ecuaciones diferenciales ordinarias, G¨ortler-H¨ammerlin extendido, propuesto por Theofilis et al. (2003) para el flujo ortogonal, que incluye los modelos previos como casos particulares y recupera los resultados del analisis global de estabilidad lineal. Se han realizado simulaciones directas con el fin de verificar los resultados del análisis de estabilidad así como para investigar los límites de validez del modelo de flujo base utilizado. En este trabajo se ha documentado el efecto del ángulo de ataque AoA en las condiciones críticas del problema no ortogonal obteniendo que el incremento del ángulo de ataque, de AoA = 0 (flujo ortogonal) hasta valores próximos a _/2, en el cual las hipótesis sobre las que se basa el flujo base dejan de ser válidas, tiende sistemáticamente a desestabilizar el flujo. Las condiciones críticas del caso no ortogonal 0 _ AoA _ _/2 pueden recuperarse a partir del caso ortogonal mediante el uso de una transformación analítica simple que implica el ángulo de ataque AoA. Estos resultados pueden ayudar a comprender los mecanismos de desestabilización que se producen en el borde de ataque de las alas de los aviones a ángulos de ataque finitos. Como tareas pendientes quedaría realizar estudios que tengan en cuenta variaciones del campo de presión en el flujo base así como la extensión de éste al caso de flujos compresibles.

Modeling impact assessment of transport policies through a multiregional input-output integrated approach

El objetivo de esta investigación es desarrollar una metodología para estimar los potenciales impactos económicos y de transporte generados por la aplicación de políticas en el sector transporte. Los departamentos de transporte y otras instituciones gubernamentales relacionadas se encuentran interesadas en estos análisis debido a que son presentados comúnmente de forma errónea por la insuficiencia de datos o por la falta de metodologías adecuadas. La presente investigación tiene por objeto llenar este vacío haciendo un análisis exhaustivo de las técnicas disponibles que coincidan con ese propósito. Se ha realizado un análisis que ha identificado las diferencias cuando son aplicados para la valoración de los beneficios para el usuario o para otros efectos como aspectos sociales. Como resultado de ello, esta investigación ofrece un enfoque integrado que incluye un modelo Input-Output de múltiples regiones basado en la utilidad aleatoria (RUBMRIO), y un modelo de red de transporte por carretera. Este modelo permite la reproducción con mayor detalle y realismo del transporte de mercancías que por medio de su estructura sectorial identifica los vínculos de las compras y ventas inter-industriales dentro de un país utilizando los servicios del transporte de mercancías. Por esta razón, el modelo integrado es aplicable a diversas políticas de transporte. En efecto, el enfoque se ha aplicado para estudiar los efectos macroeconómicos regionales de la implementación de dos políticas diferentes en el sistema de transporte de mercancías de España, tales como la tarificación basada en la distancia recorrida por vehículo-kilómetro (€/km) aplicada a los vehículos del transporte de mercancías, y para la introducción de vehículos más largos y pesados de mercancías en la red de carreteras de España. El enfoque metodológico se ha evaluado caso por caso teniendo en cuenta una selección de la red de carreteras que unen las capitales de las regiones españolas. También se ha tenido en cuenta una dimensión económica a través de una tabla Input-Output de múltiples regiones (MRIO) y la base de datos de conteo de tráfico existente para realizar la validación del modelo. El enfoque integrado reproduce las condiciones de comercio observadas entre las regiones usando el sistema de transporte de mercancías por carretera, y que permite por comparación con los escenarios de políticas, determinar las contribuciones a los cambios distributivos y generativos. Así pues, el análisis estima los impactos económicos en cualquier región considerando los cambios en el Producto Interno Bruto (PIB) y el empleo. El enfoque identifica los cambios en el sistema de transporte a través de todos los caminos de la red de transporte a través de las medidas de efectividad (MOEs). Los resultados presentados en esta investigación proporcionan evidencia sustancial de que en la evaluación de las políticas de transporte, es necesario establecer un vínculo entre la estructura económica de las regiones y de los servicios de transporte. Los análisis muestran que para la mayoría de las regiones del país, los cambios son evidentes para el PIB y el empleo, ya que el comercio se fomenta o se inhibe. El enfoque muestra cómo el tráfico se desvía en ambas políticas, y también determina detalles de las emisiones de contaminantes en los dos escenarios. Además, las políticas de fijación de precios o de regulación de los sistemas de transporte de mercancías por carretera dirigidas a los productores y consumidores en las regiones promoverán transformaciones regionales afectando todo el país, y esto conduce a conclusiones diferentes. Así mismo, este enfoque integrado podría ser útil para evaluar otras políticas y otros países en todo el mundo. The purpose of this research is to develop a methodological approach aimed at assessing the potential economic and transportation impacts of transport policies. Transportation departments and other related government parties are interested in such analysis because it is commonly misrepresented for the insufficiency of data and suitable methodologies available. This research is directed at filling this gap by making a comprehensive analysis of the available techniques that match with that purpose. The differences when they are applied for the valuation of user benefits or for other impacts as social matters have been identified. As a result, this research presents an integrated approach which includes both a random utility-based multiregional Input-Output model (RUBMRIO), and a road transport network model. This model accounts for freight transport with more detail and realism because its commodity-based structure traces the linkages of inter-industry purchases and sales that use freight services within a given country. For this reason, the integrated model is applicable to various transport policies. In fact, the approach is applied to study the regional macroeconomic effects of implementing two different policies in the freight transport system of Spain, such as a distance-based charge in vehicle-kilometer (€/km) for Heavy Goods Vehicles (HGVs), and the introduction of Longer and Heavier Vehicles (LHVs) in the road network of Spain. The methodological approach has been evaluated on a case by case basis considering a selected road network of highways linking the capitals of the Spanish regions. It has also considered an economic dimension through a Multiregional Input Output Table (MRIO) and the existing traffic count database used in the model validation. The integrated approach replicates observed conditions of trade among regions using road freight transport systems that determine contributions to distributional and generative changes by comparison with policy scenarios. Therefore, the model estimates economic impacts in any given area by considering changes in Gross Domestic Product (GDP), employment (jobs), and in the transportation system across all paths of the transport network considering Measures of effectiveness (MOEs). The results presented in this research provide substantive evidence that in the assessment of transport policies it is necessary to establish a link between the economic structure of regions and the transportation services. The analysis shows that for most regions in the country, GDP and employment changes are noticeable when trade is encouraged or discouraged. This approach shows how traffic is diverted in both policies, and also provides details of the pollutant emissions in both scenarios. Furthermore, policies, such as pricing or regulation of road freight transportation systems, directed to producers and consumers in regions will promote different regional transformations across the country, and this lead to different conclusions. In addition, this integrated approach could be useful to assess other policies and countries worldwide.

