O presente trabalho tem por objetivo propor uma metodologia para o dimensionamento da potência tracionária em função do percurso percorrido por veículos elétricos, assim como realizar o projeto básico de uma máquina elétrica para suprir a carga tracionária exigida. Os cálculos foram realizados através de um algoritmo onde o usuário insere as variáveis mecânicas do veículo e recebe a potência e conjugado necessários, e o projeto elementar da máquina que deverá equipar o veículo. Realizou-se um estudo de caso a partir de uma scooter elétrica para validar os resultados, foi projetada uma máquina brushless dc com fluxo radial e rotor externo. Os resultados experimentais atenderam as espectativas de conjugado de 46 Nm, o máximo conjugado estático atingiu 48 Nm e a rotação a vazio obteve valor de 613 rpm à tensão nominal. Os ensaios de frenagem regenerativa com supercapacitores apresentaram resultados promissores com 50% da energia mecânica recebida durante a frenagem entregue na partida do veículo. / The present study aims to propose a methodology for the design of power traction depending on the distance traveled by electrical vehicles, and perform the basic design of an electrical machine to supply the traction load required. The calculations will be performed through an algorithm where the user enters the mechanical variables of the vehicle and receives the power and torque required and the basic machine design that should equip the vehicle. A case study was conducted for an electric scooter considering the designed dc brushless machine with radial flow and external rotor in order to validate the results. The experimental results met the target of 46 Nm of torque, the maximum static torque 48 Nm was reached, and the no-load rotation obtained was 613 rpm at nominal voltage. Regenerative braking tests with super capacitors showed a promising result: 50% of the received mechanical energy during braking was delivered to the vehicle starting.

Artificial neural networks (ANNs) have been used successfully in various practical problems. Though extensive improvements on different types of ANNs have been made to improve their performance, each ANN design still experiences its own limitations. The existing digital human models are mature enough to provide accurate and useful results for different tasks and scenarios under various conditions. There is, however, a critical need for these models to run in real time, especially those with large-scale problems like motion prediction which can be computationally demanding. For even small changes to the task conditions, the motion simulation needs to run for a relatively long time (minutes to tens of minutes). Thus, there can be a limited number of training cases due to the computational time and cost associated with collecting training data. In addition, the motion problem is relatively large with respect to the number of outputs, where there are hundreds of outputs (between 500-700 outputs) to predict for a single problem. Therefore, the aforementioned necessities in motion problems lead to the use of tools like the ANN in this work. This work introduces new algorithms for the design of the radial-basis network (RBN) for problems with minimal available training data. The new RBN design incorporates new training stages with approaches to facilitate proper setting of necessary network parameters. The use of training algorithms with minimal heuristics allows the new RBN design to produce results with quality that none of the competing methods have achieved. The new RBN design, called Opt_RBN, is tested on experimental and practical problems, and the results outperform those produced from standard regression and ANN models. In general, the Opt_RBN shows stable and robust performance for a given set of training cases. When the Opt_RBN is applied on the large-scale motion prediction application, the network experiences a CPU memory issue when performing the optimization step in the training process. Therefore, new algorithms are introduced to modify some steps of the new Opt_RBN training process to address the memory issue. The modified steps should only be used for large-scale applications similar to the motion problem. The new RBN design proposes an ANN that is capable of improved learning without needing more training data. Although the new design is driven by its use with motion prediction problems, the consequent ANN design can be used with a broad range of large-scale problems in various engineering and industrial fields that experience delay issues when running computational tools that require a massive number of procedures and a great deal of CPU memory. The results of evaluating the modified Opt_RBN design on two motion problems are promising, with relatively small errors obtained when predicting approximately 500-700 outputs. In addition, new methods for constraint implementation within the new RBN design are introduced. Moreover, the new RBN design and its associated parameters are used as a tool for simulated task analysis. This work initiates the idea that output weights (W) can be used to determine the most critical basis functions that cause the greatest reduction in the network test error. Then, the critical basis functions can specify the most significant training cases that are responsible for the proper performance achieved by the network. The inputs with the most change in value can be extracted from the basis function centers (U) in order to determine the dominant inputs. The outputs with the most change in value and their corresponding key body degrees-of-freedom for a motion task can also be specified using the training cases that are used to create the network's basis functions.

