Concern regarding sustainability and climate change is increasing, which is forcing countries world-wide to take action. The Swedish government has set a goal of fossil-free tra_c until 2030. Battery Electric Buses (BEB) might be one of the solutions needed in order to reach this goal. However, currently its prevalence is at an early stage.   The purpose of this study is to investigate how the technological transition towards BEBs in Sweden a_ects the public transport operators (PTOs). Moreover, to investigate how a third party service provider of Fleet Management System (FMS) services can support the PTOs in this transition.   The research has been carried out in co-operation with a PTO and a FMS service provider. The research contributes to their current understanding of how they will be a_ected by the emerging technological transition. This thesis also contributes with new empirical data of the technological transition towards electric vehicles within public bus transport, seen as a Large Technical System. Conceptually it contributes, by exploring how external companies can support the technological transition towards BEBs, with the application of Technological Transitions theory and the Multi Layer Perspective framework.   The methodology used is a case study of the technological transition towards BEBs in Sweden. Data was collected through twelve semi-structured interviews with researchers, PTOs, public transport authorities (PTA), a BEB manufacturer and a FMS-service company. Parallel to this a questionnaire was distributed to the twenty largest PTOs in Sweden. Moreover data was collected from company visits, pilot-project results and internal documentation.   Our findings show that there are thirteen perceived barriers present among the PTOs, in the process of BEB adoption. Six of these barriers relate to component aspects of BEBs, and seven relate to managerial aspects. Perceived barriers linked to component aspects of BEBs are; Variation in solutions and lack of technical standards, the Charging infrastructure, Shorter range or decreased load capacity, Unknown functionality in cold climate, Reliability and Durability. Perceived barriers linked to managerial aspects of BEBs are; Lack of knowledge and experience, Behavioral change, Economy, Maintenance, Ownership of infrastructure and buses, Business models and Varying requirements from PTAs. The barriers FMS-service providers can address are primarily, due to the technological nature of the services, present at niche level. PTOs together with FMS-service providers are encouraged to together strive towards gaining deeper knowledge about the new emerging echnologies. Through this, PTOs could be enabled to overcome the aforementioned barriers. Three reverse salients were also identi_ed, linked to the aforementioned barriers. If the everse salients are assessed, BEB acceptance among PTOs could be increased. The three identi_ed reverse salients are, the battery technology, the charging infrastructure and the contracts/ownership. The co-operation with the commissioning PTO and FMS-service provider has led to valuable access to Swedish public transport actors, and has aided in a deeper understanding of the phenomena. Although, this co-operation might have exposed us to a risk of being influenced.

Organizations gather enormous amounts of data and analyze these data to extract insights that can be useful for them and help them to make better decisions. Predictive data analytics is a crucial subfield within data analytics that make accurate predictions. Predictive data analytics extracts insights from data by using machine learning algorithms. This thesis presents the supervised learning algorithm to perform predicative data analytics in Embedded Test System at the Nordic Engineering Partner company. Predictive Maintenance is a concept that is often used in manufacturing industries which refers to predicting asset failures before they occur. The machine learning algorithms used in this thesis are support vector machines, multi-layer perceptrons, random forests, and gradient boosting. Both binary and multi-class classifier have been provided to fit the models, and cross-validation, sampling techniques, and a confusion matrix have been provided to accurately measure their performance. In addition to accuracy, recall, precision, f1, kappa, mcc, and roc auc measurements are used as well. The prediction models that are fitted achieve high accuracy.

