The dissolved air flotation (DAF) has been used in drinking water treatment for its excellent algae and natural organic matter (NOM) removal. DAF drinking water treatability test are often conducted in a DAF jar test apparatus. Although, DAF jar test studies showed that they were able to predict NOM removals at full-scale facilities well, they have not always been successful in predicting the turbidity removals. One possible reason of the DAF jar test inaccuracy results could be associated to the small jar diameter, which may create wall effects. Therefore, the first two objectives of this research are: a) to develop and test a new, larger diameter and larger volume batch bench-scale dissolved air flotation system (LB-DAF) to better simulate turbidity removals in drinking water applications; b) to confirm these results by comparing the LB-DAF and full-scale DAF turbidity removals for two other source waters. The raw water characteristics of the three plants were quite different and the testing was performed at different times of the year. The development/optimization of the LB-DAF evaluated the impact of different variables (i.e., mixing intensity, water depth/tank diameter ratio, impeller shape, saturator pressure and recycle ratio). The results showed that the LB-DAF predicted well the full-scale DAF turbidity removals at three water treatment plants, and these predictions were better than those of DAF jar tests. For the LB-DAF design and operational variables evaluated had a limited impact on the turbidity removals. The LB-DAF predicted well DAF full-scale turbidity removals regardless of water temperature. This is an indication of the robustness of the DAF system. Ballasted sedimentation (BS) is a compact coagulation/flocculation and sedimentation process combination that has become very popular because it is very compact and because it can handle large variations in raw water turbidity and NOM. The literature survey did not initially identify studies on the BS treatment of algal impacted waters, for which DAF is considered particularly suitable. Thus, the third main objective of this dissertation was to compare the efficiency of BS with that conventional gravity settling (CGS), and that of DAF for the treatment of an algal impacted water via jar tests. These comparisons were performed at the Belleville Water Treatment Plant using Bay of Quinte water, one of the most eutrophic zones of Lake Ontario. Unfortunately, a change of weather prior to the testing resulted in raw water samples with relatively low concentrations of algae and cyanobacteria. The testing showed that DAF and BS had very similar NOM, cyanobacteria/algae (chlorophyll a and phycocyanin) removals.; however, the BS required microsand addition, polymer addition and a slightly higher alum dose. Only for turbidity removal the DAF was somewhat superior. It is suggested that these comparison experiments be repeated with waters that are more impacted by algae and cyanobacteria.

According to the regulation announced by Bureau of Standard, Metrology & Inspection (BSMI) of Ministry of Economic Affairs (MOEA), lamp fixtures must follow safety and electromagnetic compatibility (EMC) standards. However, the self-ballasted fluorescent lamps in the fixture should only be approved by the safety test but not regulated by EMC standard. Obviously, fixtures without light bulbs do not generate any electromagnetic noise. Electromagnetic interference (EMI) comes from the fluorescent light bulb embedded with an electronic ballast which included an inverter with high-frequency switching. A variety of tests demonstrate evidently that a fixture with different self-ballasted compact fluorescent lamps may possibly violate the EMC standard, revealing the absurdity of the regulation. In fact, self-ballasted fluorescent lamps use mostly self-excited electronic ballasts. The operating frequencies for this kind of electronic ballasts can not be precisely controlled due to the influence of many factors. They are not operated at a specified frequency but in a frequency range. This means that the generated EMI spectrum is hardly predicted, especially when a fixture is fitted by light bulbs from several manufacturers. This research inducts the worst cases from numerous measurements on a fixture with 1 piece to 8 pieces of light bulbs, and then attempts to design an EMI filter for all cases. As a result, a lamp fixture with the filter at the line input terminal can suppress the EMI. As long as the consumer buys the lamp fixture which are installed with the EMI filter together with any bulb in use, EMI noise can comply with standard limits.

