Redes neurais e suas aplicações em calibração multivariada

Neural Networks are a set of mathematical methods and computer programs designed to simulate the information process and the knowledge acquisition of the human brain. In last years its application in chemistry is increasing significantly, due the special characteristics for model complex systems. The basic principles of two types of neural networks, the multi-layer perceptrons and radial basis functions, are introduced, as well as, a pruning approach to architecture optimization. Two analytical applications based on near infrared spectroscopy are presented, the first one for determination of nitrogen content in wheat leaves using multi-layer perceptrons networks and second one for determination of BRIX in sugar cane juices using radial basis functions networks.

Imaging and Characterisation of Dirt Particles in Pulp and Paper

Dirt counting and dirt particle characterisation of pulp samples is an important part of quality control in pulp and paper production. The need for an automatic image analysis system to consider dirt particle characterisation in various pulp samples is also very critical. However, existent image analysis systems utilise a single threshold to segment the dirt particles in different pulp samples. This limits their precision. Based on evidence, designing an automatic image analysis system that could overcome this deficiency is very useful. In this study, the developed Niblack thresholding method is proposed. The method defines the threshold based on the number of segmented particles. In addition, the Kittler thresholding is utilised. Both of these thresholding methods can determine the dirt count of the different pulp samples accurately as compared to visual inspection and the Digital Optical Measuring and Analysis System (DOMAS). In addition, the minimum resolution needed for acquiring a scanner image is defined. By considering the variation in dirt particle features, the curl shows acceptable difference to discriminate the bark and the fibre bundles in different pulp samples. Three classifiers, called k-Nearest Neighbour, Linear Discriminant Analysis and Multi-layer Perceptron are utilised to categorize the dirt particles. Linear Discriminant Analysis and Multi-layer Perceptron are the most accurate in classifying the segmented dirt particles by the Kittler thresholding with morphological processing. The result shows that the dirt particles are successfully categorized for bark and for fibre bundles.

Paper for printed electronics and functionality

Mass-produced paper electronics (large area organic printed electronics on paper-based substrates, “throw-away electronics”) has the potential to introduce the use of flexible electronic applications in everyday life. While paper manufacturing and printing have a long history, they were not developed with electronic applications in mind. Modifications to paper substrates and printing processes are required in order to obtain working electronic devices. This should be done while maintaining the high throughput of conventional printing techniques and the low cost and recyclability of paper. An understanding of the interactions between the functional materials, the printing process and the substrate are required for successful manufacturing of advanced devices on paper. Based on the understanding, a recyclable, multilayer-coated paper-based substrate that combines adequate barrier and printability properties for printed electronics and sensor applications was developed in this work. In this multilayer structure, a thin top-coating consisting of mineral pigments is coated on top of a dispersion-coated barrier layer. The top-coating provides well-controlled sorption properties through controlled thickness and porosity, thus enabling optimizing the printability of functional materials. The penetration of ink solvents and functional materials stops at the barrier layer, which not only improves the performance of the functional material but also eliminates potential fiber swelling and de-bonding that can occur when the solvents are allowed to penetrate into the base paper. The multi-layer coated paper under consideration in the current work consists of a pre-coating and a smoothing layer on which the barrier layer is deposited. Coated fine paper may also be used directly as basepaper, ensuring a smooth base for the barrier layer. The top layer is thin and smooth consisting of mineral pigments such as kaolin, precipitated calcium carbonate, silica or blends of these. All the materials in the coating structure have been chosen in order to maintain the recyclability and sustainability of the substrate. The substrate can be coated in steps, sequentially layer by layer, which requires detailed understanding and tuning of the wetting properties and topography of the barrier layer versus the surface tension of the top-coating. A cost competitive method for industrial scale production is the curtain coating technique allowing extremely thin top-coatings to be applied simultaneously with a closed and sealed barrier layer. The understanding of the interactions between functional materials formulated and applied on paper as inks, makes it possible to create a paper-based substrate that can be used to manufacture printed electronics-based devices and sensors on paper. The multitude of functional materials and their complex interactions make it challenging to draw general conclusions in this topic area. Inevitably, the results become partially specific to the device chosen and the materials needed in its manufacturing. Based on the results, it is clear that for inks based on dissolved or small size functional materials, a barrier layer is beneficial and ensures the functionality of the printed material in a device. The required active barrier life time depends on the solvents or analytes used and their volatility. High aspect ratio mineral pigments, which create tortuous pathways and physical barriers within the barrier layer limit the penetration of solvents used in functional inks. The surface pore volume and pore size can be optimized for a given printing process and ink through a choice of pigment type and coating layer thickness. However, when manufacturing multilayer functional devices, such as transistors, which consist of several printed layers, compromises have to be made. E.g., while a thick and porous top-coating is preferable for printing of source and drain electrodes with a silver particle ink, a thinner and less absorbing surface is required to form a functional semiconducting layer. With the multilayer coating structure concept developed in this work, it was possible to make the paper substrate suitable for printed functionality. The possibility of printing functional devices, such as transistors, sensors and pixels in a roll-to-roll process on paper is demonstrated which may enable introducing paper for use in disposable “onetime use” or “throwaway” electronics and sensors, such as lab-on-strip devices for various analyses, consumer packages equipped with product quality sensors or remote tracking devices.

