Дисертації з теми "PARALLEL COUPLED"

This thesis presents a new actuator system consisting of a micro-actuator and a macro-actuator coupled in parallel via a compliant transmission. The system is called the Parallel Coupled Micro-Macro Actuator, or PaCMMA. In this system, the micro-actuator is capable of high bandwidth force control due to its low mass and direct-drive connection to the output shaft. The compliant transmission of the macro-actuator reduces the impedance (stiffness) at the output shaft and increases the dynamic range of force. Performance improvement over single actuator systems was expected in force control, impedance control, force distortion and reduction of transient impact forces. A set of quantitative measures is proposed and the actuator system is evaluated against them: Force Control Bandwidth, Position Bandwidth, Dynamic Range, Impact Force, Impedance ("Backdriveability'"), Force Distortion and Force Performance Space. Several theoretical performance limits are derived from the saturation limits of the system. A control law is proposed and control system performance is compared to the theoretical limits. A prototype testbed was built using permanenent magnet motors and an experimental comparison was performed between this actuator concept and two single actuator systems. The following performance was observed: Force bandwidth of 56Hz, Torque Dynamic Range of 800:1, Peak Torque of 1040mNm, Minimum Torque of 1.3mNm. Peak Impact Force was reduced by an order of magnitude. Distortion at small amplitudes was reduced substantially. Backdriven impedance was reduced by 2-3 orders of magnitude. This actuator system shows promise for manipulator design as well as psychophysical tests of human performance.

Dans le contexte du calcul scientique, l'équilibrage de la charge est un problème crucial qui conditionne la performance des simulations numériques parallèles. L'objectif est de répartir la charge de travail entre un nombre de processeurs donné, afin de minimiser le temps global d'exécution. Une stratégie populaire pour résoudre ce problème consiste à modéliser la simulation à l'aide d'un graphe et à appliquer des algorithmes de partitionnement. En outre, les simulations numériques tendent à se complexifier, notamment en mixant plusieurs codes représentant des physiques différentes ou des échelles différentes. On parle alors de couplage de codes multi-physiques ou multi-échelles. Dans ce contexte, le problème de l'équilibrage de charge devient également plus difficile, car il ne s'agit plus d'équilibrer chacun des codes séparément, mais l'ensemble de ces codes pris dans leur globalité. Dans ce travail, on propose de resoudre ce problème en utilisant le modèle de partitionnement à sommets fixes qui pourrait représenter efficacement les contraintes supplémentaires imposées par les codes couplés (co-partitionnement). Nous avons donc développé un algorithme direct de partitionnement de graphe qui gère des sommets fixes. L'algorithme a été implémenté dans le partitionneur Scotch et une série d'expériences ont été menées sur la collection des graphes DIMACS. Ensuite nous avons proposé trois algorithmes de co-partitionnement qui respectent les contraintes issues des codes couplés respectifs. Nous avons egalement validé nos algorithmes par une étude expérimentale en comparant nos méthodes aux strategies actuelles sur des cas artificiels ainsi que sur des codes réels couplés
Load balancing is an important step conditioning the performance of parallel applications. The goal is to distribute roughly equal amounts of computational load across a number of processors, while minimising interprocessor communication. A common approach to model the problem is based on graph structures and graph partitioning algorithms. Moreover, new challenges involve the simulation of more complex physical phenomena, where different parts of the computational domain exhibit different physical behavior. Such simulations follow the paradigm of multi-physics or multi-scale modeling approaches. Combining such different models in massively parallel computations is still a challenge to reach high performance. Additionally, traditional load balancing algorithms are often inadequate, and more sophisticated solutions should be explored. In this thesis, we propose new graph partitioning algorithms that balance the load of such simulations, refered to as co-partitioning. We formulate this problem with the use of graph partitioning with initially fixed vertices which we believe represents efficiently the additional constraints of coupled simulations. We have therefore developed a direct algorithm for graph partitioning that manages successfully problems with fixed vertices. The algorithm is implemented inside Scotch partitioner and a series of experiments were carried out on the DIMACS graph collection. Moreover we proposed three copartitioning algorithms that respect the constraints of the respective coupled codes. We finally validated our algorithms by an experimental study comparing our methods with current strategies on artificial cases and on real-life coupled simulations

