Pour l’industrie automobile, l’identification de sources vibratoires dans un but d’amélioration du confort acoustique est un enjeu majeur. L’identification d’une source vibratoire se résume le plus souvent à la connaissance de sa matrice des fonctions de transfert et des efforts vibratoires au niveau des points de fixation entre la source et son récepteur. Le moyen le plus simple pour effectuer des mesures d’efforts vibratoires est de réaliser des mesures directement à l’aide de capteurs d’effort. Cependant, pour diverses raisons pratiques (accessibilité, encombrement... ), ces efforts vibratoires sont bien souvent impossibles à mesurer de manière directe. Dans ces circonstances on a alors recours à une mesure indirecte des efforts à partir des réponses de la structure en fonctionnement et d’un modèle dynamique de la structure. La mise en œuvre d’une telle démarche est souvent difficile car les problèmes inverses sont en général mal posés au sens d’Hadamard. L’objectif du présent travail de thèse a été l’amélioration de la robustesse de l’identification de source vibratoire par méthode inverse. Le manuscrit est découpé en cinq chapitres : — Le premier chapitre aborde en détail la problématique de la reconstruction des efforts par méthode inverse. On y présente un état de l’art permettant d’appréhender les enjeux et les difficultés de la reconstruction des efforts par méthode inverse. Nous explicitons également dans ce chapitre un certain nombre de relations permettant le couplage des matrices des fonctions de transfert, la caractérisation de la réceptance d’une source vibratoire, la reconstruction des efforts transmis à l’interface, des "réponses libres" et des "forces de blocage" dans le cas général. — Dans le second chapitre sont présentées, dans un premier temps, certaines des méthodes de régularisation du problème inverse les plus utilisées puis dans un second temps nous proposons deux approches de régularisation originales : — La méthode de sélection de points de mesure "CIM". L’objectif de cette méthode est de positionner de manière optimale un ensemble de capteurs de réponse sur la structure étudiée. La méthode opère par itérations initialisées par une matrice des fonctions de transfert contenant tous les points de mesure envisageables et en éliminant progressivement ces derniers pour ne conserver au final que les points de mesure offrant le maximum d’information. — Afin d’éliminer le bruit pouvant entacher les fonctions de transfert et donc améliorer le conditionnement du problème, nous proposons de les lisser à l’aide d’une décomposition modale. Pour ce faire, nous exprimons les fonctions de transfert en fonction des paramètres modaux de la structure, obtenus à l’aide d’une analyse modale expérimentale. Ces méthodes de régularisation ont été validées sur un système numérique simple mais représentatif de la transmission d’efforts vibratoires. — Dans le troisième chapitre nous abordons la prise en compte des incertitudes lors de la reconstruction des efforts par méthode inverse. Nous y présentons en détail la méthode de propagation de l’incertitude nommée "méthode du chaos polynomial". Cette méthode formalise, par un développement en série de polynômes de variables aléatoires, la séparation entre la partie déterministe et la partie stochastique d’une fonction aléatoire. Nous proposons d’utiliser cette méthode afin de propager vers la solution du problème inverse l’incertitude entachant les réponses et le modèle de la structure. Cette stratégie de résolution a été appliquée à un problème numérique représentatif et a fait la preuve de son efficacité comparativement à une méthode de Monte Carlo. — Dans le quatrième chapitre, nous présentons une méthode originale de caractérisation de sources vibratoires que nous avons nommé la méthode du "banc observant". La caractérisation d’une source vibratoire nécessite généralement d’effectuer des mesures au niveau de son interface. [...] / For the automotive industry, the quantification of vibratory forces is a key challenge to improve acoustic comfort. The identification of a vibratory source comes down to the knowledge of its matrix of transfer functions and vibratory forces at the points of attachment between the source and its receiver. The easiest way to measure vibration forces is to take measurements directly using force sensors. However, for various practical reasons, these vibratory forces are often impossible to measure directly. In these circumstances, an indirect measure of vibratory forces is used. The implementation of this approach is often difficult because the inverse problems are generally ill posed in the sense of Hadamard. The objective of this study was to improve the robustness of vibratory source identification by inverse method. The manuscript is divided into five chapters : — The first chapter deals in detail with the problem of the measurement of vibratory forces by inverse method. It presents a state of the art to understand the difficulties of identifying vibrational forces by inverse method. We also explain the relations allowing the coupling of two matrices of transfer functions, the characterization of the receptance of a vibratory source, the reconstruction of the forces transmitted to the interface, the "free answers" and the "blocked forces" in the general case. — In the second chapter are presented, first, some of the methods of regularization of the inverse problem the most used then in a second time we propose two original approaches of regularization : — The method of selecting measuring points "CIM". The objective of this method is to optimally position a set of sensors on the studied structure. The method operates by iterations initialized by a matrix of the transfer functions containing all the possible measurement points and by progressively eliminating them so as to keep only the measurement points offering the maximum of information. — In order to eliminate the noise that could tarnish the transfer functions and thus improve the conditioning of the problem, we propose to smooth them using a modal decomposition. To do this, we write the transfer functions according to the modal parameters of the structure, obtained using an experimental modal analysis. These regularization methods have been validated on a simple numerical system but representative of the transmission of vibratory forces. — In the third chapter we take into account the uncertainties in the reconstruction of the inverse method efforts. We present in detail the method of propagation of uncertainty called "Polynomial Chaos". This method formalizes, by a series development of polynomials of random variables, the separation between the deterministic part and the stochastic part of a random function. We propose to use this method in order to propagate to the solution of the inverse problem the uncertainty affecting the responses and the model of the structure. This strategy has been applied to a representative numerical problem and has proved its effectiveness compared to a Monte Carlo method. — In the fourth chapter, we present an original method of characterization of vibratory sources that we have named the "observing bench" method. The identification of a vibratory source usually requires measurements at its interface. [...]

