Since the emergence of solid state devices onto the storage scene, improvements in capacity and price have brought them to the point where they are becoming a viable alternative to traditional magnetic storage for some applications. Current file system and device level I/O scheduler design is optimized for rotational magnetic hard disk drives. Since solid state devices have drastically different properties and structure, we may need to rethink the design of some aspects of the file system and scheduler levels of the I/O subsystem. In this thesis, we consider the current approach to I/O scheduling and show that the current scheduler design may not be ideally suited to solid state devices. We also present a framework for extracting some device parameters of solid state drives. Using the information from the parameter extraction, we present a new I/O scheduler design which utilizes the structure of solid state devices to efficiently schedule writes. The new scheduler, implemented on a 2.6 Linux kernel, shows up to 25% improvement for common workloads.

The research that stems from my doctoral dissertation focuses on addressing essential challenges in developing techniques that utilize solid-state memory technologies (with emphasis on NAND flash memory) from device, circuit, architecture, and system perspectives in order to exploit their true potential for improving I/O performance in high-performance computing systems. These challenges include not only the performance quirks arising from the physical nature of NAND flash memory, e.g., the inability to modify data in-place, read/write performance asymmetry, and slow and constrained erase functionality, but also the reliability drawbacks that limits solid state drives (SSDs) from widely deployed. To address these challenges, I have proposed, analyzed, and evaluated the I/O scheduling schemes, strategies for storage space virtualization, and data protection methods, to boost the performance and reliability of SSDs.

Mass spectrometry has long been used as a scientific tool in a wide variety of applications. A portable mass spectrometer would make many of these applications faster and more efficient. One of the key components of a mass spectrometer is its ion detection system; to make a mass spectrometer portable, this system must be small and involve as few components as possible. Single ion detection has been achieved through several methods, nearly all of which are well-known and understood. These methods, however, often require bulky vacuum and/or cooling systems in order to achieve high sensitivity. An ion detection system that can achieve high sensitivity under atmospheric pressure and normal temperature conditions would make portable mass spectrometry much more feasible. This thesis introduces a new method of detecting ions which does not require a vacuum or cold temperatures to operate: the solid-state ion detector, or SSID. Although ion detection using solid-state devices has been investigated previously, this work introduces metal-oxide-semiconductor field-effect transistors (MOSFETs) in a cascode configuration which acts as the primary detector when combined with a Faraday cup and mechanical switch. This detector is followed by a second amplifying stage which features RC-filters to help reduce noise and improve the detector's overall sensitivity. The detector is placed on a printed circuit board that was designed to fit a pre-determined system. Additional power circuitry for the mechanical switch was also designed and added to the detector circuitry. The SSID will be most sensitive when the input capacitance is made as small as possible. With this in mind, MOSFETs with a very low (< 1pF) gate capacitance were fabricated at BYU for use in the SSID. The performance of these MOSFETs was compared to a commercially available device in the same configuration. When tested, both MOSFETs had a sensitivity of hundreds of electrons when integrated in the complete SSID circuit. The commercial MOSFET demonstrated an estimated sensitivity of 150 electrons. The SSID shows much promise, and suggestions are made for further improving it to achieve even higher sensitivity levels. If made more sensitive, the next step would be to create an array of SSID detectors to be used in a portable mass spectrometer.

The thesis deals with the application of compensators andswitches based on power electronics in AC transmission anddistribution systems. The objective of the studieddevices/equipment is the power flow and voltage control intransmission systems and the mitigation of voltage sags andmomentary interruptions to critical loads in distributionsystems. For validating the power electronics based devices/equipmentdescribed in the thesis, scaled models at a real-time simulatorhave been built. Simulation results of these models arepresented and discussed in the thesis. The equipment studied in the thesis exploit the fast controlactions that can be taken by power electronics devices, whichare much faster than the speed of conventional equipment andprotection systems, based on electromechanical devices. In thisway, the power quality of distribution systems is improved,regarding duration and magnitude of voltage sags (dips) andmomentary interruptions, which are the most relevant types ofdisturbances in distribution systems. The thesis presents some compensators based onforced-commutation voltage-source converters for correctingvoltage sags and swells to critical loads. The seriesconverter, usually denoted Dynamic Voltage Restorer (DVR), hasbeen proved suitable for the task of compensating voltage sagsin the supply network. The use of solid-state devices ascircuit breakers in distribution systems has also been studiedwith the objective of achieving fast interruption or limitationof fault currents. The location and practical aspects for theinstallation of these solid-state breakers are presented. Ithas beenshown that a configuration based on shunt and seriesconnected solid-state devices with controllable turn-offcapability can also provide voltage sag mitigation, without theneed of transformers and large energy storage elements. The operation and control of two Flexible AC TransmissionSystem (FACTS) devices for voltage and power flow control intransmission systems, namely the Static Synchronous Compensator(STATCOM) and the Unified Power Flow Controller (UPFC),respectively, are also studied. A faster response compared totraditional equipment consisting of mechanically based/switchedelements is then achieved. This allows a more flexible controlof power flow and a secure loading of transmission lines tolevels nearer to their thermal limits. The behaviour of thesedevices during faults in the transmission system is alsopresented. Keywords: power electronics, power quality, voltagesags, voltage-source converters, Custom Power, FACTS, real-timesimulations, solid-state devices.