Thesis (PhD)--Stellenbosch University, 2014.
ENGLISH ABSTRACT: We investigate a model system for particle diffusion in elastically coupled one-dimensional narrow channels. The elastic coupling of the channels is such that channels mutually affect the stochastic dynamics of particles. This kind of constrained and coupled stochastic diffusion may occur in supramolecular lattices where pore occupancy by guest particles may induce a reversible mechanical deformation of the lattice hence, affecting particle evolution in neighbouring pores. The model is explored first for out-of-equilibrium conditions, where we look mainly at the kinetic properties of the system, and thereafter under equilibrium conditions, where we try to understand the nature of dynamic correlation within the coupled channel system. For an out-of-equilibrium version of the model the focus is placed on the steady state behaviour of the two elastically coupled finite channels. The channels are kept in contact with particle reservoirs at the boundaries. Three current-density regimes of different distinct behaviour are identified using a simulation experiment. The sensitivity of the system mean occupancy profile and the steady state particle flux to small and large coupling parameter strength are explored. We find that, for small coupling strength, the system steady state profile and flux behaviour can be approximated by a simple mean field theory ignoring density-density correlations. We present the analytic description of the system using a cellular automaton formalism and then we generalize the theory for a multi-coupled channel system using a hopping particle dynamics approach. For small coupling parameter values, the analytic results are confirmed by the stochastic simulation. From the equilibrium perspective, we model the elastically coupled channel system as a system of infinite narrow channels having a uniform guest particle occupancy and we calculate density fluctuation correlation functions. The elastic coupling between channels is modelled as short range interacting potential and the particle evolution is modelled through Langevin dynamics. The dynamics are cast into the functional integral formalism expressed in terms of the collective particle number density, current density and the associated density response fields. The resulting generating functional takes these fields into consideration within the random phase approximation (RPA) up to second order. For a short range interaction potential, we uncover the behaviour of the system by looking at the influence of the inter-channel interaction strength on the dynamic density-density correlation functions. We conclude that the system long time limit effective friction coefficient is reduced with increase in the coupling parameter values while the strength of thermal forces for the effective system becomes renormalized. We also find out that the RPA breaks down under certain conditions, signalling a transition to a behaviour that is no longer characterised by a homogeneous density. The work presented here provides the beginnings for microscopic insights into the filling, filtering and storage processes for which certain types of microporous materials can be utilised.
AFRIKAANSE OPSOMMING: Ons ondersoek ’n modelstelsel vir die diffusie van deeltjies in elasties gekoppelde, eendimensionele nou kanale. Die elastiese koppeling is sodanig dat die stogastiese dinamika wedersyds be¨ınvloed word. Hierdie gekoppelde en aan dwangvoorwaardes onderhewig diffusie kan in supermolekulˆere roosters gebeur waar die besetting van holtes deur deeltjies ’n omkeerbare meganiese vervorming van die kristalrooster kan veroorsaak en sodoende die tydontwikkeling in ’n aangrensende porie be¨ınvloed. Die model word eers vir nie-ewewig toestande ondersoek, waar ons hoofsaaklik die kinetiese eienskappe van die stelsel beskou, en daarna word dit vir ewewig ondersoek, waar ons die aard van die dinamiese korrelasie binne die gekoppelde kanaalstelsel probeer verstaan. Vir die nie-ekwilibrium weergawe van die model word die fokus op die gedrag van twee gekoppelde en eindige kanale se bestendige toestand gerig. Die porie¨e bly aan hulle eindpunte in kontak met reservoirs van deeltjies. Daar word drie deeltjiestroom-digtheid gebiede ge¨ıdentifiseer met behulp van ’n simulasieeksperiment. Die sensitiewiteit van die stelsel se gemiddelde deeltjiebesettingsprofiel en die deeltjievloed in ’n bestendige toestand is ondersoek vir groot en klein koppelingsparameters. Vir klein koppelingsterkte vind ons dat die stelsel se bestendige toestand deeltjiebesettingspofiel en deeltjievloed deur ’n eenvoudige gemiddelde-veld teorie beskryf kan word, waar digtheid-degtheids korrelasies verontagsaam kan word. Ons bied die analitiese beskrywing aan vir die gekoppelde stelsel deur van ’n sellulˆere outomaat-formalisme gebruik te maak, en om dan die teorie te veralgemeen vir ’n stelsel bestaande uit vele aanmekaar gekoppelde porie¨e ,deur gebruik te maak van ’n formalisme waarin deeltjies tussen holtes hop. Die analtiese resultate word vir kleine waardes van die koppelingsparameter deur ’n rekenaarsimulasie bevestig. Vir die ekwilibrium gesigspunt modelleer ons die gekoppelde stelsel van kanale as oneindig nou, met ’n homogene verdeling van die deeeltjiebesetting en ons bereken digtheids-korrelasiefunksies. Die elastiese vebinding tussen porie¨e word deur ’n kortrykwydte potensiaal gemodelleer en die deeltjies se dinamika met behulp van Langevindinamika. Die dinamika word met behulp van die funksionaal integraalformalisme uitgedruk in terme van die deeltjiegetaldigtheid-, stroomdigtheid- en meegaande antwoordvelde. Die resulterende genererende funksionaal neem hierdie hoeveelhede tot tweede order binne die “random field approximation” (RPA). Vir ’n potensiaal met kort rykwydte, ondersoek ons die stelsel se gedrag deur die dinamiese digtheid-digtheidskorrelasie te ondersoek. Ons lei af dat die stelsel se langtyd effektiewe wrywingsko¨effisi¨ent afneem met die toename in die koppelingsparameter se waardes terwyl die sterkte van die termiese kragte vir die effektiewe stelsel renormeer word. Ons vind ook dat die RPA onder sekere omstandighede sy geldigheid verloor, wat ’n oorgang kan beteken wat nie meer deur ’n homogene digtheid beskryf kan word nie. Die werk wat hier aangebied word dui die eerste stappe aan vir hoe mikroskopiese insigte vir vul-, filter- en stoorprosesse vir sekere tipes mikroporeuse materiale gebruik kan word.

As we move steadily through the multicore era, and the number of processing cores on each chip continues to rise, parallel computation becomes increasingly important. However, parallelising an application is often difficult because of dependencies between different regions of code which require cores to communicate. Communication is usually slow compared to computation, and so restricts the opportunities for profitable parallelisation. In this work, I explore the opportunities provided when communication between cores has a very low latency and low energy cost. I observe that there are many different ways in which multiple cores can be used to execute a program, allowing more parallelism to be exploited in more situations, and also providing energy savings in some cases. Individual cores can be made very simple and efficient because they do not need to exploit parallelism internally. The communication patterns between cores can be updated frequently to reflect the parallelism available at the time, allowing better utilisation than specialised hardware which is used infrequently. In this dissertation I introduce Loki: a homogeneous, tiled architecture made up of many simple, tightly-coupled cores. I demonstrate the benefits in both performance and energy consumption which can be achieved with this arrangement and observe that it is also likely to have lower design and validation costs and be easier to optimise. I then determine exactly where the performance bottlenecks of the design are, and where the energy is consumed, and look into some more-advanced optimisations which can make parallelism even more profitable.

Particle accelerators are employed in many fields of the physics, engineering and medicine science. Although the idea of acceleration of charged particles by means of time varying electromagnetic field is an almost century old, the research and development in this field is active more than ever. An efficient particle accelerator gives the opportunity to reach energy levels not yet achieved and allows the study of the deepest laws of the physics. Also a better performance of the particle accelerators is desirable in order to develop compact table-top machine for medical or industrial purpose. The work here presented takes place in this framework. A parallel coupled electron LINAC has been studied for high gradient operation in the X-band frequency. This kind of device can address some issues of high gradient accelerating LINAC. Due to decoupled propagation axes for RF power and beam it is possible to increase the structure shunt impedance without affecting the cell to cell coupling, combining the advantages of travelling and standing wave structures. Besides in case of a breakdown only the energy stored in one cell is involved in the discharge process and not the total energy stored of the section, thus the damages caused by the breakdown event are expected to be smaller. Following a discussion on how the proposed configuration was conceived, full wave simulations were carried out to numerically validate the idea and to optimize the single cell in order to minimize the probability of breakdowns. After a successful cold simulation a beam dynamics analysis has been done in order to estimate the properties of the beam moving inside the device. Due to the small irises dimension the transverse wakefield can be source of instability for the bunch propagation. Therefore this issue has been investigated by means of numerical methods. Finally, the problem of the beam alignment with respect to the accelerating section is addressed and the results of a device suitable to beam position measurement are presented.

High skirt selectivity and extended out-of-band rejection is a major challenge for the successful progress of in-line microwave filters. This thesis presents novel filter realizations with improved performance, compatible with the standard single thin all-metal insert in a split-block housing and therefore maintaining the low-cost fabrication characteristics. In addition, significant filter performance improvement is achieved. The synthesis procedure implemented for the filter concept consists of a few steps. Some preliminary steps are a rigorous characterization of a double-ridge coaxial waveguide, and the modelling of an equivalent circuit model for the parallel coupled ridge waveguide devised in the filter concept. From these elements, a full wave electromagnetic analysis shows that parallel-coupled asymmetric ridge waveguides produce strongly dispersive coupling which introduces a transmission zero. Later on this property is extended to parallel-coupled asymmetric ridge waveguide resonators, where it is demonstrated that it is possible to independently control the coupling coefficient and the frequency of the transmission zero. This allows the realization of pseudo-elliptic narrowband in-line bandpass filters in E-plane technology. A general synthesis procedure for high order filters is outlined and numerical and experimental results are presented for validation. The elements employed for the synthesis procedure of the bandpass prototypes are also applied to investigate structures suitable for different applications. In particular, stopband and dual stopband filters are presented with numerical and experimental results. Finally, the study of a microwave chemical/biochemical sensing device for the characterization and detection of cells in chemical substances and cells in solution in micro-litre volumes is also reported.

