L’étude effectuée au cours de cette thèse a permis la caractérisation expérimentale des instabilités aérodynamiques se développant dans un compresseur centrifuge et une première évaluation de l’efficacité d’une stratégie de contrôle par aspiration de couche limite. Le compresseur, développé par Safran Helicopter Engines et dénommé Turbocel, est composé d’une roue directrice d’entrée, d’un rouet centrifuge splitté, d’un diffuseur radial aubé et splitté et d’un redresseur axial. Des travaux numériques antérieurs réalisés au Laboratoire de Mécanique des Fluides et d’Acoustique ont montré, aux bas régimes de rotation, un comportement singulier caractérisé par une structure d’écoulement dite « alternée » impliquant deux canaux adjacents du diffuseur radial. L’étude stationnaire réalisée sur l’ensemble des régimes de rotation du compresseur a conduit à une ségrégation des vitesses de rotation suite à l’établissement d’une variable– le taux d’asymétrie - caractérisant l’asymétrie de l’aérodynamique du diffuseur. Ce taux, quasi nul à très basse vitesse de rotation, croît jusqu’à atteindre un maximum à vitesse de rotation intermédiaire, puis s’effondre pour ré-augmenter légèrement. Une analyse fine des données instationnaires acquises à bas régimes a permis la description de deux modes de fonctionnement du compresseur associés à des structures de décollements différentes dans le diffuseur. Le premier mode est caractérisé par l’oscillation à une fréquence de l’ordre de 42 Hz d’un décollement localisé sur la face en dépression des aubes principales du diffuseur. Le second mode, à 12Hz, associé au pompage modéré du compresseur, correspond à la mise en place d’un schéma alterné et à son oscillation sur deux canaux adjacents du diffuseur.Les origines probables de ces différents modes de fonctionnement sont discutées à partir de considérations • aérodynamiques -- la mise en place d’une recirculation en tête de rouet est suspectée d’influer sur le taux d’asymétrie en modifiant l’incidence en entrée de diffuseur,• géométriques -- le nombre et le calage des aubes du diffuseur radial ainsi que la distance inter-roue indiquent une prédisposition du diffuseur à fonctionner en régime alterné sous certaines conditions d’incidence,• aéro-acoustiques -- un accrochage des fréquences aérodynamiques avec les fréquences des ondes acoustiques du banc d’essai semble se produire. Enfin, les résultats sur le contrôle d’écoulement par aspiration de couche limite à régime partiel sont présentés. Une amélioration du rendement est observée à certains points de fonctionnement, mais aucune extension de la plage de fonctionnement du compresseur n’est mesurée. Sans l’atténuer, l’aspiration permet de contrôler sur quels canaux s’établit le régime alterné. / This thesis presents an experimental characterization of the evolution of aerodynamic instabilities in a centrifugal compressor, and a first evaluation of the effectiveness of boundary layer suction as a control strategy. The compressor used in this study is Turbocel, a centrifugal compressor developed by Safran Helicopter Engines, featuring inlet guide vanes, a backswept splittered unshrouded impeller, a splittered vaned radial diffuser and axial outlet guide vanes.Previous numerical work, conducted at the Laboratoire de Mécanique des Fluides et d’Acoustique de Lyon, revealed a unusual behaviour of the compressor at low rotational speeds characterized by a distinctive alternate flow structure in the radial diffuser that develops across two adjacent blade channels. The steady analysis, which was conducted over the full operating range of rotational speeds, led to the distinction of different operating zones, following the establishment of a new indicator variable - the asymmetry rate - characterizing the asymmetry of the diffuser aerodynamics. This rate, which is close to zero at very low rotation speed, increases until it reaches a maximum value at intermediate rotational speed, before collapsing and slightly increasing again near the nominal rotational speed.Analysis of the unsteady data acquired at low speeds allowed for the characterization of two compressor operating modes, associated with different flow phenomena in the stalled diffuser. The first mode is characterized by the oscillation of a separation at 42 Hz, on the suction side of the main blades in the diffuser. The second mode, at 12Hz, associated with mild surge of the compressor, corresponds to the emergence of an alternate pattern of unsteady flow separation that occurs across two adjacent channels of the diffuser.The probable causes for these different operating modes are discussed in the context of different considerations:• aerodynamic -- the formation of a recirculation near the tip of the impeller is suspected to affect the asymmetry rate by changing the incidence angle at the diffuser inlet.• geometric -- the number and the stagger angle of the radial diffuser blades as well as the distance between the impeller and the diffuser may result in a predisposition of the diffuser to operate in an alternating mode, under certain conditions of incidence.• aero-acoustic -- as there is evidence of a lock-in of the aerodynamic frequencies with the acoustic modes of the test rig.Finally, boundary layer suction is explored as a means of flow control at partial rotational speed. Improvements in performance were observed for some operating points, however no extension of the compressor operating range was measured. Although boundary layer suction did not allow for the intensity of the oscillating separation pattern in the diffuser to be reduced, it was found to be an effective means of controlling the location of the alternate flow structure in the diffuser.