In this work, a numerical approach based on meshless methods is proposed to obtain eigenmodes of Laplace-Beltrami operator on manifolds, and its performance is compared against existing alternative methods. Radial Basis Function (RBF)-based methods allow one to obtain interpolation and differentiation matrices easily by using scattered data points. We derive expressions for such matrices for the Laplace-Beltrami operator via so-called Reilly’s formulas and use them to solve the respective eigenvalue problem. Numerical studies of proposed methods are performed in order to demonstrate convergence on simple examples of one-dimensional curves and two-dimensional surfaces.

Accurate and fast machine learning (ML) algorithms are highly vital in artificial intelligence (AI) applications. In complex dataset problems, traditional ML methods such as radial basis function neural network (RBFN), sparse coding (SC) using dictionary learning, and particle swarm optimization (PSO) provide trivial results, large structure, slow training, and/or slow testing. This dissertation introduces four novel ML techniques: the multi-column RBFN network (MCRN), the projected dictionary learning algorithm (PDL) and the multi-column adaptive and non-adaptive particle swarm optimization techniques (MC-APSO and MC-PSO). These novel techniques provide efficient alternatives for traditional ML techniques. Compared to traditional ML techniques, the novel ML techniques demonstrate more accurate results, faster training and testing timing, and parallelized structured solutions. MCRN deploys small RBFNs in a parallel structure to speed up both training and testing. Each RBFN is trained with a subset of the dataset and the overall structure provides results that are more accurate. PDL introduces a conceptual dictionary learning method in updating the dictionary atoms with the reconstructed input blocks. This method improves the sparsity of extracted features and hence, the image denoising results. MC-PSO and MC-APSO provide fast and more accurate alternatives to the PSO and APSO slow evolutionary techniques. MC-PSO and MC-APSO use multi-column parallelized RBFN structure to improve results and speed with a wide range of classification dataset problems. The novel techniques are trained and tested using benchmark dataset problems and the results are compared with the state-of-the-art counterpart techniques to evaluate their performance. Novel techniques’ results show superiority over techniques in accuracy and speed in most of the experimental results, which make them good alternatives in solving difficult ML problems.

No description available.

Les phénomènes de bioadhésion de microorganismes sur les aciers inoxydables son fréquents et peuvent entraîner des problèmes de santé publique tels que des infections toxicologiques. En partenariat avec Ugine-Alz, nous nous sommes intéressés, à l'aide d'une chambre à flux radial (Décavé et al. Biophysical Journal, 2002, 82, 2383-95), à l'étude quantitative des cinétiques de détachement de trois microorganismes après adhésion sur des surfaces d'acier inoxydable. Saccharomyces cerevisiae (levure), Escherichia coli (Gram-), et Staphylococcus epidermidis (Gram+). Les aciers inoxydables utilisés sont des échantillons polis miroir et polis attaqués avec des tailles de grain et des profondeurs d'attaque des joints de grain différentes. L'utilisation de ces différents états de surface ne modifie pas la valeur de la contrainte nécessaire pour détacher 50% des microorganismes (seuil de détachement). Par contre, la profondeur d'attaque des joints de grain et la taille de grain accélèrent la cinétique de détachement de S. cerevisiae et ralentissent celle d'E. coli. Enfin, une adhésion préférentielle a été mise en évidence, pour les trois microorganismes, sur les grains d'orientation cristallographique <001> (Demilly et al., Colloids and Surfaces B, in press, 2006).

