京都大学 / 新制・課程博士 / 博士(工学) / 甲第23183号 / 工博第4827号 / 新制||工||1754(附属図書館) / 京都大学大学院工学研究科都市環境工学専攻 / (主査)教授 酒井 伸一, 教授 田中 宏明, 准教授 平井 康宏 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

In dieser Doktorarbeit untersuchen wir zeitliche Langsamkeit als Prinzip für die Selbstorganisation des sensorischen Kortex sowie für computer-basierte Mustererkennung. Wir beginnen mit einer Einführung und Diskussion dieses Prinzips und stellen anschliessend den Slow Feature Analysis (SFA) Algorithmus vor, der das matemathisches Problem für diskrete Zeitreihen in einem endlich dimensionalen Funktionenraum löst. Im Hauptteil der Doktorarbeit untersuchen wir zeitliche Langsamkeit als Lernprinzip für rezeptive Felder im visuellen Kortex. Unter Verwendung von SFA werden Transformationsfunktionen gelernt, die, angewendet auf natürliche Bildsequenzen, möglichst langsam variierende Merkmale extrahieren. Die Funktionen können als nichtlineare raum-zeitliche rezeptive Felder interpretiert und mit Neuronen im primären visuellen Kortex (V1) verglichen werden. Wir zeigen, dass sie viele Eigenschaften von komplexen Zellen in V1 besitzen, nicht nur die primären, d.h. Gabor-ähnliche optimale Stimuli und Phaseninvarianz, sondern auch sekundäre, wie Richtungsselektivität, nicht-orthogonale Inhibition sowie End- und Seiteninhibition. Diese Resultate zeigen, dass ein einziges unüberwachtes Lernprinzip eine solche Mannigfaltigkeit an Eigenschaften begründen kann. Für die Analyse der mit SFA gelernten nichtlinearen Funktionen haben wir eine Reihe mathematischer und numerischer Werkzeuge entwickelt, mit denen man die quadratischen Formen als rezeptive Felder charakterisieren kann. Wir erweitern sie im weiteren Verlauf, um sie von allgemeinerem Interesse für theoretische und physiologische Modelle zu machen. Den Abschluss dieser Arbeit bildet die Anwendung des Prinzips der zeitlichen Langsamkeit auf Mustererkennungsprobleme. Die fehlende zeitliche Struktur in dieser Problemklasse erfordert eine Modifikation des SFA-Algorithmus. Wir stellen eine alternative Formulierung vor und wenden diese auf eine Standard-Datenbank von handgeschriebenen Ziffern an. / In this thesis we investigate the relevance of temporal slowness as a principle for the self-organization of the visual cortex and for technical applications. We first introduce and discuss this principle and put it into mathematical terms. We then define the slow feature analysis (SFA) algorithm, which solves the mathematical problem for multidimensional, discrete time series in a finite dimensional function space. In the main part of the thesis we apply temporal slowness as a learning principle of receptive fields in the visual cortex. Using SFA we learn the input-output functions that, when applied to natural image sequences, vary as slowly as possible in time and thus optimize the slowness objective. The resulting functions can be interpreted as nonlinear spatio-temporal receptive fields and compared to neurons in the primary visual cortex (V1). We find that they reproduce (qualitatively and quantitatively) many of the properties of complex cells in V1, not only the two basic ones, namely a Gabor-like optimal stimulus and phase-shift invariance, but also secondary ones like direction selectivity, non-orthogonal inhibition, end-inhibition and side-inhibition. These results show that a single unsupervised learning principle can account for a rich repertoire of receptive field properties. In order to analyze the nonlinear functions learned by SFA in our model, we developed a set of mathematical and numerical tools to characterize quadratic forms as receptive fields. We expand them in a successive chapter to be of more general interest for theoretical and physiological models. We conclude this thesis by showing the application of the temporal slowness principle to pattern recognition. We reformulate the SFA algorithm such that it can be applied to pattern recognition problems that lack of a temporal structure and present the optimal solutions in this case. We then apply the system to a standard handwritten digits database with good performance.

