Residual-distribution methods offer several potential benefits over classical methods, such as a means of applying upwinding in a multi-dimensional manner and a multi-dimensional positivity property. While it is apparent that residual-distribution methods also offer higher accuracy than finite-volume methods on similar meshes, few studies have directly compared the performance of the two approaches in a systematic and quantitative manner. In this study, comparisons between residual distribution and finite volume are made for steady-state smooth and discontinuous flows of gas dynamics, governed by hyperbolic conservation laws, to illustrate the strengths and deficiencies of the residual-distribution method. Deficiencies which reduce the accuracy are analyzed and a new nonlinear scheme is proposed that closely reproduces or surpasses the accuracy of the best linear residual-distribution scheme. The accuracy is further improved by extending the scheme to fourth order using established finite-element techniques. Finally, the compact stencil, arithmetic workload, and data parallelism of the fourth-order residual-distribution scheme are exploited to accelerate parallel computations on an architecture consisting of both CPU cores and a graphics processing unit. Numerical experiments are used to assess the gains to efficiency and possible monetary savings that may be provided by accelerated architectures.

Residual-distribution methods offer several potential benefits over classical methods, such as a means of applying upwinding in a multi-dimensional manner and a multi-dimensional positivity property. While it is apparent that residual-distribution methods also offer higher accuracy than finite-volume methods on similar meshes, few studies have directly compared the performance of the two approaches in a systematic and quantitative manner. In this study, comparisons between residual distribution and finite volume are made for steady-state smooth and discontinuous flows of gas dynamics, governed by hyperbolic conservation laws, to illustrate the strengths and deficiencies of the residual-distribution method. Deficiencies which reduce the accuracy are analyzed and a new nonlinear scheme is proposed that closely reproduces or surpasses the accuracy of the best linear residual-distribution scheme. The accuracy is further improved by extending the scheme to fourth order using established finite-element techniques. Finally, the compact stencil, arithmetic workload, and data parallelism of the fourth-order residual-distribution scheme are exploited to accelerate parallel computations on an architecture consisting of both CPU cores and a graphics processing unit. Numerical experiments are used to assess the gains to efficiency and possible monetary savings that may be provided by accelerated architectures.

科学研究費補助金 研究種目:一般研究(C) 課題番号:02680006 研究代表者:庄司 多津男 研究期間:1990-1991年度

With the rapid increase in electrical loads, a considerable percentage of global power systems have been stretched to their designed capacity, leading to a sharp rise in the fault current. High fault current levels have severe impact on power systems. They can increase the likelihood of overcurrent damage, which may destroy electrical facilities or even cause system blackout. The resistive-type superconducting fault current limiter (RSFCL) is a device that can help to mitigate the increasing level of fault currents. It can also contribute to the performance, stability and efficiency of electricity grids. In order to promote the RSFCL more effectively, it is essential to study the device itself and conduct simulations regarding the performance and applicability from the system point of view. Chapter 1 and Chapter 2 of this thesis introduced fundamentals of superconductivity and RSFCLs, respectively. In Chapter 3, a power system model was built, and the transient analysis of short-circuit currents was given. Then, the operating principle of RSFCL was explained, and a step-resistance RSFCL model was introduced. To validate the current limiting performance of the SFCL module, wind farm protection schemes were studied under various fault scenarios. After thorough analysis it was concluded that the optimal allocating strategy of SFCLs was the installation of one SFCL at the integrating point of the system model. Chapter 4 presented a comprehensive study on the performance and optimal allocation strategy of RSFCLs. The two power system models used in this chapter were built based on the UK network standard. To assess the impact of incorporating SC material properties on the performance of SFCLs, three different models were compared throughout the study. Although computing time can be reduced when step-resistance and exponential equation models were used, such simplifications led to strong overestimations of the SFCL performance and resulted in wrong conclusion of optimal installing strategies. For both power system models, the simultaneous use of three SFCLs was the best protection strategy in terms of the performance, economic efficiency and reliability of the overall grids. To draw this conclusion, all the potential combinations of two, three, four, and five SFCLs were studied under a wide number of fault scenarios and measuring strategies. In Chapter 5, a series of experiments were performed to study the magnetic field-angular dependence of the critical current of different commercial YBCO samples. We selected ten 2G-HTS tapes with broad differences in width, fabrication process, and laminar structure. The obtained I_c (B,θ) characteristics of HTS samples were applied in the simulation of RSFCLs, showing unneglectable effect on the first peak limiting performance during faults. This study is helpful to the design and simulation of RSFCLs and other HTS applications which require superconducting wires operating inside magnetic fields. In Chapter 6, we presented a High Frequency (HF) AC-assisted quench study of the YBCO conductor. The differences in quench processes triggered by HF AC field were recorded and studied. We found that AC signals of 10 kHz can trigger quench of the HTS tape. Also, the device proved to be effective at guaranteeing uniform quench of the sample. After then, we placed the experimental device in a magnet, to study whether the quench process can be accelerated by external DC magnetic field. It was found that the DC field can reduce the quench time, but its performance was dependent on the amplitude of transport current and the characteristics of HF AC signals. In addition, the angle between the DC magnetic field and the tape surface showed a huge impact on the quench time. Finally, a comprehensive field-angular dependence study about the quench time of the YBCO sample was conducted, revealing the same pattern as J_c (θ,B) dependence of the tape, but with a greater anisotropy. To summarize, this technique showed outstanding performance regarding quench acceleration and tape protection, and therefore has huge potential to be applied in RSFCLs.