Une forte progression du transport ferroviaire est observée les dernières années dans plusieurs pays. L’augmentation de la capacité du réseau ferroviaire demande à la fois l’évaluation de l’infrastructure existante selon le trafic attendu, la vitesse des trains, la charge à l’essieu, ainsi que la réduction des interventions de maintenance. Une performance accrue de la voie ferrée par rapport à ces critères nécessite l’amélioration des normes de conception et des outils de prédiction qui puissent prendre en compte toute la durée de vie de la structure. Dans ce contexte, l’objectif de cette thèse est d’apporter un point de vue géotechnique à la modélisation numérique du comportement des voies ferrées sous charge mobile. Un modèle numérique rationnel est développé dans la thèse, composé de trois aspects principaux:(i) comportement dynamique de la voie ferrée, (ii) analyse probabiliste et (iii) comportement non linéaire des géomatériaux. Cette approche permet d’appréhender le comportement mécanique de la voie ferrée à différents instants de son cycle de vie. La première partie de cette thèse est consacrée au développement d’un modèle numérique en dynamique de la voie ferrée, adapté à l’analyse probabiliste et au comportement non linéaire. Une modélisation par Eléments Finis dans le domaine temporel est choisie pour cadre général. Ainsi, un modèle 2D en déformation plane avec épaisseur est proposé dans cette thèse, l’épaisseur hors plan étant calibrée à partir des calculs 3D en statique. Les avantages et inconvénients de cette méthodologie sont discutés selon la représentativité du champ de contraintes dans le plan et du temps de calcul associé, paramètre important pour l’analyse probabiliste. Une méthodologie pour la mise en charge est discutée et implémentée afin de réduire la génération d’ondes parasites. La réponse dynamique de la voie ferrée et l’influence croisée de la vitesse de la charge et de la rigidité de la plateforme sont évaluées sous hypothèse de comportement élastique linéaire. L’influence de la variabilité des propriétés mécaniques de la voie ferrée dans la mesure de la raideur de voie est discutée dans la deuxième partie de cette thèse. Des variations spatiales du module d’Young des couches ferroviaires sont modélisées par des champs aléatoires invariants scalaires. La densité de probabilité de la loi marginale d’ordre 1 associée au champ est obtenue grâce à une analyse statistique des mesures in situ. L’influence croisée du support discret et de la distance de corrélation des champs d’entrée dans les variations de la raideur de voie est mise en évidence à partir de différentes structures de corrélation. Afin de vérifier l’importance de chaque paramètre d’entrée sur les variations de raideur de voie, une analyse de sensibilité globale est effectuée pour différentes configurations de voie. La raideur de voie est principalement affectée par des variations de rigidité de la plateforme et des semelles. L’importance du comportement non linéaire des géomatériaux est soulignée dans la dernière partie de la thèse. Le modèle de comportement élastoplastique développé à l’Ecole Centrale Paris fournit un cadre approprié pour l’étude du comportement des géomatériaux sous chargement cyclique. Cette approche est bien adaptée au comportement des matériaux pendant leur “conditionnement initial”, ou les premiers cycles de charge, quand les tassements permanents sont plus importants et les matériaux cumulent des déformations plastiques élevées. Les paramètres du modèle sont calibrés pour les différents géomatériaux ferroviaires (ballast, couche intermédiaire, sol de la plateforme) à partir d’essais triaxiaux disponibles dans la littérature. Les résultats obtenus illustrent les mécanismes prépondérants dans cette phase : densification et augmentation de la rigidité des différents matériaux par accumulation des déformations plastiques. [...] / An increasing demand for railway transportation is observed in many countries around the world. Achieving higher network capacity requires the evaluation of the existing structure regarding the required traffic, speed and axle load, as well as the reduction of maintenance interventions. A higher track performance in terms of these metrics can be achieved by enhanced design standards and predictive tools accounting for the whole structure’s life span.Within this context, this thesis aims to provide a global framework for combining geotechnical perspective and numerical modeling for the railway infrastructure. A rational approach for railway track modeling is proposed. It is composed by three main aspects: (i) railway track dynamics, (ii) probabilistic analysis and (iii) geomaterials’ non linear behavior. This approach allows assessing the track behavior during different instants of its life span. The first step of this thesis is the development of a dynamic numerical model of the railway track for both probabilistic and non linear analysis. For this purpose, the Finite Element method in time domain is chosen as general modeling framework. A 2D planestrain model with a modified width is used in this thesis, the out-of-plane width being calibrated from 3D static analysis. The advantages and drawbacks of such methodology are discussed in the light of the representativeness of the in-plane stress field and associated computational cost for probabilistic analysis. A loading methodology for reducing spurious wave generation