Although the German framework of incentive-based energy regulation appears to be well-defined because it draws from international experience and eliminates several present drawbacks existing with revenue-cap regulation doubts of industry and politicians concerning the application of benchmarking techniques and more importantly, the future industry structure, remain. The regulation scheme is based on experience with energy sector reforms in the US due to it being a precursor in this regard, e.g. by introducing incentive-based regulation as early as 1994 and publishing the relevant data in a very detailed way. This thesis therefore brings together the two issues of certain efficiency measurement problems and the industry restructuring in Germany and the US in order to contribute to the current discussion of robust benchmarking and to provide political implications related to the industry’s structure. Therefore, the application of DEA and SFA constitutes the heart of the whole study. These techniques are applied on German electricity DSOs as well as on data for US electricity DSOs and US gas TSOs. The application to US data can then be used to derive information about business strategies and their success in a sector that is being restructured. The US experience will lead to relevant German policy implications with respect to the future structure of the energy industry in Germany. / Der deutsche Regulierungsansatz erscheint im Allgemeinen als ausgewogen, da er zum einen internationale Erfahrungen mit einbezieht und zum anderen bestehenden Schwächen der Erlös-Obergrenzenregulierung berücksichtigt. Dennoch bestehen Zweifel seitens Industrie- und Politvertretern bezüglich der Anwendung von Benchmarking-Techniken und noch bedeutsamer der zukünftigen Industriestruktur. Der regulatorische Rahmen basiert auf Erfahrungen aus der Regulierung der Energiewirtschaft aus den USA, da diese eine Vorreiterrolle in dieser Hinsicht innehaben, beispielsweise aufgrund der frühen Einführung der Anreizregulierung in 1994 und der sehr detaillierten Veröffentlichung der relevanten Daten. Diese Arbeit verknüpft daher diese zwei Aspekte zum einen der Existenz gewisser Problemfelder bei der Effizienzmessung als auch der Umstrukturierung der deutschen Industriestruktur in der Energiewirtschaft. Damit trägt sie zu der aktuellen Diskussion mit Bezug auf robustes Benchmarking bei und liefert zudem politische Implikationen bezüglich der Industriestruktur. Zu diesem Zweck steht die Anwendung der für die deutsche Regulierung der Energienetze relevanten Benchmarking-Techniken (Data Envelopment Analysis (DEA) und Stochastic Frontier Analysis (SFA)) im Zentrum der gesamten Studie. Die empirischen Analysen beziehen sich dabei auf deutsche Elektrizitätsverteilnetzbetreiber sowie auf Elektrizitätsverteilnetzbetreibern und Ferngasgesellschaften aus den USA. Die Analysen von US Daten können dann genutzt werden, um gehaltvolle Informationen über Unternehmensstrategien und deren Erfolg in einem restrukturierten Sektor zu gewinnen. Diese Erfahrungen münden dann in Schlussfolgerungen und Empfehlungen für die deutsche Regulierung und Industriestruktur der deutschen Energienetzwirtschaft.

Although the German framework of incentive-based energy regulation appears to be well-defined because it draws from international experience and eliminates several present drawbacks existing with revenue-cap regulation doubts of industry and politicians concerning the application of benchmarking techniques and more importantly, the future industry structure, remain. The regulation scheme is based on experience with energy sector reforms in the US due to it being a precursor in this regard, e.g. by introducing incentive-based regulation as early as 1994 and publishing the relevant data in a very detailed way. This thesis therefore brings together the two issues of certain efficiency measurement problems and the industry restructuring in Germany and the US in order to contribute to the current discussion of robust benchmarking and to provide political implications related to the industry’s structure. Therefore, the application of DEA and SFA constitutes the heart of the whole study. These techniques are applied on German electricity DSOs as well as on data for US electricity DSOs and US gas TSOs. The application to US data can then be used to derive information about business strategies and their success in a sector that is being restructured. The US experience will lead to relevant German policy implications with respect to the future structure of the energy industry in Germany. / Der deutsche Regulierungsansatz erscheint im Allgemeinen als ausgewogen, da er zum einen internationale Erfahrungen mit einbezieht und zum anderen bestehenden Schwächen der Erlös-Obergrenzenregulierung berücksichtigt. Dennoch bestehen Zweifel seitens Industrie- und Politvertretern bezüglich der Anwendung von Benchmarking-Techniken und noch bedeutsamer der zukünftigen Industriestruktur. Der regulatorische Rahmen basiert auf Erfahrungen aus der Regulierung der Energiewirtschaft aus den USA, da diese eine Vorreiterrolle in dieser Hinsicht innehaben, beispielsweise aufgrund der frühen Einführung der Anreizregulierung in 1994 und der sehr detaillierten Veröffentlichung der relevanten Daten. Diese Arbeit verknüpft daher diese zwei Aspekte zum einen der Existenz gewisser Problemfelder bei der Effizienzmessung als auch der Umstrukturierung der deutschen Industriestruktur in der Energiewirtschaft. Damit trägt sie zu der aktuellen Diskussion mit Bezug auf robustes Benchmarking bei und liefert zudem politische Implikationen bezüglich der Industriestruktur. Zu diesem Zweck steht die Anwendung der für die deutsche Regulierung der Energienetze relevanten Benchmarking-Techniken (Data Envelopment Analysis (DEA) und Stochastic Frontier Analysis (SFA)) im Zentrum der gesamten Studie. Die empirischen Analysen beziehen sich dabei auf deutsche Elektrizitätsverteilnetzbetreiber sowie auf Elektrizitätsverteilnetzbetreibern und Ferngasgesellschaften aus den USA. Die Analysen von US Daten können dann genutzt werden, um gehaltvolle Informationen über Unternehmensstrategien und deren Erfolg in einem restrukturierten Sektor zu gewinnen. Diese Erfahrungen münden dann in Schlussfolgerungen und Empfehlungen für die deutsche Regulierung und Industriestruktur der deutschen Energienetzwirtschaft.

