Biological remediation of Cr(VI) contaminated soil and groundwater is an emerging field. In this study, the in situ bioremediation technology for treating Cr(VI) contaminated groundwater aquifers was evaluated using a laboratory microcosm system. The study was conducted using columns with five equally spaced intermediate sampling ports along the length to facilitate finite difference modelling of the Cr(VI) concentration profile within the column. Cr(VI) concentration was continuously measured in the influent, in five equally spaced intermediate ports within the column and in the effluent port. The change or the shift in microbial community within the inoculated column was also monitored due to exposure to toxic conditions after seven weeks of operation using the 16S rRNA genotype fingerprinting method. The effect of introducing a natural carbon source (sawdust) in inoculated columns in comparison with the performance of sterile controls under various loading conditions was also evaluated. Near complete Cr(VI) removal was achieved in an inoculated carbon source reactor, whereas only 69.5% of Cr(VI) removal was achieved in an inoculated column without an added carbon source after 4 days of operation at 20 mg/L. In a sterile control reactor less than 2% of Cr(VI) was removed after 4 days of operation at 20 mg/L. Experimental cores demonstrated a successful Cr(VI) reduction process in the simulated microbial barrier system that was evaluated internally. The model that simulates Cr(VI) removal and transport in the subsoil environment was developed. The Cr(VI) mass balance model across the reactor that accounts for the flow characteristics and biological removal mechanism successfully captured the trends of Cr(VI) response profiles under quasi-steady state conditions for different loading conditions. This study demonstrate the potential of applying effective Cr(VI) reducers in the reactive barrier systems to contain or attenuate the spread of Cr(VI) contaminant in groundwater aquifer systems. The finite difference model developed in this study to evaluate the behaviour of Cr(VI) in the reactor could contribute towards improved designs of future in situ bioremediation systems that can be implemented for remediation of Cr(VI) on site. / Dissertation (MSc)--University of Pretoria, 2011. / Chemical Engineering / unrestricted

The steering arrangement of a 6x6 off-road military vehicle was investigated, with the aim to determine if a variable steering ratio between the first and second steering axle of the vehicle will make an improvement in the steady and transient state handling of the vehicle. Low speed manoeuvring was evaluated, comparing the vehicle steering geometry with Ackerman geometry. For steady state handling, a bicycle model was developed, and constant radius simulations at various track radii, vehicle speeds and steering ratios (ratio between the first and second steering axle) was performed. For transient dynamic simulations, a mathematical model was developed that included a simple driver model to steer the vehicle through a single lane change, again at various speeds and steering ratios. The vehicle was instrumented, and actual constant radii tests, as well as single lane change tests were performed. The measurements enabled the comparison of simulated and measured results. Although basic mathematical models were used, acceptable correlation was obtained for both steady state and transient dynamic behaviour. The results indicated that for this specific vehicle geometry, where the centre of mass is above the second axle, no marked improvement would be obtained by implementing a variable ratio steering system. The mathematical model was changed to simulate a vehicle with longer wheelbase and different centre of mass. With the new geometry, theoretical slip angles (and therefore tire wear) reductions were more noticeable It was concluded that a variable ratio system between the front and second axle would not be an economically viable improvement for this vehicle, since the improvement achieved will not warrant the additional cost and complexity added to the vehicle. / Dissertation (MEng (Mechanical Engineering))--University of Pretoria, 2007. / Mechanical and Aeronautical Engineering / MEng / unrestricted