Power dividers are vital components and widely used in radio technology, such as antenna arrays, power amplifiers, multiplexers and mixers. A good example is the well-known Wilkinson power divider with its distinctive feeding network characteristics. A comprehensive review indicated that limited research is carried out in the area of planar multiport and reconfigurable power dividers in terms of the power levels between output ports. The main objectives of this work were to develop a small size power divider, a planer multi-output ports power divider and a power divider with a reconfigurable power division ratio. These power dividers were designed to operate over either an ultra-wideband frequency (3.1-10.6 GHz) or WLAN bands (2.4 or 5.2 GHz). A novel multi-layered topology solved the complexity of interconnecting isolation resistors by introducing an additional layer below the ground layer. The prototype was fabricated and tested to validate the results. The measurements and simulation were in good agreement. Finally, a novel uniplanar power divider with reconfigurable output power level difference was developed. The configurability feature was achieved by tuning the quarter wave transformer using one varactor diode. The power divider was applied to improve a full duplex system cancellation performance at the receiver element caused by interference from in-site transmitting antennas. This study investigated fixed power dividers, multi-output power dividers and reconfigurable power dividers. The measurements validated by the simulation results and applications proved the designed power dividers could be used in practical applications. / Higher Committee for Education Development (HCED), Iraq

Electricity spot prices are difficult to predict since they depend on different unstable and erratic parameters, and also due to the fact that electricity is a commodity that cannot be stored efficiently. This results in a volatile, highly fluctuating behavior of the prices, with many peaks. Machine learning algorithms have outperformed traditional methods in various areas due to their ability to learn complex patterns. In the last decade, deep learning approaches have been introduced in electricity spot price prediction problems, often exceeding their predecessors. In this thesis, several deep learning models were built and evaluated for their ability to predict the spot prices 10-days ahead. Several conclusions were made. Firstly, it was concluded that rather simple neural network architectures can predict prices with high accuracy, except for the most extreme sudden peaks. Secondly, all the deep networks outperformed the benchmark statistical model. Lastly, the proposed LSTM and CNN provided forecasts which were statistically, significantly superior and had the lowest errors, suggesting they are the most suitable for the prediction task. / Elspotspriser är svåra att förutsäga eftersom de beror på olika instabila och oregelbundna faktorer, och också på grund av att elektricitet är en vara som inte kan lagras effektivt. Detta leder till ett volatilt, fluktuerande beteende hos priserna, med många plötsliga toppar. Maskininlärningsalgoritmer har överträffat traditionella metoder inom olika områden på grund av deras förmåga att lära sig komplexa mönster. Under det senaste decenniet har djupinlärningsmetoder introducerats till problem inom elprisprognostisering och ofta visat sig överlägsna sina föregångare. I denna avhandling konstruerades och utvärderades flera djupinlärningsmodeller på deras förmåga att förutsäga spotpriserna 10 dagar framåt. Den första slutsatsen är att relativt simpla nätverksarkitekturer kan förutsäga priser med hög noggrannhet, förutom för fallen med de mest extrema, plötsliga topparna. Vidare, så övertränade alla djupa neurala nätverken den statistiska modellen som användes som riktmärke. Slutligen, så gav de föreslagna LSTM- och CNN-modellerna prognoser som var statistiskt, signifikant överlägsna de andra och hade de lägsta felen, vilket tyder på att de är bäst lämpade för prognostiseringsuppgiften.

Computer simulation is a commonly applied technique for studying electronic warfare duels. This thesis aims to apply machine learning techniques to convert simulation output data into knowledge and insights regarding defensive actions for a ship facing multiple hostile missiles. The analysis may support tactical decision-making, hence the interpretability aspect of predictions is necessary to allow for human evaluation and understanding of impacts from the explanatory variables. The final distance for the threats to the target and the probability of the threats hitting the target was modeled using a multi-layer perceptron model with a multi-task approach, including custom loss functions. The results generated in this study show that the selected methodology is more successful than a baseline using regression models. Modeling the outcome with artificial neural networks results in a black box for decision making. Therefore the concept of interpretable machine learning was applied using a post-hoc approach. Given the learned model, the features considered, and the multiple threats, the feature contributions to the model were interpreted using Kernel SHapley Additive exPlanations (SHAP). The method consists of local linear surrogate models for approximating Shapley values. The analysis primarily showed that an increased seeker activation distance was important, and the increased time for defensive actions improved the outcomes. Further, predicting the final distance to the ship at the beginning of a simulation is important and, in general, a guidance of the actual outcome. The action of firing chaff grenades in the tracking gate also had importance. More chaff grenades influenced the missiles' tracking and provided a preferable outcome from the defended ship's point of view.