Le protocole pour le greffage irréversible du copolymère amphiphile polystyrène-b-poly (acrylate de sodium) PS-b-PANa, sur un substrat de mica et de silice hydrophobe a été développé, en utilisant la méthode de greffage à partir de solution. Les propriétés de surface du bloc chargé ont été évaluées. L’effet de la force ionique sur le gonflement des chaînes a été investigué par ellipsométrie. Les forces d’interaction entre les surfaces recouvertes du copolymère ont été évaluées par la technique SFA. Les profils de force ont démontré être stables et nettement répulsifs en compression et décompression, montrant l’irréversibilité du greffage. Les forces de frottement entre les brosses de PANa sont élevées, mais aucune évidence d’endommagement de la surface n’a été observée. La comparaison entre le comportement à la surface des chaînes de l’acide polyacrylique PAA et celles du PANa, obtenues par deux méthodes de greffage différentes, est également investiguée. / A protocol for irreversibly grafting of amphiphilic copolymer polystyrene-b-poly (sodium acrylate) PS-b-PANa onto hydrophobized mica and silica was developed, using the grafting to approach. Surface properties of charge block were evaluated. The swelling of chains and force ionic effect were studied by ellipsometry. The interaction forces and frictional forces were evaluated by SFA technique. Forces profiles were stable and clearly repulsive in loading and receding, indicating an irreversible grafting. High friction forces onto PANa brushes were determinate without evidence of damage at the surface. The comparison between solution behavior of polyacrylic acid, PAA and PANa brushes, obtained via different grafting methods, was equally investigated.

Les recherches présentées dans ce mémoire s'intéressent à la désorption du tensioactif dans les films de latex en cours de séchage. Les travaux concernent d'une part l'élaboration d'un système expérimental modélisant le rapprochement de deux particules de latex recouvertes de tensioactif pour accéder aux paramètres influençant la désorption, et d'autre part l'étude de ce phénomène directement sur un latex. Un protocole pertinent de fabrication d'un film polymère mince déposé sur du mica, de faible rugosité, et résistant à une immersion prolongée dans un milieu aqueux a été obtenu. L'adsorption d'un tensioactif anionique, le dodécyl sulfate de sodium (SDS) sur un film polymère de poly (méthacrylate d'éthyle) (PEMA) est investiguée par microscopie à force atomique et indique la présence d'agrégats de surface dont la densité croit lorsque la concentration en tensioactif dans la solution dépasse la concentration micellaire critique (CMC). Le profil des interactions s'exerçant entre deux surfaces de polymère immergées en solutions aqueuses de tensioactif en fonction de la séparation entre les interfaces est obtenu par la technique de mesure des forces de surface (surface force apparatus, SFA), et permet de répondre à la question de l'occurrence ou non de la désorption du tensioactif (SDS) lorsque le milieu aqueux intermédiaire séparant les interfaces de PEMA est réduit jusqu'au contact physique, comme dans le mécanisme de séchage lorsque les colloïdes de latex se rapprochent au fur et à mesure que l'eau disparaît. L'étude de la désorption du tensioactif dans un système réel, à savoir lors du séchage d'un film de latex, est menée par mesures d'anisotropie de fluorescence et mesures par résonateur de torsion et montre la simultanéité entre la perte de mobilité de la sonde fluorescente hydrophile et l'évaporation de l'eau, augurant d'un comportement analogue pour le tensioactif.

