Modulation Techniques for Multi-Phase Converters and Control Strategies for Multi-Phase Electric Drives

The ever-increasing spread of automation in industry puts the electrical engineer in a central role as a promoter of technological development in a sector such as the use of electricity, which is the basis of all the machinery and productive processes. Moreover the spread of drives for motor control and static converters with structures ever more complex, places the electrical engineer to face new challenges whose solution has as critical elements in the implementation of digital control techniques with the requirements of inexpensiveness and efficiency of the final product. The successfully application of solutions using non-conventional static converters awake an increasing interest in science and industry due to the promising opportunities. However, in the same time, new problems emerge whose solution is still under study and debate in the scientific community During the Ph.D. course several themes have been developed that, while obtaining the recent and growing interest of scientific community, have much space for the development of research activity and for industrial applications. The first area of research is related to the control of three phase induction motors with high dynamic performance and the sensorless control in the high speed range. The management of the operation of induction machine without position or speed sensors awakes interest in the industrial world due to the increased reliability and robustness of this solution combined with a lower cost of production and purchase of this technology compared to the others available in the market. During this dissertation control techniques will be proposed which are able to exploit the total dc link voltage and at the same time capable to exploit the maximum torque capability in whole speed range with good dynamic performance. The proposed solution preserves the simplicity of tuning of the regulators. Furthermore, in order to validate the effectiveness of presented solution, it is assessed in terms of performance and complexity and compared to two other algorithm presented in literature. The feasibility of the proposed algorithm is also tested on induction motor drive fed by a matrix converter. Another important research area is connected to the development of technology for vehicular applications. In this field the dynamic performances and the low power consumption is one of most important goals for an effective algorithm. Towards this direction, a control scheme for induction motor that integrates within a coherent solution some of the features that are commonly required to an electric vehicle drive is presented. The main features of the proposed control scheme are the capability to exploit the maximum torque in the whole speed range, a weak dependence on the motor parameters, a good robustness against the variations of the dc-link voltage and, whenever possible, the maximum efficiency. The second part of this dissertation is dedicated to the multi-phase systems. This technology, in fact, is characterized by a number of issues worthy of investigation that make it competitive with other technologies already on the market. Multiphase systems, allow to redistribute power at a higher number of phases, thus making possible the construction of electronic converters which otherwise would be very difficult to achieve due to the limits of present power electronics. Multiphase drives have an intrinsic reliability given by the possibility that a fault of a phase, caused by the possible failure of a component of the converter, can be solved without inefficiency of the machine or application of a pulsating torque. The control of the magnetic field spatial harmonics in the air-gap with order higher than one allows to reduce torque noise and to obtain high torque density motor and multi-motor applications. In one of the next chapters a control scheme able to increase the motor torque by adding a third harmonic component to the air-gap magnetic field will be presented. Above the base speed the control system reduces the motor flux in such a way to ensure the maximum torque capability. The presented analysis considers the drive constrains and shows how these limits modify the motor performance. The multi-motor applications are described by a well-defined number of multiphase machines, having series connected stator windings, with an opportune permutation of the phases these machines can be independently controlled with a single multi-phase inverter. In this dissertation this solution will be presented and an electric drive consisting of two five-phase PM tubular actuators fed by a single five-phase inverter will be presented. Finally the modulation strategies for a multi-phase inverter will be illustrated. The problem of the space vector modulation of multiphase inverters with an odd number of phases is solved in different way. An algorithmic approach and a look-up table solution will be proposed. The inverter output voltage capability will be investigated, showing that the proposed modulation strategy is able to fully exploit the dc input voltage either in sinusoidal or non-sinusoidal operating conditions. All this aspects are considered in the next chapters. In particular, Chapter 1 summarizes the mathematical model of induction motor. The Chapter 2 is a brief state of art on three-phase inverter. Chapter 3 proposes a stator flux vector control for a three- phase induction machine and compares this solution with two other algorithms presented in literature. Furthermore, in the same chapter, a complete electric drive based on matrix converter is presented. In Chapter 4 a control strategy suitable for electric vehicles is illustrated. Chapter 5 describes the mathematical model of multi-phase induction machines whereas chapter 6 analyzes the multi-phase inverter and its modulation strategies. Chapter 7 discusses the minimization of the power losses in IGBT multi-phase inverters with carrier-based pulse width modulation. In Chapter 8 an extended stator flux vector control for a seven-phase induction motor is presented. Chapter 9 concerns the high torque density applications and in Chapter 10 different fault tolerant control strategies are analyzed. Finally, the last chapter presents a positioning multi-motor drive consisting of two PM tubular five-phase actuators fed by a single five-phase inverter.

