Cette thèse s'inscrit dans le cadre d'un projet de recherche européen (Transfeu), composé de 21 partenaires, dédié à l'estimation des effets du feu sur les personnes présentes dans un train et basée sur une méthodologie de sécurité incendie. L'objectif est de modéliser le comportement au feu d'un siège (matériaux multicouches) et d'un panneau vertical (composite) d'une voiture de train indépendamment de la source de feu et du design de la voiture. Le principe est basé sur l'étude de la réaction au feu de deux produits par une approche multi-échelle (de l'échelle de la matière à l'échelle réelle). A chaque échelle, les données expérimentales et numériques sont comparées et valident les processus de décomposition thermique et de combustion mis en jeu. Les données d'entrée sont estimées selon des essais normalisés ou à partir de la littérature et sont identiques pour toutes les échelles. Les phénomènes de décomposition thermique et de combustion sont simulés à partir d'un modèle de pyrolyse et d'un modèle de combustion à fraction de mélange. Cette comparaison, à complexité croissante, permet d'observer la capacité des modèles de FDS de reproduire des simulations réalistes. Les résultats permettent de souligner les limites des modèles et de les dépasser en proposant des solutions alternatives. / This thesis work comes within the framework of a collaborative European research program (Transfeu), dedicated to estimate the fire effects on people into a train coach, based on fire safety methodology. The objective is to model the fire behaviour of a multilayer seat materials and a composite wall panel in a coach independently of the fire source in order to estimate the safety level of a coach design scenario. The principle is based on a multi-scale approach of the reaction-to-fire of two products from the raw matter to the real scale. At each scale, experimental and numerical data are compared and then validated on the thermal decomposition and the combustion processes. The used input data are the same for all scales and estimated according to standard experimental tests or literature reviews. The thermal decomposition and the combustion phenomena are simulated from a pyrolysis and a mixture fraction combustion models. This comparison at increasing complexity allows observing the ability of the models to reproduce realistic simulations. The results allow to highlight the limits of the models and then to propose better solutions in order to overcome them.

Background and Scope: The scope of this thesis is to investigate the thermal behaviour of building components as hydronic radiator and multilayer walls subjected to dynamic conditions. The modelling and simulation of these building components provide information on how these components thermally interact among each other. The thermal interaction is fundamental to know how the energy is used in buildings. In particular, the thermal energy used in rooms can be expressed as the efficiencies for emission in a space heating system. This thesis analyzes the efficiencies for emission of a space heating system equipped with hydronic radiator for Swedish buildings by providing a comprehensive and detailed approach on this topic. Methodology: The methods used in this thesis are: experiment, modelling of multilayer wall and hydronic radiator, the dynamic simulation of the building and the efficiencies for emission of a space heating system. Here, the experiment, known as step response test, shows the heating up process of a hydronic radiator. The observation of the qualitative measurements suggests the most suitable technique of modelling the radiator known as transient modelling with multiple storage elements. The multilayer wall has been discretized both in space and time variable with a Finite Difference Method. Dynamic simulation of the building provides the efficiencies for emission of a space heating system. Findings: The experimental results show how the radiator performs the charging phase. The performance of the transient model is compared with lumped steady state models in terms of temperature of exhaust flow and total heat emitted. Results of the dynamic simulation show how buildings located in a Northern climate use the energy in a better way than Southern climates in Sweden. Heavy active thermal mass provides higher efficiencies for emission than light thermal mass. Radiators with connection pipes located on the same side react faster at the thermodynamic changing of the mass flow rate by providing higher efficiencies for emission than radiators with connection pipes located on the opposite side. Conclusion and Outlook: This thesis increases the knowledge about the modelling and simulation of hydronic radiators and multilayer walls. More research is needed on this topic to encompass modelling details of building components often ignored. The modelling and simulation of building components are the key to understand how building components thermally interact with each other. The thermal interaction among building components is a fundamental parameter for the assessment of efficiencies of emission of the space heating system. In the near future, the concept of efficiencies of emission can be implemented in National Building Code, therefore, this study provides guidelines on how to assess these efficiencies. / <p>Advisors: Ronny Östin and Mohsen Soleimanni Mohseni, Department of Applied Physics and Electronics, Umeå University</p>

The design of biocompatible synthetic surfaces is an important issue for medical applications. Surface modification techniques provide good approaches to control the interactions between living systems and implanted materials by modifying the surface characteristics. This thesis work demonstrates the feasibility and effectiveness of the novel and low-cost electrostatic self-assembly (ESA) technique for the manufacturing of biocompatible thin film coatings. The ESA process is based on the alternating adsorption of molecular layers of oppositely charged polymers/nanoparticles, and can be applied in the fabrication of well-organized multilayer thin films possessing various biocompatible properties. ESA multilayer assemblies incorporating various biomaterials including metal oxides and polymers were fabricated, the uniformity, thickness, layer-by-layer linearity, and surface morphology of the films were characterized by UV/vis spectroscopy, ellipsometry, and AFM imaging. Preliminary biocompatibility testing was conducted, concentrating on contact angle surface characterization and the in vitro measurements of protein adsorption. The use of Fourier Transform Infrared Reflection-Absorption Spectroscopy (FT-IRAS) for the investigation of the protein adsorption behavior upon the ESA multilayer films is presented. / Master of Science

