As the circuit density and clock rate in SIP or SOC getting higher, the crosstalk interference between interconnects becomes more and more serious. This results in the degradation of signal integrity. Full wave simulation softwares such as HFSS are often used to analyze the characteristics of the transmission system, but the computation time is relatively long and it is difficult to integrate with active circuits. On the other hand, circuit simulation softwares such as ADS has the advantages of simplicity and needs less simulation time. The circuit simulation softwares is so far the best tool of designing high speed circuits. Usually obtain scalable equivalent macro-model in the existing literature with the technology of the frequency-domain, but using the technology of the time-domain to construct scalable equivalent macro-model have more direct, convenient and low-cost. Therefore, considering cost and practicability, this thesis develops a systematic method using the method of Pencil Matrix in time domain to obtain the scalable broadband equivalent macro-model of the components of high-speed interconnect structures. The developed macro-models can be applied to existing simulation softwares for a fast and accurate analysis of systems such as SIP.

Austria, Finland and Sweden - all small highly developed industrial and rich countries - entered the EU in 1995. Their macroeconomic performance since then was quite different. Real GDP in Finland und Sweden increased faster than in EU average, while those of Austria fell back. Austria lost its second rank in GDP per capita (at PPS) and is now the fourth richest EU country; Sweden fell back from the seventh to the eight rank, while Finland improved its position from rank 11 to nine. In a referendum in September 2003 Sweden refused to take over the Euro, whereas the other two countries are members of the Euro area. Ex post model simulations indicate that Finland appears to have profited most from EU membership (0.7 percentage point greater annual GDP growth since 1995), followed by Austria (+0.4 percent) and Sweden (+0.3 percent). / Series: EI Working Papers / Europainstitut

- 4 - Abstract In this research we explored role of financial variables in macro modeling and their performance in case of Mongolia. We employed two different models for assessing performance of financial variables in macro modeling, structural VAR model and small scale macro model (SSMM). In doing so, we performed different analysis such as impulse response for seeing how financial variables fit into system and forecasting performance for how accurate model performs after introducing financial variables. So our result suggested that financial variables have substantial role on macro modeling and inclusion of financial variable is performing very good result in terms of forecasting in both models. JEL Classification C01, C51, C53, E12, E52, G17 Keywords Financial markets, Small scale macro model, Structural VAR, Impulse response, Mean absolute errors. Author's e-mail batnyamd@gmail.com Supervisor's e-mail roman.horvath@gmail.com

Les mousses de polyuréthane (PU) hautement poreuses sont largement utilisées dans différentes industries pour dissiper l’énergie des ondes sonores et vibratoires. La propagation des ondes acoustiques dans ces matériaux poroélastiques est décrite à partir d’un ensemble de paramètres physiques connus sous le nom de paramètres de Biot (pour les matériaux isotropes, ils sont composés de 5 paramètres non acoustiques et de 4 paramètres mécaniques). Il est bien connu que les propriétés macroscopiques dépendent intrinsèquement des propriétés de la microstructure de la mousse. Ainsi, une compréhension claire des corrélations entre la structure interne des mousses de PU et leurs paramètres de Biot ainsi que la contribution de chaque paramètre, soit microscopique ou macroscopique, sur l’indicateur vibroacoustique désiré est d’un intérêt majeur au stade précoce de la conception et de l’optimisation de ces matériaux poroélastiques. Le développement d’un modèle micromacro qui corrèle les propriétés de la microstructure aux paramètres macroscopiques de Biot est donc nécessaire. Récemment, un modèle qui corrèle les propriétés de la microstructure des mousses PU hautement poreuses à leurs propriétés non acoustiques a été présenté par Doutres et coll. [24, 25]. Dans cette étude, les propriétés de la microstructure (dimensions de la cellule et taux de