Competition between evaporation and fragmentation in nuclear reactions at 15-20 AMeV beam energy

The reactions 32S+58,64Ni are studied at 14.5 AMeV. From this energy on, fragmentation begins to be a dominant process, although evaporation and fission are still present. After a selection of the collision mechanism, we show that important even-odd effects are present in the isotopic fragment distributions when the excitation energy is small. The staggering effect appears to be a universal feature of fragment production, slightly enhanced when the emission source is neutron poor. A closer look at the behavior of isotopic chains reveals that odd-even effects cannot be explained by pairing effects in the nuclear mass alone, but depend in a more complex way on the de-excitation chain.

Vortex dynamics in superconductors in the presence of anisotropic pinning

Vortex dynamics in two different classes of superconductors with anisotropic unidirected pinning sites was experimentally investigated by magnetoresistivity measurements: YBCO−films with unidirected twins and Nb-films deposited on faceted $mathrm Al_2O_3$ substrate surfaces. For the interpretation of the experimental results a theoretical model based on the Fokker-Planck equation was used. It was proved by X-ray measurements that YBCO films prepared on (001) $mathrm NdGaO_3$ substrates exhibit only one twin orientation in contrast to YBCO films grown on (100) $mathrm SrTiO_$3 substrates. The magnetoresistivity measurements of the YBCO films with unidirected twin boundaries revealed the existence of two new magnetoresistivity components, which is a characteristic feature of a guided vortex motion: an odd longitudinal component with respect to the magnetic field sign reversal and an even transversal component. However, due to the small coherence length in YBCO and the higher density of point-like defects comparing to high-quality YBCO single crystals, the strength of the isotropic point pinning was comparable with the strength of the pinning produced by twins. This smeared out all effects caused by the pinning anisotropy. The behaviour of the odd longitudinal component was found to be independent of the transport current direction with respect to the twin planes. The magnetoresistivity measurements of faceted Nb films demonstrated the appearance of an odd longitudinal and even transversal component of the magnetoresistivity. The temperature and magnetic field dependences of all relevant magnetoresistivity components were measured. The angles between the average vortex velocity vector and the transport current direction calculated from the experimental data for the different transport current orientations with respect to the facet ridges showed that the vortices moved indeed along the facet ridges. An anomalous Hall effect, i.e. a sign change of the odd transversal magnetoresistivity, has been found in the temperature and magnetic field dependences of the Hall resisitivity of the samples. The theory developed by V.~A.~Shklovskij was used for the explanation of the experimental data. It shows very good agreement with the experiment. The temperature dependence of the even longitudinal magnetoresistivity component of the samples could be very well fitted within the theoretical approach, using for the isotropic and anisotropic pinning potential simple potential with a symmetric triangular potential wells whose depths were estimated from the experimental data.