Hydroelectricity plays an important role to balance the stability of grid network.  In order to improve the stability of presently high loaded grids, hydropower plants are being operated over a wide range of operations and experiencing frequent start-stop, load rejection, and load acceptance. The turbines need to sustain sudden change in their operating condition to balance the grid frequency. Francis turbines have been widely used because of their wider operating range and higher stability in operation during rapid load variation. This has resulted in severe damage to the turbines as they are not normally designed to operate under such transient conditions. Several low and high frequency pressure fluctuations prevail during transients operating conditions. Generally, wall pressure measurements are performed which may not provide sufficient information to investigate the flow instabilities related to these fluctuations. Thus, the main objective of the present work is to simplify and perform optical measurements in a turbine during transient operating conditions to investigate the flow field. The measurements have been performed at the Water Power Laboratory using a high head model Francis turbine. The turbine is a 1:5.1 scale down model of a prototype operating at the Tokke Power Plant, Norway. The model runner diameter, net head, and discharge at the best efficiency point (BEP) were 0.349 m, 12 m, and 0.2 m3 s-1, respectively. A total ten pressure sensors were mounted at different locations namely, turbine inlet, vaneless space, and draft tube. The data were acquired at a sampling rate of 5 kHz. The instruments and sensors have been calibrated according to guidelines available in IEC standards. The determined total uncertainty in the measurement of hydraulic efficiency was ±0.15% at BEP condition. The velocity measurements in the draft tube cone were performed using a 2D PIV system and the images were sampled at a rate of 40 Hz.      Steady state measurements were carried out considering the realistic design and off-design operating conditions of the prototype turbine. Therefore, the angular speed of the runner was maintained constant for all steady state conditions during the measurements. The maximum hydraulic efficiency (92.4%) was observed at nED = 0.18, QED = 0.15, and a = 9.8º, which is named BEP. It is observed that the turbine experiences significant pressure fluctuations at the vaneless space, runner, and the draft tube. The fluctuations due to rotor-stator interaction (RSI) were observed to be most dominating at high load condition, however, fluctuations due to the rotating vortex rope (RVR) at part load (PL) condition. Two different modes (synchronous and asynchronous) modes of vortex rope are observed at PL condition of the turbine. An asymmetry in the flow leaving the runner was detected at both design and off-design conditions, with a stronger effect during off-design operating condition. Numerical simulations of the model turbine were carried out at PL operating condition. The simulations were performed using two turbulence models, standard k-ε and SST k-ω, with high-resolution advection scheme. The numerical pressure values obtained with both standard k-ε model and SST k-ω showed a small difference with the experimental values. The amplitudes of numerical pressure values were higher (~2.8%) in the vaneless space and lower (~5.0%) in the draft tube than the experimental values. The frequencies of the RSI and RVR were well captured in the turbine but the amplitudes were overestimated.   During load rejection from BEP to PL, the plunging mode of the vortex rope was observed to appear first in the system than that of the rotating mode. Whereas during the load acceptance from PL to BEP, both the modes were observed to disappear simultaneously from the system. In the velocity data, the axial velocity only contributed to the development of the plunging mode and radial velocity to the rotating mode. The region of low velocity, stagnation point, flow separation, recirculation, oscillating flow and high axial velocity gradients were well captured in the system during the transients. The induced high-velocity gradients during the load acceptance from BEP to HL was observed to develop a vortex core in the draft tube. During startup and shutdown, the guide vanes angular position was moved from one to another steady state condition to achieve the minimum load condition of the turbine. At this condition, the generator of the turbine was magnetized at the synchronous speed during startup and shutdown, respectively. The frequency of wave propagation was observed to vary with the runner angular speed during startup and complete shutdown of the turbine. Comparatively high-pressure fluctuations in the draft tube were observed during the guide vane movement from the high discharge conditions. Some unsteady phenomena such as the formation of dead velocity zone, backward flow, and flow oscillations were observed during startup and shutdown of the turbine.   The current work has been also used to continue a series of workshops, i.e., Francis-99. The first workshop was held on December 2014 with the cooperation of LTU and NTNU. The measurements performed in this work were used for the second workshop which was held on December 2016. The investigations presented in this thesis will be further explored in the third workshop scheduled for December 2018.