Tesis por compendio / This dissertation focused in the comprehensive analysis of the capabilities of some non-tested types of Artificial Neural Networks, specifically: the Probabilistic Neural Networks (PNN) and the Multi-Layer Perceptron (MLP) Ensembles. The analysis of the capabilities of these techniques was performed using the native brown trout (Salmo trutta; Linnaeus, 1758), the bermejuela (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) and the redfin barbel (Barbus haasi; Mertens, 1925) as target species. The analyses focused in the predictive capabilities, the interpretability of the models and the effect of the excess of zeros in the training datasets, which for presence-absence models is directly related to the concept of data prevalence (i.e. proportion of presence instances in the training dataset). Finally, the effect of the spatial scale (i.e. micro-scale or microhabitat scale and meso-scale) in the habitat suitability models and consequently in the e-flow assessment was studied in the last chapter. / Esta tesis se centra en el análisis comprensivo de las capacidades de algunos tipos de Red Neuronal Artificial aún no testados: las Redes Neuronales Probabilísticas (PNN) y los Conjuntos de Perceptrones Multicapa (MLP Ensembles). Los análisis sobre las capacidades de estas técnicas se desarrollaron utilizando la trucha común (Salmo trutta; Linnaeus, 1758), la bermejuela (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) y el barbo colirrojo (Barbus haasi; Mertens, 1925) como especies nativas objetivo. Los análisis se centraron en la capacidad de predicción, la interpretabilidad de los modelos y el efecto del exceso de ceros en las bases de datos de entrenamiento, la así llamada prevalencia de los datos (i.e. la proporción de casos de presencia sobre el conjunto total). Finalmente, el efecto de la escala (micro-escala o escala de microhábitat y meso-escala) en los modelos de idoneidad del hábitat y consecuentemente en la evaluación de caudales ambientales se estudió en el último capítulo. / Aquesta tesis se centra en l'anàlisi comprensiu de les capacitats d'alguns tipus de Xarxa Neuronal Artificial que encara no han estat testats: les Xarxes Neuronal Probabilístiques (PNN) i els Conjunts de Perceptrons Multicapa (MLP Ensembles). Les anàlisis sobre les capacitats d'aquestes tècniques es varen desenvolupar emprant la truita comuna (Salmo trutta; Linnaeus, 1758), la madrilla roja (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) i el barb cua-roig (Barbus haasi; Mertens, 1925) com a especies objecte d'estudi. Les anàlisi se centraren en la capacitat predictiva, interpretabilitat dels models i en l'efecte de l'excés de zeros a la base de dades d'entrenament, l'anomenada prevalença de les dades (i.e. la proporció de casos de presència sobre el conjunt total). Finalment, l'efecte de la escala (micro-escala o microhàbitat i meso-escala) en els models d'idoneïtat de l'hàbitat i conseqüentment en l'avaluació de cabals ambientals es va estudiar a l'últim capítol. / Muñoz Mas, R. (2016). Multivariate approaches in species distribution modelling: Application to native fish species in Mediterranean Rivers [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/76168 / Compendio

Les systèmes de communications sans fils actuels imposent des contraintes très sévères en termes de la capacité du canal, la qualité de transmission tout en gardant les niveaux d'interférences et multi-trajets assez faibles. De telles contraintes ont rendu les antennes multifaisceaux un élément essentiel dans ces systèmes. Parmi les techniques permettant de réaliser une antenne multifaisceaux (sans avoir recours aux systèmes à balayages électroniques), un réseau d'antennes élémentaires est associé à un réseau d'alimentation (une matrice) à formation de faisceau (Beam Forming Network-BFN). Parmi les différents types de ces matrices, la matrice de Butler a reçu une attention particulière. Ceci est dû au fait qu'elle est théoriquement sans pertes et qu'elle emploie un nombre minimum de composants (coupleurs et déphaseurs) afin de générer l'ensemble de faisceaux orthogonaux demandé (avec l'hypothèse que le nombre de faisceau est une puissance de 2). Néanmoins, la matrice de Butler a un problème de conception majeur. Ce problème réside dans la structure de la matrice qui renferme des croisements ce qui a été adressé par différents travaux de recherches dans la littérature. Les Guide Intégré au Substrat (GIS) offrent des caractéristiques intéressants pour la conception des composants microondes et millimétriques faciles à intégrer sur un même support avec d'autres composants planaires. Les composants à base de GIS combinent les avantages des guides d'ondes