Rheology and Fourier-transform rheology on water-based systems

The influence of shear fields on water-based systems was investigated within this thesis. The non-linear rheological behaviour of spherical and rod-like particles was examined with Fourier-Transform rheology under LAOS conditions. As a model system for spherical particles two different kinds of polystyrene dispersions, with a solid content higher than 0.3 each, were synthesised within this work. Due to the differences in polydispersity and Debye-length, differences were also found in the rheology. In the FT-rheology both kinds of dispersions showed a similar rise in the intensities of the magnitudes of the odd higher harmonics, which were predicted by a model. The in some cases additionally appearing second harmonics were not predicted. A novel method to analyse the time domain signal was developed, that splits the time domain signal up in four characteristic functions. Those characteristic functions correspond to rheological phenomena. In some cases the intensities of the Fourier components can interfere negatively. FD-virus particles were used as a rod-like model system, which already shows a highly non-linear behaviour at concentrations below 1. % wt. Predictions for the dependence of the higher harmonics from the strain amplitude described the non-linear behaviour well at large, but no so good at small strain amplitudes. Additionally the trends of the rheological behaviour could be described by a theory for rod-like particles. An existing rheo-optical set-up was enhanced by reducing the background birefringence by a factor of 20 and by increasing the time resolution by a factor of 24. Additionally a combination of FT-rheology and rheo-optics was achieved. The influence of a constant shear field on the crystallisation process of zinc oxide in the presence of a polymer was examined. The crystallites showed a reduction in length by a factor of 2. The directed addition of polymers in combination with a defined shear field can be an easy way for a defined change of the form of crystallites.

Polymorphs solvates and co-crystals of molecular materials

The scope of my research project is to produce and characterize new crystalline forms of organic compounds, focusing the attention on co-crystals and then transferring these notions on APIs to produce co-crystals of potential interest in the pharmaceutical field. In the first part of this work co-crystallization experiments were performed using as building blocks the family of aliphatic dicarboxylic acids HOOC-(CH2)n-COOH, with n= 2-8. This class of compounds has always been an object of study because it is characterized by an interesting phenomenon of alternation of melting points: the acids with an even number of carbon atoms show a melting point higher than those with an odd one. The acids were co-crystallized with four dipyridyl molecules (formed by two pyridine rings with a different number of bridging carbon atoms) through the formation of intermolecular interactions N•••(H)O. The bases used were: 4,4’-bipyridine (BPY), 1,2-bis(4-pyridyl)ethane (BPA), 1,2-(di-4-pyridyl)ethylene (BPE) and 1,2-bis(4-pyridyl)propane (BPP). The co-crystals obtained by solution synthesis were characterized by different solid-state techniques to determine the structure and to see how the melting points in co-crystals change. In the second part of this study we tried to obtain new crystal forms of compounds of pharmaceutical interest. The APIs studied are: O-desmethylvenlafaxine, Lidocaine, Nalidixic Acid and Sulfadiazine. Each API was subjected to Polymorph Screening and Salt/Co-crystal Screening experiments to identify new crystal forms characterized by different properties. In a typical Salt/Co-crystal Screening the sample was made to react with a co-former (solid or liquid) through different methods: crystallization by solution, grinding, kneading and solid-gas reactions. The new crystal forms obtained were characterized by different solid state techniques (X-ray single crystal diffraction, X-ray powder diffraction, Differential Scanning Calorimetry, Thermogravimetric Analysis, Evolved gas analysis, FT-IR – ATR, Solid State N.M.R).

Gleichmäßig dünne Basen zweiter Ordnung

Eine Menge B nicht negativer ganzer Zahlen heißt Basis h-ter Ordnung, wenn jede nicht negative ganze Zahl Summe von höchstens h Elementen von B ist. Eine der großen Fragen der additiven Zahlentheorie ist die nach der effektivsten Basis h-ter Ordnung für ein gegebenes h>=2. Im Fokus des Interesses steht dabei der immer noch offene Fall h=2. Bezeichnet B(x) die Anzahl der Elemente b aus B mit 0= af(x), wobei f die Wurzelfunktion bezeichne. Andererseits gibt es Basen B zweiter Ordnung mit B(x) <= cf(x). Daher kann man den Limes superior S(B), den Limes inferior s(B) sowie im Falle der Existenz den Limes d(B) des Quotienten B(x) / f(x) als Dichtefunktionen von Basen zweiter Ordnung betrachten. J. W. S. Cassels konstruierte 1957 eine Basis C zweiter Ordnung mit d(C)=5,196…. G. Hofmeister gab 2001 eine Basis H zweiter Ordnung mit asymptotischer Wurzeldichte d(H)=4,638… an. In der vorliegenden Arbeit wird eine Basis S zweiter Ordnung mit asymptotischer Wurzeldichte d(S)=3,464… konstruiert. Darüber hinaus wird für die von J. W. S. Cassels, für die von G. Hofmeister und für die in dieser Arbeit verwendete Klasse von Basen zweiter Ordnung gezeigt, dass die asymptotische Wurzeldichte innerhalb der jeweiligen Klasse nicht mehr zu verbessern ist. Bisher war die Frage nach möglichen Verbesserungen innerhalb der jeweiligen Konstruktionsprinzipien offen geblieben.