We study diffusion-reaction-advection models describing population dynamics of aquatic organisms subject to a constant drift, with reflecting upstream and outflow downstream boundary conditions. We consider three different models: single logistically growing species, two and three competing species. In the case of a single population, we determine conditions for existence, uniqueness and stability of non-trivial steady-state solutions. We analyze the dependence of such solutions on advection speed, growth rate and length of the habitat. Such analysis offers a possible explanation of the "drift paradox" in our context. We also introduce a spatially implicit ODE (nonspatial approximation) model which captures the essential behavior of the original PDE model. In the case of two competing species, we use a diffusion-advection version of the Lotka-Volterra competition model. Combining numerical and analytical techniques, in both the spatial and nonspatial approximation settings, we describe the effect of advection on competitive outcomes. Finally, in the case of three species, we use the nonspatial approximation approach to analyze and classify the possible scenarios as we change the flow speed in the habitat.

The current behavioural dead region (DR) diagnosis methods such as psychophysical tuning curves and the threshold-equalising noise test require extensive subject co-operation. These present methods cannot be applied to infants. The work presented in the thesis aimed to develop a fast objective DR diagnosis method that could be applied to sleeping hearing-impaired infants. A novel fast objective electrophysiological method of recording response amplitude curves (RACs) which could enable objective DR diagnosis was developed.RACs were derived by recording auditory steady-state response amplitudes using modulated signals in the presence of narrow-band maskers. Two RAC methods were investigated. In the swept method, RACs were recorded in a single test run by recording the response amplitudes across the frequency range of a continuously swept-frequency narrow-band masker. In the fixed method, response amplitudes of eight separate test runs, each in the presence of differing fixed-frequency narrow-band maskers, were recorded.RACs were recorded in normally hearing adult subjects. The results showed that for normally hearing subjects in condition 1 (swept masker), the mean recorded RAC tip for a 2-kHz signal was 2250 Hz and the repeatability coefficient of two repeated recordings in each subject was 389 Hz; in condition 2 (fixed masker), the respective values were 2251 Hz and 342 Hz. These results indicated that the swept masking method is a viable and fast way to record RACs in normally hearing adults.RACs and psychophysical tuning curves (PTCs) were recorded in hearing-impaired adult subjects in order to asses the tip-frequency agreement between the tests. In some cases there were difficulties in using the required signal and masker levels due to maximum sound level limits. The RACs were poorly shaped and had poor repeatability. These findings indicate that the RAC method that was successfully applied to normally hearing subjects requires further development for use with the hearing impaired. The possible causes for the differences in the accuracy of the method between normally hearing and hearing-impaired subjects are discussed. The work presented in this thesis provides the basis upon which further research can be taken forward. It is envisaged that this work, together with further research, will lead to a clinically-effective objective DR diagnosis method.

This paper describes the IBM Power8 system in comparison to the Intel Xeon processors, widely used in today’s solutions. The performance is not evaluated only on the whole system level but also on the level of threads, cores and a memory. Different metrics are demonstrated on typical optimized algorithms. The benchmarked Power8 processor provides extremely fast memory providing sustainable bandwidth up to 145 GB/s between main memory and processor, which Intel is unable to compete. Computation power is comparable (Matrix multiplication) or worse (N-body simulation, division, more complex algorithms) in comparison with current Intel Haswell-EP. The IBM Power8 is able to compete Intel processors these days and it will be interesting to observe the future generation of Power9 and its performance in comparison to current and future Intel processors.