El modelado se ha convertido en los últimos años en una herramienta esencial en el diseño de motores de combustión interna alternativos, ya que permite reducir considerablemente el tiempo y los costes de desarrollo. Las metodologías de diseño clásicas se basan en la fabricación de prototipos y la realización de pruebas de ensayo y error. Actualmente, la mayoría de estas pruebas han sido sustituidas por cálculos numéricos, de modo que sólo las opciones de diseño más prometedoras se prueban en realidad en banco motor. Durante años, los códigos unidimensionales de dinámica de gases en el dominio del tiempo han sido suficientes para modelar tanto las prestaciones y el consumo del motor como el ruido de admisión y escape. Sin embargo, para un nivel más exigente de diseño, una representación 1D puede no ser suficiente para describir con precisión el flujo en ciertos elementos. Esto es especialmente importante en el caso de silenciadores, donde la hipótesis unidimensional sólo se puede aplicar a geometrías simples. En el caso de las uniones de conductos es la existencia de estructuras tridimensionales de flujo complejas lo que establece el límite de la aplicabilidad de una descripción simple cero-dimensional. En vista de estas limitaciones, la primera opción sería el uso de un modelo de dinámica de fluidos computacional (CFD); sin embargo, su aplicación conllevaría un tiempo de cálculo excesivo. Una posible solución de compromiso viene dada por los modelos cuasi-3D, basados en esquemas tridimensionales, pero con ciertas simplificaciones capaces de reducir significativamente el tiempo de cálculo sin afectar excesivamente a la precisión. Tales soluciones se han convertido en estándar en los códigos comerciales y se han aplicado con éxito a los silenciadores, tanto para excitaciones acústicas en el régimen lineal como en condiciones reales de motor, típicamente no lineales. Esta tesis tiene como objetivo el desarrollo de un nuevo método numérico cuasi-3D en una malla escalonada, basado en la simplificación de la ecuación de la cantidad de movimiento, para ser incluido en un código unidimensional existente. Tal método, sin embargo, no está libre de inconvenientes. En particular, se ve afectado por la aparición de oscilaciones no físicas, especialmente en gradientes de presión significativos. De la revisión bibliográfica se determina que este comportamiento es típico en esquemas de segundo orden y se puede ver acentuado por las simplificaciones adoptadas. Tras estudiar las posibles soluciones aplicables a este problema, se desarrollan tres limitadores de flujo diferentes, basados en las metodologías MDT, FCT y TVD. Una vez definido el método numérico y asegurada su estabilidad, es necesario desarrollar las condiciones de contorno adecuadas que permitan su utilización. Con este objetivo, se desarrollan las condiciones de pulso de presión de entrada y de extremo anecoico, los cuales permiten simular un banco de impulso. No hay que olvidar, sin embargo, que el objetivo final es la conexión con un código unidimensional, por lo que hay que comprobar que el método numérico cuasi-3D creado es compatible con los unidimensionales existentes, mostrando algunos resultados preliminares. Finalmente, con el método ya completamente operativo, se procede a su validación en las aplicaciones para las que ha sido diseñado principalmente, las cuales son, modelado de silenciadores y uniones de conductos. Para el caso de los silenciadores, se modelan dispositivos de complejidad creciente, pasando por geometrías de sección constante hasta sistemas con geometrías reales. Los resultados obtenidos se validan con otras herramientas tanto lineales como no lineales. En el caso de las uniones de conductos, el objetivo principal es el de establecer el potencial del nuevo método numérico frente a los tradicionales unidimensionales, por lo que los resultados de ambos se comparan con datos experiment / Engine modelling has become an essential tool in the design of internal combustion engines, allowing considerable reductions in development time and cost. Classical design methodologies are based on prototype manufacturing and trial-and-error tests, but currently, most of those tests have been replaced by numerical computations, so that only the most promising design options are actually tested on engine bench. For years, one-dimensional gas dynamics codes in the time domain have offered sufficiently good solutions for modelling both engine performance and intake and exhaust noise. However, for a more demanding level of design, a 1D representation may not be sufficient to describe accurately the flow in certain elements. This is especially important in the case of silencers. In the case of duct junctions, the existence of complex 3D flow structures is what sets the applicability limit for a simple zero-dimensional description. In view of these limitations, the first option would typically be the use of a computational fluid dynamics (CFD) model; however, the application of such a model to a complete intake or exhaust system entails an excessive computational time. A possible compromise solution is given by quasi-3D models, based on three-dimensional schemes, but with certain simplifications able to significantly reduce the calculation time without excessively affecting the accuracy. Such solutions have become standard in commercial codes and have been successfully applied to silencers with perforated tubes and absorbing material, both in the linear acoustic regime and in real engine conditions, typically non-linear. The objective of this thesis is the development a new quasi-3D numerical method in a staggered-grid, based on the simplification of the momentum equation, to be included in an existing one-dimensional code. Such method however, is not hassle free. In particular, it is affected by the appearance of non-physical oscillations, specially near significant pressure gradients. From the literature review it is determined that this behaviour is typical among second-order schemes and it can be aggravated by the simplifications adopted. After researching the possible solutions to face this problem, three different flux limiters are developed, based on the MDT, FCT and TVD methodologies. In the case of the two latter methods, its effectiveness is well established for finite differences schemes, thus defining a clear improving line for quasi-3D models. Once the numerical method is defined and its stability assured, proper boundary conditions that allow its use must be developed. With this objective, a pressure pulse inlet and an anechoic termination boundary condition are developed, which allow the simulation of an impulse test rig. It should not be forgotten, however, that the ultimate objective is the connection with a one-dimensional code, therefore the compatibility of the quasi-3D numerical method created with the existing one-dimensional methods has to be tested, showing some preliminary results. Eventually, with a fully operative method, the validation process for the applications which it has been mainly developed for, takes place, namely, mufflers and duct junctions modelling. In the case of mufflers, increasingly complex devices are modelled, from constant section geometries to real geometry systems. The results obtained are validated with both linear and non-linear tools. In the case of duct junctions, the main objective is to establish the potential of the new numerical method against the traditional one-dimensional schemes, consequently, results from both approaches are compared to experimental measures, obtaining promising results. / El modelatge s'ha convertit en els últims anys en una eina essencial en el disseny de motors de combustió interna alternatius, ja que permet reduir considerablement el temps i els costos de desenvolupament. Les metodologies de disseny clàssiques es basen en la fabricació de prototips i la realització de proves d'assaig i error. Actualment, la majoria d'aquestes proves han sigut substituïdes per càlculs numèrics, de manera que només les opcions de disseny més prometedores es proven en realitat en banc motor. Durant anys, els codis unidimensionals de dinàmica de gasos en el domini del temps han sigut suficients per a modelar tant les prestacions i el consum del motor com el soroll d'admissió i escapament. No obstant això, per a un nivell més exigent de disseny, una representació 1D pot no ser prou per a descriure amb precisió el flux en certs elements. Açò és especialment important en el cas de silenciadors, on la hipòtesi unidimensional només es pot aplicar a geometries simples. En el cas de les unions de conductes és l'existència d'estructures tridimensionals de flux complexes el que establix el límit de l'aplicabilitat d'una descripció simple zero-dimensional. En vista d'estes limitacions, la primera opció seria típicament l'ús d'un model de dinàmica de fluids computacional (CFD); no obstant això, l'aplicació comporta un temps de càlcul excessiu. Una possible solució de compromís ve donada pels models quasi-3D, basats en esquemes tridimensionals, però amb certes simplificacions capaços de reduir significativament el temps de càlcul sense afectar excessivament la precisió. Tals solucions s'han convertit en estàndard en codis comercials i s'han aplicat amb èxit als silenciadors, tant per a excitacions acústiques en el règim lineal com en condicions reals de motor, típicament no lineals. Aquesta tesi té com a objectiu el desenvolupament d'un nou mètode numèric quasi-3D en una malla escalonada, basat en la simplificació de l'equació de la quantitat de moviment, per a ser inclòs en un codi unidimensional existent. Tal mètode, però, no està lliure d'inconvenients. En particular, es veu afectat per l'aparició d'oscil·lacions no físiques, especialment en gradients de pressió significatius. De la revisió bibliogràfica es determina que aquest comportament és típic en esquemes de segon ordre i es pot veure accentuat per les simplificacions adoptades. Després d'estudiar les possibles solucions aplicables a aquest problema, es desenvolupen tres limitadors de flux diferents, basats en les metodologies MDT, FCT i TVD. En el cas dels dos últims mètodes, la seua efectivitat està ben establida per als esquemes de diferències finites, la qual cosa definix una clara via de millora per als models quasi-3D. Una vegada definit el mètode numèric i assegurada la seua estabilitat, és necessari desenvolupar les condicions de contorn adequades que permeten la seua utilització. Amb aquest objectiu, es desenvolupen les condicions de pols de pressió d'entrada i d'extrem anecoic, els quals permeten simular un banc d'impuls. No cal oblidar que l'objectiu final és la connexió amb un codi unidimensional, per la qual cosa cal comprovar que el mètode numèric cuasi-3D creat és compatible amb els unidimensionals existents, mostrant alguns resultats preliminars. Finalment, es procedix a la seua validació en les aplicacions per a les que ha sigut dissenyat principalment, les quals són, modelatge de silenciadors i unions de conductes. Per al cas dels silenciadors, es modelen dispositius de complexitat creixent, passant per geometries de secció constant fins a sistemes amb geometries reals. Els resultats obtinguts es validen amb altres eines tant lineals com no lineals. En el cas de les unions de conductes, l'objectiu principal és el d'establir el potencial del nou mètode numèric front als unidimensionals tradicionals, per la qual cosa els resultats d'ambdós es comparen amb dades experim / Hernández Marco, M. (2018). A non-linear quasi-3D model for air management modelling in engines [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/103683 / TESIS