Les aciers inoxydables sont des matériaux couramment utilisés dans les industries agroalimentaires, la restauration collective mais également dans le milieu médical et hospitalier. Dans ces domaines, les phénomènes de bioadhésion de microorganismes sur ces substrats sont fréquents et peuvent entraîner des problèmes de santé publique tels que des infections toxicologiques ou des contaminations chez des sujets porteurs de matériau étranger (cathéter, prothèse).<br />Dans ce travail, effectué en partenariat avec le groupe sidérugique Ugine-Alz, nous nous sommes intéressés, à l'aide d'une chambre à flux radial, à l'étude quantitative des cinétiques de détachement de trois microorganismes après adhésion sur des surfaces d'acier inoxydable. Saccharomyces cerevisiae, une levure, est représentative des eucaryotes, Escherichia coli est une bactérie Gram- et Staphylococcus epidermidis, une bactérie Gram+. Les états de surface des aciers inoxydables utilisés sont des échantillons polis miroir (PM) et polis attaqués (PA) avec des tailles de grain différentes (20, 40 et 100 µm) et différentes profondeurs d'attaque des joints de grain (100 ± 50 nm, 150 ± 50 nm, 650 ± 150 nm). L'utilisation de ces différents états de surface ne modifie pas la valeur de la contrainte nécessaire pour détacher 50 % des microorganismes (seuil de détachement). Par contre, la profondeur d'attaque des joints de grain et, dans une moindre mesure, la taille de grain (20, 40 et 100 µm) accélèrent la cinétique de détachement de S. cerevisiae, grosse cellule non déformable et ralentissent celle d'E. coli, petite cellule déformable. Enfin, une adhésion préférentielle a été mise en évidence, pour les trois types de cellules, sur les grains d'orientation cristallographiques <001>.<br />Les observtions expérimentales ont par ailleurs été rationalisées à l'aide du modèle de B. Fourcade, D. Garrivier et E. Décavé (Décavé et al., 2002) et par le biais d'une analyse qualitative des interactions cellule-substrat en fonction de la taille et de la déformabilité des cellules.

The thesis presents the results of our research on symmetry for some semilinear elliptic problems and on existence of solution for quasilinear problems involving singularities. The text is composed of two parts, each of which begins with a specific introduction. The first part is devoted to symmetry and symmetry-breaking results. We study a class of partial differential equations involving radial weights on balls, annuli or $R^N$ --where these weights are unbounded--. We show in particular that on unbounded domains, focusing on symmetric functions permits to recover compactness, which implies existence of solutions. Then, we stress the fact that symmetry-breaking occurs on bounded domains, depending both on the weights and on the nonlinearity of the equation. We also show that for the considered class of problems, the multibumps-solution phenomenon appears on the annulus as well as on the ball. The second part of the thesis is devoted to partial and ordinary differential equations with singularities. Using concentration-compactness tools, we show that a rather large class of functionals is lower semi-continuous, leading to the existence of a ground state solution. We also focus on the unicity of solutions for such a class of problems.