Modern man’s addiction to fossil fuels or non-renewable energy is the key reason behind the unprecedented economic growth experienced globally over the past 100 years. However, by definition these energy resources are not only finite, but their widespread use is causing massive environmental damage through air pollution and its associated impact on people’s health, as well as the emission of greenhouse gases which are attributed to the unprecedented rate of global warming - And it is for this reason that international initiatives such as the Kyoto Protocol, (which South Africa is a signatory of), aim to mitigate global warming by reducing member countries’ CO2 emissions.Simultaneously, South Africa (SA) is experiencing its own electricity supply problems due to under investment in the sector. While new power plants are being built, they utilize non-renewable energy sources and will take time to build (up to 5 years). It is also important to note that due to large coal reserves, South Africans enjoy amongst the lowest electricity tariffs in the world, but SA is amongst this planet’s biggest per capita polluters.The research thus aims to identify whether high income households are wasteful users of electricity - due to historic low prices, lack of knowledge regarding energy efficiency and the impact that electricity generation has on the environment - while at the same time determining the group's perception of domestic solar water heaters (DSWH), given our country’s favourable climaticconditions. / Dissertation (MBA)--University of Pretoria, 2010. / Gordon Institute of Business Science (GIBS) / unrestricted

No description available.

In the present work, an inter principle research is carried out on complex fluid flow and heat transfer, using an innovative technology Nuclear Magnetic Resonance (NMR), from the Department of Physics. To enhance heat transfer performance of complex fluid flow, the present work mainly focuses on two different parts, one is the adoption of nanofluid; the other is flow forces analysis through bionic engineering studies on plant water migration system. Nanofluids are attracting considerable attention from both academic and industrial communities. Comparing with conventional pure fluids or solutions, nanofluids have higher thermal conductivity, due to the high surface to volume ratio of nanoparticle and liquid interface, which exhibits a great potential in enhancing heat transfer performance in various occasions. It is believed that different types and concentrations of nanofluid could strongly affect the thermal performance, and a great number of papers have been published to illustrate the phenomenon. However, none has really focused on the possible concentration change of nanofluid while flowing. Otherwise, the thermal performance of nanofluid flow could never be quantified. In the present work, a novel method to measure the dynamic concentration of nanofluid is proposed, using NMR technology. The experiments were carried out with ferrofluid under different concentration and temperature. A new parameter T2* was introduced in the study. T2* is a relaxation time of the signals that is released by hydrogen atoms after Radio Frequency (RF). And this signal intensity can be strongly affected by nanofluids. Experiments were carried out to obtain the T2* of nanofluid in the pipe. An empirical equation based on T2* and temperature was proposed to calculate the concentration of nanoparticles. Then, experiments were carried out with flowing ferrofluid in pipe. The dynamic concentration was calculated with the empirical equation. And with the series of experiment, it is confirmed that the flowing nanofluid consists of an obvious concentration gradient, and thus cause different layers of thermal performance from boundary to central line of a laminar pipe flow. Furthermore, the experiment result also gives out a chance to investigate the mechanism of nanoparticle movement in laminar flow with the concentration gradient along radius. Bionic Engineering is another research field that has been more and more interesting to researches from various fields. Since life has been evolving for over millions of years, many functions in lives has become extremely high efficiency and adaptive. These functions can be very worthy for researchers to study and utilize in industries. For heat transfer in fluid flow, it is very important to enhance the flow pattern. And thus water migration system in plants become very attractive. Plant can take water from soil up to several metres high. Learning from the water migration process in plants has been attracting interests from scientist for over a hundred years. The water migration in plant stem, especially xylem, involves various driving forces including capillary effect, osmosis effect, Marangoni effect and transpiration effect, etc. This present work mainly focuses on the water transport process within xylem. As xylem system is simplified as micro channel, a mathematic model is presented based on micro channel theory, with critical analysis and simplification. With a simplified micro channel from xylem structure and the calculation using the model of water migration in xylem, the relationship between various forces and water migration velocity is identified. The velocity of water migration within the plant stem is considered as detail as possible using all major forces involved. And a full mathematical model is proposed to calculate and predict the velocity of water migration in plants. Comparison between the calculated result and experimental one is made, to confirm the accuracy of the mathematical model. The present work proved that the mathematical model should be enough to predict the water migration in plants, and could also be critical for future water transport prediction in complex fluid flow in industry applications such as heat pipe.