Les creixents demandes de mercat són l'esperó per a les millores tecnològiques i la innovació en el sector de les telecomunicacions. La recent arribada de la cinquena generació de mòbil 5G promet millors serveis de comunicació, amb velocitats de dades 20 vegades més ràpides que la 4G, proporcionant un nivell de connectivitat mai abans vist. El compliment dels objectius de 5G estan suposant un gran desafiament per a cada component dels dispositius i xarxes de telefonia mòbil. Els serveis de la nova generació necessiten més ample de banda per uns dispositius sense fils que han de coexistir en un entorn complex, en què hi haurà un major nombre de components, amb millor rendiment i menor grandària. Aquesta és la realitat dels sistemes sense fil, especialment, els sistemes de filtrat. La tecnologia més adequada per als filtres de comunicacions mòbils és la basada en resonares acústics. Les estrictes limitacions tecnològics estan empenyent a aquesta tecnologia a assolir noves cotes de fiabilitat, aconseguir un major rendiment i millorar la supressió d'interferències en un entorn amb un espectre radioelèctric saturat. A causa de la complexitat de el procés de disseny de filtres, una part crucial és confiada a eines d'optimització. No obstant això, és una tasca que necessita una quantitat de temps considerable, i pot perllongar massa el procés de disseny. El principal objectiu d'aquesta tesi és oferir un conjunt d'eines matemàtiques per proporcionar la topologia necessària directament, i també per obtenir el millor disseny inicial, un punt de partida que faciliti a l'programari d'optimització la recerca d'una solució adequada. La topologia de filtres més comú és la ladder. Avui en dia, les metodologies de síntesi permeten definir les característiques fonamentals d'un filtre amb un cost computacional molt baix. Per obtenir dissenys que compleixin amb els requisits tecnològics i de rendiment, sovint cal modificar l'estructura clàssica de l'ladder afegint elements reactius o creant enllaços interns que un filtre ladder no contempla, i que cap mètode de síntesi és capaç de proporcionar. Aquests acoblaments també podrien aparèixer a causa de defectes de fabricació, provocant un greu deteriorament de la resposta de l'filtre. L'objectiu d'aquest treball és desenvolupar tècniques de síntesi avançades amb què crear topologies que incloguin acoblaments creuats i elements externs per crear configuracions de filtres no convencionals més útils. Aprofitant aquestes metodologies, s'han proposat noves topologies per fer front a les limitacions tecnològiques. Un altre aspecte de el procés de disseny de filtres és el control de la fase. Amb ell, és possible modificar els valors dels elements de la xarxa, adaptar filtres per a ser capaços de treballar en configuracions de multiplexors, i assegurar la seva viabilitat després de la desnormalització pas banda. A causa de la gran sensibilitat a qualsevol variació de la fase introduïda sobre el filtre original, tal control és crucial per a tecnologies acústiques. Aquest document descriu un mètode per controlar de manera efectiva la fase de la xarxa per tal d'aconseguir adequadament els objectius marcats per a un filtre determinat.
Las crecientes demandas del mercado son el acicate para las mejoras tecnológicas y la innovación en el sector de las telecomunicaciones. La reciente llegada de la quinta generación de móvil 5G promete mejores servicios de comunicación, con velocidades de datos 20 veces más rápidas que la 4G, proporcionando un nivel de conectividad nunca antes visto. El cumplimiento de los objetivos de 5G están suponiendo un gran desafío para cada componente de los dispositivos y redes de telefonía móvil. Los servicios de la nueva generación necesitan más ancho de banda para unos dispositivos inalámbricos que han de coexistir en un entorno complejo, en el que habrá un mayor número de componentes, con mejor rendimiento y menor tamaño. Esa es la realidad de los sistemas inalámbricos, especialmente, los sistemas de filtrado. La tecnología más adecuada para los filtros de comunicaciones móviles es la basada en resonares acústicos. Las estrictas limitaciones tecnológicos están empujando a esta tecnología a alcanzar nuevas cotas de fiabilidad, lograr un mayor rendimiento y mejorar la supresión de interferencias en un entorno con un espectro radioeléctrico saturado. Debido a la complejidad del proceso de diseño de filtros, una parte crucial es confiada a herramientas de optimización. Sin embargo, es una tarea que necesita una cantidad de tiempo considerable, y puede prolongar demasiado el proceso de diseño. El principal objetivo de esta tesis es ofrecer un conjunto de herramientas matemáticas para proporcionar la topología necesaria directamente, y también para obtener el mejor diseño inicial, un punto de partida que facilite al software de optimización la búsqueda de una solución adecuada. La topología de filtros más común es la ladder. Hoy en día, las metodologías de síntesis permiten definir las características fundamentales de un filtro con un coste computacional muy bajo. Para obtener diseños que cumplan con los requisitos tecnológicos y de rendimiento, a menudo es necesario modificar la estructura clásica del ladder agregando elementos reactivos o creando acoplamientos internos que un filtro ladder no contempla, y que ningún método de síntesis es capaz de proporcionar. Dichos acoplamientos también podrían aparecer debido a defectos de fabricación, provocando un grave deterioro de la respuesta del filtro. El objetivo de este trabajo es desarrollar técnicas de síntesis avanzadas con las que crear topologías que incluyan acoplamientos cruzados y elementos externos para crear configuraciones de filtros no convencionales más útiles. Aprovechando estas metodologías, se han propuesto nuevas topologías para hacer frente a las limitaciones tecnológicas. Otro aspecto del proceso de diseño de filtros es el control de la fase. Con él, es posible modificar los valores de los elementos de la red, adaptar filtros para ser capaces de trabajar en configuraciones de multiplexores, y asegurar su viabilidad después de la desnormalización paso banda. Debido a la gran sensibilidad a cualquier variación de la fase introducida sobre el filtro original, tal control es crucial para tecnologías acústicas. Este documento describe un método para controlar de forma efectiva la fase de la red a fin de lograr adecuadamente los objetivos marcados para un filtro determinado.
The increasing demands of the market are the stimulus that spurs on technological improvement and innovation in the telecommunication sector. The recent arrival of 5G wireless technology promises better communication services with data rates 20 times faster than 4G, providing a level of connectivity never seen before. The fulfillment of 5G goals is imposing a great challenge on every building block of mobile handsets and cellular networks. The services in the new generation require more bandwidth, and the wireless devices will coexist in a complex environment that includes more components with better performance and smaller size. That is a reality, especially for filtering systems. The best-suited technology for resonators in mobile communications filters is the acoustic wave. The stringent technological constraints are pushing this technology to be more reliable, achieve better performance, and improve higher interference mitigation in an already overcrowded spectrum. Due to the complexity of the filter design process, a vital part of it is entrusted to optimization software tools. However, It is a time-consuming task that may prolong the process for too long before obtaining the final network. The main objective of this thesis is to offer mathematical tools to provide the required topology straightforwardly and also to obtain the best initial design, a starting point that facilitates the optimization software to find the filtering solution easily. The most common filter topology is the ladder network. To the date, synthesis methodologies allow defining their fundamental characteristics during the design process with a very low computational cost. To obtain filter responses that fulfill technological and performance design, it is often necessary to modify the classic ladder structure by adding reactive elements or creating couplings between some blocks of the filter that a ladder structure does not contemplate, and no synthesis method can provide. In other circumstances, those couplings may appear due to manufacturing defects, causing severe deterioration of the filter response. The aim of this work is to develop advanced synthesis techniques to generate topologies that include cross-couplings and external elements to address the most useful non-standard filter network configurations. Leveraged on these methodologies, new topologies are proposed to cope with technological limitations. Another aspect of the filter design process is the phase control network. With it, it is possible to modify the network element values, adapt standalone filters for multiplexer configurations, and assure the feasibility of the filter after bandpass denormalization. Such control is crucial for acoustic wave technology, which is very sensitive to any phase variations introduced with respect to the originally designed. This document describes an effective method to control the network phase to achieve the mentioned goals adequately.