réticulation) sont d’abord caractérisées par un microscope électronique à balayage (SEM). Ensuite, l’effet du taux de réticulation (mesurant le pourcentage de fenêtres ouvertes), de la taille des cellules et de la densité relative sur les propriétés mécaniques de la mousse de polyuréthane a été élucidé à l’aide d’un modèle numérique. Se basant sur ce modèle, un modelé analytique existant, qui corrèle les propriétés de la microstructure de mousses PU entièrement réticulées à ses propriétés mécaniques, a été revu et corrigé pour tenir compte de l’effet important du taux de réticulation. En combinant le modèle de Doutres avec le modèle mécanique développé dans cette thèse, un modèle micro-macro complet est ainsi obtenu. Utilisant ce modèle, l’impact de la variabilité de la microstructure et la contribution de chacun des paramètres microstructuraux à la réponse vibroacoustique ont été étudiés utilisant une méthode d’analyse de sensibilité globale (FAST). La méthode FAST a été utilisée pour identifier l’impact de la microstructure sur, premièrement, les paramètres de Biot-Allard et, deuxièmement, sur les indicateurs vibroacoustiques (absorption et perte par transmission) des mousses de polyuréthane poroélastiques. Une fois les modèles micro-macro et la contribution des propriétés de la microstructure connus, la performance vibroacoustique de la mousse a été optimisée. Ainsi nous avons testé numériquement la performance acoustique de mousses homogènes et de mousses graduellement structurées (variation de propriétés suivant l’épaisseur de la mousse). Cette étude ouvre ainsi de nouvelles portes pour concevoir des mousses PU innovantes avec une microstructure modifiée et des performances vibroacoustique améliorées. / Abstract : Highly porous polyurethane (PU) foams are widely used in different industries to dissipate the energy of sound and vibration waves. Propagation of acoustic waves in such poroelastic materials is explained based on a set of physical parameters known as the Biot's parameters (for isotropic materials these are comprised of 5 non-acoustical parameters and 4 mechanical parameters). These macroscopic properties are inherently dependent on the microstructure properties of the foam. Hence, a clear understanding of correlations between the internal structure of PU foams and their Biot's parameters and the contribution of each parameter, either microscopic or macroscopic, to classical vibro-acoustic indicators is of utmost interest at the early stage of design and optimization of such poroelastic materials. In consequence, a micro macro model that correlates microstructure properties to macroscopic Biot's parameters is needed. Recently, a model that correlates the microstructure properties of highly porous PU foams to their non-acoustical properties was presented by~\citet{Doutres2011,Doutres2013}. In this study, micro-structure properties (strut length, strut thickness, and open pore content) are first characterized using a Scanning Electron Microscope (SEM). Then, a numerical study is performed to elucidate the effect of open pore content (known as reticulation rate), cell size, and relative density on the mechanical properties of polyurethane foam. Based on this study, an existing analytical model~\cite{Gong2005} that correlates fully reticulated unit cell microstructure properties of PU foams to its mechanical properties is corrected and updated to account for these important parameters. Combined with Doutres’ model, the proposed extension lead to a full micro-macro model for predicting the acoustic performance of PU foams from its microstructure. Using this model, the contribution of the unit cell parameters and effect of their variability on classical vibro-acoustic indicators (absorption and transmission loss) is investigated using a global sensitivity analysis method (FAST). The FAST method is used to identify the impact of microstructure role on, first, the Biot-Allard parameters and, second, on vibro-acoustical indicators of poroelastic polyurethane foams. Based on this sensitivity analysis study, the developed micro-macro model, is used to design both optimum homogeneous foam and functionally graded foams (properties optimally varnish along the thickness of the foam) targeting specific in absorption and/or transmission loss problems. This study opens thus a new door to design innovative PU foams with modified micro-structure and improved vibro-acoustical performance.