This thesis is concerned with the physics of extragalactic radio - sources. Radio observations of a given source provide a powerful method with which to make direct measurements of the physical processes occuring within the la~ge-scale radio lobes and hotspots . Multi-frequency observations with up to five frequencies have been made for a number of ~ources using the Cambridge 5-km telescope and the Very Large Array. These data have been used to determine the change in radio spectra across the source . to obtain information about the way in which the electron population has evolved. Combining this with inferred magnetic field directions, the dynamics of seven sources have been examined. Additionally the hotspot advance speeds for a statistically representative sample of sources have -been derived. For 3C234 it was possible to test the assumptions used throughout the analysis. They were found to h6ld good to within the limits of error. The results are best interpreted within the framework of a fluid-dynamical model lobes as have recently been examined in numerical experiments. The general distibution of radiating plasma can be understood in terms of the flow of material out of the main accelerating regions (the hotspots - a basic beam model � is assumed thr6ughout) and its subsequent interaction with the surrounding medium, coupled with some precession of the radio jet. As well as looking at source dynamics , the detailed spectra have been used to search for areas of particle acceleration not associated with hotspots. Three sources have been found to have such regions, but large scale acceleration within the radio-lobes is not required . The derivation of ho tspot advance speed for a statistically representative sample has enabled statistical correlations to be investigated . Advance speed is found to be strongly correlated with power and this is discussed in the wider context of overall source models.

Observations have been carried out to survey the northern galactic plane for sources of highly variable radio emission at 5 GHz. To over—come the biases of previous searches for variability, the survey is conducted by making repeated and systematic observations of the survey region, consisting of the area within about ±2° of the galactic plane in the longitude interval of 40° to 220°. This thesis presents the analysis and results of the first three years of observations, comprising 40% of the total survey. The observations were carried out for a total of 3 months, in August of 1977, 1978 and 1979, and cover an area of over 200 square degrees, with a resolution of 3'. Within this area, a total of 806 compact radio sources have been detected. The catalogue includes sources with flux density as low as 15 mJy, and is complete down to 70 mJy. Of these sources, 758 have been examined for variation on a time scale of a few days (short—term), and 434 for variations on time scales of one or two years (long—term). Twenty—three new variable radio sources have been discovered; 12 short—term and 11 long—term. An additional 18 sources are possibly variable. The amplitudes of the long—term variations are similar to those of known extragalactic variables. A number of short—term variables exhibit much larger variations. The longitude distribution suggest that the majority of short—term variables are galactic, with luminosities in the range 10³⁰—10³⁵ ergs—s⁻¹. This luminosity range is similar to that of the strong X—ray binaries sources, such as Cyg X-3, SS 433 and Sco X-1. To date, extensive follow—up observations have been carried out for only one of the variable sources discovered. This source (GT0236+610) is periodic, undergoing a radio outburst every 26.52 days. The source is positionally co-incident with the BO star LS 1+61°303 and this identification has been confirmed by Gregory et al. (1979). GT0236+610 is an X-ray source (Share et al. 1978, Bignami et al. 1980) and, in addition, is the most probable counterpart of the COS B ƴ-ray source CG135+01 (Gregory and Taylor 1978, Pollock et al. 1981 ). Another highly variable source (GT2116+493) is also found to be co—incident with a stellar object. This source is probably an RS CVn type binary at a distance of about 300 pc. Comparison of source counts from the catalogue of compact sources, to extragalactic results, show that >200 of the compact sources, with flux density less than 60 mJy, are galactic. The non—variable galactic sources are likely to be small HII regions within 6 kpc of the sun. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