FPLA: a modeling framework for describing flexible software product line architecture

Nowadays, Software Product Line (SPL) engineering [1] has been widely-adopted in software development due to the significant improvements that has provided, such as reducing cost and time-to-market and providing flexibility to respond to planned changes [2]. SPL takes advantage of common features among the products of a family through the systematic reuse of the core-assets and the effective management of variabilities across the products. SPL features are realized at the architectural level in product-line architecture (PLA) models. Therefore, suitable modeling and specification techniques are required to model variability. In fact, architectural variability modeling has become a challenge for SPLE due to the fact that PLA modeling requires not only modeling variability at the level of the external architecture configuration (see [3,4] literature reviews), but also at the level of internal specification of components [5]. In addition, PLA modeling requires preserving the traceability between features and PLAs. Finally, it is important to take into account that PLA modeling should guide architects in modeling the PLA core assets and variability, and in deriving the customized products. To deal with these needs, we present in this demonstration the FPLA Modeling Framework.

Knowledge representation tool for cognitiveprocesses modeling

In the last decades, neuropsychological theories tend to consider cognitive functions as a result of the whole brainwork and not as individual local areas of its cortex. Studies based on neuroimaging techniques have increased in the last years, promoting an exponential growth of the body of knowledge about relations between cognitive functions and brain structures [1]. However, so fast evolution make complicated to integrate them in verifiable theories and, even more, translated in to cognitive rehabilitation. The aim of this research work is to develop a cognitive process-modeling tool. The purpose of this system is, in the first term, to represent multidimensional data, from structural and functional connectivity, neuroimaging, data from lesion studies and derived data from clinical intervention [2][3]. This will allow to identify consolidated knowledge, hypothesis, experimental designs, new data from ongoing studies and emerging results from clinical interventions. In the second term, we pursuit to use Artificial Intelligence to assist in decision making allowing to advance towards evidence based and personalized treatments in cognitive rehabilitation. This work presents the knowledge base design of the knowledge representation tool. It is compound of two different taxonomies (structure and function) and a set of tags linking both taxonomies at different levels of structural and functional organization. The remainder of the abstract is organized as follows: Section 2 presents the web application used for gathering necessary information for generating the knowledge base, Section 3 describes knowledge base structure and finally Section 4 expounds reached conclusions.

Multiscale modeling of a small punch test on nanostructured CP titanium

Small punch (SP) test techniques are typically used to study the mechanical properties of materials or components from miniature size specimens. This kind of test was originally developed to assess ductility loss in steel caused by irradiation or thermal treatment, particularly when the amount of metal was limited, but it soon proved to be a powerful method to estimate several properties.

Modeling Commercial Knowledge to Develop Advanced Agent-based Marketplaces for E-commerce

This paper argues about the utility of advanced knowledge-based techniques to develop web-based applications that help consumers in finding products within marketplaces in e-commerce. In particular, we describe the idea of model-based approach to develop a shopping agent that dynamically configures a product according to the needs and preferences of customers. Finally, the paper summarizes the advantages provided by this approach.