After the review of single-phase converter topologies and detailed examination of several candidates, a non-resonantly coupled parallel resonant DCIDC converter is identified as being the most appropriate high frequency linked topology. The steady-state characteristics of the chosen converter with both variable frequency and phase shift control have been examined using state-plane analysis. These theoretical predictions are confirmed with time-domain simulation and comprehensive measurements from a 3.0 kVA prototype. The power losses resulted from real devices and components account for the discrepancies. Hybrid control has been employed for stand-alone applications. By contrast, phase-shift control only at a reduced switching frequency provides better output performance and higher conversion efficiency. Under light loads, a look-up correction table has been developed to improve the output performance further. The prototype efficiency is measured to be 94% under full load. Integral compensator and phase shift control only have been utilized for grid-connected applications. A damping resistor has been added to damp the resonant oscillation of the output filter. Similarly, a correction table is introduced under light load. An additional compensator, increased DC gains and phase margin compensation have been introduced to improve the output performance further. The impact of the grid-connected converter on the grid voltage is also examined.

Dans l'industrie aéronautique, la génération d'énergie dépend presque exclusivement de la combustion d'hydrocarbures. La meilleure façon d'améliorer le rendement de ces systèmes et de contrôler leur impact environnemental, est d'optimiser le processus de combustion. Avec la croissance continue du de la puissance des calculateurs, la simulation des systèmes complexes est devenue abordable. Jusqu'à très récemment dans les applications industrielles le rayonnement des gaz et la conduction de chaleur dans les solides ont été négligés. Dans ce travail les outils nécessaires à la résolution couplée des trois modes de transfert de chaleur ont été développés et ont été utilisés pour l'étude d'une chambre de combustion d'hélicoptère. On montre que l'inclusion de tous les modes de transfert de chaleur peut influencer la distribution de température dans le domaine. Les outils numériques et la méthodologie de couplage développés ouvrent maintenant la voie à un bon nombre d'applications tant scientifiques que technologiques
In the aeronautical industry, energy generation relies almost exclusively in the combustion of hydrocarbons. The best way to improve the efficiency of such systems, while controlling their environmental impact, is to optimize the combustion process. With the continuous rise of computational power, simulations of complex combustion systems have become feasible, but until recently in industrial applications radiation and heat conduction were neglected. In the present work the numerical tools necessary for the coupled resolution of the three heat transfer modes have been developed and applied to the study of an helicopter combustion chamber. It is shown that the inclusion of all heat transfer modes can influence the temperature repartition in the domain. The numerical tools and the coupling methodology developed are now opening the way to a good number of scientific and engineering applications

Made available in DSpace on 2014-12-17T14:56:13Z (GMT). No. of bitstreams: 1 KayoGS_DISSERT.pdf: 4975392 bytes, checksum: 5d113169a6356e5e7704aec116237caf (MD5) Previous issue date: 2013-03-25
Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior
This paper analyzes the performance of a parallel implementation of Coupled Simulated Annealing (CSA) for the unconstrained optimization of continuous variables problems. Parallel processing is an efficient form of information processing with emphasis on exploration of simultaneous events in the execution of software. It arises primarily due to high computational performance demands, and the difficulty in increasing the speed of a single processing core. Despite multicore processors being easily found nowadays, several algorithms are not yet suitable for running on parallel architectures. The algorithm is characterized by a group of Simulated Annealing (SA) optimizers working together on refining the solution. Each SA optimizer runs on a single thread executed by different processors. In the analysis of parallel performance and scalability, these metrics were investigated: the execution time; the speedup of the algorithm with respect to increasing the number of processors; and the efficient use of processing elements with respect to the increasing size of the treated problem. Furthermore, the quality of the final solution was verified. For the study, this paper proposes a parallel version of CSA and its equivalent serial version. Both algorithms were analysed on 14 benchmark functions. For each of these functions, the CSA is evaluated using 2-24 optimizers. The results obtained are shown and discussed observing the analysis of the metrics. The conclusions of the paper characterize the CSA as a good parallel algorithm, both in the quality of the solutions and the parallel scalability and parallel efficiency
O presente trabalho analisa o desempenho paralelo de uma implementa??o do Simulated Annealing Acoplado (CSA, do ingl?s Coupled Simulated Annealing) para otimiza??o de vari?veis cont?nuas sem restri??es. O processamento paralelo ? uma forma eficiente de processamento de informa??o com ?nfase na explora??o de eventos simult?neos na execu??o de um software. Ele surge principalmente devido ?s elevadas exig?ncias de desempenho computacional e ? dificuldade em aumentar a velocidade de um ?nico n?cleo de processamento. Apesar das CPUs multiprocessadas, ou processadores multicore, serem facilmente encontrados atualmente, diversos algoritmos ainda n?o s?o adequados para executar em arquiteturas paralelas. O algoritmo do CSA ? caracterizado por um grupo de otimizadores Simulated Annealing (SA) trabalhando em conjunto no refinamento da solu??o. Cada otimizador SA ? executado em uma ?nica thread, e essas executadas por diferentes processadores. Na an?lise de desempenho e escalabilidade paralela, as m?tricas investigadas foram: o tempo de execu??o; o speedup do algoritmo com respeito ao aumento do n?mero de processadores; e a efici?ncia na utiliza??o de elementos de processamento com rela??o ao aumento da inst?ncia do problema tratado. Al?m disso, foi verificada a qualidade da solu??o final. Para o estudo, esse trabalho analisa uma vers?o paralela do CSA e sua vers?o serial equivalente. Ambos algoritmos foram analisados sobre 14 fun??es de refer?ncia. Para cada uma dessas fun??es, o CSA ? avaliado utilizando de 2 a 24 otimizadores. Os resultados obtidos s?o exibidos e comentados observando-se as m?tricas de an?lise. As conclus?es do trabalho caracterizam o CSA como um bom algoritmo paralelo, seja na qualidade das solu??es como na escalabilidade e efici?ncia paralela

