Thesis (M. Soc. Sc.)--University of Hong Kong, 1996. / Includes bibliographical references (leaf 130-134).

Thesis (Ph. D.)--Northern Illinois University, 1995. / Includes bibliographical references (leaves [313]-333).

Thesis (M.S.)--Ohio University, March, 2006. / Title from PDF t.p. Includes bibliographical references (p. 101-105)

Thesis (M.Soc.Sc.)--University of Hong Kong, 1996. / Includes bibliographical references (leaf 130-134) Also available in print.

L’augmentation de la proportion d’énergie renouvelable sur le réseau électrique amène de nouveaux défis dans le domaine des réseaux électriques dont notamment la détection de l’îlotage non-intentionnel. Il s’agit de la situation pendant laquelle une partie du réseau de distribution, bien que déconnectée du réseau global après l’ouverture d’une protection, reste alimentée par la production locale. Ce phénomène, autrefois improbable, engendre de nombreux problèmes car la fréquence et la tension dans l’îlot ne sont plus imposées par le réseau global et deviennent donc incontrôlées. La probabilité d’apparition de ce phénomène semble augmenter du fait de l’insertion grandissante de la proportion de production renouvelable sur le réseau. De plus, de nouvelles réglementations tendent à élever le seuil de fréquence des protections de découplage, augmentant ainsi la probabilité que les producteurs ne détectent pas de situation anormale. Les travaux dans le cadre de cette thèse sont une contribution à une meilleure compréhension globale de ce phénomène et à sa détection. Des expériences en laboratoire ont été menées et ont notamment permis de prouver l’existence du phénomène. Ensuite, une étude de données réelles a permis de définir la durée typique de présence tension après l’ouverture d’une protection en tête de départ HTA sur lequel se trouvent des installations de production. Deux cas d’étude sont ensuite présentés. Le premier consiste à analyser la probabilité qu’un défaut monophasé sur la HTA ne soit pas détecté par la production en BT. Le second concerne des simulations numériques et montre l’impact du modèle de charge sur le comportement de l’îlotage non-intentionnel. La troisième étape de ces travaux concerne l’analyse de l’impact de nouvelles régulations dans les onduleurs photovoltaïques sur la taille de la zone de non-détection des protections de découplage. Enfin, la dernière partie présente une étude de l’impact de la dispersion naturelle des réglages des protections de découplage sur le comportement du système électrique, lors d’un événement de grande ampleur. / A high penetration of distributed generators (DGs) on the electrical networks brings new challenges for distribution system operators (DSO). Unintentional islanding is one of them; it is a situation when a part of the network is disconnected from the main grid but remains supplied by local generation. Though this was unlikely to occur before, the likelihood of this situation seems to increase with a higher penetration of DGs, and it leads to several issues since the frequency and the voltage in the island are not governed by the main grid anymore. Moreover, new requirements tend to increase the upper frequency threshold of the interface protection to prevent a system wide event. This could lead to an increase of the likelihood of undetected islanding. This thesis is a contribution to a better understanding of this phenomenon and its detection. First, laboratory experiences were conducted to prove its existence. Then, real data have been analyzed in order to define the typical duration of voltage presence after the trip of a protection feeder. Two case studies are introduced in the next part. The first one consists in assessing the likelihood of a single-phase fault on medium voltage networks that would generally not be detected by low voltage (LV) generation. The second one concerns numerical simulations and shows the impact of the load model on the frequency behavior of the island. The third part of this work assesses the impact of new regulations in the photovoltaic inverters on the non-detection zones of interface protections. Finally, the last part introduces a study regarding the impact of the natural dispersion of interface protection settings on the stability of an electrical system when a system wide event occurs.