Microstructure-based numerical modeling of the mechanical behavior of Mg alloys

Dentro de los materiales estructurales, el magnesio y sus aleaciones están siendo el foco de una de profunda investigación. Esta investigación está dirigida a comprender la relación existente entre la microestructura de las aleaciones de Mg y su comportamiento mecánico. El objetivo es optimizar las aleaciones actuales de magnesio a partir de su microestructura y diseñar nuevas aleaciones. Sin embargo, el efecto de los factores microestructurales (como la forma, el tamaño, la orientación de los precipitados y la morfología de los granos) en el comportamiento mecánico de estas aleaciones está todavía por descubrir. Para conocer mejor de la relación entre la microestructura y el comportamiento mecánico, es necesaria la combinación de técnicas avanzadas de caracterización experimental como de simulación numérica, a diferentes longitudes de escala. En lo que respecta a las técnicas de simulación numérica, la homogeneización policristalina es una herramienta muy útil para predecir la respuesta macroscópica a partir de la microestructura de un policristal (caracterizada por el tamaño, la forma y la distribución de orientaciones de los granos) y el comportamiento del monocristal. La descripción de la microestructura se lleva a cabo mediante modernas técnicas de caracterización (difracción de rayos X, difracción de electrones retrodispersados, así como con microscopia óptica y electrónica). Sin embargo, el comportamiento del cristal sigue siendo difícil de medir, especialmente en aleaciones de Mg, donde es muy complicado conocer el valor de los parámetros que controlan el comportamiento mecánico de los diferentes modos de deslizamiento y maclado. En la presente tesis se ha desarrollado una estrategia de homogeneización computacional para predecir el comportamiento de aleaciones de magnesio. El comportamiento de los policristales ha sido obtenido mediante la simulación por elementos finitos de un volumen representativo (RVE) de la microestructura, considerando la distribución real de formas y orientaciones de los granos. El comportamiento del cristal se ha simulado mediante un modelo de plasticidad cristalina que tiene en cuenta los diferentes mecanismos físicos de deformación, como el deslizamiento y el maclado. Finalmente, la obtención de los parámetros que controlan el comportamiento del cristal (tensiones críticas resueltas (CRSS) así como las tasas de endurecimiento para todos los modos de maclado y deslizamiento) se ha resuelto mediante la implementación de una metodología de optimización inversa, una de las principales aportaciones originales de este trabajo. La metodología inversa pretende, por medio del algoritmo de optimización de Levenberg-Marquardt, obtener el conjunto de parámetros que definen el comportamiento del monocristal y que mejor ajustan a un conjunto de ensayos macroscópicos independientes. Además de la implementación de la técnica, se han estudiado tanto la objetividad del metodología como la unicidad de la solución en función de la información experimental. La estrategia de optimización inversa se usó inicialmente para obtener el comportamiento cristalino de la aleación AZ31 de Mg, obtenida por laminado. Esta aleación tiene una marcada textura basal y una gran anisotropía plástica. El comportamiento de cada grano incluyó cuatro mecanismos de deformación diferentes: deslizamiento en los planos basal, prismático, piramidal hc+ai, junto con el maclado en tracción. La validez de los parámetros resultantes se validó mediante la capacidad del modelo policristalino para predecir ensayos macroscópicos independientes en diferentes direcciones. En segundo lugar se estudió mediante la misma estrategia, la influencia del contenido de Neodimio (Nd) en las propiedades de una aleación de Mg-Mn-Nd, obtenida por extrusión. Se encontró que la adición de Nd produce una progresiva isotropización del comportamiento macroscópico. El modelo mostró que este incremento de la isotropía macroscópica era debido tanto a la aleatoriedad de la textura inicial como al incremento de la isotropía del comportamiento del cristal, con valores similares de las CRSSs de los diferentes modos de deformación. Finalmente, el modelo se empleó para