Hypersurfaces with many singularities

1. Teil: Bekannte Konstruktionen. Die vorliegende Arbeit gibt zunächst einen ausführlichen Überblick über die bisherigen Entwicklungen auf dem klassischen Gebiet der Hyperflächen mit vielen Singularitäten. Die maximale Anzahl mu^n(d) von Singularitäten auf einer Hyperfläche vom Grad d im P^n(C) ist nur in sehr wenigen Fällen bekannt, im P^3(C) beispielsweise nur für d<=6. Abgesehen von solchen Ausnahmen existieren nur obere und untere Schranken. 2. Teil: Neue Konstruktionen. Für kleine Grade d ist es oft möglich, bessere Resultate zu erhalten als jene, die durch allgemeine Schranken gegeben sind. In dieser Arbeit beschreiben wir einige algorithmische Ansätze hierfür, von denen einer Computer Algebra in Charakteristik 0 benutzt. Unsere anderen algorithmischen Methoden basieren auf einer Suche über endlichen Körpern. Das Liften der so experimentell gefundenen Hyperflächen durch Ausnutzung ihrer Geometrie oder Arithmetik liefert beispielsweise eine Fläche vom Grad 7 mit $99$ reellen gewöhnlichen Doppelpunkten und eine Fläche vom Grad 9 mit 226 gewöhnlichen Doppelpunkten. Diese Konstruktionen liefern die ersten unteren Schranken für mu^3(d) für ungeraden Grad d>5, die die allgemeine Schranke übertreffen. Unser Algorithmus hat außerdem das Potential, auf viele weitere Probleme der algebraischen Geometrie angewendet zu werden. Neben diesen algorithmischen Methoden beschreiben wir eine Konstruktion von Hyperflächen vom Grad d im P^n mit vielen A_j-Singularitäten, j>=2. Diese Beispiele, deren Existenz wir mit Hilfe der Theorie der Dessins d'Enfants beweisen, übertreffen die bekannten unteren Schranken in den meisten Fällen und ergeben insbesondere neue asymptotische untere Schranken für j>=2, n>=3. 3. Teil: Visualisierung. Wir beschließen unsere Arbeit mit einer Anwendung unserer neuen Visualisierungs-Software surfex, die die Stärken mehrerer existierender Programme bündelt, auf die Konstruktion affiner Gleichungen aller 45 topologischen Typen reeller kubischer Flächen.

Chirale Störungsrechnung im Sektor ungerader innerer Parität bis zur chiralen Ordnung O(q 6)

Die vorliegende Dissertation behandelt den anomalen Sektor bzw. den Sektor ungerader innerer Parität in mesonischer chiraler Störungsrechnung (mesonische ChPT) bis zur chiralen Ordnung O(q^6). Auf eine Einführung in die Quantenchromodynamik (QCD) und ihrer Verknüpfung mit der chiralen Symmetrie folgt die Betrachtung der mesonischen ChPT im Sektor gerader sowie ungerader innerer Parität bis zur Ordnung O(q^4). Der sogenannte Wess-Zumino-Witten Term, welcher den Einfluss der axialen Anomalie bezogen auf die ChPT widerspiegelt, wird studiert. Anschließend wird die allgemeinste Lagrangedichte der Ordnung O(q^6) im Sektor ungerader innerer Parität detailiert analysiert. Sie enthält in ihrer SU(3)-Formulierung 23 Niederenergiekonstanten(low-energy constant=LEC). Aus Sicht der ChPT sind diese LECs freie Parameter, die auf irgendeine Art und Weise fixiert werden müssen. Es wird herausgearbeitet, bei welchen Prozessen und in welchen Kombinationen die jeweiligen LECs auftreten. Daraufhin wird versucht so viele dieser LECs wie möglich mittels Vektormesondominanz (VMD) sowie experimenteller Daten abzuschätzen und anzupassen. Hierfür wird zuerst die Vorgehensweise einer konsistenten Rechnung im Sektor ungerader innerer Parität bis zur Ordnung O(q^6) studiert, gefolgt von der Berechnung von insgesamt vierzehn geeigneten Prozessen im Rahmen der ChPT bis zur Ordnung O(q^6). Unter Verwendung experimenteller Daten werden dreizehn der LECs angepasst, wobei gegenwärtig nicht bei allen betrachteten Prozessen experimentelle Daten zur Verfügung stehen. Die Ergebnisse werden diskutiert und Unterschiede bzw. Übereinstimmungen mit anderen Rechnungen herausgearbeitet. Zusammenfassend erhält man einen umfassenden Einblick in den Sektor ungerader innerer Parität in mesonischer ChPT bis zur Ordnung O(q^6).