This work details the development of a computational tool that can accurately model strongly-coupled fluid-structure-interaction (FSI) problems, with a particular focus on thin-walled structures undergoing large, geometrically non-linear deformations, which has a major interest in, amongst others, the aerospace and biomedical industries. The first part of this work investigates improving the efficiency with which a stable and robust in-house code, Elemental, models thin structures undergoing dynamic fluid-induced bending deformations. Variations of the existing finite volume formulation as well as linear and higher-order finite element formulations are implemented. The governing equations for the solid domain are formulated in a total Lagrangian or undeformed conguration and large geometrically non-linear deformations are accounted for. The set of equations is solved via a single-step Jacobi iterative scheme which is implemented such as to ensure a matrix-free and robust solution. Second-order accurate temporal discretisation is achieved via dual-timestepping, with both consistent and lumped mass matrices and with a Jacobi pseudo-time iteration method employed for solution purposes. The matrix-free approach makes the scheme particularly well-suited for distributed memory parallel hardware architectures. Three key outcomes, not well documented in literature, are highlighted: the issue of shear locking or sensitivity to element aspect ratio, which is a common problem with the linear Q4 finite element formulation when subjected to bending, is evaluated on the finite volume formulations; a rigorous comparison of finite element vs. finite volume methods on geometrically non-linear structures is done; a higher-order finite volume solid mechanics procedure is developed and evaluated. The second part of this work is concerned with fluid-structure interaction (FSI) modelling. It considers the implementation and coupling of a higher order finite element structural solver with the existing finite volume fluid-flow solver in Elemental. To the author’s knowledge, this is the first instance in which a strongly-coupled hybrid finite element–finite volume FSI formulation is developed. The coupling between the fluid and structural components with non-matching nodes is rigorously assessed. A new partitioned fluid-solid interface coupling methodology is also developed, which ensures stable partitioned solution for strongly-coupled problems without any additional computational overhead. The solver is parallelised for distributed memory parallel hardware architectures. The developed technology is successfully validated through rigorous temporal and mesh independent studies of representative two-dimensional strongly-coupled large-displacement FSI test problems for which analytical or benchmark solutions exist.
Dissertation (MEng)--University of Pretoria, 2012.
Mechanical and Aeronautical Engineering
unrestricted

Accurate correlated electronic structure methods involve a significant amount of computations and can be only employed to small molecular systems. For example, the coupled-cluster singles, doubles, and perturbative triples model (CCSD(T)), which is known as the ``gold standard" of quantum chemistry for its accuracy, usually can treat molecules with 20-30 atoms. To extend the reach of accurate correlated electronic structure methods to larger molecular systems, we work towards two directions: parallel computing and reduced-cost/scaling approaches. Parallel computing can utilize more computational resources to handle systems that demand more substantial computational efforts. Reduced-cost/scaling approaches, which introduce approximations to the existing electronic structure methods, can significantly reduce the amount of computation and storage requirements. In this work, we introduce a new distributed-memory massively parallel implementation of standard and explicitly correlated (F12) coupled-cluster singles and doubles (CCSD) with canonical bigO{N^6} computational complexity ( C. Peng, J. A. Calvin, F. Pavov{s}evi'c, J. Zhang, and E. F. Valeev, textit{J. Phys. Chem. A} 2016, textbf{120}, 10231.), based on the TiledArray tensor framework. Excellent strong scaling is demonstrated on a multi-core shared-memory computer, a commodity distributed-memory computer, and a national-scale supercomputer. We also present a distributed-memory implementation of the density-fitting (DF) based CCSD(T) method. (C. Peng, J. A. Calvin, and E. F. Valeev, textit{in preparation for submission} ) An improved parallel DF-CCSD is presented utilizing lazy evaluation for tensors with more than two unoccupied indices, which makes the DF-CCSD storage requirements always smaller than those of the non-iterative triples correction (T). Excellent strong scaling is observed on both shared-memory and distributed-memory computers equipped with conventional Intel Xeon processors and the Intel Xeon Phi (Knights Landing) processors. With the new implementation, the CCSD(T) energies can be evaluated for systems containing 200 electrons and 1000 basis functions in a few days using a small size commodity cluster, with even more massive computations possible on leadership-class computing resources. The inclusion of F12 correction to the CCSD(T) method makes it converge to basis set limit much more rapidly. The large-scale parallel explicitly correlated coupled-cluster program makes the accurate estimation of the coupled-cluster basis set limit for molecules with 20 or more atoms a routine. Thus, it can be used rigorously to test the emerging reduced-scaling coupled-cluster approaches. Moreover, we extend the pair natural orbital (PNO) approach to excited states through the equation-of-motion coupled cluster singles and doubles (EOM-CCSD) method. (C. Peng, M. C. Clement, and E. F. Valeev, textit{submitted}) We simulate the PNO-EOM-CCSD method using an existing massively parallel canonical EOM-CCSD program. We propose the use of state-averaged PNOs, which are generated from the average of the pair density of excited states, to span the PNO space of all the excited states. The doubles amplitudes in the CIS(D) method are used to compute the state-averaged pair density of excited states. The issue of incorrect states in the state-averaged pair density, caused by an energy reordering of excited states between the CIS(D) and EOM-CCSD, is resolved by simply computing more states than desired. We find that with a truncation threshold of $10^{-7}$, the truncation error for the excitation energy is already below 0.02 eV for the systems tested, while the average number of PNOs is reduced to 50-70 per pair. The accuracy of the PNO-EOM-CCSD method on local, Rydberg and charge transfer states is also investigated.
Ph. D.

In questa tesi sono stati apportati due importanti contributi nel campo degli acceleratori embedded many-core. Abbiamo implementato un runtime OpenMP ottimizzato per la gestione del tasking model per sistemi a processori strettamente accoppiati in cluster e poi interconnessi attraverso una network on chip. Ci siamo focalizzati sulla loro scalabilità e sul supporto di task di granularità fine, come è tipico nelle applicazioni embedded. Il secondo contributo di questa tesi è stata proporre una estensione del runtime di OpenMP che cerca di prevedere la manifestazione di errori dati da fenomeni di variability tramite una schedulazione efficiente del carico di lavoro.