analizar el efecto de la temperatura en el comportamiento del cristal de la aleación de Mg-Mn-Nd. La introducción en el modelo de los efectos non-Schmid sobre el modo de deslizamiento piramidal hc+ai permitió capturar el comportamiento mecánico a temperaturas superiores a 150_C. Esta es la primera vez, de acuerdo con el conocimiento del autor, que los efectos non-Schmid han sido observados en una aleación de Magnesio. The study of Magnesium and its alloys is a hot research topic in structural materials. In particular, special attention is being paid in understanding the relationship between microstructure and mechanical behavior in order to optimize the current alloy microstructures and guide the design of new alloys. However, the particular effect of several microstructural factors (precipitate shape, size and orientation, grain morphology distribution, etc.) in the mechanical performance of a Mg alloy is still under study. The combination of advanced characterization techniques and modeling at several length scales is necessary to improve the understanding of the relation microstructure and mechanical behavior. Respect to the simulation techniques, polycrystalline homogenization is a very useful tool to predict the macroscopic response from polycrystalline microstructure (grain size, shape and orientation distributions) and crystal behavior. The microstructure description is fully covered with modern characterization techniques (X-ray diffraction, EBSD, optical and electronic microscopy). However, the mechanical behaviour of single crystals is not well-known, especially in Mg alloys where the correct parameterization of the mechanical behavior of the different slip/twin modes is a very difficult task. A computational homogenization framework for predicting the behavior of Magnesium alloys has been developed in this thesis. The polycrystalline behavior was obtained by means of the finite element simulation of a representative volume element (RVE) of the microstructure including the actual grain shape and orientation distributions. The crystal behavior for the grains was accounted for a crystal plasticity model which took into account the physical deformation mechanisms, e.g. slip and twinning. Finally, the problem of the parametrization of the crystal behavior (critical resolved shear stresses (CRSS) and strain hardening rates of all the slip and twinning modes) was obtained by the development of an inverse optimization methodology, one of the main original contributions of this thesis. The inverse methodology aims at finding, by means of the Levenberg-Marquardt optimization algorithm, the set of parameters defining crystal behavior that best fit a set of independent macroscopic tests. The objectivity of the method and the uniqueness of solution as function of the input information has been numerically studied. The inverse optimization strategy was first used to obtain the crystal behavior of a rolled polycrystalline AZ31 Mg alloy that showed a marked basal texture and a strong plastic anisotropy. Four different deformation mechanisms: basal, prismatic and pyramidal hc+ai slip, together with tensile twinning were included to characterize the single crystal behavior. The validity of the resulting parameters was proved by the ability of the polycrystalline model to predict independent macroscopic tests on different directions. Secondly, the influence of Neodymium (Nd) content on an extruded polycrystalline Mg-Mn-Nd alloy was studied using the same homogenization and optimization framework. The effect of Nd addition was a progressive isotropization of the macroscopic behavior. The model showed that this increase in the macroscopic isotropy was due to a randomization of the initial texture and also to an increase of the crystal behavior isotropy (similar values of the CRSSs of the different modes). Finally, the model was used to analyze the effect of temperature on the crystal behaviour of a Mg-Mn-Nd alloy. The introduction in the model of non-Schmid effects on the pyramidal hc+ai slip allowed to capture the inverse strength differential that appeared, between the tension and compression, above 150_C. This is the first time, to the author's knowledge, that non-Schmid effects have been reported for Mg alloys.