A description and the results of two surveys to search for highly variable radio emission in the galactic plane are presented. The surveys were carried out by making daily beamswitched maps of the survey regions and comparing the results of each day to the average of all days for evidence of variability. The first survey, at 21 cm, maps an area of 86 sq. deg. of the local spiral arm, in the direction of the constellation Cygnus, and searches for variability to a sensitivity of ~ 0.3 Jy. The second survey, at 6 cm, covers an area of 56 sq. deg. in the galactic longitude range (ℓˡˡ) 40°- 140°. The study of variability in this survey is not yet complete but variations at a level of ~10 mJy are expected to be measurable. It is concluded that large intensity variations (greater than ~ 1 Jy) on a time scale of days are extremely rare; only one possible variation at a level of 0.9 Jy was detected during the 52 days of the 21 cm survey. A preliminary examination of the 6 cm data reveals one highly variable radio source, designated GT 0236+610, that reached a maximum flux density of 285 mJy. The results of further observations of this source are presented. The 6 cm observations were averaged over the complete survey period and the data were searched for compact sources down to a level of 10 - 20 mJy. The positions and 6 cm flux density for the resulting 196 compact sources are listed. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

Efficient bale stack location, infield bale logistics, and equipment track impacted area were conducted in three different studies using simulation in R. Even though the geometric median produced the best logistics, among the five mathematical grouping methods, the field middle was recommended as it was comparable and easily accessible in the field. Curvilinear method developed (8-259 ha), incorporating equipment turning (tractor: 1 and 2 bales/trip, automatic bale picker (ABP): 8-23 bales/trip, harvester, and baler), evaluated the aggregation distance, impacted area, and operation time. The harvester generated the most, followed by the baler, and the ABP the least impacted area and operation time. The ABP was considered as the most effective bale aggregation equipment compared to the tractor. Simple specific and generalized prediction models, developed for aggregation logistics, impacted area, and operation time, have performed well (0.88?R2?0.99). An ABP of 8 bales capacity, also capable of 11 bales/trip, was recommended. / North Dakota State University. Teaching Assistant Scholarship

Les particules atmosphériques ont des impacts économiques, sanitaires et environnementaux importants à l'échelle mondiale. La réduction de leurs émissions afin d’améliorer la qualité de l'air ambiant exige une connaissance approfondie de leurs sources, qui peut être apportée par l'étude de leur composition chimique. La présente thèse vise à identifier les principales sources et origines géographiques des PM10 sur 5 sites d'échantillonnage (3 urbains, 1 trafic et 1 distant) représentatifs du Nord de la France, fréquemment soumis à des dépassements des valeurs réglementaires. La méthodologie a consisté en : (i) une caractérisation chimique complète des PM10 sur les 5 sites ; (ii) l’identification et la quantification des sources primaires et des processus secondaires pour chaque site par factorisation par matrice positive (PMF) ; (iii) l’estimation des champs de concentration (CF) des sources lointaines identifiées par PMF par combinaison avec les rétrotrajectoires de masses d’air, afin de localiser les zones d’émissions.La matière organique et le nitrate d'ammonium prédominent, avec quelques différences selon le type de site. De fortes contributions en aérosols secondaires (riches en nitrate, sulfate et oxalate) ont été observées, ainsi qu’un impact significatif du trafic et de la combustion de biomasse. Les zones d’émission identifiées pour les particules naturelles sels marins frais et vieillis (principalement de l'océan Atlantique) et biogéniques marins (principalement de la mer du Nord), et les contributions continentales anthropiques de particules secondaires riches en nitrate et sulfate, valident l’approche utilisée. / Airborne particles have significant economical, health and environmental impacts at a global scale. Mitigating their emissions to improve ambient air quality demands a deep knowledge on their sources, which can be determined by investigating their chemical composition. The present thesis aims at identifying major PM10 sources and geographical origins at 5 sampling sites (3 urban background, 1 traffic and 1 remote) representative of Northern France, which is frequently submitted to exceedances of limit values as defined by European Directives. The first step of this study included a comprehensive chemical characterization of PM10 filter samples collected every third day at the 5 sites. Organic matter and ammonium nitrate were confirmed as the main PM10 species in the investigated area with some differences according to the site type. Positive Matrix Factorization (PMF) was used to identify and quantify the contributions of primary sources as well as secondary processes impacting each sampling site. Large contributions of secondary aerosols (mainly grouped as nitrate-, sulfate- and oxalate-rich aerosols) were obtained at all sites, as well as significant traffic and biomass burning. The use of specific tracers like MSA and polyols also proved to be useful to identify marine and continental biogenic aerosols, respectively. Statistical trajectory-based model was applied to PMF outputs to calculate Concentration Field maps and locate emission sources of marine particles, namely fresh and aged sea salts (primarily from the Atlantic Ocean) and marine biogenic aerosols (mainly from the North Sea), as well as continental contributions of nitrate-and sulfate-rich secondary particles.

No description available.