Geometry of BV quantization and Mathai-Quillen formalism

Il formalismo Mathai-Quillen (MQ) è un metodo per costruire la classe di Thom di un fibrato vettoriale attraverso una forma differenziale di profilo Gaussiano. Lo scopo di questa tesi è quello di formulare una nuova rappresentazione della classe di Thom usando aspetti geometrici della quantizzazione Batalin-Vilkovisky (BV). Nella prima parte del lavoro vengono riassunti i formalismi BV e MQ entrambi nel caso finito dimensionale. Infine sfrutteremo la trasformata di Fourier “odd" considerando la forma MQ come una funzione definita su un opportuno spazio graduato.

Functional and neural mechanisms of intertemporal choice

People are daily faced with intertemporal choice, i.e., choices differing in the timing of their consequences, frequently preferring smaller-sooner rewards over larger-delayed ones, reflecting temporal discounting of the value of future outcomes. This dissertation addresses two main goals. New evidence about the neural bases of intertemporal choice is provided. Following the disruption of either the medial orbitofrontal cortex or the insula, the willingness to wait for larger-delayed outcomes is affected in odd directions, suggesting the causal involvement of these areas in regulating the value computation of rewards available with different timings. These findings were also supported by a reported imaging study. Moreover, this dissertation provides new evidence about how temporal discounting can be modulated at a behavioral level through different manipulations, e.g., allowing individuals to think about the distant time, pairing rewards with aversive events, or changing their perceived spatial position. A relationship between intertemporal choice, moral judgements and aging is also discussed. All these findings link together to support a unitary neural model of temporal discounting according to which signals coming from several cortical (i.e., medial orbitofrontal cortex, insula) and subcortical regions (i.e., amygdala, ventral striatum) are integrated to represent the subjective value of both earlier and later rewards, under the top-down regulation of dorsolateral prefrontal cortex. The present findings also support the idea that the process of outcome evaluation is strictly related to the ability to pre-experience and envision future events through self-projection, the anticipation of visceral feelings associated with receiving rewards, and the psychological distance from rewards. Furthermore, taking into account the emotions and the state of arousal at the time of decision seems necessary to understand impulsivity associated with preferring smaller-sooner goods in place of larger-later goods.

q-Hook length formulas for colored labeled forests

The major index has been deeply studied from the early 1900s and recently has been generalized in different directions, such as the case of labeled forests and colored permutations. In this thesis we define new types of labelings for forests in which the labels are colored integers. We extend the definition of the flag-major index for these labelings and we present an analogue of well known major index hook length formulas. Finally, this study (which has just apparently a simple combinatoric nature) allows us to show a notion of duality for two particular families of groups obtained from the product G(r,n)×G(r,m).