Enhancing regression models for complex systems using evolutionary techniques for feature engineering

This work proposes an automatic methodology for modeling complex systems. Our methodology is based on the combination of Grammatical Evolution and classical regression to obtain an optimal set of features that take part of a linear and convex model. This technique provides both Feature Engineering and Symbolic Regression in order to infer accurate models with no effort or designer's expertise requirements. As advanced Cloud services are becoming mainstream, the contribution of data centers in the overall power consumption of modern cities is growing dramatically. These facilities consume from 10 to 100 times more power per square foot than typical office buildings. Modeling the power consumption for these infrastructures is crucial to anticipate the effects of aggressive optimization policies, but accurate and fast power modeling is a complex challenge for high-end servers not yet satisfied by analytical approaches. For this case study, our methodology minimizes error in power prediction. This work has been tested using real Cloud applications resulting on an average error in power estimation of 3.98%. Our work improves the possibilities of deriving Cloud energy efficient policies in Cloud data centers being applicable to other computing environments with similar characteristics.

Server power modeling for run-time energy optimization of cloud computing facilities

As advanced Cloud services are becoming mainstream, the contribution of data centers in the overall power consumption of modern cities is growing dramatically. The average consumption of a single data center is equivalent to the energy consumption of 25.000 households. Modeling the power consumption for these infrastructures is crucial to anticipate the effects of aggressive optimization policies, but accurate and fast power modeling is a complex challenge for high-end servers not yet satisfied by analytical approaches. This work proposes an automatic method, based on Multi-Objective Particle Swarm Optimization, for the identification of power models of enterprise servers in Cloud data centers. Our approach, as opposed to previous procedures, does not only consider the workload consolidation for deriving the power model, but also incorporates other non traditional factors like the static power consumption and its dependence with temperature. Our experimental results shows that we reach slightly better models than classical approaches, but simul- taneously simplifying the power model structure and thus the numbers of sensors needed, which is very promising for a short-term energy prediction. This work, validated with real Cloud applications, broadens the possibilities to derive efficient energy saving techniques for Cloud facilities.

Multiscale modeling of the plastic behaviour in single crystal tungsten: from atomistic to crystal plasticity simulations