Earthquakes have been historically perceived as one of the most damaging natural hazards. Seismic soil liquefaction is often one of the major sources of damage and disruptions, and has been observed to severely affect key lifelines. Settlement and tilting of shallow foundations resting on saturated sandy/silty soils has been repeatedly observed throughout the world as a consequence of liquefaction or softening of the foundation soil. Such settlements and tilts can render structures unusable, and homes uninhabitable, causing significant economic losses. Despite the undoubted relevance of this phenomenon, field data on the liquefaction induced settlement of shallow foundations are scarce. New data from 24 buildings that suffered settlement and tilting as a consequences of soil liquefaction during the February 27th 2010 Maule earthquake in Chile, are presented in this work to supplement the existing field cases database. Due to the complexity of this phenomenon, field data are not suffcient to fully understand the mechanisms controlling the settlement of structures resting on liquefied or softened ground.In this framework, centrifuge modelling provides a valuable tool for research by reproducing field conditions in a controlled environment. A series of 10 dynamic centrifuge tests were performed as part of this work. Thanks to the University of Dundee newly installed centrifuge-mounted servohydraulic earthquake simulator, scaled version of field earthquake motions were reproduced in the models tested, enhancing the reliability of experimental results. Particular attention was given to the effect of key parameters on the observed foundation settlement. These parameters are the bearing pressure of the foundation, the thickness of the liquefied soil layer and the soil's relative density. The effect of the soil layering pattern was also investigated, with particular attention to the effect of a low permeability soil crust overlying the liquefied soil. Results suggest that the excess pore pressure generation in the foundation soil is significantly influenced by the stress distribution due to the presence of the foundation itself. In particular, lower excess pore pressure where measured in soil subjected to high static shear stresses (i.e. below the edge of a footing). The soil stratification pattern, and the relative thicknesses of the liquefied and un-liquefied portions of the soil profile, were also found to play a crucial role in determining the seismic demand at foundation level and the type of failure mechanism leading to foundation settlement. Observed differences between centrifuge (i.e. field) and element testing soil response are also discussed. Experimental results are compared to field observations, with the aim of improving the current understanding of the behaviour of structures built on shallow foundations in the eventuality of seismic induced liquefaction of their foundation soil.

Shallow embankment slopes are commonly used to support elements of transport infrastructure in seismic regions. In this thesis, the seismic performance of such slopes in non-liquefiable granular soils has been investigated and an extensive programme of centrifuge testing was conducted to quantify the improvements to seismic slope performance which can be achieved by installing a row of discretely spaced vertical precast concrete piles. This study focussed on permanent movement and dynamic response at different positions within the slope, especially at the crest, which would form key inputs into the aseismic design of supported infrastructure. In contrast to previous studies, the evolution of this behaviour under multiple sequential strong ground motions is studied through dynamic centrifuge modelling, analytical (sliding-block) and numerical (Finite Element) models. This thesis makes three major contributions. Firstly, an improved sliding-block (‘Newmark’) approach is developed for estimating permanent deformations of unreinforced slopes during preliminary design phases, in which the formulation of the yield acceleration is fully strain-dependent, incorporating the effects of both material hardening/softening and geometric hardening (re-grading). This is supported by the development of numerical (Finite Element) models which can additionally predict the settlement profile at the crest of the slope and also the dynamic ground motions at this point, for detailed seismic design were also developed. It is shown that these new models considerably outperform existing state-of-the art models which do not incorporate the geometric changes for the case of an earthquake on a virgin slope. It is further shown that only the improved models can correctly capture the behaviour under further earthquakes (e.g. strong aftershocks) and therefore can be used to determine the whole-life performance of a slope under a suite of representative ground motions that the slope may see during its design life, and allow improved estimates of the seismic performance of slopes beyond their design life. The finite element models can accurately replicate the settlement profile at the crest (important for highway or rail infrastructure) and quantify the dynamic motions which would be input to supported structures, though these were generally over-predicted. Secondly, the principles of physical modelling have been used to