A brief analysis of the nonresonant-coupled parallel resonant converter is performed. The converter is modeled and a reference classical analog controller is designed and simulated. Infrastructure required for digital control of the converter (including anti-aliasing filters and a modulator) is designed and a classical digital controller is designed and simulated, yielding a ~30% degradation in control bandwidth at the worst-case operating point as compared with the analog controller. Based on the strong relationship observed between low-frequency converter gain and operating point, a gain-scheduled digital controller is proposed, designed, and simulated, showing 4:1 improved worst-case control bandwidth as compared with the analog controller. A complete prototype is designed and built which experimentally validates the results of the gain-scheduled controller simulation with good correlation. The three approaches that were investigated are compared and conclusions are drawn. Suggestions for further research are presented.
Master of Science

Thesis (M. S.)--Computing, Georgia Institute of Technology, 2009.
Committee Chair: Fujimoto, Richard; Committee Member: Bader, David; Committee Member: Perumalla, Kalyan; Committee Member: Riley, George; Committee Member: Vuduc, Richard.

En aéronautique, les contraintes sont telles que la masse des filtres peut représenterjusqu'à 50% de la masse totale du convertisseur. Ces dernières années, les convertisseursmulticellulaires parallèles entrelacés et magnétiquement couplés ont conduit à améliorer lesperformances des convertisseurs (densité de puissance, efficacité, dynamique,...). Denombreuses topologies de filtrages entrelacés existent, l'objectif principal de cette étude est detrouver parmi ces topologies celles qui sont les mieux adaptées à la réalisation d'un onduleurde tension 110Veff / 400Hz triphasé 25kVA. Il est démontré que le choix du type de matériaumagnétique a un impact déterminant sur le poids, le volume et les pertes du convertisseur. Quidit parallélisation, dit multiplication du nombre de semi-conducteurs. Ces nouvelles structuresdoivent garantir à la fois un rendement élevé, une masse faible et une continuté defonctionnement, même en cas de panne d'un semiconducteur de puissance ou de sacommande. Mais coupler les phases entre elles, impose un lien indissociable qui peut êtrenuisible au fonctionnement de la structure en cas de dysfonctionnement. Des solutionsexistent et sont abordées dans l'étude. Enfin, la réalisation pratique d'un prototype semiindustrielde convertisseur triphasé utilisant des transformateurs interphases est présentée. Ils'agit d'un onduleur réseau avionique triphasé avec reconstruction de neutre pour fonctionneren déséquilibré. Les résultats expérimentaux démontrent l'avantage d'un convertisseur

Le but de ce travail était de développer des filtres passe-bande RF dans la technologie de PCB, avec trois objectifs principaux. Le premier objectif était de développer des formules de synthèse tosimplify la procédure de conception du filtre. Le deuxième était de parvenir à un rejet wideout bande sans modifier les caractéristiques de la bande de filtrage. Le troisième objectif est de contrôler la fréquence centrale du filtre en utilisant diode varicap
The purpose of this work was to develop RF bandpass filters in PCB technology,with three main objectives. The first objective was to develop synthesis formulas tosimplify the design procedure of the filter. The second was to achieve wideout-of-band rejection without modifying the in-band filtering characteristics. Thethird objective was to control the center frequency of the filter by using varactordiode.The bandpass filter topology treated in this thesis is based on Stub-LoadedResonators (SLR). The main features of this filter topology were treated. Equivalentcircuits based on J-inverters and susceptance parameters were derived. Based onthese equivalent circuits, synthesis formulas were developed. Simulations werepresented to validate the synthesis theory. For a proof-of-concept, third orderstripline bandpass filters were designed and fabricated based on this synthesis.Analysis technique using odd- and even- mode was achieved on the SLR. Thusresonance odd- and even-mode conditions were derived. These conditions aim toeasily control the first spurious frequency. Moreover, to go further in improving theout-of-band rejection a new technique, called “U corner structure”, was developedand design rules were derived. Based on these design rules an extended out-of-bandrejection was achieved without any modification in the passband and by maintainingthe compactness of the filter. A first spurious frequency was localized at up to ninetimes the working frequency in the case of the Parallel-coupled Stub-Loadedresonator (PC-SLR) filter. Also, by applying this technique into the classicalparallel-coupled filter the first and second spurious frequencies were rejected. Toaddress the issue of tunable filters, the SLRs were correctly loaded by variablecapacitors (varactor diode). The center frequency of the PC-SLR filter was easilycontrolled by maintaining a large out-of-band rejection

碩士
國立交通大學
電信工程研究所
104
In this thesis, a stepped-impedance coupled line (SICLN) structure is proposed to design the parallel coupled line filters. The proposed SICLN can be equivalent to a conventional quarter-wave coupled line section at the center frequency. The proposed SICLN half- and quarter-wave parallel-coupled filters are implemented by changing the corresponding conventional quarter-wave coupled-line sections in filters to the SICLNs. The proposed filters sizes are reduced while the first spurious responses can be pushed away over three times from the fundamental frequency. Therefore, the proposed filters show wide stopband and good stopband rejection.

碩士
國立臺灣大學
電信工程學研究所
94
Compact parallel-coupled bandpass filters with good selectivity and better stopband rejection are proposed in this study. They are designed with quarter-wavelength resonators such that there is no repeated passband at twice the center frequency, and the lengths of coupled-line sections can be reduced by using additional capacitors or inductors. In addition, by introducing the capacitive or inductive cross-coupling effect, transmission zeros can be achieved to improve the selectivity. The open stub which works as a K-inverter can also introduce a transmission zero at high frequency so as to suppress the spurious passband. Suitable equivalent models and design formulas are establish, from which the proposed filter can be design and investigate. Finally, compact 4th and 6th order bandpass filters with better selectivity and wider stopband up to four times the center frequency are accomplished. Comparing to the conventional parallel-coupled bandpass filter, the size reduction is over 60%.

博士
長庚大學
電子工程學系
105
This thesis is classified into two research topics regarding dual-band coupled-line couplers and rat-race couplers with various parallel coupled-line structures. The first topic is designs of dual-band couplers with wide separation between bands, in which four novel parallel coupled-line structures are presented. The first coupler (Coupler A) consists of three cascaded coupled-line sections with the advantage of high frequency ratio. The coupling levels of 10.6 and 10.8 dB, isolated levels of 22.3 and 16.8 dB, and phase differences of 90.5 and 89.6 are observed at 0.84 and 5.17 GHz, respectively. The second coupler (Coupler B) is composed of coupled lines with stubs tapped at their center, from which 11.3 and 11.1 dB couplings, 33.43 and 24.9 dB isolatings, and phase differences of 90.9 and 87.1 can be obtained at 0.7 and 2.6 GHz, respectively. Coupler B has advantages of a high band ratio, compact size, and easy fabrication. The third one (Coupler C) consists of the coupled-uncoupled-coupled lines with stubs tapped at the junctions. The proposed structure has the advantages of high band ratio and strong coupling because of the constructed repeat-coupled structure. An experimental Coupler C with operational frequencies of 0.7 and 2.6 GHz is fabricated and achieved coupling levels of 2.96 and 3.85 dB, isolated levels of 30.14 and 20.47 dB, and phase differences of 90.16 and 87.83, respectively. Coupler D is composed of uncoupled-coupled-uncoupled lines and has the advantages of high band ratio, strong coupling level, forward wave, and compactness. An experimental Coupler D is observed to obtain coupling levels of 3.46 and 3.3 dB, isolated levels of 25.77 and 16.8 dB, and phase differences of 91.01 and 86.75 at 0.7 and 2.55 GHz, respectively. The detailed design equations and parameter analysis are introduced in this thesis. The measured results are in good agreement with the simulated results. The second topic is focused on the dual-band transmission lines, which has the advantages of −90/+90 at two arbitrary frequencies and realizable high equivalent characteristic impedances. The proposed structure is composed of diagonal-shorted coupled lines with stubs tapped at their center. This structure can be applied to replace the conventional 270(3/4) line with dual-band operation, high equivalent characteristic impedances, and compactness. An experimental 2.4/5.1 GHz rat-race coupler with 1:4 power-split ratio is fabricated. The measured results show excellent dual-band performance and the coupling levels of −6.9 and −7.1 dB and isolated levels of 29.6 and 18.4 dB are observed at 2.4 and 5.1 GHz, respectively. The phase differences S21−S31 are −2.3 and 1.5 and S24−S34 are 177.2 and −178.7 at 2.4 and 5.1 GHz, respectively.