En una planta de fusión, los materiales en contacto con el plasma así como los materiales de primera pared experimentan condiciones particularmente hostiles al estar expuestos a altos flujos de partículas, neutrones y grandes cargas térmicas. Como consecuencia de estas diferentes y complejas condiciones de trabajo, el estudio, desarrollo y diseño de estos materiales es uno de los más importantes retos que ha surgido en los últimos años para la comunidad científica en el campo de los materiales y la energía. Debido a su baja tasa de erosión, alta resistencia al sputtering, alta conductividad térmica, muy alto punto de fusión y baja retención de tritio, el tungsteno (wolframio) es un importante candidato como material de primera pared y como posible material estructural avanzado en fusión por confinamiento magnético e inercial. Sin embargo, el tiempo de vida del tungsteno viene controlado por diversos factores como son su respuesta termo-mecánica en la superficie, la posibilidad de fusión y el fallo por acumulación de helio. Es por ello que el tiempo de vida limitado por la respuesta mecánica del tungsteno (W), y en particular su fragilidad, sean dos importantes aspectos que tienes que ser investigados. El comportamiento plástico en materiales refractarios con estructura cristalina cúbica centrada en las caras (bcc) como el tungsteno está gobernado por las dislocaciones de tipo tornillo a escala atómica y por conjuntos e interacciones de dislocaciones a escalas más grandes. El modelado de este complejo comportamiento requiere la aplicación de métodos capaces de resolver de forma rigurosa cada una de las escalas. El trabajo que se presenta en esta tesis propone un modelado multiescala que es capaz de dar respuestas ingenieriles a las solicitudes técnicas del tungsteno, y que a su vez está apoyado por la rigurosa física subyacente a extensas simulaciones atomísticas. En primer lugar, las propiedades estáticas y dinámicas de las dislocaciones de tipo tornillo en cinco potenciales interatómicos de tungsteno son comparadas, determinando cuáles de ellos garantizan una mayor fidelidad física y eficiencia computacional. Las grandes tasas de deformación asociadas a las técnicas de dinámica molecular hacen que las funciones de movilidad de las dislocaciones obtenidas no puedan ser utilizadas en los siguientes pasos del modelado multiescala. En este trabajo, proponemos dos métodos alternativos para obtener las funciones de movilidad de las dislocaciones: un modelo Monte Cario cinético y expresiones analíticas. El conjunto de parámetros necesarios para formular el modelo de Monte Cario cinético y la ley de movilidad analítica son calculados atomísticamente. Estos parámetros incluyen, pero no se limitan a: la determinación de las entalpias y energías de formación de las parejas de escalones que forman las dislocaciones, la parametrización de los efectos de no Schmid característicos en materiales bcc,etc. Conociendo la ley de movilidad de las dislocaciones en función del esfuerzo aplicado y la temperatura, se introduce esta relación como ecuación de flujo dentro de un modelo de plasticidad cristalina. La predicción del modelo sobre la dependencia del límite de fluencia con la temperatura es validada experimentalmente con ensayos uniaxiales en tungsteno monocristalino. A continuación, se calcula el límite de fluencia al aplicar ensayos uniaxiales de tensión para un conjunto de orientaciones cristalográticas dentro del triángulo estándar variando la tasa de deformación y la temperatura de los ensayos. Finalmente, y con el objetivo de ser capaces de predecir una respuesta más dúctil del tungsteno para una variedad de estados de carga, se realizan ensayos biaxiales de tensión sobre algunas de las orientaciones cristalográficas ya estudiadas en función de la temperatura.-------------------------------------------------------------------------ABSTRACT ----------------------------------------------------------Tungsten and tungsten alloys are being considered as leading candidates for structural and functional materials in future fusion energy devices. The most attractive properties of tungsten for the design of magnetic and inertial fusion energy reactors are its high melting point, high thermal conductivity, low sputtering yield and low longterm disposal radioactive footprint. However, tungsten also presents a very low fracture toughness, mostly associated with inter-granular failure and bulk plasticity, that limits its applications. As a result of these various and complex conditions of work, the study, development and design of these materials is one of the most important challenges that have emerged in recent years to the scientific community in the field of materials for energy applications. The plastic behavior of body-centered cubic (bcc) refractory metals like tungsten is governed by the kink-pair mediated thermally activated motion of h¿ (\1 11)i screw dislocations on the atomistic scale and by ensembles and interactions of dislocations at larger scales. Modeling this complex behavior requires the application of methods capable of resolving rigorously each relevant scale. The work presented in this thesis proposes a multiscale model approach that gives engineering-level responses to the technical specifications required for the use of tungsten in fusion energy reactors, and it is also supported by the rigorous underlying physics of extensive atomistic simulations. First, the static and dynamic properties of screw dislocations in five interatomic potentials for tungsten are compared, determining which of these ensure greater physical fidelity and computational efficiency. The large strain rates associated with molecular dynamics techniques make the dislocation mobility functions obtained not suitable to be used in the next steps of the multiscale model. Therefore, it is necessary to employ mobility laws obtained from a different method. In this work, we suggest two alternative methods to get the dislocation mobility functions: a kinetic Monte Carlo model and analytical expressions. The set of parameters needed to formulate the kinetic Monte Carlo model and the analytical mobility law are calculated atomistically. These parameters include, but are not limited to: enthalpy and energy barriers of kink-pairs as a function of the stress, width of the kink-pairs, non-Schmid effects ( both twinning-antitwinning asymmetry and non-glide stresses), etc. The function relating dislocation velocity with applied stress and temperature is used as the main source of constitutive information into a dislocation-based crystal plasticity framework. We validate the dependence of the yield strength with the temperature predicted by the model against existing experimental data of tensile tests in singlecrystal tungsten, with excellent agreement between the simulations and the measured data. We then extend the model to a number of crystallographic orientations uniformly distributed in the standard triangle and study the effects of temperature and strain rate. Finally, we perform biaxial tensile tests and provide the yield surface as a function of the temperature for some of the crystallographic orientations explored in the uniaxial tensile tests.