is also discussed and implemented. With the developed model, the track structural response and the crossed influence of speed and subgrade stiffness are first analyzed under linear elasticity hypothesis.The influence of track properties variability in the track stiffness measurement is discussed in the second part of this thesis. Spatial variations are introduced by considering the rigidity of each track layer as an invariant scalar random field. The first-order marginal probability distributions are calibrated from statistical analysis of in situ measurements. By considering different theoretical correlation structures, the crossed influence of the discrete sleeper support and the input correlation length on the track stiffness field is highlighted. In order to verify the importance of each input parameter in the track stiffness’ variability, a global sensitivity analysis is conducted for different track configurations. It is shown that track stiffness variations are primarily caused by variations of subgrade stiffness and possible variations of rail pad stiffness. Furthermore, the importance of geomaterials’ non linear behavior is discussed in the last part of the thesis. A suitable framework for the description of geomaterials’ behavior under cyclic loading, for a large range of stress paths, is provided by a fully elastoplastic multimechanism model. This approach is well adapted for assessing the track behavior during the so-called “conditioning phase”, or the the first cycles when high track settlements are observed and materials cumulate high plastic strains. The model parameters are calibrated from triaxial test results available in the literature for different track materials (ballast, interlayer, subgrade soil). The model is able to capture the main mechanisms acting during the conditioning phase: densification and increase in stiffness of the different materials by accumulation of plastic strains. The load transfer mechanisms and the stress-strain response of the materials are then analyzed. Different stress-strain paths and plastic strains are observed in the ballast layer according to the position of the control point relative to the sleepers. The load speed influence on track permanent settlement and ballast stress-strain response is also studied. Finally, the influence of both interlayer and subgrade behavior on the track response is assessed via a parametric analysis.

Les voies ferrées n'ont cessé d'évoluer depuis l'apparition du premier rail en fer laminé (1789). Les traverses en bois laissent maintenant place à des traverses en béton alors que dans certains pays et dans les tunnels, le ballast est remplacé par des dalles continues en béton armé ou non. Ce type de voie nécessite cependant l'intégration de systèmes supports de rails permettant d'assurer le confort des passagers. Le système "low vibration track" développé par la société Sonneville fait partie de ce type de système. Il se compose d'un blochet en béton reposant sur une semelle micro-cellulaire le tout étant intégré dans un "chausson" en caoutchouc noyé dans la dalle en béton non armé. Ce système équipe aujourd'hui un peu plus de 1100 km de voies de par le monde dont les voies du tunnel sous la Manche, tunnel ayant la section sous-marine la plus longue du monde.Ce travail rentre dans le cadre de la chaire de recherche commune à l'Ecole des Ponts ParisTech et au groupe Eurotunnel et s'intéresse, d'une part à l'étude du comportement dynamique et, d'autre part à la dégradation de la voie ferroviaire du tunnel sous la Manche.La première partie de ce travail présente les modèles existants appliqués aux voies ballastées et propose de développer des modèles pour les voies non ballastées. En utilisant une condition périodique en régime stationnaire, on démontre que la distribution des efforts sur les systèmes supports est gouvernée par une relation caractéristique qui ne dépend pas du comportement des supports. Ce résultat permet de proposer un "système équivalent" de la voie représenté par un ressort caractérisé par sa rigidité et une pré-force qui ne dépendent que des caractéristiques du rail et du chargement. Ce modèle analytique permet alors de calculer rapidement la réponse dynamique de la voie quel que soit le comportement des supports (linéaires ou non linéaires), ce modèle ayant été validé par confrontation avec les résultats des mesures réalisées en tunnel. Ce modèle a enfin été développé pour une voie dont les supports présentent des propriétés mécaniques différentes (supports sains ou endommagés) et a été implanté dans un logiciel (logiciel "EuroAnalyse") qui permet de calculer la réponse dynamique de la voie en intégrant l'évolution des caractéristiques mécaniques de la voie.La deuxième partie est consacrée à l'étude de l'endommagement des blochets en béton du système support. La première phase de cette étude porte sur la compréhension de ces endommagements à l'aide d'une modélisation éléments finis 3D du blochet, la seconde étant consacrée à la mise au point d'une méthode d'identification de ces endommagements en tunnel. Pour ce faire, ce travail met expérimentalement en évidence l'existence d'une variation des deux premières fréquences propres et du coefficient d'amortissement des blochets en fonction de leur état de dégradation. Les mesures sur blochet montrent en effet qu’il existe une corrélation entre l'état de dégradation des blochets et leur signature vibratoire, ces