Kollineare Laserspektroskopie an radioaktiven Praseodymionen und Cadmiumatomen

Die Analyse optischer Spektren liefert einen kernmodellunabhängigen Zugang zur Bestimmung der Kernspins, Ladungsradien und elektromagnetischen Momente von Atomkernen im Grundzustand und von langlebigen Isomeren. Eine der vielseitigsten Methoden zur optischen Spektroskopie an kurzlebigen Isotopen ist die kollineare Laserspektroskopie. Im Rahmen dieser Arbeit wurde zum einen die TRIGA-LASER Strahlstrecke am Institut für Kernchemie der Universität Mainz durch die Implementierung einer neuen offline Oberflächenionenquelle für hohe Verdampfungstemperaturen und eines Strahlanalysesystems weiterentwickelt. Zum anderen wurde kollineare Laserspektroskopie an kurzlebigen Praseodym- und Cadmiumisotopen an ISOLDE/CERN durchgeführt. Die neue Ionenquelle ermöglichte dabei den Test der kollinearen Laserspektroskopie an Praseodymionen am TRIGA-LASER Experiment. Die Spektroskopie der Prasdeodymionen motivierte sich aus der Beobachtung einer zeitlichen Modulation der EC-Zerfallsrate von wasserstoffähnlichem 140Pr58+. Für die Erklärung dieser sogenannten GSI Oszillationen wird unter anderem das magnetische Moment des Kerns benötigt, welches bislang noch nicht experimentell bestimmt wurde. Zudem wurde für wasserstoffähnliches 140Pr58+ überraschenderweise eine kleinere EC-Zerfallskonstante gemessen als für heliumähnliches 140Pr57+. Die Erklärung dieses Phänomens setzt ein positives magnetisches Moment voraus. Bei der Spektroskopie am COLLAPS Experiment wurden erstmals die magnetischen Momente von 135Pr, 136Pr und 137Pr vermessen. Aufgrund zu geringer Produktionsraten war die Spektroskopie des gewünschten Isotops 140Pr jedoch nicht erfolgreich. Die Untersuchung der Cadmiumisotope ist kernphysikalisch motiviert. In der Zinnregion erstrecken sich die Isotope über die beiden magischen Zahlen N=50 und N=82 bei gleichzeitiger Nähe des Z=50 Schalenabschlusses. Hier können verschiedene Kernmodelle getestet werden, die sich beispielsweise hinsichtlich der Stärke des N=82 Schalenabschlusses widersprechen. Diese Arbeit berichtet über erste Ergebnisse der Spektroskopie an Cadmiumatomen, die sich über die Isotope 106−124,126Cd sowie die zugehörigen langlebigen I=11/2− Isomere erstreckt. Die zuvor experimentell bekannten oder aus dem erweiterten Schalenmodell abgeleiteten Kernspins konnten für alle Isotope bis auf 119Cd bestätigt werden. Der Kernspin von 119Cd wurde eindeutig zu I=1/2 bestimmt. Die elektrischen Quadrupolmomente der Isomere zeigen ein bemerkenswert lineares Verhalten in Abhängigkeit von der Neutronenzahl und dies über die eigentliche Kapazität der 1h11/2 Unterschale hinaus. Die Änderungen der mittleren quadratischen Ladungsradien zeigen den auch an Indium- und Zinnisotopen beobachteten stetigen Verlauf. Der lineare Anteil passt sehr gut zu den Berechnung des Tröpfchenmodells in der Parametrisierung nach Berchidevsky und Tondeur. Die Differenzen der mittleren quadratischen Ladungsradien zwischen Grund- und isomeren Zustand der ungeraden Cadmiumisotope zeigen einen interessanten parabolischen Verlauf mit einem Minimum zwischen A=117 und A=119.