Computational modeling of stented coronary arteries

The application of computational fluid dynamics (CFD) and finite element analysis (FEA) has been growing rapidly in the various fields of science and technology. One of the areas of interest is in biomedical engineering. The altered hemodynamics inside the blood vessels plays a key role in the development of the arterial disease called atherosclerosis, which is the major cause of human death worldwide. Atherosclerosis is often treated with the stenting procedure to restore the normal blood flow. A stent is a tubular, flexible structure, usually made of metals, which is driven and expanded in the blocked arteries. Despite the success rate of the stenting procedure, it is often associated with the restenosis (re-narrowing of the artery) process. The presence of non-biological device in the artery causes inflammation or re-growth of atherosclerotic lesions in the treated vessels. Several factors including the design of stents, type of stent expansion, expansion pressure, morphology and composition of vessel wall influence the restenosis process. Therefore, the role of computational studies is crucial in the investigation and optimisation of the factors that influence post-stenting complications. This thesis focuses on the stent-vessel wall interactions followed by the blood flow in the post-stenting stage of stenosed human coronary artery. Hemodynamic and mechanical stresses were analysed in three separate stent-plaque-artery models. Plaque was modeled as a multi-layer (fibrous cap (FC), necrotic core (NC), and fibrosis (F)) and the arterial wall as a single layer domain. CFD/FEA simulations were performed using commercial software packages in several models mimicking the various stages and morphologies of atherosclerosis. The tissue prolapse (TP) of stented vessel wall, the distribution of von Mises stress (VMS) inside various layers of vessel wall, and the wall shear stress (WSS) along the luminal surface of the deformed vessel wall were measured and evaluated. The results revealed the role of the stenosis size, thickness of each layer of atherosclerotic wall, thickness of stent strut, pressure applied for stenosis expansion, and the flow condition in the distribution of stresses. The thicknesses of FC, and NC and the total thickness of plaque are critical in controlling the stresses inside the tissue. A small change in morphology of artery wall can significantly affect the distribution of stresses. In particular, FC is the most sensitive layer to TP and stresses, which could determine plaque’s vulnerability to rupture. The WSS is highly influenced by the deflection of artery, which in turn is dependent on the structural composition of arterial wall layers. Together with the stenosis size, their roles could play a decisive role in controlling the low values of WSS (<0.5 Pa) prone to restenosis. Moreover, the time dependent flow altered the percentage of luminal area with WSS values less than 0.5 Pa at different time instants. The non- Newtonian viscosity model of the blood properties significantly affects the prediction of WSS magnitude. The outcomes of this investigation will help to better understand the roles of the individual layers of atherosclerotic vessels and their risk to provoke restenosis at the post-stenting stage. As a consequence, the implementation of such an approach to assess the post-stented stresses will assist the engineers and clinicians in optimizing the stenting techniques to minimize the occurrence of restenosis.