mesures étant en accord avec les calculs éléments finis effectués. La méthode de mesure des fréquences propres basée sur l'analyse modale a été implantée dans un logiciel (logiciel "EuroDetection") qui fera partie intégrante du système automatique de détection des endommagements des blochets en tunnel que la société Eurotunnel utilisera dans le cadre de ses campagnes de maintenance de la voie / Railways have been evolving since the onset of the first rail rolled iron (1789). Wooden sleepers now give way to concrete sleepers while in some countries and commonly in tunnels ballast is replaced by continuous concrete slabs. This type of track, however, requires the integration of rail supports systems to ensure passenger comfort. The "low vibration track" (LVT) system developed by the company Sonneville is such a system. It contains a concrete block based on a micro-cellular pad all integrated in a rubber boot embedded in the slab of unreinforced concrete. This system today covers more than 1100km of railway track around the world, including the Channel tunnel.This work takes place within the framework of joint research between Ecole des Ponts ParisTech and the Eurotunnel group, that investigates the railway dynamic behavior and the support system degradation in the Channel tunnel.The first part of this work presents existing models for ballasted tracks and proposes extended models for non-ballasted tracks. Using a periodic condition of the railway track in steady state it can be shown that the distribution of loads on the support systems is governed by a characteristic relation that does not depend on the support behavior. This result allows us to offer an equivalent system represented by a preloaded spring. This analytical model is then used to calculate the dynamic response of the track by combining with support behavior (linear or nonlinear). This model has been validated by comparison with measurement data in the Channel tunnel. Lastly, a model is presented for a track whose supports have different properties. This model allows rapid calculation of track leveling due to support damage and has been integrated into the new software "EuroAnalyse".The second part of this work deals with the degradation of the LVT concrete block. Firstly, the research focuses on the damage mechanism by calculating the block stress with the help of the finite element method. The result shows that the concrete experiences tension stresses which can lead to damage by fatigue loads. Secondly, a method of identification is developed to monitor the concrete blocks in a tunnel. We show that the eigen frequencies of the block depend on its damage level and this is demonstrated by data and simulations. The method of measuring the eigen frequencies based on modal analysis has been integrated into the software "EuroDetection" which will be part of the track monitoring system for Eurotunnel

No description available.

Environmental awareness has strongly increased these last years, especially in the developed countries where societies have become increasingly preoccupied by the natural resource depletion and environmental degradation. At the same time, the increasing mass transportation demand throughout the European Union requires the development of new infrastructures. Life Cycle Assessment is increasingly used to provide environmental information for decision-makers, when a choice is to be made about the transportation mode to be implemented on a given route. In a life-cycle perspective, not only the environmental pressure of the operation of vehicles but also the burden from the infrastructure, in particular bridges as key links of the road and railway networks, has to be assessed when comparing transportation modes. Based on an extensive literature review, a simplified quantitative LCA is implemented in order to compare the environmental performance of two railway bridge designs. It is meant to be useful at an early stage in the design process, when no detailed information about the bridge is available, and when rough environmental estimations are needed. The Excel based model covers the entire life-cycle of the bridge, from raw material extraction to construction materials recycling and disposal. Various assumptions and omissions are made to narrow the scope of the analysis. For instance, processes that are found insignificant in the literature are omitted, and only a limited set of relevant emissions and impacts to the environment is considered. The model provides fully transparent results at the inventory and impact assessment level. The streamlined approach is tested by comparing the environmental burden throughout the life-cycle of a steel-concrete composite railway bridge on a single span, equipped with either a ballasted or a fixed concrete single track. The results show that the environmental impacts of the fixed track alternative are lower than that of the ballasted track alternative, for every impact categories. In a sustainable development perspective, it would thus have been preferable to install a fixed track over the bridge to reduce its overall impact on the environment by about 77%. The raw material phase is found decisive in the life-cycle of both alternatives. The frequency of the replacement of the track is identified as a key environmental parameter, since the road traffic emissions during bridge closure nearly overwhelmed the other life-cycle stages.