Vu que les tags RFID sont actuellement en phase de large déploiement dans le cadre de plusieurs applications (comme les paiements automatiques, le contrôle d'accès à distance, et la gestion des chaînes d’approvisionnement), il est important de concevoir des protocoles de sécurité garantissant la protection de la vie privée des détenteurs de tags RFID. Or, la conception de ces protocoles est régie par les limitations en termes de puissance et de calcul de la technologie RFID, et par les modèles de sécurité qui sont à notre avis trop forts pour des systèmes aussi contraints que les tags RFID. De ce fait, on limite dans cette thèse le modèle de sécurité; en particulier, un adversaire ne peut pas observer toutes les interactions entre tags et lecteurs. Cette restriction est réaliste notamment dans le contexte de la gestion des chaînes d’approvisionnement qui est l’application cible de ce travail. Sous cette hypothèse, on présente quatre protocoles cryptographiques assurant une meilleure collaboration entre les différents partenaires de la chaîne d’approvisionnement. D’abord, on propose un protocole de transfert de propriété des tags RFID, qui garantit l’authentification des tags en temps constant alors que les tags implémentent uniquement des algorithmes symétriques, et qui permet de vérifier l'authenticité de l’origine des tags. Ensuite, on aborde le problème d'authenticité des produits en introduisant deux protocoles de sécurité qui permettent à un ensemble de vérificateurs de vérifier que des tags “sans capacité de calcul” ont emprunté des chemins valides dans la chaîne d’approvisionnement. Le dernier résultat présenté dans cette thèse est un protocole d’appariement d’objets utilisant des tags “sans capacité de calcul”, qui vise l’automatisation des inspections de sécurité dans la chaîne d’approvisionnement lors du transport des produits dangereux. Les protocoles introduits dans cette thèse utilisent les courbes elliptiques et les couplages bilinéaires qui permettent la construction des algorithmes de signature et de chiffrement efficaces, et qui minimisent donc le stockage et le calcul dans les systèmes RFID. De plus, la sécurité de ces protocoles est démontrée sous des modèles formels bien définis qui prennent en compte les limitations et les contraintes des tags RFID, et les exigences strictes en termes de sécurité et de la protection de la vie privée des chaines d’approvisionnement. / While RFID systems are one of the key enablers helping the prototype of pervasive computer applications, the deployment of RFID technologies also comes with new privacy and security concerns ranging from people tracking and industrial espionage to produ ct cloning and denial of service. Cryptographic solutions to tackle these issues were in general challenged by the limited resources of RFID tags, and by the formalizations of RFID privacy that are believed to be too strong for such constrained devices. It follows that most of the existing RFID-based cryptographic schemes failed at ensuring tag privacy without sacrificing RFID scalability or RFID cost effectiveness. In this thesis, we therefore relax the existing definitions of tag privacy to bridge the gap between RFID privacy in theory and RFID privacy in practice, by assuming that an adversary cannot continuously monitor tags. Under this assumption, we are able to design sec ure and privacy preserving multi-party protocols for RFID-enabled supply chains. Namely, we propose a protocol for tag ownership transfer that features constant-time authentication while tags are only required to compute hash functions. Then, we tackle the problem of product genuineness verification by introducing two protocols for product tracking in the supply chain that rely on storage only tags. Finally, we present a solution for item matching that uses storage only tags and aims at the automation of safety inspections in the supply chain.The protocols presented in this manuscript rely on operations performed in subgroups of elliptic curves that allow for the construction of short encryptions and signatures, resulting in minimal storage requirements for RFID tags. Moreover, the privacy and the security of these protocols are proven under well defined formal models that take into account the computational limitations of RFID technology and the stringent privacy and security requirements of each targeted supply chain application.

This project was carried out to develop a better understanding of "Electrochemical Chloride Removal (also known as Desalination), a relatively new technique for rehabilitating reinforced concrete structures suffering from chloride induced corrosion. The factors influencing chloride removal were investigated and a mathematical model for predicting the spatial distribution of chloride in the concrete at the end of the treatment was developed. To investigate the factors influencing chloride removal, concrete prisms containing varying levels of admixed sodium chloride and number of steel bars were tested. The amount of chloride removed during the treatment was assessed by analysing the anolyte. It was found that chloride removal increased with increasing applied potential, number of reinforcing bars at a particular depth, water/cement ratio, chloride contamination depth and initial chloride content. Chloride removal was unaffected by concrete strength but decreased with increasing percentage of cement replacement material. A greater percentage of chloride was removed from prisms where the thickness of the chloride bearing layer of concrete was less than the depth of cover to the reinforcement. Where the thickness of the chloride bearing layer exceeded the cover to the reinforcement, the use of an external cathode significantly increased the total amount of chloride removed. Chloride removal from a face remote from the source of the chloride contamination (soffit desalination) was shown to be feasible. In prisms containing 2, 3 and 4% Cl (by weight of cement), it was found that the amount of chloride remaining in the prisms reached a limiting value irrespective of the initial admixed chloride content. This value was believed to be approximately equal to the amount of bound chloride in the concrete. Analysis of ground samples of concrete showed, however, that in a few cases bound chloride was removed locally to the reinforcing bars. Even though significant amounts of chloride remained in the concrete after completion of the treatment, the reinforcement remained passive for periods exceeding 30 months. Towards the second aim of predicting the spatial chloride distribution in the concrete, a mathematical model for electrochemical chloride removal from concrete based on the Nernst-Planck and Laplace equations was developed. The model relies on experimentally derived chloride transport number profiles. The predicted quantities of chloride removed into the anolyte and chloride remaining in concrete correlate well with data obtained by testing prisms containing one bar and dosed with 2, 3 and 4% Cl. The model shows that all parts of the concrete prisms undergo chloride removal albeit at different rates. Concrete directly between the anode and cathode undergoes the highest rate of chloride removal while concrete below the cathode, the lowest. It would appear, therefore, that there are no 'dark spots' in the concrete where chloride removal is prevented.