rectangulaires, comme leur grand facteur de qualité Q, leur faibles pertes tout en étant compatible avec les technologies à faibles coûts comme le PCB et le LTCC. Vus ses caractéristiques attrayants, la technologie GIS devient un bon candidat pour la réalisation des matrices multifaisceaux faciles à intégrer avec d'autres systèmes en technologies planaires ou à base de guide GIS. Dans cette thèse, de nouveaux composants passifs sont développés en exploitant la technologie GIS en multicouches en vue de la réalisation d'une matrice de Butler 4x4 compacte et large bande. Les composants recherchés sont donc des coupleurs et des déphaseurs ayant des performances large bande en termes des amplitudes des coefficients de transmissions et les phases associés tout en gardant de faibles niveaux de pertes et de bonnes isolations. Différents techniques pour l'implémentation de déphaseurs large bande en technologie GIS sont présentés. Une nouvelle structure à base d'une propagation composite : main gauche main droite (Composite Right/Left- Handed, CRLH) dans un guide d'onde est proposée. La structure consiste d'un guide d'onde monocouche ayant des fenêtres inductives et des fentes transversales à réactances capacitives pour synthétiser l'inductance parallèle et la capacité série main gauche, respectivement. La structure est adaptée pour les réalisations de déphaseurs compacts en technologie GIS. Bien que les pertes d'insertions restent dans le même ordre de grandeur de celles des structures CRLH à base d'éléments non-localisés, ces niveaux de pertes restent relativement grands par rapport aux applications nécessitant plusieurs déphaseurs. Les déphaseurs à bases de GIS ayant des longueurs égales et des largeurs variables sont ensuite abordés. Ce type de déphaseur est effectivement très adapté à la technologie GIS qui permet des réalisations de parcours avec différentes formes (parcours droits, courbés, coudés, ..) tout en assurant des différences de phase large bande. Afin de satisfaire de faibles pertes d'insertions pour une large dynamique de phase, la longueur de ces déphaseurs est en compromis avec les variations progressives des différentes largeurs associées aux valeurs de déphasages requises. Une transition large bande, double couche et à faible perte est ainsi proposée. La transition est analysée à partir de son circuit électrique équivalent afin d'étudier les performances en termes de l'amplitude et la phase du coefficient de transmission par rapport aux différents paramètres structurels de la transition. Cette transition est ensuite exploitée pour développer un déphaseur à trois couches, large bande, en GIS. La structure consiste effectivement d'un guide d'onde replié à plusieurs reprises sur luimême selon la longueur dans une topologie trois couches à faibles pertes. De nouveaux coupleurs double couche en GIS sont également proposés. Pour les applications BFNs, une structure originale d'un coupleur large bande est développée. La structure consiste de deux guides d'onde parallèles qui partagent leur grand mur ayant une paire de fentes inclinées et décalées par rapport au centre de la structure. Une étude paramétrique détaillée est faite pour étudier l'impact des différents paramètres des fentes sur l'amplitude et la phase du coefficient de transmission. Le coupleur proposé a l'avantage d'assurer une large dynamique de couplage ayant des performances larges bandes en termes des amplitudes et les phases des coefficients de transmission avec de faibles pertes et de bonnes isolations entre le port d'entré et celui isolé. D'autre part, contrairement à d'autres travaux antérieurs et récents qui souffraient d'une corrélation directe entre la phase en transmission et le niveau de couplage, la structure proposée permet de contrôler le niveau de couplage en maintenant presque les mêmes valeurs de phase en transmission pour différents niveaux de couplage. Ceci le rend un bon candidat pour les BFNs déployant différents coupleurs telle la matrice de Nolen. Une deuxième structure originale d’un coupleur bibande est également proposée. La structure consiste de deux coupleurs concentriques en guide nervuré intégré au substrat avec un motif innovant de démultiplexage à base de GIS. Ce coupleur a été développé conjointement avec M. Tarek Djerafi de l’Ecole Polytechnique de Montréal dans un cadre de collaboration avec le Prof. Ke Wu. Finalement, pour l'implémentation de la matrice de Butler, la topologie double couche est explorée à deux niveaux. Le premier consiste à optimiser les caractéristiques électriques de la matrice, tandis que le second concerne l'optimisation de la surface occupée afin de rendre la matrice la plus compacte possible sans dégrader ses