With the introduction of high-speed trains over the past few years, as well as withan increase in traffic loads and volumes, the attractiveness and competitiveness ofthe railway transport system have increased. This leads to the need to develop newcost-effective ways in railway construction, providing an increase in the life cycle ofthe railway track and reducing the need for maintenance.In several European countries, Japan, and the USA, asphalt mixtures, also called“bituminous mixtures”, are used as such an alternative solution in both traditionalballasted and ballastless railway designs. International experience has shown thatbituminous mixtures have good technical characteristics and can meet many of therequirements of modern railway construction, especially when it comes to the useof asphalt materials in ballasted tracks. However, there is a lack of knowledge andexperience in using bituminous mixtures in railway construction in the northernregions and harsh winter conditions.In this thesis, the development of two- and three-dimensional models of ballastedrailway track with an asphalt sub-ballast layer is presented. The analysis of thestatic and dynamic characteristics of the whole model and the asphalt layer exposedto different temperatures is carried out using the finite element method (FEM) inthe COMSOL software. Simulation for all conditions is carried out with a trainspeed of 200 km/h. The relationship between various temperatures from 􀀀30 °Cto +30 °C and the stress-strain behaviour of the asphalt sub-ballast layer is alsodetermined. The results of the study show that both under the influence of lowand high temperatures, the stresses and strains in the asphalt layer are quite small.Besides, the use of asphalt sub-ballast in the track structure reduces deformationsand stresses acting on the subgrade surface.

A via permanente representa um elemento imprescindível na composição do transporte ferroviário e seu desempenho deve ser adequado, de forma a garantir tanto segurança quanto conforto. Assim, diversos aspectos devem ser analisados ainda na fase de projeto, através de dimensionamentos que confrontem diferentes parâmetros da resposta da via e os limites estabelecidos. Dessa forma, o conhecimento do comportamento mecânico da via, devido aos esforços impostos pela passagem do material rodante, passa a ser essencial no projeto de uma estrutura que garanta os requisitos necessários, sem ser inviável economicamente. Visto que esse comportamento mecânico é muito sensível à rigidez vertical da estrutura, o presente trabalho apresenta análises da influência desse parâmetro na resposta da via e, consequentemente, no seu dimensionamento. Nesse contexto, o trabalho abrange tanto o caso de vias em lastro solicitadas por trens de carga, quanto o caso de vias em laje solicitadas por trens de passageiros em meios urbanos. No primeiro caso são realizados estudos paramétricos, por meio de modelos clássicos e um modelo mecanicista, para a análise de momentos fletores e deflexões nos trilhos, bem como tensões verticais nas camadas de lastro, sub-lastro e subleito. Já no segundo caso, são realizados estudos paramétricos relativos à transmissibilidade e à atenuação de vibrações causadoras de ruído secundário. Também é feita uma análise da influência da rigidez vertical na amplificação dinâmica das cargas estáticas, que pode ser aplicada a ambos os casos citados e até extrapolada para casos de vias de alta velocidade. Os resultados mostraram que aumentos de rigidez vertical resultam em ganhos do ponto de vista de momentos fletores e deflexões nos trilhos, além de maior resistência e capacidade de dissipação de tensões verticais nas camadas de lastro, sub-lastro e subleito. Por outro lado, esses aumentos também levaram a maiores tensões nas camadas subjacentes à grade citadas, além de atenuações de vibrações em menores intervalos de frequência e maiores amplificações dinâmicas das cargas estáticas em vias de alta velocidade. Assim, é mostrado que a influência da rigidez vertical, tanto da via como um todo quanto de alguns elementos específicos, não deve ser analisada de forma genérica, pois, dependendo do parâmetro da resposta da via considerado no dimensionamento, seu aumento pode representar uma influência positiva ou negativa. / The railway track is an essential element in the composition of rail transport and its performance should be adequate to ensure both safety and comfort. Therefore, several aspects should be analyzed still in the project phase, through project designs that compare the response of the track from the point of view of different parameters and some stablished limits. Thus, the knowledge of the mechanical behavior of the track due to stresses imposed by the passage of rolling