Maintenance and rehabilitation of existing masonry and reinforced concrete structures are of great importance in the field of civil engineering. Due to deterioration and severe environment, numerous structures fail to meet functional or safety requirements, and as a result, they should be strengthened. Several methods have been utilized to repair the structures, including steel plate bonding, cable post-tensioning, and section enlargement. However, these methods bring disadvantages, such as significant added dead load and high labour cost. Therefore, externally bonding with composite materials has attracted considerable attention recently. Externally bonded fibre-reinforced polymer (FRP) sheets have been widely used to strengthen reinforced concrete and masonry structures. FRP has been a common method to provide a higher service life for structures for several decades. However, strengthening structural members with FRP introduces certain drawbacks, such as their poor performance in fire scenarios caused by the rapid softening of the polymer-based resin. An alternative strengthening system known as a fabric-reinforced cementitious matrix (FRCM) has been developed to address this issue by replacing resin-based material with an inorganic cementitious-based matrix. Nonetheless, the performance of FRCM at high temperatures has not been investigated sufficiently so far. Hence, this research focused on the mechanical behaviour of FRCM at high temperatures. This experimental research investigates the tensile performance of carbon FRCM at high temperatures. First, the temperature distribution within the specimens during heating was studied using nine specimens with one, two, or three layers to reveal the required time for the inner fabric to reach a steady temperature. Then, the tension and stiffness degradation of FRCM coupons were studied at different temperatures. A total of 84 FRCM coupons were fabricated and tested in tension; 60 of the tests were conducted at steady-state conditions in which temperature was held constant and load increased, and 24 specimens were carried out in transient-state tests, in which load was constant, and temperature grew. In order to provide a more comprehensive knowledge concerning the FRCM composite, some key variables were included in this research. These parameters are the number of layers (1, 2, 3) leading to different thicknesses (20, 30, 40 mm), the orientation of the fabric layer (unidirectional and bidirectional), target temperature (ambient, 100, 200, 300, 400°C), and heating condition (steady-state, transient state). These tests aimed to reveal the primary mechanical characteristics such as ultimate strength and cracked elastic modulus at different temperatures and compare them with control specimens tested at room temperature. With the increase in the number of fabric grids from one to two and three, the stress at failure decreased by about 11 and 18%, respectively. With regards to cracked elastic modulus two and three-layered specimens showed 18 and 20% reduction in value. It is also noteworthy to mention that overall load capacity of specimens rose with the increase in number of layers; however, due to the more significant increase in area, the stress was reduced. The same decreases in the cracked elastic modulus and ultimate strength were observed as the target temperature increased. Increasing the temperature to 400°C led to a decrease in ultimate strength and cracked elastic modulus of approximately 60 to 70%. Furthermore, the bidirectional specimens showed a better behaviour than unidirectional specimens in terms of ultimate strength; however, their cracked elastic moduli were almost the same. With regards to the transient-state tests, as the material became thicker, the failure temperature increased considerably. For instance, a 20-mm specimen failed at 467°C with a 20% sustained load, while a 30-mm specimen failed at 558°C. Another vital parameter studied in transient-state tests was the decrease in temperature with the increase in sustained load. An example of this is the 20-mm specimens which failed at 352 and 258°C, while they were preloaded to 40 and 60% of their capacities. The conclusions of this study suggest that FRCM materials do retain a non-negligible strength capacity at high temperatures. However, further investigations to reveal FRCM bond behaviour and retrofitted structural members at high temperatures are still required to provide comprehensive knowledge.

Rationale: Schizophrenia patients consistently show deficits in sensory-evoked broadband gamma oscillations and click-evoked entrainment at 40 Hz, called the 40-Hz auditory steady-state response (ASSR). Since such evoked oscillations depend on cortical N-methyl D-aspartic acid (NMDA)-mediated network activity, they can serve as pharmacodynamic biomarkers in the preclinical and clinical development of drug candidates engaging these circuits. However, there are few test-retest reliability data in preclinical species, a prerequisite for within-subject testing paradigms. Objective: We investigated the long-term psychometric stability of these measures in a rodent model. Methods: Female rats with chronic epidural implants were used to record tone- and 40 Hz click-evoked responses at multiple time points and across six sessions, spread over 3 weeks. We assessed reliability using intraclass correlation coefficients (ICC). Separately, we used mixed-effects ANOVA to examine time and session effects. Individual subject variability was determined using the coefficient of variation (CV). Lastly, to illustrate the importance of long-term measure stability for within-subject testing design, we used low to moderate doses of an NMDA antagonist MK801 (0.025–0.15 mg/kg) to disrupt the evoked response. Results: We found that 40-Hz ASSR showed good reliability (ICC=0.60–0.75), while the reliability of tone-evoked gamma ranged from poor to good (0.33–0.67). We noted time but no session effects. Subjects showed a lower variance for ASSR over tone-evoked gamma. Both measures were dose-dependently attenuated by NMDA antagonism. Conclusion: Overall, while both evoked gamma measures use NMDA transmission, 40-Hz ASSR showed superior psychometric properties of higher ICC and lower CV, relative to tone-evoked gamma.