This thesis has the dual purpose of raising awareness of the importance of the mixing protocol on the end products of polyelectrolyte-oppositely charged surfactant systems, and to contribute to a better understanding of the properties of bottle-brush polyelectrolytes when adsorbed onto interfaces. In the first part of this thesis work, the effects of the mixing protocol and the mixing procedure on formed polyelectrolyte-oppositely charged surfactant aggregates were investigated. It was shown that the initial properties of the aggregates were highly dependent on the mixing parameters, and that the difference between the resulting aggregates persisted for long periods of time. The second part of the studies was devoted to the surface properties of a series of bottle-brush polyelectrolytes made of charged segments and segments bearing poly(ethylene oxide) side chains; particular attention was paid to the effect of side chain to charge density ratio of the polyelectrolytes. It was shown that the adsorbed mass of the polyelectrolytes, and the corresponding number of poly(ethylene oxide) bearing segments at the interface, went through a maximum as the charge density of the polyelectrolyte was increased. Also, it was found that bottle-brush polyelectrolyte layers were desorbed quite easily when subjected to salt solutions. This observation was rationalized by the unfavourable excluded volume interactions between the side chains and the entropic penalty of confining them at an interface, which weaken the strength of the binding of the polyelectrolytes to the interface. However, it was shown that the same side chains effectively protect the adsorbed layer against desorption when the layer is exposed to solutions containing an oppositely charged surfactant. Investigation of the lubrication properties of the bottle-brush polyelectrolytes in an asymmetric (mica-silica) system also related the observed favourable frictional properties to the protective nature of the side chains. The decisive factor for achieving very low coefficients of friction was found to be the concentration of the side chains in the gap between the surfaces. Interestingly, it was shown that a brush-like conformation of the bottle-brush polyelectrolyte at the interface has little effect on achieving favourable lubrication properties. However, a brush-like conformation is vital for the resilience of the adsorbed layer against the competitive adsorption of species with a higher surface affinity. / QC 20100830

Waste management has gone through a history of shiftingproblems, demands, and strategies over the years. In contrastto the long prevailing view that the problem could be solved byhiding or moving it, waste is now viewed as a problem rangingfrom local to global concern, and as being an integral part ofseveral sectors in society. Decisive for this view has beensociety’s increasing complexity and thus the increasingcomplexity of waste, together with a general development ofenvironmental consciousness, moving from local focus on pointemission sources, to regional and global issues of more complexnature. This thesis is about the development and application orware;a model for computer aided environmental systems analysis ofmunicipal waste management. Its origin is the hypothesis thatwidened perspectives are needed in waste managementdecision-making to avoid severe sub-optimisation ofenvironmental performance. With a strong foundation in lifecycle assessment (LCA), orware aims to cover the environmentalimpacts over the entire life cycle of waste management. It alsoperforms substance flow analysis (SFA) calculations at a ratherdetailed level of the system. Applying orware has confirmed the importance of applyingsystems perspective and of taking into account site specificdifferences in analysis and planning of waste manage-ment,rather than relying on overly simplified solutions. Somefindings can be general-ised and used as guidelines to reduceenvironmental impact of waste management. Recovery of materialand energy resources from waste generally leads to netreductions in energy use and environmental impact, because ofthe savings this brings about in other sectors. Waste treatmentwith low rate of energy and materials recovery should thereforebe avoided. The exact choice of technology however depends onwhat products can be recovered andhow they are used. Despite the complexity of the model and a certain degree ofuser unfriendliness, involved stakeholders have expressed thevalue of participating in orware case studies. It providesimproved decision-basis, but also wider understanding of thecomplexity of waste management and of environmental issues ingeneral. The thesis also contains a first suggestion of a frameworkto handle uncertainty in orware, based on a review of types ofuncertainty in LCA and tools to handle it. / QC 20100413