stock becomes essential to the design of a structure which ensures the necessary requirements, without being uneconomic. Since this mechanical behavior is very sensitive to vertical stiffness of the structure, this thesis presents some analyses about the influence of this parameter in the track response and, consequently, in its design. In this context, the research covers both the case of ballasted tracks with freight trains and the case of slab tracks with passenger trains in urban areas. In the first case, parametric studies are performed by means of classical and mechanistic models with analyses of bending moments and deflections in rails, as well as the vertical stresses on ballast, sub-ballast and subgrade layers. In the second, it was analyzed in the parametric studies the transmission and attenuation of vibrations causing ground borne vibration. In addition to these, there is also an analysis of the influence of the vertical stiffness in dynamic amplification of the static loads, which can be applied to both the mentioned cases and even extrapolated to cases of high-speed railways. The results showed that stiffness increases result in positive gains from the point of view of bending moments and deflections in rails, as well as higher strength and dissipation of vertical stresses on ballast, sub-ballast and subgrade layers. On the other hand, this increase of stiffness also lead to higher pressures on the layers cited, in addition to attenuation of vibrations in lower frequency ranges and higher dynamic amplifications of static loads in the case of high-speed operation. Thus, the influence of vertical stiffness should not be analyzed in a generic way, because, depending on the parameter of the track response considered, it can represent a positive or negative influence.

A via permanente representa um elemento imprescindível na composição do transporte ferroviário e seu desempenho deve ser adequado, de forma a garantir tanto segurança quanto conforto. Assim, diversos aspectos devem ser analisados ainda na fase de projeto, através de dimensionamentos que confrontem diferentes parâmetros da resposta da via e os limites estabelecidos. Dessa forma, o conhecimento do comportamento mecânico da via, devido aos esforços impostos pela passagem do material rodante, passa a ser essencial no projeto de uma estrutura que garanta os requisitos necessários, sem ser inviável economicamente. Visto que esse comportamento mecânico é muito sensível à rigidez vertical da estrutura, o presente trabalho apresenta análises da influência desse parâmetro na resposta da via e, consequentemente, no seu dimensionamento. Nesse contexto, o trabalho abrange tanto o caso de vias em lastro solicitadas por trens de carga, quanto o caso de vias em laje solicitadas por trens de passageiros em meios urbanos. No primeiro caso são realizados estudos paramétricos, por meio de modelos clássicos e um modelo mecanicista, para a análise de momentos fletores e deflexões nos trilhos, bem como tensões verticais nas camadas de lastro, sub-lastro e subleito. Já no segundo caso, são realizados estudos paramétricos relativos à transmissibilidade e à atenuação de vibrações causadoras de ruído secundário. Também é feita uma análise da influência da rigidez vertical na amplificação dinâmica das cargas estáticas, que pode ser aplicada a ambos os casos citados e até extrapolada para casos de vias de alta velocidade. Os resultados mostraram que aumentos de rigidez vertical resultam em ganhos do ponto de vista de momentos fletores e deflexões nos trilhos, além de maior resistência e capacidade de dissipação de tensões verticais nas camadas de lastro, sub-lastro e subleito. Por outro lado, esses aumentos também levaram a maiores tensões nas camadas subjacentes à grade citadas, além de atenuações de vibrações em menores intervalos de frequência e maiores amplificações dinâmicas das cargas estáticas em vias de alta velocidade. Assim, é mostrado que a influência da rigidez vertical, tanto da via como um todo quanto de alguns elementos específicos, não deve ser analisada de forma genérica, pois, dependendo do parâmetro da resposta da via considerado no dimensionamento, seu aumento pode representar uma influência positiva ou negativa. / The railway track is an essential element in the composition of rail transport and its performance should be adequate to ensure both safety and comfort. Therefore, several aspects should be analyzed still in the project phase, through project designs that compare the response of the track from the point of view of different parameters and some stablished limits. Thus, the knowledge of the mechanical behavior of the track due to stresses imposed by the passage of rolling stock becomes essential to the design of a