performances électriques. D'une part, la structure double couche présente une solution intrinsèque au problème de croisement permettant ainsi une plus grande flexibilité pour la compensation de phase sur une large bande de fréquence. Ceci est réalisé par une conception adéquate de la surface géométrique sur chaque couche de substrat et optimiser les différentes sections de GIS avec les différents parcours adoptés. La deuxième étape consiste effectivement à optimiser la surface sur chaque couche en profitant de la technologie GIS. Ceci consiste à réaliser des murs latéraux communs entre différents chemin électrique de la matrice en vue d'une compacité optimale. Les deux prototypes de matrices de Butler 4x4 sont optimisés, fabriqués et mesurés. Les résultats de mesures sont en bon accord avec ceux de la simulation. Des niveaux d'isolations mieux que - 15 dB avec des niveaux de réflexions inférieurs à -12 dB sont validés expérimentalement sur plus de 24% de bande autour de 12.5 GHz. Les coefficients de transmission montrent de faibles dispersions d'environ 1 dB avec une moyenne de -6.8 dB, et 10° par rapport aux valeurs théoriques, respectivement, sur toute la bande de fréquence. / Multibeam antennas have become a key element in nowadays wireless communication systems where increased channel capacity, improved transmission quality with minimum interference and multipath phenomena are severe design constraints. These antennas are classified in two main categories namely adaptive smart antennas and switched-beam antennas. Switched-beam antennas consist of an elementary antenna array connected to a Multiple Beam Forming Network (M-BFN). Among the different M-BFNs, the Butler matrix has received particular attention as it is theoretically lossless and employs the minimum number of components to generate a given set of orthogonal beams (provided that the number of beams is a power of 2). However, the Butler matrix has a main design problem which is the presence of path crossings that has been previously addressed in different research works. Substrate Integrated Waveguide (SIW) features interesting characteristics for the design of microwave and millimetre-wave integrated circuits. SIW based components combine the advantages of the rectangular waveguide, such as the high Q factor (low insertion loss) and high power capability while being compatible with low-cost PCB and LTCC technologies. Owing to its attractive features, the use of SIW technology appears as a good candidate for the implementation of BFNs. The resulting structure is therefore suitable for both waveguide-like and planar structures. In this thesis, different novel passive components (couplers and phase shifters) have been developed exploring the multi-layer SIW technology towards the implementation of a two-layer compact 4×4 Butler matrix offering wideband performances for both transmission magnitudes and phases with good isolation and input reflection characteristics. Different techniques for the implementation of wideband fixed phase shifters in SIW technology are presented. First, a novel waveguide-based CRLH structure is proposed. The structure is based on a single-layer waveguide with shunt inductive windows (irises) and series transverse capacitive slots, suitable for SIW implementations for compact phase shifters. The structure suffers relatively large insertion loss which remains however within the typical range of non-lumped elements based CRLH implementations. Second, the well-known equal length, unequal width SIW phase shifters is discussed. These phase shifters are very adapted for SIW implementations as they fully exploit the flexibility of the SIW technology in different path shapes while offering wideband phase characteristics. To satisfy good return loss characteristics with this type of phase shifters, the length has to be compromised with respect to the progressive width variations associated with the required phase shift values. A twolayer, wideband low-loss SIW transition is then proposed. The transition is analyzed using its equivalent circuit model bringing a deeper understanding of its transmission characteristics for both amplitude and phase providing therefore the basic guidelines for electromagnetic optimization. Based on its equivalent circuit model, the transition can be optimized within the well equal-length SIW phase shifters in order to compensate its additional phase shift within the frequency band of interest. This twolayer wideband phase shifter scheme has been adopted in the final developed matrix architecture.This transition is then exploited to develop a three-layer, multiply-folded waveguide structure as a good candidate for compensated-length, variable