<p>The thesis deals with the application of compensators andswitches based on power electronics in AC transmission anddistribution systems. The objective of the studieddevices/equipment is the power flow and voltage control intransmission systems and the mitigation of voltage sags andmomentary interruptions to critical loads in distributionsystems.</p><p>For validating the power electronics based devices/equipmentdescribed in the thesis, scaled models at a real-time simulatorhave been built. Simulation results of these models arepresented and discussed in the thesis.</p><p>The equipment studied in the thesis exploit the fast controlactions that can be taken by power electronics devices, whichare much faster than the speed of conventional equipment andprotection systems, based on electromechanical devices. In thisway, the power quality of distribution systems is improved,regarding duration and magnitude of voltage sags (dips) andmomentary interruptions, which are the most relevant types ofdisturbances in distribution systems.</p><p>The thesis presents some compensators based onforced-commutation voltage-source converters for correctingvoltage sags and swells to critical loads. The seriesconverter, usually denoted Dynamic Voltage Restorer (DVR), hasbeen proved suitable for the task of compensating voltage sagsin the supply network. The use of solid-state devices ascircuit breakers in distribution systems has also been studiedwith the objective of achieving fast interruption or limitationof fault currents. The location and practical aspects for theinstallation of these solid-state breakers are presented. Ithas beenshown that a configuration based on shunt and seriesconnected solid-state devices with controllable turn-offcapability can also provide voltage sag mitigation, without theneed of transformers and large energy storage elements.</p><p>The operation and control of two Flexible AC TransmissionSystem (FACTS) devices for voltage and power flow control intransmission systems, namely the Static Synchronous Compensator(STATCOM) and the Unified Power Flow Controller (UPFC),respectively, are also studied. A faster response compared totraditional equipment consisting of mechanically based/switchedelements is then achieved. This allows a more flexible controlof power flow and a secure loading of transmission lines tolevels nearer to their thermal limits. The behaviour of thesedevices during faults in the transmission system is alsopresented. Keywords: power electronics, power quality, voltagesags, voltage-source converters, Custom Power, FACTS, real-timesimulations, solid-state devices.</p>

Ce travail de thèse a été réalisé au Laboratoire de Physique Subatomique et de Cosmologie (IN2P3) deGrenoble en collaboration avec le groupe Thalès avec pour objectif le développement d’une nouvellegénération de plasmas micro-onde fonctionnant sur une gamme de pression étendue allant de 0,5 mtorrà 10 torr en argon. La travail présenté porte donc en : i) la conception des applicateurs basés sur destronçons de longueur λ/4 faisant office de transformateurs d’impédance entre le générateur et leplasma d’impédance supposée donnée (adaptation d’impédance approchée); ii) la déterminationexpérimentale de l’impédance réelle du plasma (partie réelle et partie imaginaire) par mesure dumodule et de la phase du coefficient de réflexion dans des conditions opératoires définies; iii) leredimensionnement des différents tronçons de l’applicateur par simulation numérique en tenantcompte de l’impédance réelle du plasma; iv) la validation expérimentale de l’adaptation d’impédanceentre générateur et plasma. Les résultats obtenus démontrent clairement qu’il est possible, à fréquencedonnée (2.45 GHz dans le cas présent), de concevoir et de dimensionner une source plasma avec uneefficacité énergétique supérieure à 80% pour des fenêtres en pression (d’au moins une décade)équivalentes à des fenêtres opératoires en termes de paramètres plasma. Ces sources individuelles àabsorption localisée de micro-ondes peuvent être utilisées en nombre pour la réalisation des plasmasuniformes de grandes dimensions par leur distribution selon des réseaux à deux dimensions (sourcesplanes) ou à trois dimensions (volumes de plasma), et donc pour des applications industrielles auxtraitements de surface. / This work was done in the « Laboratoire de Physique Subatomique et de Cosmologie (IN2P3,Grenoble) » during a collaboration with Thales. The aim of the project was the development of a newgeneration of microwave plasma with extended operating conditions in the pressure range 0.5 mtorr to10 torr in argon. The presented work consists of: i) designing applicators based on sections of λ/4length serving as impedance transformers between the generator and the plasma with impedance ofgiven assumed value (approximate impedance adaptation); ii) experimentally determine the realplasma impedance (the real part and the imaginary part) for given operating conditions from themeasurement of modulus and phase of the reflection coefficient S11; iii) resize the different sections ofthe applicator by digital simulation taking the real plasma impedance into account; iv) finally, verifyexperimentally that the impedance adaptation between the generator and the plasma is correct. Theobtained results clearly demonstrate that it is possible, at a given frequency (here 2.45 GHz), to designand size a plasma source with an efficiency greater than 80 % for a window in pressure (at least onedecade) equivalent to an operating window in terms of plasma parameters. These individual sourceswith localized absorption of microwaves can be used in numbers to achieve uniform plasmas via theirdistribution over two-dimensional (planar sources) or tri-dimensional (volume plasma) networks, andthus for industrial surface treatments.