structure which ensures the necessary requirements, without being uneconomic. Since this mechanical behavior is very sensitive to vertical stiffness of the structure, this thesis presents some analyses about the influence of this parameter in the track response and, consequently, in its design. In this context, the research covers both the case of ballasted tracks with freight trains and the case of slab tracks with passenger trains in urban areas. In the first case, parametric studies are performed by means of classical and mechanistic models with analyses of bending moments and deflections in rails, as well as the vertical stresses on ballast, sub-ballast and subgrade layers. In the second, it was analyzed in the parametric studies the transmission and attenuation of vibrations causing ground borne vibration. In addition to these, there is also an analysis of the influence of the vertical stiffness in dynamic amplification of the static loads, which can be applied to both the mentioned cases and even extrapolated to cases of high-speed railways. The results showed that stiffness increases result in positive gains from the point of view of bending moments and deflections in rails, as well as higher strength and dissipation of vertical stresses on ballast, sub-ballast and subgrade layers. On the other hand, this increase of stiffness also lead to higher pressures on the layers cited, in addition to attenuation of vibrations in lower frequency ranges and higher dynamic amplifications of static loads in the case of high-speed operation. Thus, the influence of vertical stiffness should not be analyzed in a generic way, because, depending on the parameter of the track response considered, it can represent a positive or negative influence.

Pour améliorer la durabilité des voies ferrées, il a été introduit sur la LGV BPL (Bretagne- Pays de Loire), une couche bitumineuse sous ballast, destinée entre autre à diminuer l’amplitude des accélérations produites par le passage des TGV, ce qui est une cause majeure du tassement de ballast. Afin d'étudier les réponses dynamiques des différentes structures, trois sections avec sous couche bitumineuse, et une section avec sous couche granulaire de la voie BPL ont fait l’objet d’une instrumentation, avec suivi à distance des mesures. L’objectif de ce travail de thèse est d’analyser et de modéliser la réponse des sections instrumentées, et de comparer leur comportement durant la phase de montée en vitesse puis durant la phase d’exploitation commerciale. Les résultats ont indiqué un bon comportement des sections bitumineuses. La comparaison des accélérations obtenues sur les différentes sections a confirmé l’effet positif de la couche bitumineuse sur la stabilité de la voie. Sur la base des mesures de la phase de montée en vitesse, des modélisations de la réponse mécanique des sections de la LGV BPL à l’aide du logiciel ViscoRail ont confirmé la fiabilité du logiciel à reproduire le comportement mécanique des structures. Les mesures de température, de teneur en eau, et de tassements verticaux ont également confirmé le bon comportement saisonnier des structures bitumineuses. Enfin, une extension de ViscoRail est mise en place, pour comparer l’effet de différents types de semelles sous rail, linéaire et non linéaire, sur la réponse de la voie. / To improve railways durability, it has been introduced on the HSL BPL (Bretagne-Pays de Loire), a bituminous layer under ballast, intended among other things to reduce the amplitude of the accelerations produced by highspeed trains circulations, which is a major cause of ballast settlement. In order to study the dynamic responses of the different structures, three sections with bituminous underlayment, and a section with a granular sub-layer of BPL track were instrumented. Measurements are monitored remotely. The objective of this thesis is to analyze and model the response of the instrumented sections, and to compare their behavior during the speed up phase and during the commercial service phase. The results indicated good behavior of bituminous sections. The comparison of the accelerations obtained on different sections confirmed the positive effect of the asphalt concrete layer on the stability of the track. Based on the speed up phase measurements, modeling the mechanical response of the various HSL sections using ViscoRail has confirmed the software’s reliability to reproduce the mechanical behavior of railway structures. Temperature measurements, water content and vertical settlements have also confirmed the good seasonal behavior of bituminous structures. Finally, an extension of ViscoRail is put in place, to compare the influence of different types of rail pads, linear and nonlinear, on the response of the track.