Method for automatic detection of wheezing in lung sounds

The present report describes the development of a technique for automatic wheezing recognition in digitally recorded lung sounds. This method is based on the extraction and processing of spectral information from the respiratory cycle and the use of these data for user feedback and automatic recognition. The respiratory cycle is first pre-processed, in order to normalize its spectral information, and its spectrogram is then computed. After this procedure, the spectrogram image is processed by a two-dimensional convolution filter and a half-threshold in order to increase the contrast and isolate its highest amplitude components, respectively. Thus, in order to generate more compressed data to automatic recognition, the spectral projection from the processed spectrogram is computed and stored as an array. The higher magnitude values of the array and its respective spectral values are then located and used as inputs to a multi-layer perceptron artificial neural network, which results an automatic indication about the presence of wheezes. For validation of the methodology, lung sounds recorded from three different repositories were used. The results show that the proposed technique achieves 84.82% accuracy in the detection of wheezing for an isolated respiratory cycle and 92.86% accuracy for the detection of wheezes when detection is carried out using groups of respiratory cycles obtained from the same person. Also, the system presents the original recorded sound and the post-processed spectrogram image for the user to draw his own conclusions from the data.

Leak-proof Heat Sealing of Press-Formed Paperboard Trays

Three-dimensional (3D) forming of paperboard and heat sealing of lidding films to trays manufactured by the press forming process are investigated in this thesis. The aim of the work was to investigate and recognize the factors affecting the quality of heat sealing and the leak resistance (tightness) of press-formed, polymer-coated paperboard trays heatsealed with a multi-layer polymer based lidding film. One target was to achieve a solution that can be used in food packaging using modified atmosphere packaging (MAP). The main challenge in acquiring adequate tightness properties for the use of MAP is creases in the sealing area of the paperboard trays which can act as capillary tubes and prevent leak-proof sealing. Several experiments were made to investigate the effect of different factors and process parameters in the forming and sealing processes. Also different methods of analysis, such as microscopic analysis and 3D-profilometry were used to investigate the structure of the creases in the sealing area, and to analyse the surface characteristics of the tray flange of the formed trays to define quality that can be sealed with satisfactory tightness for the use of MAP. The main factors and parameters that had an effect on the result of leak-proof sealing and must be adjusted accordingly were the tray geometry and dimensions, blank holding force in press forming, surface roughness of the sealing area, the geometry and depth of the creases, and the sealing pressure. The results show that the quality of press-formed, polymer-coated paperboard trays and multi-layer polymer lidding films can be satisfactory for the use of modified atmosphere packaging in food solutions. Suitable tools, materials, and process parameters have to be selected and used during the tray manufacturing process and lid sealing process, however. Utilizing these solutions and results makes it possible for a package that is made mostly from renewable and recyclable sources to be a considerable alternative for packages made completely from oil based polymers, and to achieve a greater market share for fibre-based solutions in food packaging using MAP.

Planar Topological Defects in Unconventional Superconductors

In this work, we consider the properties of planar topological defects in unconventional superconductors. Specifically, we calculate microscopically the interaction energy of domain walls separating degenerate ground states in a chiral p-wave fermionic superfluid. The interaction is mediated by the quasiparticles experiencing Andreev scattering at the domain walls. As a by-product, we derive a useful general expression for the free energy of an arbitrary nonuniform texture of the order parameter in terms of the quasiparticle scattering matrix. The thesis is structured as follows. We begin with a historical review of the theories of superconductivity (Sec. 1.1), which led the way to the celebrated Bardeen-Cooper- Schrieffer (BCS) theory (Sec. 1.3). Then we proceed to the treatment of superconductors with so-called "unconventional pairing" in Sec. 1.4, and in Sec. 1.5 we introduce the specific case of chiral p-wave superconductivity. After introducing in Sec. 2 the domain wall (DW) model that will be considered throughout the work, we derive the Bogoliubov-de Gennes (BdG) equations in Sec. 3.1, which determine the quasiparticle excitation spectrum for a nonuniform superconductor. In this work, we use the semiclassical (Andreev) approximation, and solve the Andreev equations (which are a particular case of the BdG equations) in Sec. 4 to determine the quasiparticle spectrum for both the single- and two-DW textures. The Andreev equations are derived in Sec. 3.2, and the formal properties of the Andreev scattering coefficients are discussed in the following subsection. In Sec. 5, we use the transfer matrix method to relate the interaction energy of the DWs to the scattering matrix of the Bogoliubov quasiparticles. This facilitates the derivation of an analytical expression for the interaction energy between the two DWs in Sec. 5.3. Finally, to illustrate the general applicability our method, we apply it in Sec. 6 to the interaction between phase solitons in a two-band s-wave superconductor.