width, low-loss, compact wideband phase shifters in SIW technology. Novel two-layer SIW couplers are also addressed. For BFNs applications, an original structure for a two-layer 90° broadband coupler is developed. The proposed coupler consists of two parallel waveguides coupled together by means of two parallel inclined-offset resonant slots in their common broad wall. A complete parametric study of the coupler is carried out including the effect of the slot length, inclination angle and offset on both the coupling level and the transmission phase. The first advantage of the proposed coupler is providing a wide coupling dynamic range by varying the slot parameters allowing the design of wideband SIW Butler matrix in two-layer topology. In addition, previously published SIW couplers suffer from direct correlation between the transmission phase and the coupling level, while the coupler, hereby proposed, allows controlling the transmission phase without significantly affecting the coupling level, making it a good candidate for BFNs employing different couplers, such as, the Nolen matrix. A novel dual-band hybrid ring coupler is also developed in multi-layer Ridged SIW (RSIW) technology. This coupler has been jointly developed with Tarek Djerafi in a collaboration scenario with Prof. Ke Wu from the Ecole Polytechnique de Montréal. The coupler has an original structure based on two concentric rings in RSIW topology with the outer ring periodically loaded with radial, stub-loaded transverse slots. A design procedure is presented based on the Transverse Resonance Method (TRM) of the ridged waveguide together with the simple design rules of the hybrid ring coupler. A C/K dual band coupler with bandwidths of 8.5% and 14.6% centered at 7.2 GHz and 20.5 GHz, respectively, is presented. The coupler provides independent dual band operation with low-dispersive wideband operation. Finally, for the Butler matrix design, the two-layer SIW implementation is explored through a two-fold enhancement approach for both the matrix electrical and physical characteristics. On the one hand, the two-layer topology allows an inherent solution for the crossing problem allowing therefore more flexibility for phase compensation over a wide frequency band. This is achieved by proper geometrical optimization of the surface on each layer and exploiting the SIW technology in the realization of variable width waveguides sections with the corresponding SIW bends. On the other hand, the two-layer SIW technology is exploited for an optimized space saving design by implementing common SIW lateral walls for the matrix adjacent components seeking maximum size reduction. The two corresponding 4×4 Butler matrix prototypes are optimized, fabricated and measured. Measured results are in good agreement with the simulated ones. Isolation characteristics better than -15 dB with input reflection levels lower than -12 dB are experimentally validated over 24% frequency bandwidth centered at 12.5 GHz. Measured transmission magnitudes and phases exhibit good dispersive characteristics of 1dB, around an average value of -6.8 dB, and 10° with respect to the theoretical phase values, respectively, over the entire frequency band.

Die vorliegende Arbeit beschreibt die Übertragung der aus der Siliziumtechnologie bekannten Präzision der Strukturierung und die Zuverlässigkeit der Verbindungstechnologie auf andere Materialien wie Kupfer und PMMA. Diese Untersuchung ist auf die Entwicklung der Teiltechnologien Strukturierung und Integration fokussiert und konzentriert sich insbesondere auf die Kombination von Mikrostrukturierung und dreidimensionalen Aufbautechniken einschließlich vertikaler fluidischer Durchkontaktierungen bei den Materialien Silizium, Kupfer und Kunststoff (PMMA). Eine begleitende Charakterisierung und messtechnische Bewertung gestattet die Weiterentwicklung während der Experimentedurchführung und erweitert den Stand der Wissenschaft hinsichtlich der genannten Kombinationen. / The work describes the transfer of well known high precisive and reliable micro technologies for patterning and packaging of Silicon to new materials like Copper and PMMA. This investigation is focused on special patterning technologies and system integration aspects. Furthermore the development of material-dependent micro patterning technologies and multi layer packaging techniques including vertical fluidic interconnects using materials like Silicon, Copper, and PMMA (polymer) is shown. An accompanying characterization and measurement-based evaluation enables the ongoing development while performing experimental analysis. At least a higher state of the art for these complex combinations is reached.