Theoretischer Hintergrund: Störungen des Körperbildes stellen einen der zentralen Risikofaktoren für die Entstehung und Aufrechterhaltung von Essstörungen dar. Gemäß der kognitiv-behavioralen Theorie für Körperbildstörungen führen körperbezogene Stimuli zu einer Aktivierung körperbezogener Schemata, welche zu Verzerrungen in kognitiven Prozessen führen können. Im Zuge der Untersuchung kognitiver Verzerrungen blieb bislang die Frage offen, ob sich kognitive Verzerrungen primär auf den eigenen Körper oder auf Körper generell erstrecken. Innerhalb dieser Dissertation wurde daher untersucht, ob die Identität eines Körpers evaluative, attentionale und perzeptive Prozesse der Körperverarbeitung beeinflusst. In den Studien 1 bis 3 wurde mittels eines Forschungsparadigmas zur Manipulation der Identität geprüft, ob Doppelstandards bei der Bewertung von Körpern in unterschiedlichen Populationen vorliegen. In Studie 4 wurde außerdem der attentionale Verlauf für den eigenen Körper mit dem für einen fremden Körper und in Studie 5 die Erkennung des eigenen Körpers und fremder Körper, sowie deren neuronale Korrelate, bei Frauen mit hohen und niedrigen Figur- und Gewichtssorgen verglichen. Methode: Für die Studien 1 bis 3 wurden weibliche und männliche Körper mittels einer Software erstellt und N = 104 Frauen ohne Essstörungen (Studie 1), n = 34 Frauen mit Anorexia nervosa und n = 31 Frauen mit Bulimia nervosa (Studie 2), sowie N = 93 Männern ohne Essstörungen (Studie 3) präsentiert. Die Probanden und Probandinnen sollten nach der Präsentation jedes Körperstimulus einordnen, wie sie sich gefühlt haben, und die Körper bewerten. Um die Identität der Körper zu manipulieren, wurden die Körper mit demselben Geschlecht der Versuchsperson sowohl mit dem Kopf einer fremden Person als auch mit dem Kopf der jeweiligen Versuchsperson präsentiert. In Studie 4 wurde ein bereits etabliertes EEG-Paradigma zur Untersuchung von Aufmerksamkeitsverzerrungen bei Körpern unterschiedlicher Identität und Gewichtsausprägungen angewendet. N = 20 Frauen mit hohen Figur- und Gewichtssorgen und n = 24 Frauen mit niedrigen Figur- und Gewichtssorgen wurden Bilder des eigenen Körpers der jeweiligen Probandin oder eines fremden Körpers präsentiert, während sie sich mit einer Detektionsaufgabe befassten. Durch die Messung eines Steady-State visuell evozierten Potentials (SSVEP) konnte abgeleitet werden, wie stark die einzelnen Körper von der Detektionsaufgabe ablenkten. In Studie 5 wurden n = 25 Frauen mit hohen Figur- und Gewichtssorgen und n = 26 Frauen mit niedrigen Figur- und Gewichtssorgen Morphfilme präsentiert, bei denen sich ein Körper allmählich in den eigenen Körper der jeweiligen Probandin oder einen fremden Körper verwandelte. Es wurde die Performanz in der Erkennung der Körper, sowie als neuronales Korrelat ein SSVEP miterhoben. Nach der SSVEP-Aufgabe wurden die einzelnen Morphs in einer weiteren Aufgabe ohne EEG präsentiert und die Erkennungsleistung erfasst. Ergebnisse: In Studie 1 zeigte sich, dass Frauen ohne Essstörungen keine körperübergreifenden Doppelstandards in der Körperbewertung