produce realistically damageable model piles using a new model reinforced concrete (both a designed section specifically detailed to carry the bending moments induced by the slipping soil mass and a nominally reinforced section with low moment capacity). This was used to investigate how piles can stabilise slopes under earthquake events and how the permanent deformation and the dynamic response of stabilised slope are strongly influenced by the pile spacing (S/B) especially at the minimum pile spacing (i.e. S/B=3.5). This is consistent with previous suggestions made for the optimal S/B ratio for encouraging soil arching between piles at maximum spacing both under monotonic conditions, and for numerical investigations of the seismic problem. These were supported by further centrifuge tests on conventional ‘elastic’ piles which were instrumented to measure seismic soil-pile interaction. The importance of reinforcement detailing was also highlighted, with the nominally reinforced section yielding early in the earthquake; the damaged piles subsequently only offer a small (though measureable) reduction in seismic slope performance compared to the unreinforced case. It was demonstrated that both permanent deformations at the slope crest (e.g. settlement) and dynamic ground motions at the crest can be significantly reduced as pile spacing reduced. Finally, a coupled P-y and elastic continuum approach for modelling soil-pile interaction has been used to develop a Newmark procedure applicable for pile-reinforced slopes. It was observed that the single pile resistance is mobilising at beginning of the earthquake’s time and it is strongly influenced by pile stiffness properties, pile spacing and the depth of the slip surface. It was observed also that the depth of the slip surface and pile spacing (S/B) play an important role in the determination of the permanent deformation of the slope. The results show great agreement to centrifuge test data in term of the permanent deformation (settlement at the crest of the slope) with slight differences between the measured (centrifuge) and calculated (this procedure) maximum bending moments.

[Truncated abstract] Seismic tomography has been extensively used in geophysics for different purposes such as geological mapping and prospecting for oil and gas. In geophysics, ultrasound or electromagnetic waves are normally used to provide the tomographic information. In the geotechnical area, seismic tomography is emerging as a promising technique that can be used to determine the spatial variability of shear wave velocities and hence the small strain stiffness of geomaterials. Although some studies have been undertaken to incorporate seismic measurement into centrifuge modelling, there has been to date no attempt to build a complete seismic tomography facility with high resolution for use in a geotechnical centrifuge. Such a powerful facility can help in better understanding of soil behaviour by providing a complete picture of the spatial variation of the soil property of concern. The main aim of this study was to develop a high-resolution seismic tomography (ST) system for the beam centrifuge at the University of Western Australia (UWA) by which the shear wave velocity and hence maximum shear modulus could be determined anywhere in the centrifuge model. ... This limitation was the requirement to use an a priori model. The exact solutions in the different examples presented in this chapter were known, and they were used as a priori models into the inversion process. However, in practice the exact solution is unknown, and the aim of the tomographic inversion is to obtain a solution that best describes the measured data. Carrying out inversion without using an a priori model can yield an output model that hints at the nature of the model. This output can then be used as the starting point in an iterative process, in which the output from one step is used as an a priori model for reinverting the original data in a subsequent step. In this case, this process slightly improved the output tomogram and decreased the value of root mean squares of travel time residuals (Rrms). An alternative inversion strategy was proposed based on the results obtained in this study. It involves using a searching algorithm. A searching process can be carried out based on the output from the first iteration (without using an a priori model). The search can involve varying the parameters that describe buried anomalies, such as the size of the anomaly, the velocity value in the anomaly, and the location of the anomaly. The aim is to search for the combination of anomaly parameters that minimises the resulting error parameters (mainly Rrmx, but also the average error and the standard deviation of the error). For more subtle cases, such as the velocity model under a footing, where inversion without using an a priori model did not recover the input model, a searching algorithm involving applying perturbations to the exact Boussinesq model can be performed. Not only can the searching procedure involve adding perturbation to the velocity values in the Boussinesq model, but it can also add perturbation to the shape of the velocity distribution below the footing. The searching process can continue until a model that fits the data with a minimum error is found, i.e., minimising Rrms.