In this doctoral project, a relatively new form of carbon material, with unique narrow pore size distribution around 7 Å and with uniform structure, has been electrochemically characterised using the single particle microelectrode technique. The carbon has been used as electrode material for supercapacitors. This type of capacitors is used as high power energy buffers in hybrid vehicles and for stationary power backup. The principle for the microelectrode technique consists of connecting a carbon particle with a carbon fibre by means of a micromanipulator. The single particle and carbon fibre together form a microelectrode. Combination of this technique with electroanalytical methods such as cyclic voltammetry and potential step measurements allows for the survey of electrochemical phenomena and for the determination of ion transport parameters inside the nanopores. A mathematical model based on Fick’s second law, for diffusion of ions inside the nanopores at non steady state, was used for the determination of effective diffusion coefficients (Deff). The coefficients were calculated from an asymptotic solution of Fick’s equation, applied for a thin layer adjacent to the external surface of the carbon particles and valid for the current response in a short time region. Another asymptotic solution was obtained, using spherical geometry and valid for the current response in a long time region. In this doctoral work, the carbon particles have been exposed to potential cycling, which mimics that of large electrodes during operation of a double layer capacitor. The potential-current response, E-I, for the nanoporous carbon, shows a pure capacitive behaviour between –0.5 V and 0.1 V vs. the Hg|HgO reference electrode. The detection of the faradaic processes beyond these potentials was possible by lowering of the voltammometric sweep rate. The electrochemical processes occurring at positive and at negative potential were investigated separately. Cyclic voltammometric measurements showed that the chemisorption of hydroxyl groups, occurring between 0.1 and 0.3 V, leads to a mild oxidation of the carbon structure, resulting in surface groups containing an oxygen atom at a specific carbon site (e.g., phenolic or quinine type). These oxygen-containing surface groups caused an increase of the specific capacitance, which remained constant throughout a number of voltammometric cycles. The Deff decreased on the other hand with the number of cycles. The Deff decreases also with the positive potential. The evaluation of Deff indicates adsorption of hydroxyl groups and an increase of the effective tortuosity of the pore system. The oxidation of the carbon particles, between 0 and 0.5 V, leads to more extensive oxidation and to surface groups containing two oxygen atoms at a single carbon site, followed by formation of carbonate ions. The oxygen-containing surface groups and carbonate ions formed at these potentials do not contribute to the specific capacitance and drastically retard or obstruct the ion transport inside the nanopores. At negative potentials the carbon particles show a dominantly capacitive behaviour. The faradaic processes taking place below –0.5 V vs. Hg|HgO reference electrode are generation and adsorption of hydrogen. These processes do not perturb significantly the electrochemical and ion transport properties of the nanoporous carbon particles. It was found that hydrogen generation occurs at –0.5 V vs. Hg|HgO and that two hydrogen oxidation processes take place at positive potentials. The results indicate that the weakly adsorbed hydrogen undergoes oxidation between 0 and 0.1 V and that the strongly adsorbed hydrogen is oxidised at more positive potentials. The single particle technique was adapted for the determination of diffusion coefficients of an organic electrolyte. The different size of the anions and cations caused different transport characteristics at negative and positive potentials. Slow cycling was found important for ion penetration inside the nanopores and for the evaluation of the effective diffusion coefficients. The effective diffusion coefficients for the nanoporous carbon using aqueous 6M KOH and 0.1M TEABF4 in acetonitrile were estimated to 1.4 (±0.8).10-9 cm2 s-1 and 1.3 (±0.4) 10-8 cm2 s-1, respectively.

The electronic and geometrical structure of free neutral clusters and liquids have been studied using synchrotron-radiation based photoelectron and Auger electron spectroscopy in combination with lineshape modeling. A novel experimental setup has been developed for studies of liquids, based on the liquid microjet technique. Theoretical lineshapes have been computed using both classical (molecular dynamics) and quantum mechanical (mainly density functional theory) methods. Clusters are finite ensembles of atoms or molecules, ranging in size from a few to several thousand atoms. Apart from being fundamentally interesting, clusters are also promising as building blocks for nano-technology. In this thesis results are presented for rare-gas and molecular clusters, ranging from weakly van-deer-Waals bonded to hydrogen bonded. It is shown that the combination of core-level photoelectron spectroscopy (XPS) and lineshape modeling can be used to estimate the sizes of clusters. A model for treating the effect of inter-molecular nuclear relaxation upon ionization is proposed. The structure of single-component molecular clusters are investigated by molecular dynamics simulations, validated against XPS data. Finally, the radial structure of a two-component molecular cluster is investigated by XPS. Liquids have been studied for centuries, but still many questions remain regarding the microscopic properties. With the recent development of the liquid microjet technique, new insight into the atomic structure can be obtained. In this thesis we study aqueous solutions using photoelectron and Auger electron spectroscopy (AES). We investigate the structure of surface active molecules by XPS, study the Auger decay after core-level ionization in aqueous potassium chloride (KCl), and follow the changes in molecular structure of glycine as a function of pH.