Fault management techniques for systems with SRAM-based FPGAs

Las Field-Programmable Gate Arrays (FPGAs) SRAM se construyen sobre una memoria de configuración de tecnología RAM Estática (SRAM). Presentan múltiples características que las hacen muy interesantes para diseñar sistemas empotrados complejos. En primer lugar presentan un coste no-recurrente de ingeniería (NRE) bajo, ya que los elementos lógicos y de enrutado están pre-implementados (el diseño de usuario define su conexionado). También, a diferencia de otras tecnologías de FPGA, pueden ser reconfiguradas (incluso en campo) un número ilimitado de veces. Es más, las FPGAs SRAM de Xilinx soportan Reconfiguración Parcial Dinámica (DPR), la cual permite reconfigurar la FPGA sin interrumpir la aplicación. Finalmente, presentan una alta densidad de lógica, una alta capacidad de procesamiento y un rico juego de macro-bloques. Sin embargo, un inconveniente de esta tecnología es su susceptibilidad a la radiación ionizante, la cual aumenta con el grado de integración (geometrías más pequeñas, menores tensiones y mayores frecuencias). Esta es una precupación de primer nivel para aplicaciones en entornos altamente radiativos y con requisitos de alta confiabilidad. Este fenómeno conlleva una degradación a largo plazo y también puede inducir fallos instantáneos, los cuales pueden ser reversibles o producir daños irreversibles. En las FPGAs SRAM, los fallos inducidos por radiación pueden aparecer en en dos capas de arquitectura diferentes, que están físicamente superpuestas en el dado de silicio. La Capa de Aplicación (o A-Layer) contiene el hardware definido por el usuario, y la Capa de Configuración contiene la memoria de configuración y la circuitería de soporte. Los fallos en cualquiera de estas capas pueden hacer fracasar el sistema, lo cual puede ser ás o menos tolerable dependiendo de los requisitos de confiabilidad del sistema. En el caso general, estos fallos deben gestionados de alguna manera. Esta tesis trata sobre la gestión de fallos en FPGAs SRAM a nivel de sistema, en el contexto de sistemas empotrados autónomos y confiables operando en un entorno radiativo. La tesis se centra principalmente en aplicaciones espaciales, pero los mismos principios pueden aplicarse a aplicaciones terrenas. Las principales diferencias entre ambas son el nivel de radiación y la posibilidad de mantenimiento. Las diferentes técnicas para la gestión de fallos en A-Layer y C-Layer son clasificados, y sus implicaciones en la confiabilidad del sistema son analizados. Se proponen varias arquitecturas tanto para Gestores de Fallos de una capa como de doble-capa. Para estos últimos se propone una arquitectura novedosa, flexible y versátil. Gestiona las dos capas concurrentemente de manera coordinada, y permite equilibrar el nivel de redundancia y la confiabilidad. Con el objeto de validar técnicas de gestión de fallos dinámicas, se desarrollan dos diferentes soluciones. La primera es un entorno de simulación para Gestores de Fallos de C-Layer, basado en SystemC como lenguaje de modelado y como simulador basado en eventos. Este entorno y su metodología asociada permite explorar el espacio de diseño del Gestor de Fallos, desacoplando su diseño del desarrollo de la FPGA objetivo. El entorno incluye modelos tanto para la C-Layer de la FPGA como para el Gestor de Fallos, los cuales pueden interactuar a diferentes niveles de abstracción (a nivel de configuration frames y a nivel físico JTAG o SelectMAP). El entorno es configurable, escalable y versátil, e incluye capacidades de inyección de fallos. Los resultados de simulación para algunos escenarios son presentados y comentados. La segunda es una plataforma de validación para Gestores de Fallos de FPGAs Xilinx Virtex. La plataforma hardware aloja tres Módulos de FPGA Xilinx Virtex-4 FX12 y dos Módulos de Unidad de Microcontrolador (MCUs) de 32-bits de propósito general. Los Módulos MCU permiten prototipar Gestores de Fallos de C-Layer y A-Layer basados en software. Cada Módulo FPGA implementa un enlace de A-Layer Ethernet (a través de un switch Ethernet) con uno de los Módulos MCU, y un enlace de C-Layer JTAG con el otro. Además, ambos Módulos MCU intercambian comandos y datos a través de un enlace interno tipo UART. Al igual que para el entorno de simulación, se incluyen capacidades de inyección de fallos. Los resultados de pruebas para algunos escenarios son también presentados y comentados. En resumen, esta tesis cubre el proceso completo desde la descripción de los fallos FPGAs SRAM inducidos por radiación, pasando por la identificación y clasificación de técnicas