Spins, moments and radii of cd isotopes

The complex nature of the nucleon-nucleon interaction and the wide range of systems covered by the roughly 3000 known nuclides leads to a multitude of effects observed in nuclear structure. Among the most prominent ones is the occurence of shell closures at so-called ”magic numbers”, which are explained by the nuclear shell model. Although the shell model already is on duty for several decades, it is still constantly extended and improved. For this process of extension, fine adjustment and verification, it is important to have experimental data of nuclear properties, especially at crucial points like in the vicinity of shell closures. This is the motivation for the work performed in this thesis: the measurement and analysis of nuclear ground state properties of the isotopic chain of 100−130Cd by collinear laser spectroscopy.rnrnThe experiment was conducted at ISOLDE/CERN using the collinear laser spectroscopy apparatus COLLAPS. This experiment is the continuation of a run on neutral atomic cadmium from A = 106 to A = 126 and extends the measured isotopes to even more exotic species. The required gain in sensitivity is mainly achieved by using a radiofrequency cooler and buncher for background reduction and by using the strong 5s 2S1/2 → 5p 2P3/2 transition in singly ionized Cd. The latter requires a continuous wave laser system with a wavelength of 214.6 nm, which has been developed during this thesis. Fourth harmonic generation of an infrared titanium sapphire laser is achieved by two subsequent cavity-enhanced second harmonic generations, leading to the production of deep-UV laser light up to about 100 mW.rnrnThe acquired data of the Z = 48 Cd isotopes, having one proton pair less than the Z = 50 shell closure at tin, covers the isotopes from N = 52 up to N = 82 and therefore almost the complete region between the neutron shell closures N = 50 and N = 82. The isotope shifts and the hyperfine structures of these isotopes have been recorded and the magnetic dipole moments, the electric quadrupole moments, spins and changes in mean square charge radii are extracted. The obtained data reveal among other features an extremely linear behaviour of the quadrupole moments of the I = 11/2− isomeric states and a parabolic development in differences in mean square nuclear charge radii between ground and isomeric state. The development of charge radii between the shell closures is smooth, exposes a regular odd-even staggering and can be described and interpreted in the model of Zamick and Thalmi.

Reservoir characterization of the Cardium Formation in the Ferrier Oilfield, west-central Alberta, Canada.

This thesis has the main aim of defining the lithostratigraphy, depositional architecture, post-depositional modifications and reservoir characteristics of the Cardium Formation in the Ferrier Oilfield, and how these characteristics can have great impact over production rates, GOR and produced fluid discrimination. In the Ferrier area, the Cardium Formation is composed by a NE prograding clastic sequence made up of offshore to shoreface deposits sealed by marine shales. The main reservoir is composed by sandstones and conglomerates interpreted to have deposited in a shoreface depositional environment. Lithofacies and net reservoir thickness mapping led to more detailed understanding of the 3D reservoir architecture, and cross-sections shed light on the Cardium depositional architecture and post-deposition sediment erosion in the Ferrier area. Detailed core logging, thin section, SEM and CL analyses were used to study the mineralogy, texture and pore characterization of the Cardium reservoir, and three main compartments have been identified based on production data and reservoir characteristics. Finally, two situations showing odd production behaviour of the Cardium were resolved. This shed light on the effect of structural features and reservoir quality and thickness over hydrocarbon migration pathways. The Ferrier example offers a unique case of fluid discrimination in clastic reservoirs due both to depositional and post-depositional factors, and could be used as analogue for similar situations in the Western Canadian Sedimentary Basin.

A sequential Cox approach for estimating the causal effect of treatment in the presence of time-dependent confounding applied to data from the Swiss HIV Cohort Study

When estimating the effect of treatment on HIV using data from observational studies, standard methods may produce biased estimates due to the presence of time-dependent confounders. Such confounding can be present when a covariate, affected by past exposure, is both a predictor of the future exposure and the outcome. One example is the CD4 cell count, being a marker for disease progression for HIV patients, but also a marker for treatment initiation and influenced by treatment. Fitting a marginal structural model (MSM) using inverse probability weights is one way to give appropriate adjustment for this type of confounding. In this paper we study a simple and intuitive approach to estimate similar treatment effects, using observational data to mimic several randomized controlled trials. Each 'trial' is constructed based on individuals starting treatment in a certain time interval. An overall effect estimate for all such trials is found using composite likelihood inference. The method offers an alternative to the use of inverse probability of treatment weights, which is unstable in certain situations. The estimated parameter is not identical to the one of an MSM, it is conditioned on covariate values at the start of each mimicked trial. This allows the study of questions that are not that easily addressed fitting an MSM. The analysis can be performed as a stratified weighted Cox analysis on the joint data set of all the constructed trials, where each trial is one stratum. The model is applied to data from the Swiss HIV cohort study.