aufwiesen, einen übergewichtigen Körper aber deutlich negativer bewerteten, wenn dieser die eigene Identität im Vergleich zu einer fremden Identität trug. In Studie 2 wurde erkenntlich, dass Frauen mit Anorexia nervosa und Frauen mit Bulimia nervosa mehr selbstschädigende Doppelstandards aufwiesen als Frauen ohne Essstörungen. Über alle Körpertypen hinweg bewerteten sie die Körper als dicker, wenn diese die eigene Identität im Vergleich zur fremden Identität trugen und unterschieden sich hierhin von Frauen ohne Essstörungen. Frauen mit Anorexia nervosa zeigten darüber hinaus signifikant deutlichere selbstschädigende Doppelstandards in den anderen Variablen als Frauen ohne Essstörungen. Studie 3 zeigte auf, dass Männer ohne Essstörungen unattraktive Körper negativer bewerteten, wenn sie die eigene Identität aufwiesen, und den idealen muskulösen Körper positiver bewerteten, wenn dieser die eigene Identität trug. In Studie 4 konnte darüber hinaus gezeigt werden, dass sich Frauen mit hohen Figur- und Gewichtssorgen stärker und andauernd durch Bilder ihres eigenen Körpers von einer Aufgabe ablenken ließen als durch Bilder fremder Körper. Dies traf auf Frauen mit niedrigen Figur- und Gewichtssorgen nicht zu. Studie 5 ergab, dass sich Frauen mit hohen und mit niedrigen Figur- und Gewichtssorgen generell nicht in der Erkennungsleistung und im SSVEP unterschieden. Allerdings brauchten Frauen mit hohen Figur- und Gewichtssorgen mehr Anteile des eigenen Körpers, um diesen in der Vermischung mit einem dünnen Körper zu erkennen. Diskussion: Die Ergebnisse verweisen auf die Aktivierung differentieller körperbezogener Schemata bei Körpern der eigenen Identität im Vergleich zu Körpern einer fremden Identität, die daher zu divergierenden Bewertungen dieser Körper führen. Dabei scheinen solche Doppelstandards in der Körperbewertung selbstschädigender zu sein, wenn eine Essstörungspathologie vorliegt. Außerdem können Männer im Gegensatz zu Frauen sich auch selbstdienlich bewerten, was zu den oftmals nachgewiesenen Geschlechtsunterschieden im Körperbild beitragen könnte. Darüber hinaus lassen sich Frauen mit hohen Figur- und Gewichtssorgen deutlicher durch ihren eigenen Körper als durch fremde Körper ablenken, was nahelegt, dass neben Körperbewertungen auch Aufmerksamkeitsverzerrungen von der Identität eines Körpers abhängig sind. Perzeptive Prozesse für die Differenzierung des eigenen Körpers von fremden Körpern erschienen gleichermaßen funktional bei Frauen mit hohen und niedrigen Figur- und Gewichtssorgen. Allerdings könnte auch hier der Einfluss dysfunktionaler körperbezogener Schemata zu einer schlechteren Erkennungsleistung bei Frauen mit hohen Figur- und Gewichtssorgen bei der Vermischung des eigenen Körpers mit einem dünnen Körper geführt haben. Insgesamt konnten mit dieser Dissertation einzelne Aspekte der kognitiv-behavioralen Theorie für Körperbildstörungen empirisch untermauert werden und es konnte gezeigt werden, dass kognitive Verzerrungen in der Körperverarbeitung von der Identität des Körpers und dem Ausmaß einer Körperbildstörung abhängig sind.