Dans la première partie de cette thèse, nous étudions un modèle de marche aléatoire centrifuge. Nous démontrons une loi du logarithme itéré pour sa norme, et nous obtenons la loi asymptotique des fluctuations de sa direction. Nous donnons ensuite un encadrement du taux de décroissance exponentielle de la probabilité qu'elle se trouve à l'instant n dans un compact fixé en montrant que la probabilité qu'une marche aléatoire centrée classique retourne dans un compact à l'instant n sans quitter un cône ne décroît pas à vitesse exponentielle. Dans la seconde partie, nous étudions le mouvement brownien de dimension quelconque, conditionné à rester dans un cône de révolution pendant une unité de temps, et nous en déduisons un principe d'invariance pour une marche aléatoire conditionnée à rester dans un cône.

abstract: A series of experiments were conducted to support validation of a numerical model for the performance of geomembrane liners subject to waste settlement and seismic loading. These experiments included large scale centrifuge model testing of a geomembrane-lined landfill, small scale laboratory testing to get the relevant properties of the materials used in the large scale centrifuge model, and tensile tests on seamed geomembrane coupons. The landfill model in the large scale centrifuge test was built with a cemented sand base, a thin film NafionTM geomembrane liner, and a mixture of sand and peat for model waste. The centrifuge model was spun up to 60 g, allowed to settle, and then subjected to seismic loading at three different peak ground accelerations (PGA). Strain on the liner and settlement of the waste during model spin-up and subsequent seismic loading and accelerations throughout the model due to seismic loading were acquired from sensors within the model. Laboratory testing conducted to evaluate the properties of the materials used in the model included triaxial compression tests on the cemented sand base, wide-width tensile testing of the thin film geomembrane, interface shear testing between the thin film geomembrane and the waste material, and one dimensional compression and cyclic direct simple shear testing of the sand-peat mixture used to simulate the waste. The tensile tests on seamed high-density polyethylene (HDPE) coupons were conducted to evaluate strain concentration associated with seams oriented perpendicular to an applied tensile load. Digital image correlation (DIC) was employed to evaluate the strain field, and hence seam strain concentrations, in these tensile tests. One-dimensional compression tests were also conducted on composite sand and HDPE samples to evaluate the compressive modulus of HDPE. The large scale centrifuge model and small scale laboratory tests provide the necessary data for numerical model validation. The tensile tests on seamed HDPE specimens show that maximum tensile strain due to strain concentrations at a seam is greater than previously suggested, a finding with profound implications for landfill liner design and construction quality control/quality assurance (QC/QA) practices. The results of the one-dimensional compression tests on composite sand-HDPE specimens were inconclusive. / Dissertation/Thesis / Doctoral Dissertation Civil and Environmental Engineering 2016

L'étude effectuée au cours de cette thèse s'inscrit dans le contexte du contrôle des écoulements en turbomachines. Suite à l'intérêt avéré de l'aspiration de la couche limite en compresseurs axiaux pour en améliorer les performances, elle vise à évaluer si cette méthodologie de contrôle est pertinente en géométrie radiale. Elle s'appuie pour cela sur un compresseur centrifuge transsonique Turbomeca. Le fonctionnement du compresseur est d'abord analysé sur la base de simulations numériques stationnaires et instationnaires chorochroniques, réalisées grâce au code de calcul elsA développé par l'Onera et en utilisant le modèle de turbulence k-l de Smith. La validité des résultats numériques est vérifiée par comparaison avec les données expérimentales disponibles : performances mono-dimensionnelles et signaux de pressions instationnaires. Ces données sont issues à la fois de la caractérisation expérimentale du compresseur effectuée par Turbomeca et de campagnes de mesures menées a l'ISAE, ou le compresseur est monté sur un banc de recherche dédié à l'étude des interactions entre le rouet et le diffuseur radial. L'analyse détaillée des résultats numériques révèle le développement d'un décollement de coin dans le diffuseur, entre la paroi moyeu et la face en dépression des aubages, lorsque le point de fonctionnement se rapproche du pompage. L'étude topologique du spectre du frottement pariétal permet d'en préciser le lieu et la structure, et de localiser les points singuliers ainsi que la ligne de séparation principale. En stationnaire, la croissance du décollement aboutit au décrochage du diffuseur, et compromet les performances de l'étage de compression et finalement la stabilité numérique. L'exploitation des résultats instationnaires permet de préciser le comportement temporel de l'écoulement : l'étendue du décollement fluctue sous l'effet du défilement des ondes de pression générées par l'interaction rouet-diffuseur, mais les trajectoires des particules fluides impliquées dans le décollement sont en accord avec les lignes de courant du champ de l'écoulement moyenné temporellement. Le décollement instationnaire est fixe, et sa topologie correspond à celle prévue par les simulations stationnaires. En