With the development of big data computing, the speech recognition has been popular for serving human’s life. However, when place the speech recognition system into noisy environments, the background noises greatly degrades the speech recognition system accuracy as it adds in unuseful information into the desired speech. Thus for a speech recognition system, obtaining a good performance under noises has become a vital issue. To tackle the noise effect problem of automatic speech recognition (ASR), a method to reduce the noise effect is essential. Recently, multiple of methods have been developed to enhance the speech signal, they usually follow the principle of suppressing the noise in a noisy speech signal. This thesis simulated the popular techniques for speech recogniton and speech enhancement, which are the multilayer perceptron and the spectral subtraction. The aim of this work is to use MATLAB to build an automatic speech recognition system that can be used in noisy environment. MATLAB simulations are used to verify the success of recognition with clean speech and show the system performance improvements after applying speech enhancement method in seven kinds of noisy environments. The result is presented by using comparative histograms between noisy signals and corresponding denoised signals. It shows that, using denoised signal will obtain a higher recognition rate, thus we can say the system performance is improved in noisy environments.

Electromagnetic scattering from a rough two dimensional homogeneous scatterer has been computationally modeled. The scatterer is intended to simulate reflection from a two interface multilayer. The rough scatterer was created from Gaussian random points centered about an ideal interface. The points were connected with a third order spline interpolant which accounts for correlation between neighboring surface atoms. The scalar electric field integral equation (EFIE) and magnetic field integral equation (MFIE) were solved using the Nystrom method to obtain the reflected intensity as a function of observation angle. Verification of the accuracy of the code was obtained by means of comparison with well-known analytic solutions and approximations. The predicted Nevot-Croce factor drop in reflectance was found to be in general agreement with the computed decrease in reflectance due to surface roughness. However, an angle dependent difference was also noticed, indicating the Nevot-Croce factor might need revision. The code is being modified to run on a supercomputing cluster where longer, more realistic surfaces can be analyzed to determine whether an improved roughness correction factor is needed.

The objectives of this study were to carry out research to better understand of the formation, stability and properties of multilayer emulsions containing nano-laminated biopolymer coatings, and to utilize this information to develop food-grade delivery systems. The effect of various preparation parameters on the formation and stability of multilayer emulsions was investigated: droplet concentration; mean droplet diameter; droplet charge; biopolymer concentration. β-lactoglobulin (β-Lg) stabilized emulsions (0.5 – 10 wt% oil) containing different pectin concentrations (0 to 0.5 wt%) were prepared at pH 7 (where lipid droplets and pectin molecules were both anionic) and pH 3.5 (where lipid droplets were cationic and pectin molecules anionic) and “stability maps” were constructed. At pH 3.5, pectin adsorbed to the droplet surfaces, and the emulsions were unstable to bridging flocculation at intermediate pectin concentrations and unstable to depletion flocculation at high pectin concentrations. At certain droplet and pectin concentrations stable multilayer emulsions could be formed consisting of protein-coated lipid droplets surrounded by a pectin layer. An in situ electro-acoustic (EA) technique was introduced to monitor the adsorption of charged polysaccharides onto oppositely charged protein-coated lipid droplets. The possibility of controlling interfacial and functional characteristics of multilayer emulsions by using mixed polysaccharides (pectin/carrageenan or pectin/gum arabic) was then examined. Emulsions containing different types of polysaccharides had different interfacial characteristics and aggregation stabilities: carrageenan had the highest charge density and affinity for the protein-coated lipid droplets, but gave the poorest emulsion stability. The possibility of assembling protein-rich coatings around lipid droplets was examined using the electrostatic deposition method, with the aim of producing emulsions with novel functionality. Protein-rich biopolymer coatings consisting of β-Lg and pectin were formed around lipid droplets using the electrostatic deposition method. The composite particles formed had relatively small diameters (d < 500 nm) and were stable to gravitational separation. They also remained stable after they were heated above the thermal denaturation temperature of the globular protein and had better stability to aggregation at high salt concentrations (50 – 200 mM NaCl) than conventional emulsions stabilized by only protein. The effect of a polysaccharide coating on the displacement of adsorbed globular proteins by non-ionic surfactants from lipid droplet surfaces was examined to simulate situations where competitive adsorption occurs. Oil-in-water emulsions stabilized by β- Lg were prepared containing either no pectin (1º emulsions) or different amounts of pectin (2º emulsions). At pH 3.5, where pectin forms a coating around the β-Lg stabilized lipid droplets, the amount of desorbed protein was much less for the 2º emulsion (3%) than for the 1º emulsion (39%), which indicated that the pectin coating inhibited protein desorption by surface active agents. Knowledge gained from this research will provide guidelines for rationally designing emulsion-based delivery systems that are resistant to environmental stresses or with controlled release properties. These delivery systems could be used to encapsulate, protect and release functional components in various industrial products, such as foods, pharmaceuticals, cosmetics, and personal care products.

No description available.

No description available.

No description available.

No description available.