The structure and dynamics of the smectic-A liquid crystal 8CB (4 cyano-4 octylbiphenyl) when sheared and confined to mesoscale gaps (with crossed cylindrical geometry and mica confining surfaces) were studied using a Surface Forces Apparatus (SFA). Triangular shear patterns with frequencies of 0.01, 0.1, 1.0 and 10 Hz, and amplitudes of 62.5 nm, 625 nm and 6.25 m were applied to samples at gap sizes of 0.5 and 5.0 m. The study was performed at room temperature (20.5C) and at two higher temperatures (22C and 27C). In order to minimize the thermal fluctuations within the test chamber and hence to allow for the rapid re-initialization of test runs, the SFA was modified to allow for quick, precise and remote control of the confining surfaces. The procedure maximized the number of tests that could be undertaken with a single pair of surfaces so that a single gap geometry could be maintained for the duration of the test run. In order to run the SFA remotely, scripts written with a commercial software package, LabVIEW, were used to control of the SFA components, its FECO-monitoring camera and all its peripheral electronic equipment as well. Samples were agitated to disrupt any shear-induced liquid crystal domain alignment from previous testing following each shear test, and methodologies were developed to ascertain the extent of confinement quickly and remotely following agitation. Separate methods were developed for gap sizes at each extreme of the mesoscale regime, where the transition from bulklike structure and dynamics to nano-confinement occurs (between 1 and 10 microns for smectic-A 8CB). The results revealed that the greater amplitude-gap aspect ratio and surface-to-domain contact associated with smaller gaps facilitated reorientation of the domains in the shear direction. Evidence was also presented of domains at the higher end or outside of the mesoscale regime that, while straining and accreting, were unable to reorient and thereby led to an overall increase of viscoelastic response. The effective viscosity was found to obey a simple power law with respect to shear rate, , and the flow behaviour indices, n, slightly in excess of unity indicate shear thickening occurs with large enough shear amplitude, and that the viscosity reached a plateau near unity over shear rates of 0.005 to 500 s-1 within the mesoscale regime. Different K and n values were observed depending on the shear amplitude used. Unlike bulk smectic 8CB, whose domains do not align well in the shear direction with large shear-strain amplitude, at mesoscale levels of confinement large amplitude shearing (up to 12.5 shear strain amplitude) was found to be very effective at aligning domains. In general domain reorientation is found to be much more rapid within the mesoscale regime than has been reported in bulk. Aggressive shearing was found to result in a complete drop in viscoelastic response within seconds, while gentler shearing is found to produce a very gradual increase that persists for more than six hours, with individual shear periods exhibiting frequent and significant deviations from the expected smooth shear path that may be a product of discrete domain reorientations. From these findings, certain traits of the smectic 8CB domain structures under mesoscale confinement were deduced, including how they respond to shear depending on the level of confinement, and how their reorientation due to shear varies not only with shear rate but also independently with shear amplitude. An equation describing the viscosity change as a function of both shear rate and shear amplitude is proposed. The shear amplitude dependence introduces the notion of shearing beyond the proposed smectic 8CB “viscoelastic limit”, which was shown to exhibit behaviour in accordance with Large Amplitude Oscillatory Shear (LAOS) techniques developed for Fourier Transform rheology. The findings provided an understanding of the behavioural changes that occur as one reduces the level of confinement of smectic materials from bulk to nanoconfinement.

Ce travail a pour objectif d'étudier, de manière expérimentale et numérique, la condition<br />limite hydrodynamique pour les écoulements de liquides simples en présence d'une paroi<br />solide.<br />Dans une première partie nous avons étudié ce problème pour différents systèmes, à<br />l'aide d'une machine à forces de surfaces dynamique. Nous avons mis en évidence que l'état<br />des surfaces et la nature de l'interaction solide-liquide sont des paramètres déterminants<br />pour la condition limite hydrodynamique. Nous n'obtenons du glissement que dans le cas de<br />liquides non mouillants. Pour un fluide confiné entre des surfaces hydrophobes lisses, nous<br />montrons que l'écoulement est très bien décrit par une condition limite de glissement dont<br />la longueur caractéristique est indépendante du confinement et du taux de cisaillement.<br />Dans une deuxième partie nous nous sommes intéressés à étudier, par des simulations de<br />dynamique moléculaire, l'effet conjugué de la rugosité et de la mouillabilité sur la condition<br />limite. Nous avons montré que selon la pression et la forme de la rugosité, la présence<br />de cette dernière peut augmenter fortement le glissement ou au contraire le diminuer.<br />Enfin nous avons étudié l'influence