particulier, le lieu du décollement est prévu de façon similaire par les modèles stationnaires et instationnaires. Une stratégie d'aspiration est ensuite développée sur la base de ces observations : une fente de prélèvement est positionnée au voisinage du col de séparation principal, dont le lieu correspond à celui du maximum du gradient de pression adverse stationnaire. Cette stratégie est implémentée au sein des modèles numériques stationnaires et instationnaires. En stationnaire, un prélèvement de 1 % de la valeur totale du débit traversant le rouet permet un contrôle total du décollement dans le diffuseur, et conduit à une augmentation significative de la plage de stabilité numérique. La réduction du débit de prélèvement à 0,3 % du débit total permet un contrôle partiel du décollement, et conduit également à une augmentation de la plage de stabilité numérique. En instationnaire, le décollement de coin initial est contrôlé. Mais les résultats mettent en évidence le rôle majeur joue par le défilement des ondes de pression le long des aubages du diffuseur. Celles-ci se renforcent au passage du col du diffuseur, et engendrent l'existence d'un important gradient de pression adverse instantané. Ce maximum du gradient de pression instantané conduit au développement d'un nouveau décollement de coin, en aval de la fente de prélèvement, et la plage de stabilité des calculs instationnaires n'est pas augmentée. Ces résultats mettent en évidence la possibilité d'agir sur les décollements se produisant dans les diffuseurs radiaux à l'aide de la technique d'aspiration de la couche limite. […] / The study presented in this manuscript takes place in the context of flow control within tur - bomachinery. Following the demonstration of the effectiveness of the boundary layer suction technique in axial compressors for improving their performance, the present work aims at determining if this same technique may be relevant in radial compressors. Boundary layer suction is then applied on a centrifugal compressor stage, designed and built by Turbomeca, Safran group. The working of the compressor is first analyzed thanks to steady-state and unsteady numerical simulations, performed using the elsA solver developed by Onera, the French Aerospace Laboratory. The turbulence is modeled with the two equations k-l model of Smith. The validity of the numerical results is ensured by comparison with available experimental measurements results: one-dimensional performance coefficients and time-dependent pressure signals. This data were obtained by both Turbomeca, during the characterization of the compressor, and the fluid mechanics laboratory of ISAE, Université de Toulouse, were the compressor is mounted in an experimental test rig dedicated for studying the impeller-diffuser interactions. The detailed analysis of the numerical results reveals the growth of a corner separation within the diffuser between the hub endwall and the vane suction side when the operating point moves toward surge. The precise location and the internal structure of that boundary layer separation are then elucidated thanks to a topological study, which allows to identify the singular points and the separating lines of the skin-friction pattern. In steady-state numerical simulations, the development of that corner separation leads to the stall of the diffuser, which compromises the compressor stage performance and finally the stability of the numerical model. Unsteady numerical simulations results allow to specify the temporal behavior of the corner separation: the extent of the separated zone is modulated by the scrolling of pressure waves created by the impeller-diffuser interaction, but the trajectories of fluid particles within the separation match with the streamlines of the time-averaged flow field. The unsteady separation is then fixed, and its topology is in agreement with the prediction of the steady-state numerical simulations. In particular, the location of the separation is predicted similarly by the steady-state and the unsteady numerical models. Afterward, a control strategy using suction technique is developed thanks to the previous conclusions. The suction slot is set in the neighborhood of the main saddle of the separation, which corresponds to the location of the maximum of the time-averaged adverse pressure gradient. This strategy is implemented within both the steady-state and unsteady numerical models. The steady-state numerical model predicts the complete control of the separated zone with a removal of 1% of the total massflow through the compressor. This leads to a significant increase of the numerical stable range. With a removal of 0.3% of the compressor total massflow, the numerical model predicts only a partial control of the corner separation, but it also leads to a significant increase of the numerical stable range. With the unsteady numerical model, the initial hub corner separation is again controlled thanks to aspiration. But the results also highlight the major role played by the scrolling of pressure waves. They reinforce when crossing the diffuser throat, and generates a strong instantaneous adverse pressure gradient. This maximum provokes a new boundary layer separation, further downstream of the suction slot. The stable range of the unsteady numerical model is not increased. These results put into evidence the possibility to act on boundary layer separation that occur in radial diffusers thanks to the boundary layer suction technique. [...]