de gestión de fallos, y por la propuesta de arquitecturas de Gestores de Fallos, para finalmente validarlas por simulación y pruebas. El trabajo futuro está relacionado sobre todo con la implementación de Gestores de Fallos de Sistema endurecidos para radiación. ABSTRACT SRAM-based Field-Programmable Gate Arrays (FPGAs) are built on Static RAM (SRAM) technology configuration memory. They present a number of features that make them very convenient for building complex embedded systems. First of all, they benefit from low Non-Recurrent Engineering (NRE) costs, as the logic and routing elements are pre-implemented (user design defines their connection). Also, as opposed to other FPGA technologies, they can be reconfigured (even in the field) an unlimited number of times. Moreover, Xilinx SRAM-based FPGAs feature Dynamic Partial Reconfiguration (DPR), which allows to partially reconfigure the FPGA without disrupting de application. Finally, they feature a high logic density, high processing capability and a rich set of hard macros. However, one limitation of this technology is its susceptibility to ionizing radiation, which increases with technology scaling (smaller geometries, lower voltages and higher frequencies). This is a first order concern for applications in harsh radiation environments and requiring high dependability. Ionizing radiation leads to long term degradation as well as instantaneous faults, which can in turn be reversible or produce irreversible damage. In SRAM-based FPGAs, radiation-induced faults can appear at two architectural layers, which are physically overlaid on the silicon die. The Application Layer (or A-Layer) contains the user-defined hardware, and the Configuration Layer (or C-Layer) contains the (volatile) configuration memory and its support circuitry. Faults at either layers can imply a system failure, which may be more ore less tolerated depending on the dependability requirements. In the general case, such faults must be managed in some way. This thesis is about managing SRAM-based FPGA faults at system level, in the context of autonomous and dependable embedded systems operating in a radiative environment. The focus is mainly on space applications, but the same principles can be applied to ground applications. The main differences between them are the radiation level and the possibility for maintenance. The different techniques for A-Layer and C-Layer fault management are classified and their implications in system dependability are assessed. Several architectures are proposed, both for single-layer and dual-layer Fault Managers. For the latter, a novel, flexible and versatile architecture is proposed. It manages both layers concurrently in a coordinated way, and allows balancing redundancy level and dependability. For the purpose of validating dynamic fault management techniques, two different solutions are developed. The first one is a simulation framework for C-Layer Fault Managers, based on SystemC as modeling language and event-driven simulator. This framework and its associated methodology allows exploring the Fault Manager design space, decoupling its design from the target FPGA development. The framework includes models for both the FPGA C-Layer and for the Fault Manager, which can interact at different abstraction levels (at configuration frame level and at JTAG or SelectMAP physical level). The framework is configurable, scalable and versatile, and includes fault injection capabilities. Simulation results for some scenarios are presented and discussed. The second one is a validation platform for Xilinx Virtex FPGA Fault Managers. The platform hosts three Xilinx Virtex-4 FX12 FPGA Modules and two general-purpose 32-bit Microcontroller Unit (MCU) Modules. The MCU Modules allow prototyping software-based CLayer and A-Layer Fault Managers. Each FPGA Module implements one A-Layer Ethernet link (through an Ethernet switch) with one of the MCU Modules, and one C-Layer JTAG link with the other. In addition, both MCU Modules exchange commands and data over an internal UART link. Similarly to the simulation framework, fault injection capabilities are implemented. Test results for some scenarios are also presented and discussed. In summary, this thesis covers the whole process from describing the problem of radiationinduced faults in SRAM-based FPGAs, then identifying and classifying fault management techniques, then proposing Fault Manager architectures and finally validating them by simulation and test. The proposed future work is mainly related to the implementation of radiation-hardened System Fault Managers.