Since mid-80‟s cathodic protection (CP) has been recognised as the „‟only technique known to stop corrosion regardless of the levels of chloride contamination in concrete‟‟ (FHWA, 1982) and is proved to be the most cost effective means to extend the useful life of the structure. Cathodic protection is an electrochemical technique to stop/mitigate corrosion by supplying „current‟ from an external source in order to suppress the „internally generated‟ current flow due to corrosion processes. The „external‟ current source could be obtained simply by coupling the steel to another electrochemically more active metal, e.g. zinc; alternatively the „external‟ current may be derived from a mains operated low voltage DC power source, viz. transformer/rectifier unit. These two different approaches to supply „external‟ current to stop corrosion are generically termed as: „Sacrificial Anode Cathodic Protection (SACP)‟ system and „Impressed Current Cathodic Protection (ICCP)‟ system, respectively. Both approaches have proved to be feasible, but the impressed current CP system offers greater flexibility with regard to its ability to provide the necessary current in situations where concrete resistivity is relatively high and variable. The sacrificial anode system is most effective if the concrete resistivity is very low or the anode is placed in a very low resistivity environment such as soil with low resistivity, as the inherent driving voltage is low e.g. the potential difference between zinc and corroding steel in concrete is limited to approximately 0.7 volts. Other contra-distinction between the two approaches are that the design life of the sacrificial anode systems are usually range between 10 -15 years; on the other hand the design life of the ICCP systems could be well in excess of 60+ years ( depending on the type of anode system). Page 2 Following the successful application of first CP system, based on impressed current CP (ICCP), on a bridge deck in California, USA 1973, the technology has advanced significantly, particularly the anode systems (which is the main arbiter of a CP system) to deliver the protection current efficiently providing adequate protection (i.e. meeting the criteria recommended in BS EN ISO 12696: 2012 and other International Standards). Most of the CP installations worldwide are operating in ICCP mode. However, due to the escalating cost of anode systems and associated external power supply as well as monitoring/control units for ICCP installation has led researchers to actively pursue different means of developing low cost anode systems. Researchers have mainly focused on sacrificial anode CP (SACP) systems, as SACP does not require an external power supply and control units, but the drawback to this anode system is that it has a shorter life span (usually 10 -15 years compared to 60+ years for ICCP anodes). This work describes the development of an ICCP anode system design utilising commercially available zinc rich paint (ZRP) as a primary anode material offering an innovative but considerably low cost alternative to currently used materials for ICCP anode systems. It also describes the development of a simple and low cost „multifunctional‟ probe for monitoring the performance of the installed CP system, among other functions, such as LPR measurements, macrocell corrosion current measurement, E-log I tests for assessing the current requirements for CP design. For these functions both laboratory investigations and field trial on real life structure were employed.

No description available.

本文根據32個國家的資料，觀察是否代理問題嚴重的公司傾向發放較多股利。以股權集中度作為代理問題之代理變數，發現保護好的國家，以股權分散的公司居多，而在保護不好的國家，大部分的公司股權結構十分集中。所以在保護好的國家，主要是股東與經理人之間的代理問題，股權越分散，代理問題越大；而保護不好的國家，其主要的代理問題是介於大股東與小股東之間，股權越集中，代理問題越大。 最後結論得出：在投資人保護好的國家，股利發放與股權集中度呈現正向關係；而在投資人保護差的國家，股利發放與股權集中度則呈現負向關係。代表代理問題嚴重的公司傾向發放較多的股利，由於公司有外部融資的需求，儘管在投資人保護差的國家，公司仍會藉由發放股利來減輕代理問題。 / We use a sample of companies from 32 countries around the world to shed light on the relationship between dividend payouts and insider holdings. Schooley and Barney (1994) indicate that negative relationship between dividend payouts and ownership exists before a turning point; once managerial ownership exceeds the turning point, it switches to a positive one. Also, as La Porta et al. (2000) point out, legal protection causes that widely-held firms dominate in a protected environment and closely-held firms in an unprotected one. So, most research addresses a negative relationship between dividend payouts and ownership. It is because their samples are usually composed of firms in a protected environment, where widely-held firms are the normality. By contrast, in countries with poor shareholder protection, closely-held companies prevail, and most of their ownership is coupled with control. Dividends thus become a tool for controlling shareholders to signal the moderation of expropriation and to reduce agency costs. Because firms with more concentrated ownership have less cost to pay dividends and need to allay the concerns about expropriation, it implies that more dividends are paid as the insider holdings increase. We surmise a positive relationship between ownership and payouts can be observed in an unprotected environment. The empirical evidence supports our hypothesis.