L’augmentation de l’efficacité des turbines fonctionnant sur un cycle de Brayton récupéré passe par l’augmentation de la température des gaz d’admission. Cependant, les pales métalliques doivent être massivement refroidies afin de supporter des températures dépassant les 1600 K. Une nouvelle configuration utilisant des pales en céramiques supportées par une jante de composite et un moyeu flexible supporte des températures de plus de 1600 K sans refroidir les pales. La configuration maintient les pales en compression durant la rotation de la turbine évitant le bris de la céramique lorsque soumise à la tension. La jante de composite ne doit pas dépasser 600 K, nécessitant un système de refroidissement. L’utilisation d’un refroidissement par film sur la paroi de bout de pales permet de réduire de 30 % le flux thermique traversant la paroi. Dans une configuration classique d’une turbine, les pales ne sont pas supportées par leur bout, faisant en sorte qu’aucune recherche ne porte sur le refroidissement par film de la paroi de bout de pales. Les multiples recherches faites sur le refroidissement par film des différentes composantes d’une turbine montrent que la rotation de la turbine et la courbure des surfaces à protéger influence l’efficacité du refroidissement par film. Ce projet de maitrise vise à expliquer comment le champ centrifuge créé par la rotation de la turbine et la courbure de la paroi de bout de pales affectent le refroidissement par film. Une analyse de stabilité de gaz stratifiés, confinés entre deux parois et ayant une injection de gaz à l’une des parois explique théoriquement comment la vitesse des gaz, le champ centrifuge et la vitesse d’injection influencent la stabilité des gaz stratifiés. Le champ centrifuge augmente la plage de longueur d’onde de perturbations que peut tolérer le système sans être déstabilisé, alors qu’une différence de vitesse entre les gaz déstabilise le système. La vitesse d’injection doit être réduite au minimum puisqu’elle déstabilise le système. Les résultats de cette analyse ont été vérifiés expérimentalement dans un canal courbé. Les tests montrent que l’injection normale à l’écoulement ne permet pas de stratifier deux gaz, alors qu’une injection parallèle entraine une stratification qui prend plus de 37 𝑚𝑚 à se mélanger, soit plus du double de la longueur du rotor de la turbine étudiée dans ce projet. Cependant, en raison du couplage entre la vitesse des gaz et le champ centrifuge qu’impose un canal courbé, il est impossible d’observer l’effet du champ centrifuge et de la différence de vitesse entre les gaz indépendamment, indiquant la nécessité de concevoir un banc d’essai rotatif pour de futurs tests. Une analyse thermique du refroidissement par film de la paroi de bout de pales relie le flux thermique traversant la paroi au débit massique de gaz de refroidissement et à la température de la paroi. Cette analyse, fait le lien entre le refroidissement par film et le refroidissement par ailettes séparant la jante de composite de la paroi de bout de pales.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第23476号 / 工博第4888号 / 新制||工||1764(附属図書館) / 京都大学大学院工学研究科社会基盤工学専攻 / (主査)教授 渦岡 良介, 教授 三村 衛, 教授 肥後 陽介 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

A vast majority of coal and mineral cleaning and upgrading processes involve the addition of water. The water allows the movement of particles throughout the processing plant and the upgrading of the material. When the process is complete the finished product must be dewatered. This is due to storage concerns, in which the water takes up a majority of the space, and high transportation costs, in which no compensation is obtained from the buyer for the shipment of the liquid. Dewatering is accomplished by many devices, with the two most common pieces of equipment being the screen-bowl centrifuge and disk filter. This thesis tests and compares the effect of reagents on dewatering using the screen-bowl centrifuge and disk filter. Coal was obtained from the Upper Banner, Pittsburgh No. 8, Taggart, and Dorchester seams, crushed and ground to the desired size, and run through the dewatering circuits. The results showed that the moisture content of the product can be greatly reduced in the disk filter while being only slightly reduced in the screen-bowl centrifuge. It was also shown that the recovery can be slightly increased in the screen-bowl centrifuge. Overall, with the addition of reagents, the disk filter outperformed the centrifuge in both recovery and moisture content. A model was also developed for the screen-bowl centrifuge. The results from the screen-bowl tests helped in the development of this model. This model can be used to predict the moisture content of the product, the recovery, particle size distribution of the effluent and particle size distribution of the product. The model also predicted how the product moisture and recovery were affected by changing the feed flow rate, feed percent solids, centrifuge speed, and particle size distribution. / Master of Science