Composite thin Pd/Ag alloy membranes with long-term thermal and chemical stabilities have potential applications for H2 separation via catalytic membrane reactors and may be one of the key determinants to achieve the 21st century's global hydrogen economy. This work provides a detailed microstructure characterization study and a better understanding of the fundamental principles involved in the synthesis of a novel Pd/Ag intermetallic diffusion barrier formed by the bi-metal multi-layer (BMML) deposition technique. The BMML deposition technique formed an extremely effective Pd/Ag intermetallic diffusion barrier and significantly improved the thermal and long-term stability of the composite Pd and Pd/alloy membranes over a temperature range of 500-600oC. In addition, high temperature annealing studies over a temperature range of 500-800oC in H2 atmosphere led a thorough understanding of the surface interactions and the phase changes between the Pd and Ag metals and the porous metal support elements (Fe, Cr and Ni) and it was shown by the SEI, EDX and X-ray phase analyses that the Ag/Fe and Ag/Ni binary systems exerted complete immiscibility compared to the completely miscible solid solutions of Pd/Fe and Pd/Ni phases. A novel characterization method of in-situ time-resolved high temperature X-ray diffraction (HTXRD) analysis was used to elucidate the mechanistic details of the isothermal nucleation and growth kinetics of the Pd/Ag alloy phase over a temperature range of 500-600oC in H2. The nucleation of the Pd/Ag alloy phase was instantaneous where the growth mechanism was through diffusion-controlled one-dimensional thickening of the Pd/Ag alloy layer. The Pd/Ag alloy phase growth was strongly dependent upon the deposition morphology of the as-synthesized Pd and Ag layers due to the presence of the heterogeneous nucleation sites. Based on the empirical rate constants derived from the solid-state reaction models, the estimated activation energies for the Pd/Ag alloy phase transformation were 236.5 and 185.6 kJ/mol and in good agreement with the literature values of 183-239.5 kJ/mol. The successful utilization of surface modification techniques and modified plating conditions led to the synthesis of several dense Pd/Ag layers, which were as thin as 5-15 µm with a bulk Ag content in the 10-40 wt% range. The long-term testing of the composite Pd/Ag membranes (5-15 µm) at 500oC showed stable hydrogen permeances as high as 30 to 54 m3/m2-h-atm0.5 with H2/He selectivities ranging from 200 to 14000. Furthermore, the atomic absorption flame analysis was used for the first time to elucidate the effects of temperature, initial metal ion concentration, initial hydrazine concentration and bath agitation on the electroless plating rates of Pd and Ag. The electroless plating of both Pd and Ag were strongly affected by the external mass transfer in the absence of bath agitation. The external mass transfer limitations for both Pd and Ag deposition have been overcome at or above an agitation rate of 400 rpm, resulting in a maximum conversion of the plating reaction and dramatically shortened plating times with the added advantage of uniform deposition morphology as evidenced by the SEI micrographs. Finally, the agitation rate of 400 rpm was successfully employed for the synthesis of composite Pd and Pd/Ag membranes. The H2 permeance for a 4.7 µm thick pure-Pd membrane at 400oC was as high as 61 m3/m2-h-atm0.5 with H2/He selectivity over 310 after a total testing period of 690 hours.

Cette thèse s’intéresse à la problématique de la prédiction dans le cadre du design de systèmes énergétiques et des problèmes d’opération, et en particulier, à l’évaluation de l’adéquation de systèmes de production d’énergie renouvelables. L’objectif général est de développer une approche empirique pour générer des prédictions avec les incertitudes associées. En ce qui concerne cette direction de la recherche, une approche non paramétrique et empirique pour estimer les intervalles de prédiction (PIs) basés sur les réseaux de neurones (NNs) a été développée, quantifiant l’incertitude dans les prédictions due à la variabilité des données d’entrée et du comportement du système (i.e. due au comportement stochastique des sources renouvelables et de la demande d'énergie électrique), et des erreurs liées aux approximations faites pour établir le modèle de prédiction. Une nouvelle méthode basée sur l'optimisation multi-objectif pour estimer les PIs basée sur les réseaux de neurones et optimale à la fois en termes de précision (probabilité de couverture) et d’information (largeur d’intervalle) est proposée. L’ensemble de NN individuels par deux nouvelles approches est enfin présenté comme un moyen d’augmenter la performance des modèles. Des applications sur des études de cas réels démontrent la puissance de la méthode développée. / This Ph.D. work addresses the problem of prediction within energy systems design and operation problems, and particularly the adequacy assessment of renewable power generation systems. The general aim is to develop an empirical modeling framework for providing predictions with the associated uncertainties. Along this research direction, a non-parametric, empirical approach to estimate neural network (NN)-based prediction intervals (PIs) has been developed, accounting for the uncertainty in the predictions due to the variability in the input data and the system behavior (e.g. due to the stochastic behavior of the renewable sources and of the energy demand by the loads), and to model approximation errors. A novel multi-objective framework for estimating NN-based PIs, optimal in terms of both accuracy (coverage probability) and informativeness (interval width) is proposed. Ensembling of individual NNs via two novel approaches is proposed as a way to increase the performance of the models. Applications on real case studies demonstrate the power of the proposed framework.