碩士
國立臺灣科技大學
電子工程系
89
Recently, because of the booming of global communication market, RF circuits have been applied in many of the communication systems such as IEEE 802.11G, Bluetooth, and GSM. Furthermore, RF circuits and some system components like RF Amplifier, Filter, Mixer, and Oscillator compose most of the fundamental elements of the communication systems. Thus, the RF circuit technologies play an important role in the development of the communication systems. Considering the RF circuit technologies, the micro-strip line is the basic and popular transmission line because of its simple architecture, cost effective, and easy to apply in the print circuit technology. Therefore, we chose the micro-strip line to design the band pass filter and the thesis is written for improving the filter characteristic. The design of the band pass filter is based on the coupling theory of the parallel-coupled microstrip line structure. In the following chapters, methods for reducing the second pass band response were proposed. All e examples in the thesis were performed in real circuits and circuit tested data were compared with the related simulation results and related theories.

碩士
逢甲大學
資電碩士在職專班
101
This paper presents a new structure of Parallel-Coupled Bandpass filter. This structure offers a simple and easy method to manufacture Bandpass filter. Parallel-Coupled design uses impedance ratio formed by the discontinuity surface in Microstrip Line to achieve conditioning Spurious Response and improvement of characteristic for Selectivity; The central frequency of the Bandpass filter is 2.4GHz. This Bandpass filter features Symmetrical Suppression characteristic, and it has better Harmonic Suppression and higher Selectivity. This paper includes the content of how to bending electric circuit and hence shrinks area to 61% for compact electric circuit.

碩士
國立臺灣大學
電信工程學研究所
93
In this study, parallel-coupled bandpass filters based on lumped-distributed coupled-line sections have been proposed to possess the merits of compact size, good selectivity, and good stopband rejection. Novel parallel-coupled bandpass filters are proposed, by using additional lumped capacitors or inductors to reduce the length of coupled-line sections. In addition, by introducing the capacitive or inductive cross-coupling effect, transmission zeros can be achieved to improve the filter selectivity. The locations of transmission zeros may be simply adjusted by varying the cross-coupling capacitance or inductance. Simple equivalent-circuit models and design formulas are also established, from which the proposed parallel-coupled bandpass filters based on lumped-distributed coupled-line sections may be design and investigated. Furthermore, the approach to suppress the spurious response is also discussed so as to enhance the stopband characteristics.

碩士
國立海洋大學
電機工程學系
87
Abstract In the design of microwave multiplexer, the number of channel, the center frequency, the insertion and return losses and fractional bandwith of each channel are important parameters to be considered. Besides, if the parallel-coupled structures are used, the microwave multiplexers with more channels can be easily designed and realized. In this thesis, the parallel-coupled structures are used to design the mitered-bend microwave multiplexer firstly, and then the radial-bend microwave multiplexer with 50Ωmicrostrip lines is designed to improve the return loss. In this thesis, the theory of the design of parallel-coupled multiplexer is analyzed firstly. Next, two kinds of parallel-coupled multiplexers are designed and realized, respectively. Finally, the characteristics of the multiplexers are measured and compared with the simulation results.

碩士
國立臺灣大學
電機工程研究所
84
This thesis presents the design of three new coplanar waveguide filters, including the fork type、grounded type bandstop filters and the fork grounded type bandpass filter. The filters are built from cascading several sections of quarter wavelength coupled transmission lines. After a suitable analysis based on the model decomposition theory, each section is maded equivalent to an admittance inverter essential in the filter design. Several design charts are provided in yielding the geometrical parameters of each section to implement the desired admittance inverter. By neglecting the eqasi-static equivalent lumped circuit models of the discontinuities near the points of two sections, the equivalent models of coupled transmission lines are evaluated and incorporated into SPICE simulation to investigate more thoroughly the filter characteristics. The agreement between simulation results and experimental data justifies the design procedure and validates the present analysis approach.

Many signal processing and neural network algorithms can be mathematically described in terms of vector matrix multiplication. This thesis introduces two new architectures for computing high-speed vector matrix multiplication using charge coupled devices. These integrated circuits have been designed to accept optical matrix input as well as direct electrical matrix input. In both architectures, the matrix elements are stored as analog charge packets in CCD wells while the vectors are communicated to and from the integrated circuits by electrical means. The first architecture accomplishes the vector matrix product using a semiparallel computation scheme that requires N clock cycles of the device to complete one vector matrix multiplication where N is the length of the input vector. An analysis of the linearity and charge transfer induced errors is given. The circuit represents an advance over other analog signal processors in density and speed but has serious shortcomings in accuracy, particularly the limited precision of the input vectors. The second architecture is based on charge injection device (CID) imager arrays and addresses many of the inadequacies of the semiparallel architecture. A fully parallel circuit, the CID has similar density and much higher computation speed and accuracy. A novel digital input method is introduced that extends the input vector precision significantly. In addition, accuracy issues related to charge transfer efficiency are resolved. An analysis of linearity and accuracy is provided showing the advantages of the architecture over previous implementations.

碩士
國立聯合大學
電子工程學系碩士班
101
Microwave filter design frequently uses a lumped-element lowpass prototype, and a series of skills of transformation and approximation are then applied to achieve a distributed circuit that is suitable for layout realization. During the design process, filter’s design specifications are degraded,especially filter bandwidth. In order to enhance the bandwidth forecast, new method was proposed in a literature to reduce the shrinkage of bandwidth, and it was applied for the design of wideband parallel-coupled microstrip bandpass filter. In this paper, the proposed method is used to design two versions of filter that centered at 4GHz and having 50% fractional bandwidth. The circuit parameters of one of the two designs are determined by classical method, and the other one is determined by the literature-proposed method. Layouts of both designed are examined by electromagnetic simulator, implemented and verified by measurements.