<p>Water infrastructure forms a critical sector of our social system and provides goods and services for public health, the natural environment, economic safety, various businesses, and government operations. In the United States (US), drinking water is supplied nationally through one million miles of pipes, most of which were installed in the early to mid-20th century with a life span of 75 to 100 years. Along with this fact, water bills which are rising faster than inflation, result in communities grappling with aging water systems, fewer water resources, and extreme weather. The federal government’s share of capital investment for water infrastructure has fallen from 31% in 1977 to 4% in 2017. Regional and state expenditure has accounted for a much larger share as federal aid for water infrastructure capital needs has declined. This has led to water rates rising to cover the costs of replacing and upgrading water infrastructure in many communities across the country. They are struggling to meet such costs through local rates and fees.</p><p>Over the next 20 years, more than 56 million new users are expected to connect to centralized treatment systems, and $271 billion is needed to meet current and future demands. However, the investment in critical water infrastructure is currently only meeting a fraction of the funding need. In 2019, the total capital spending on water infrastructure at all levels was $48 billion, while investment needs totaled $129 billion, creating an $81 billion gap. As such, the most recent American Society of Civil Engineers’ Infrastructure Report Card assigned a D to the drinking water infrastructure and a D+ to the nation’s wastewater infrastructure. Ineffectual and wasteful investment in the water sector has caused an adverse effect on grades in the infrastructure report card for water infrastructures. Moreover, this may negatively impact water-reliant sectors and water-related infrastructures due to the economic ripple effect.</p><p>This research has developed a data-driven strategic investment decision support system to close the existing water infrastructure investment gap and reduce the vulnerability of aging water infrastructure. The first phase of this study was to determine the causes affecting the grades in the infrastructure report card for drinking water and wastewater infrastructure and contributing to any latent threats and prolonged risks. It uses data-driven approaches based on analysis of existing ineffective improvement methods and recommendations. It attempts to leverage a data-driven supervised statistical learning method to capture the complex relationships between new challenges and the growing demand for water infrastructure needs. The ultimate outcome of this phase is a research approach to minimize water and wastewater vulnerability and close the investment gap to help create a paradigm shift in the current state of practice. Furthermore, improving the resiliency of and increasing investments in the water and wastewater infrastructure will lead to a resilient, efficient, and reliable water future and protect the public health of future generations.</p><p>The second phase of this study was to predict the economic benefits of additional federal support in water infrastructure among interdependent sectors within an economic system to facilitate the federal government’s share of capital investment. It conducts ripple effects analysis, which predicts the effectiveness of water infrastructure capital investment using historical economic data. It explores how federal capital investment in water infrastructure spreads economic benefits within an interdependent system. This phase was conducted at the federal level using the interindustry-macro model that analyzes macroeconomic data, including over 400 sectors. Investments that are coordinated at the federal, state, and local level will help control and stabilize rising water rates across the US.</p><p>The third phase of this study was to conduct a cost-benefit assessment in terms of private, financial, economic, and efficiency considerations using nominal and real terms to maximize the benefit of investing in the water sector and reduce the vulnerability of water infrastructures. In order to measure the costs and benefits of a strategy to maximize the efficiency of limited budgets and resources, this phase conducts a cost-benefit analysis due to the investment costs for rehabilitating and improving water infrastructures using historical economic and financial data. The long-term financial framework, including considerations of deep uncertainties so that decision-makers can understand the benefit of investing assets for an optimal level versus the cost of doing nothing and allowing the asset to run to failure is developed using the cost-benefit assessment.</p><p>Finally, a data-driven strategic investment decision support system that helps governments make water infrastructure development plans and infrastructure investment decisions in the water sector is presented. It can help governments with designing a novel system or modifying existing ineffective assessment methods and recommendations aimed at minimizing the mismatch in the water infrastructure investment gap between current spending levels and funding needs. Furthermore, minimizing the risks of ineffectual and wasteful water sector investment through rehabilitating and improving water infrastructures in a rational manner will lead to improved grades in the infrastructure report card and the resiliency of interrelated infrastructures and sectors.</p>

Le chapitre 1 introduit les modèles et donne les principaux résultats obtenus. Dans le chapitre 2, on présente des simulations numériques d'un modèle macroscopique en deux dimensions. La méthode de discrétisation par éléments finis utilisée est décrite. Pour le cas test de l'écoulement autour d'un cylindre, les phénomènes en jeu dans les fluides vieillissants sont observés. Le chapitre 3 concerne l'étude mathématique de la version unidimensionnelle du système d'équations aux dérivées partielles utilisé pour les simulations. On montre que le problème est bien posé et on examine le comportement en temps long de la solution. Dans le dernier chapitre, des équations macroscopiques sont dérivées à partir d'une équation mésoscopique. L'analyse mathématique de cette équation mésoscopique est également menée / This thesis is devoted to the modelling, the mathematical analysis and the simulation of non-Newtonian fluids. Some fluids in an intermediate liquid-solid phase are particularly considered: aging fluids. Modelling scales are macroscopic and mesoscopic. In Chapter 1, we introduce the models and give the main results obtained. In Chapter 2, we present numerical simulations of a macroscopic two-dimensional model. The finite element method used for discretization is described. For the flow past a cylinder test-case, phenomena at play in aging fluids are observed. The Chapter 3 contains a mathematical analysis of the one-dimensional version of the system of partial differential equations used for the simulations. We show well-posedness and investigate the longtime behaviour of the solution. In the last chapter, macroscopic equations are derived from a mesoscopic equation. The mathematical analysis of this mesoscopic equation is also carried out