This thesis treats of problems of nonlinear circuits. In introduction the basic possibilities of solving nonlinear circuits are described. Then the basic information on modern current active elements is mentioned, for example current follower, current operational amplifier and current conveyor. In the next part of this thesis the problem of real characteristics of semiconductor diodes is described, major attention is focused on reverse recovery time. Three various types of circuits with nonlinear elements are described in this masters thesis. The first type of circuits are diode precision rectifiers, the full-wave rectifiers containing voltage operational amplifier are most explored, current follower and other active components are described in previous chapters. Simulations in PSpice were executed for each of these circuits. Then the attention was devoted to increase of maximum working frequence. The minimum necessary amplitude of input signal was also analyzed. In the following part two designs of diode function converters are processed, function converter with voltage operational amplifier and function converter with current follower. For each of the diode function converters designed, simulations were executed to find their real properties. The variety of diode voltage limiters and diode current limiters are the third type of nonlinear circuits described, from which the upper limiter and double limiter are investigated in more detail. Simulations were executed for all limiters and their functionality and problems connected with their usage for limitation of high frequency signals were analyzed. In the last part of the thesis there is described the practical realization of precise rectifier, which showed the best ability to process high frequency signals of all investigated circuits. This circuit was realized so that it is able to operate in current and mixed mode. Measured values of real circuits were compared with values gained from the simulations.