碩士
國立交通大學
電信工程系
90
This thesis aims at designing microstrip bandpass filters with compact size. Coupled stages with folded lines are used. To make the effectiveness in reducing the circuit size as efficient as possible, both the width and gap of the coupled microstrip lines are chosen as small as possible. We analyze and compensate the folded resonators and coupled folded resonators, so that bandpass filters can be constructed with coupled multiple folded lines. A scheme for design coupled stages with an arbitrary coupled length is also incorporated into the design, for the convenience of resonator folding. Several Chebyshev filters of orders three and five are fabricated and measured to demonstrate the idea.

碩士
大同大學
通訊工程研究所
99
This thesis presents a parallel feed line which couples diamond-shaped slot to ring and cross-slot to patch. The feed line is used to feed the two parts in parallel. Then, the structure can easily be operated as a dual-band circularly polarized antenna to cover both (0.902GHz~0.928GHz and 2.4GHz~2.5GHz) bands. Simulated and measured results indicate that the proposed antenna can achieve good circularly polarized radiation performance.

碩士
國立交通大學
電信工程系所
94
The inherent zero of a microstrip coupled stage near twice the design frequency (2fo) is found tunable by varying its coupling length. This zero is used to suppress the unwanted response of paralle-coupled line filters at this frequency. The above idea is extended to design over-coupled middle stages of the filter for the spurious |S21| peaks at 3fo and 4fo, so that the upper stopband can be greatly enhanced to 5fo. The passband preserves a response as good as the traditional design. Measured results have a good agreement with simulation data and show that the idea works very well. Corrugated coupled lines are also devised to design microstrip bandpass filters with multispuriois suppression. Quarter-wave corrugated stages are tuned to allocate inherent transmission zeros at 2fo and 4fo. Stages with proper coupling lengths are arranged to cancel the unwanted peaks at 3fo, 6fo and 7fo, and the tapped input/output scheme is employed to tackle that at 5fo. The measured data of designed bandpass filters show rejection levels better than 30dB in the upper stopband. Three circuits are fabricated and measured to demonstrate the idea. Periodic stepped-impedance resonators (PSIRS) are proposed to design bandpass filters for multispurious suppression. Denoted as PSIRN, a PSIR of ��/2 long at design frequency fo consists of N periods of hi-Z and low-Z sections. A PSIRN coupled section shows transmission zeros at various frequencies. The zero can be tuned by changing the impedance ratio of the hi-Z and low-Z sections, and be adopted to suppress the spurious peaks. Responses of the PSIR filters show good rejection in the upper stopband.

In this research we consider developing a reservoir simulator capable of simulating complex coupled poromechanical processes on massively parallel computers. A variety of problems arising from petroleum and environmental engineering inherently necessitate the understanding of interactions between fluid flow and solid mechanics. Examples in petroleum engineering include reservoir compaction, wellbore collapse, sand production, and hydraulic fracturing. In environmental engineering, surface subsidence, carbon sequestration, and waste disposal are also coupled poromechanical processes. These economically and environmentally important problems motivate the active pursuit of robust, efficient, and accurate simulation tools for coupled poromechanical problems. Three coupling approaches are currently employed in the reservoir simulation community to solve the poromechanics system, namely, the fully implicit coupling (FIM), the explicit coupling, and the iterative coupling. The choice of the coupling scheme significantly affects the efficiency of the simulator and the accuracy of the solution. We adopt the fixed-stress iterative coupling scheme to solve the coupled system due to its advantages over the other two. Unlike the explicit coupling, the fixed-stress split has been theoretically proven to converge to the FIM for linear poroelasticity model. In addition, it is more efficient and easier to implement than the FIM. Our computational results indicate that this approach is also valid for multiphase flow. We discretize the quasi-static linear elasticity model for geomechanics in space using the continuous Galerkin (CG) finite element method (FEM) on general hexahedral grids. Fluid flow models are discretized by locally mass conservative schemes, specifically, the mixed finite element method (MFE) for the equation of state compositional flow on Cartesian grids and the multipoint flux mixed finite element method (MFMFE) for the single phase and two-phase flows on general hexahedral grids. While both the MFE and the MFMFE generate cell-centered stencils for pressure, the MFMFE has advantages in handling full tensor permeabilities and general geometry and boundary conditions. The MFMFE also obtains accurate fluxes at cell interfaces. These characteristics enable the simulation of more practical problems. For many reservoir simulation applications, for instance, the carbon sequestration simulation, we need to account for thermal effects on the compositional flow phase behavior and the solid structure stress evolution. We explicitly couple the poromechanics equations to a simplified energy conservation equation. A time-split scheme is used to solve heat convection and conduction successively. For the convection equation, a higher order Godunov method is employed to capture the sharp temperature front; for the conduction equation, the MFE is utilized. Simulations of coupled poromechanical or thermoporomechanical processes in field scales with high resolution usually require parallel computing capabilities. The flow models, the geomechanics model, and the thermodynamics model are modularized in the Integrated Parallel Accurate Reservoir Simulator (IPARS) which has been developed at the Center for Subsurface Modeling at the University of Texas at Austin. The IPARS framework handles structured (logically rectangular) grids and was originally designed for element-based data communication, such as the pressure data in the flow models. To parallelize the node-based geomechanics model, we enhance the capabilities of the IPARS framework for node-based data communication. Because the geomechanics linear system is more costly to solve than those of flow and thermodynamics models, the performance of linear solvers for the geomechanics model largely dictates the speed and scalability of the coupled simulator. We use the generalized minimal residual (GMRES) solver with the BoomerAMG preconditioner from the hypre library and the geometric multigrid (GMG) solver from the UG4 software toolbox to solve the geomechanics linear system. Additionally, the multilevel k-way mesh partitioning algorithm from METIS is used to generate high quality mesh partitioning to improve solver performance. Numerical examples of coupled poromechanics and thermoporomechanics simulations are presented to show the capabilities of the coupled simulator in solving practical problems accurately and efficiently. These examples include a real carbon sequestration field case with stress-dependent permeability, a synthetic thermoporoelastic reservoir simulation, poroelasticity simulations on highly distorted hexahedral grids, and parallel scalability tests on a massively parallel computer.
text

碩士
國立交通大學
電信工程系
90
An overlap scheme at both end stages of parallel coupled microstrip filters is investigated to improve the passband symmetry and suppress the parasitic passband at twice the center frequency. The overlap scheme extends the phase length for the odd mode so that the electric lengths of the entire filter for both odd and even modes can be identical, so that a transmission zero can be located near twice the passband frequency. Simulation results show that the required overlap length may change as the filter specification or the substrate is changed. It is found that the spurious response at twice the passband frequency can be suppressed down under —40 dB. Good agreement between measured and simulated results shows that the idea works very well.