Les fluctuations électriques des composants sont une limitation à la miniaturisation des circuits. Malgré des procédés de fabrications en continuelle évolution, les variations des caractéristiques électriques dues au désappariement entre deux dispositifs limitent les performances des circuits. Concernant les applications à faible consommation, ces fluctuations locales peuvent devenir très critiques. Dans le contexte du développement d’une technologie CMOS 90nm avec mémoire Flash embarquée pour des applications basse consommation, l’appariement de transistors MOS est étudié. Une analyse de l’impact du dopage de grille des transistors NMOS est menée. L’étude se focalise sur l’appariement en tension des paires différentielles polarisées dans la zone de fonctionnement sous le seuil. Il est démontré que cet appariement peut être dégradé à cause de l’effet « hump », c'est-à-dire la présence de transistors parasites en bord d’active. Un macro-modèle permettant aux concepteurs de modéliser cet effet est présenté. Il est étudié au niveau composant, au niveau circuit et en température. Enfin, une étude de la dégradation de l’appariement des transistors MOS sous stress porteurs chauds est réalisée, validant un modèle de dégradation. Des transistors octogonaux sont proposés pour supprimer l’effet « hump » et donnent d’excellents résultats en termes d’appariement ainsi qu’en fiabilité. / Electrical fluctuations of devices limit chip miniaturization. Despite manufacturing processes in continuous evolution, circuit performances are limited by electrical characteristics variations due to mismatch between two devices. Concerning low power applications, local fluctuations can become very critical. In the context of development of a 90nm CMOS technology with Embedded Flash memory for low power applications, MOS transistors matching is studied. A study of NMOS transistors gate doping impact is conducted. Study focuses on voltage matching of differential pairs biased under threshold. It is demonstrated that this matching can be degraded due to « hump » effect, meaning presence of parasitic devices on active edge. A macro-model allowing designers to model this effect is presented. It is studied at device level, circuit level and for different temperatures. Finally, a degradation study of MOS transistors mismatch under Hot Carriers Injection stress is performed, validating a degradation model. Octagonal devices are proposed to suppress « hump » effect and give good results in terms of matching as well as reliability.

The performance of solid oxide fuel cells (SOFCs) at the cell and system levels is studied using computer simulation. At the cell level, a new model combining the cell micro and macro models is developed. Using this model, the microstructural variables of porous composite electrodes can be linked to the cell performance. In this approach, the electrochemical performance of porous composite electrodes is predicted using a micro-model. In the micro-model, the random-packing sphere method is used to estimate the microstructural properties of porous composite electrodes from the independent microstructural variables. These variables are the electrode porosity, thickness, particle size ratio, and size and volume fraction of electron-conducting particles. Then, the complex interdependency among the multi-component mass transport, electron and ion transports, and the electrochemical and chemical reactions in the microstructure of electrodes is taken into account to predict the electrochemical performance of electrodes. The temperature distribution in the solid structure of the cell and the temperature and species partial pressure distributions in the bulk fuel and air streams are predicted using the cell macro-model. In the macro-model, the energy transport is considered for the cell solid structure and the mass and energy transports are considered for the fuel and air streams. To demonstrate the application of the cell level model developed, entitled the combined micro- and micro-model, several anode-supported co-flow planar cells with a range of microstructures of porous composite electrodes are simulated. The mean total polarization resistance, the mean total power density, and the temperature distribution in the cells are predicted. The results of this study reveal that there is an optimum value for most of the microstructural variables of the electrodes at which the mean total polarization resistance of the cell is minimized. There is also an optimum value for most of the microstructural variables of the electrodes at which the mean total power density of the cell is maximized. The microstructure of porous composite electrodes also plays a significant role in the mean temperature, the temperature difference between the hottest and coldest spots, and the maximum temperature gradient in the solid structure of the cell. Overall, using the combined micro- and micro-model, an appropriate microstructure for porous composite electrodes to enhance the cell performance can be designed. At the system level, the full load operation of two SOFC systems is studied. To model these systems, the basic cell model is used for SOFCs at the cell level, the repeated-cell stack model is used for SOFCs at the stack level, and the thermodynamic model is used for the balance of plant components of the system. In addition to these models, a carbon deposition model based on the thermodynamic equilibrium assumption is employed. For the system level model, the first SOFC system considered is a combined heat and power (CHP) system that operates with biogas fuel. The performance of this system at three different configurations is evaluated. These configurations are different in the fuel processing method to prevent carbon deposition on the anode catalyst. The fuel processing methods considered in these configurations are the anode gas recirculation (AGR), steam reforming (SR), and partial oxidation reformer (POX) methods. The application of this system is studied for operation in a wastewater treatment plant (WWTP) and in single-family detached dwellings. The evaluation of this system for operation in a WWTP indicates that if the entire biogas produced in the WWTP is used in the system with AGR or SR fuel processors, the electric power and heat required to operate the plant can be completely supplied and the extra electric power generated can be sold to the electrical grid. The evaluation of this system for operation in single-family detached dwellings indicates that, depending on the size, location, and building type and design, this system with all configurations studied is suitable to provide the domestic hot water and electric power demands. The second SOFC system is a novel portable electric power generation system that operates with liquid ammonia fuel. Size, simplicity, and high electrical efficiency are the main advantages of this environmentally friendly system. Using a sensitivity analysis, the effects of the cell voltage at several fuel utilization ratios on the number of cells required for the SOFC stack, system efficiency and voltage, and excess air required for thermal management of the SOFC stack are studied.