Understanding deep architectures and the effect of unsupervised pre-training

Cette thèse porte sur une classe d'algorithmes d'apprentissage appelés architectures profondes. Il existe des résultats qui indiquent que les représentations peu profondes et locales ne sont pas suffisantes pour la modélisation des fonctions comportant plusieurs facteurs de variation. Nous sommes particulièrement intéressés par ce genre de données car nous espérons qu'un agent intelligent sera en mesure d'apprendre à les modéliser automatiquement; l'hypothèse est que les architectures profondes sont mieux adaptées pour les modéliser. Les travaux de Hinton (2006) furent une véritable percée, car l'idée d'utiliser un algorithme d'apprentissage non-supervisé, les machines de Boltzmann restreintes, pour l'initialisation des poids d'un réseau de neurones supervisé a été cruciale pour entraîner l'architecture profonde la plus populaire, soit les réseaux de neurones artificiels avec des poids totalement connectés. Cette idée a été reprise et reproduite avec succès dans plusieurs contextes et avec une variété de modèles. Dans le cadre de cette thèse, nous considérons les architectures profondes comme des biais inductifs. Ces biais sont représentés non seulement par les modèles eux-mêmes, mais aussi par les méthodes d'entraînement qui sont souvent utilisés en conjonction avec ceux-ci. Nous désirons définir les raisons pour lesquelles cette classe de fonctions généralise bien, les situations auxquelles ces fonctions pourront être appliquées, ainsi que les descriptions qualitatives de telles fonctions. L'objectif de cette thèse est d'obtenir une meilleure compréhension du succès des architectures profondes. Dans le premier article, nous testons la concordance entre nos intuitions---que les réseaux profonds sont nécessaires pour mieux apprendre avec des données comportant plusieurs facteurs de variation---et les résultats empiriques. Le second article est une étude approfondie de la question: pourquoi l'apprentissage non-supervisé aide à mieux généraliser dans un réseau profond? Nous explorons et évaluons plusieurs hypothèses tentant d'élucider le fonctionnement de ces modèles. Finalement, le troisième article cherche à définir de façon qualitative les fonctions modélisées par un réseau profond. Ces visualisations facilitent l'interprétation des représentations et invariances modélisées par une architecture profonde.

Lien entre les matrices de transfert de spins et de boucles du modèle de Potts sur le tore

Le lien entre le spectre de la matrice de transfert de la formulation de spins du modèle de Potts critique et celui de la matrice de transfert double-ligne de la formulation de boucles est établi. La relation entre la trace des deux opérateurs est obtenue dans deux représentations de l'algèbre de Temperley-Lieb cyclique, dont la matrice de transfert de boucles est un élément. Le résultat est exprimé en termes des traces modifiées, qui correspondent à des traces effectuées dans le sous-espace de l'espace de représentation des N-liens se transformant selon la m ième représentation irréductible du groupe cyclique. Le mémoire comporte trois chapitres. Dans le premier chapitre, les résultats essentiels concernant les formulations de spins et de boucles du modèle de Potts sont rappelés. Dans le second chapitre, les propriétés de l'algèbre de Temperley-Lieb cyclique et de ses représentations sont étudiées. Enfin, le lien entre les deux traces est construit dans le troisième chapitre. Le résultat final s'apparente à celui obtenu par Richard et Jacobsen en 2007, mais une nouvelle représentation n'ayant pas été étudiée est aussi investiguée.