This thesis deals with non-linear curcuit structures with current active elements. In its introduction this work deals with a description of the current active elements, such as current conveyor, transconductance operational amplifier and current follower. Further the text pays attention to the possiblities of the circuits for modification signals in analogue technology. First are described circuits of diode limiters and transducer. Great attention is paid to the amplifiers with current active elements. Are propřed simulated and practically realized circuits of universal precise full-wave rectifiers using operational amplifier and current conveyor and made comparisons of known and proposed circuits. The paper also discussed other circuit structures using of current active elements, such as multiplier, divider, a triangular signal generator and oscillators.

Electrical stimulation on nerves is a relatively new area of research and has been proved to speed up recoveryfrom nerve damage. In this work, the efficiency and stability of antennas integrated on printed circuit boards provided by the department of electrical engineering are examined. An automated test bench containing a stepmotor with a slider and an Arduino is created. Different setups were used when measuring on the boards, which resulted in that the largest antenna gave the most stable output despite the distance between transmitterand receiver. The conclusion was that the second best antenna and the smallest one would be suitable as well,and the better choice if it is to be implemented under the skin. A physical setup consisting of LEDs, an Arduino, a computer, and a function generator was created to examinethe voltage control functionality, where colored LEDs were lit depending on the voltage level. The functionality was then implemented in a circuit that in the future shall be integrated on the printed circuit board. To control high voltages a limiter circuit was examined and implemented. The circuit was simulated and tested, with a realization that a feature covering voltage enlargement is needed for the future.

No description available.

The master´s thesis is focused on design of ADSR filter and voltage controlled amplifier (VCA). Three additional circuits performing analog signal processing are added. Functionality of designed circuits is verified in simulation program. All designed circuits are practically realized. Thesis includes complete design of the mentioned circuits and all necessary informations for its practical realization. All designed circuits are measured and the results are presented.