The performance of solid oxide fuel cells (SOFCs) at the cell and system levels is studied using computer simulation. At the cell level, a new model combining the cell micro and macro models is developed. Using this model, the microstructural variables of porous composite electrodes can be linked to the cell performance. In this approach, the electrochemical performance of porous composite electrodes is predicted using a micro-model. In the micro-model, the random-packing sphere method is used to estimate the microstructural properties of porous composite electrodes from the independent microstructural variables. These variables are the electrode porosity, thickness, particle size ratio, and size and volume fraction of electron-conducting particles. Then, the complex interdependency among the multi-component mass transport, electron and ion transports, and the electrochemical and chemical reactions in the microstructure of electrodes is taken into account to predict the electrochemical performance of electrodes. The temperature distribution in the solid structure of the cell and the temperature and species partial pressure distributions in the bulk fuel and air streams are predicted using the cell macro-model. In the macro-model, the energy transport is considered for the cell solid structure and the mass and energy transports are considered for the fuel and air streams. To demonstrate the application of the cell level model developed, entitled the combined micro- and micro-model, several anode-supported co-flow planar cells with a range of microstructures of porous composite electrodes are simulated. The mean total polarization resistance, the mean total power density, and the temperature distribution in the cells are predicted. The results of this study reveal that there is an optimum value for most of the microstructural variables of the electrodes at which the mean total polarization resistance of the cell is minimized. There is also an optimum value for most of the microstructural variables of the electrodes at which the mean total power density of the cell is maximized. The microstructure of porous composite electrodes also plays a significant role in the mean temperature, the temperature difference between the hottest and coldest spots, and the maximum temperature gradient in the solid structure of the cell. Overall, using the combined micro- and micro-model, an appropriate microstructure for porous composite electrodes to enhance the cell performance can be designed. At the system level, the full load operation of two SOFC systems is studied. To model these systems, the basic cell model is used for SOFCs at the cell level, the repeated-cell stack model is used for SOFCs at the stack level, and the thermodynamic model is used for the balance of plant components of the system. In addition to these models, a carbon deposition model based on the thermodynamic equilibrium assumption is employed. For the system level model, the first SOFC system considered is a combined heat and power (CHP) system that operates with biogas fuel. The performance of this system at three different configurations is evaluated. These configurations are different in the fuel processing method to prevent carbon deposition on the anode catalyst. The fuel processing methods considered in these configurations are the anode gas recirculation (AGR), steam reforming (SR), and partial oxidation reformer (POX) methods. The application of this system is studied for operation in a wastewater treatment plant (WWTP) and in single-family detached dwellings. The evaluation of this system for operation in a WWTP indicates that if the entire biogas produced in the WWTP is used in the system with AGR or SR fuel processors, the electric power and heat required to operate the plant can be completely supplied and the extra electric power generated can be sold to the electrical grid. The evaluation of this system for operation in single-family detached dwellings indicates that, depending on the size, location, and building type and design, this system with all configurations studied is suitable to provide the domestic hot water and electric power demands. The second SOFC system is a novel portable electric power generation system that operates with liquid ammonia fuel. Size, simplicity, and high electrical efficiency are the main advantages of this environmentally friendly system. Using a sensitivity analysis, the effects of the cell voltage at several fuel utilization ratios on the number of cells required for the SOFC stack, system efficiency and voltage, and excess air required for thermal management of the SOFC stack are studied.