Le développement récent du transport ferroviaire a conduit à une forte augmentation du trafic durant la dernière décennie, augmentation qui devrait sepoursuivre dans la décennie à venir avec les futurs développements de ce mode de transport. L'augmentation du trafic se traduit pour l'infrastructure par une augmentation des sollicitations subies par les voies ferrées. La qualité des voies ferrées est garante de la performance du mode de transport ferroviaire,mais également de la sécurité des matériels roulants et des usagers de l'infrastructure. Conserver la qualité du réseau ferré est donc une priorité pour les gestionnaires d'infrastructures ferroviaires. Le ballast ferroviaire a pour rôle de répartir les efforts induits par la circulation des trains et d'assurer la bonne géométrie de l'infrastructure. La surveillance de la dégradation du ballast est donc primordiale pour assurer la qualité del'infrastructure.L'objectif de cette thèse est de proposer un outil numérique permettant d'anticiper la dégradation et le comportement d'une infrastructure ferroviaire sous trafic. Cet outil doit permettre de simuler efficacement lecomportement du ballast ferroviaire sous un grand nombre de cycles de chargements mobiles.Une première partie détaille la composition des infrastructures ferroviaires, précise les rôles du ballast et identifie les grandes lignes de son comportement.Il est fait le choix d'utiliser une méthode de représentation du ballast par élément fini et un modèle de comportement élastoplastique. Le ballast est alors traité comme un géomatériau continu répondant à des lois de comportementélastoplastique.Dans une seconde partie, les bases de la modélisation élastoplastique sont rappelées. Puis, plusieurs modèles de comportement issus de la littérature et comportant des éléments intéressants sont identifiés et étudiés.Dans une troisième partie, un nouveau modèle de comportement frottant compactant est proposé. Il est composé d’un critère nouveau et d'une loi d'écoulement originale. Enfin, le modèle est complété par la prise encompte, de la dégradation du matériau via une diminution de l'angle de frottement interne.Dans une quatrième partie, les méthodes stationnaires sont rappelées. Ces méthodes sont spécifiquement développées pour la modélisation des problèmes incluant des charges en mouvement. Les différents algorithmes sont implémentés àl'aide d'un code de calcul développé dans le logiciel Matlab et à l'aide du logiciel d'éléments finis COMSOL Multiphysics. En particulier, la méthode stationnaire à double échelle de temps, qui est un développement nouveau, permet la modélisation rapide d'un grand nombre de cycles de chargements et la prise en compte des effets d'usure du matériau.Dans une dernière partie, les deux principaux modèles tridimensionnels utilisés sont présentés. Sur un faible nombre de cycles de chargements, divers résultats sont établis selon les différents modèles de comportement utilisé. Sur un très grand nombre de cycles de chargements, les résultats obtenus montrent l'évolution des déformations irréversibles de la structure et de la dégradation du matériau. / Recent development in rail transportation has led traffic increasing during last decade, an increase which should continue in next decade due to the future developments of railway transport. Traffic growing induces infrastructure's solicitation increasing. Railways quality is warrant of rail transport performance, but also ensure security of the rolling stock and safety of the infrastructure users. Maintaining quality of the rail network is major priority for railway infrastructure managers. Purpose of rail ballast is to distribute load of the moving trains and to ensure track geometry. Monitoring of ballast degradation is therefore essential to ensure the quality of the infrastructure.The aim of this thesis is to develop a numerical tool to anticipate the degradation and behaviour of a railway infrastructure under traffic. This tool is intended to effectively simulate the behaviour of railway ballast under numerous mobile loads cycles.A first part details the composition of railway infrastructures, specifies roles of the ballast and identifies the main lines of its behaviour.Choice is made to use finite element method to represent the ballast and an elastic-plastic behaviour model. The ballast is then treated as a continuous geomaterial responding to elastic-plastic behaviour. Therefore, in a second part, the basics of elastic-plastic modelling are exposed. Then, several existing behaviours models which present interesting elements are identified and studied.In a third part, a new behaviour model, called "frottant-compactant" is presented. It is composed with a new criterion and an original flow rule. Finally, the model is completed by taking account of materials degradation via internal friction angle decreasing.In a fourth part, the steady-states methods are exposed. These methods are specifically developed for modelling problems including moving loads. Different algorithms are implemented using Matlab coding software and using the finites elements method software COMSOL Multiphysics. Principally, the double time-scale steady-state method, newly developed, allow rapid modelling of numerous loading cycles and representation of material wear effects.In a final part, the two three-dimensional models used are presented. On a few loading cycles, various results are established depending on different behaviour models. On a very large number of load cycles, the results obtained show the evolution of plastic strains in the structure and the degradation of the material.