La structure de Jordan des matrices de transfert des modèles de boucles et la relation avec les hamiltoniens XXZ

Les modèles sur réseau comme ceux de la percolation, d’Ising et de Potts servent à décrire les transitions de phase en deux dimensions. La recherche de leur solution analytique passe par le calcul de la fonction de partition et la diagonalisation de matrices de transfert. Au point critique, ces modèles statistiques bidimensionnels sont invariants sous les transformations conformes et la construction de théories des champs conformes rationnelles, limites continues des modèles statistiques, permet un calcul de la fonction de partition au point critique. Plusieurs chercheurs pensent cependant que le paradigme des théories des champs conformes rationnelles peut être élargi pour inclure les modèles statistiques avec des matrices de transfert non diagonalisables. Ces modèles seraient alors décrits, dans la limite d’échelle, par des théories des champs logarithmiques et les représentations de l’algèbre de Virasoro intervenant dans la description des observables physiques seraient indécomposables. La matrice de transfert de boucles D_N(λ, u), un élément de l’algèbre de Temperley- Lieb, se manifeste dans les théories physiques à l’aide des représentations de connectivités ρ (link modules). L’espace vectoriel sur lequel agit cette représentation se décompose en secteurs étiquetés par un paramètre physique, le nombre d de défauts. L’action de cette représentation ne peut que diminuer ce nombre ou le laisser constant. La thèse est consacrée à l’identification de la structure de Jordan de D_N(λ, u) dans ces représentations. Le paramètre β = 2 cos λ = −(q + 1/q) fixe la théorie : β = 1 pour la percolation et √2 pour le modèle d’Ising, par exemple. Sur la géométrie du ruban, nous montrons que D_N(λ, u) possède les mêmes blocs de Jordan que F_N, son plus haut coefficient de Fourier. Nous étudions la non diagonalisabilité de F_N à l’aide des divergences de certaines composantes de ses vecteurs propres, qui apparaissent aux valeurs critiques de λ. Nous prouvons dans ρ(D_N(λ, u)) l’existence de cellules de Jordan intersectorielles, de rang 2 et couplant des secteurs d, d′ lorsque certaines contraintes sur λ, d, d′ et N sont satisfaites. Pour le modèle de polymères denses critique (β = 0) sur le ruban, les valeurs propres de ρ(D_N(λ, u)) étaient connues, mais les dégénérescences conjecturées. En construisant un isomorphisme entre les modules de connectivités et un sous-espace des modules de spins du modèle XXZ en q = i, nous prouvons cette conjecture. Nous montrons aussi que la restriction de l’hamiltonien de boucles à un secteur donné est diagonalisable et trouvons la forme de Jordan exacte de l’hamiltonien XX, non triviale pour N pair seulement. Enfin nous étudions la structure de Jordan de la matrice de transfert T_N(λ, ν) pour des conditions aux frontières périodiques. La matrice T_N(λ, ν) a des blocs de Jordan intrasectoriels et intersectoriels lorsque λ = πa/b, et a, b ∈ Z×. L’approche par F_N admet une généralisation qui permet de diagnostiquer des cellules intersectorielles dont le rang excède 2 dans certains cas et peut croître indéfiniment avec N. Pour les blocs de Jordan intrasectoriels, nous montrons que les représentations de connectivités sur le cylindre et celles du modèle XXZ sont isomorphes sauf pour certaines valeurs précises de q et du paramètre de torsion v. En utilisant le comportement de la transformation i_N^d dans un voisinage des valeurs critiques (q_c, v_c), nous construisons explicitement des vecteurs généralisés de Jordan de rang 2 et discutons l’existence de blocs de Jordan intrasectoriels de plus haut rang.

A sensitive fibre optic pH sensor using multiple sol–gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements.

A sensitive fibre optic pH sensor using multiple sol–gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements

A sensitive fibre optic pH sensor using multiple sol–gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements

A sensitive fibre optic pH sensor using multiple sol-gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements.

Quantum wire with periodic serial structure

Electron wave motion in a quantum wire with periodic structure is treated by direct solution of the Schrödinger equation as a mode-matching problem. Our method is particularly useful for a wire consisting of several distinct units, where the total transfer matrix for wave propagation is just the product of those for its basic units. It is generally applicable to any linearly connected serial device, and it can be implemented on a small computer. The one-dimensional mesoscopic crystal recently considered by Ulloa, Castaño, and Kirczenow [Phys. Rev. B 41, 12 350 (1990)] is discussed with our method, and is shown to be a strictly one-dimensional problem. Electron motion in the multiple-stub T-shaped potential well considered by Sols et al. [J. Appl. Phys. 66, 3892 (1989)] is also treated. A structure combining features of both of these is investigated