La perspective lagrangienne, décrivant un écoulement selon les trajectoires de traceurs fluides, est une approche naturelle pour étudier les phénomènes de dispersion dans les écoulements turbulents. En turbulence de paroi, le mouvement des traceurs est influencé par le cisaillement moyen et par une forte inhomogénéité et anisotropie en proche paroi. On étudie les propriétés lagrangiennes d’un écoulement de canal turbulent par simulation numérique directe à un nombre de Reynolds modéré. Les statistiques d’accélération lagrangienne sont comparées aux expériences de suivi de particules réalisées en parallèle à ce travail. Comme en turbulence homogène isotrope (THI), les composantes d’accélération le long des trajectoires lagrangiennes se décorrèlent sur des temps comparables aux plus petites échelles de l’écoulement, tandis que la norme de l’accélération reste corrélée plus longtemps. La persistance d’anisotropie à petite échelle loin de la paroi est constatée par l’existence d’une corrélation croisée non nulle entredeux composantes de l’accélération. On montre que, en conséquence des flux moyens d’énergie cinétique en turbulence de paroi, près des parois les traceurs se déplacent et s’étalent sur des plus grandes distances quand ils sont suivis en arrière dans le temps qu’en avant. La dispersion relative de paires de traceurs est aussi étudiée. Aux temps courts, la séparation des paires est balistique pour toutes les distances à la paroi. Comme en THI, les traceurs se séparent plus rapidement lorsqu’ils sont suivis en arrière dans le temps. Aux temps plus longs, le cisaillement moyen accélère la séparation dans la direction de l’écoulement moyen. Un modèle de cascade balistique initialement proposé pour la THI est adapté aux écoulements inhomogènes / The Lagrangian perspective, describing a flow from the trajectories of fluid tracers, isa natural framework for studying dispersion phenomena in turbulent flows. In wall-boundedturbulence, the motion of fluid tracers is affected by mean shear and by strong inhomogeneityand anisotropy near walls. We investigate the Lagrangian properties of a turbulent channel flowusing direct numerical simulations at a moderate Reynolds number. Lagrangian accelerationstatistics are compared to particle tracking experiments performed in parallel to this work. Asin homogeneous isotropic turbulence (HIT), the acceleration components along Lagrangianpaths decorrelate over time scales representative of the smallest scales of the flow, while theacceleration norm stays correlated for much longer. The persistence of small-scale anisotropy farfrom the wall is demonstrated in the form of a non-zero cross-correlation between accelerationcomponents. As a result of the average fluxes of kinetic energy in wall turbulence, tracers initiallylocated close to the wall travel and spread over longer distances when tracked backwardsin time than forwards. The relative dispersion of tracer pairs is finally investigated. At shorttimes, pair separation is ballistic for all wall distances. As in HIT, relative dispersion is timeasymmetric, with tracers separating faster when tracked backwards in time. At longer times,mean shear dominates leading to rapid separation in the mean flow direction. A ballisticcascade model previously proposed for HIT is adapted to inhomogeneous flows

L'hydroélectricité est la première des énergies renouvelables électriques. Sa production repose en partie sur des centrales au fil de l'eau dont les capacités de modulation sont encore faiblement exploitées. Les capacités d'ajustement des centrales hydrauliques sont d'autant plus essentielles aujourd'hui que la pénétration d'énergies intermittentes dans un mix énergétique décarboné est indispensable.Dans cette thèse, nous nous intéressons aux centrales au fil de l'eau turbinant le débit de cours d'eau aux marnages limités. Les enjeux de sûreté, notamment liés au multi-usage de l'eau, ainsi que la perspective de moduler les débits turbinés nous ont amenés à considérer le problème de la modélisation des écoulements dans les canaux reliant les centrales au fil de l'eau. Les équations de Saint-Venant sont les plus pertinentes pour ce type de modélisation. Nous avons proposé plusieurs approches à partir de ces dernières pour caractériser analytiquement la dynamique de l'écoulement à des variations de débits turbinés. Nous avons considéré la dynamique du système autour d'un régime fluvial stationnaire non-uniforme caractéristique des configurations hydroélectriques. La première approche est basée sur une approximation basses fréquences. La seconde approche est basée sur une méthode de réduction de modèle avec une paramétrisation selon le débit support. Une troisième approche est proposée en considérant explicitement la recherche d'une solution approximée des équations de Saint-Venant linéarisées autour d'une configuration hydroélectrique. Un critère spatio-fréquentiel est alors introduit, l'existence d'un biais en basses fréquences nous conduit à proposer un modèle d'ordre réduit dont la dynamique basses fréquences est imposée selon les résultats de la première approche. La solution exprimée sous forme de fonctions de transfert, comme pour les deux précédentes approches, met en évidence explicitement la présence de modes de résonance/anti-résonance. Finalement, nous illustrons les résultats vis à vis de simulations non-linéaires et de données réelles et proposons une régulation de niveau basée sur cette dernière approche. / New challenges arise from energy transition toward a more sustainable energy mix. Hydropower is already the main source of renewable electricity. In order to integrate a massive increase in generation of renewable intermitent energies, improving the flexibility of run-of-the-river hydropower plants becomes essential. In this thesis, we focus on run-of-the river power plants facing water level constraints. Safety issues, partly due to the multiple uses of water, and the opportunity to modulate turbined flow rates have led us to adress the problem of flow modelisation in open channels that connect run-of-the rivers facilities with each others. An accurate model is provided by the Saint-Venant equations. From these latters, we have proposed different approaches to characterize analytically the flow dynamics in response to turbined flow variations. The system dynamics have been considered around a subcritical stationary non-uniform regime typical of hydroelectric configuration. The first approach is based on a low frequency approximation. The second approach is based on a parametric model reduction technique. By seeking explicitly an approximate solution to the linearized Saint-Venant equations around an hydroelectric configuration, we have proposed a third approach. A space-frequency criterion is introduced, which shows a bias in low frequency. Results of the first approach are then used to propose a reduced order model asymptotically exact in low frequency. As for the two other approaches, the solution takes the shape of parametric transfer functions. Resonance/anti-resonance modes explicitly appear. Finally, comparisons with non-linear simulations taking into account actual real data are discussed and a water level controller is developed based on the last approach.

Flow modelling in a compound channel is a complex matter. Indeed, due to the smaller velocities in the floodplains than in the main channel, shear layers develop at the interfaces between these subsections, and the channel conveyance is affected by a momentum transfer corresponding to this shear layer, but also to possible geometrical changes in a non-prismatic reach. In this work, a one-dimensional approach, the Exchange Discharge Model (EDM), is proposed for such flows. The EDM accounts for the momentum transfer between channel subsections, estimated as proportional to the velocity gradient and to the discharges exchanged through the interface; where two main processes are identified : (1) the turbulent exchange, due to the shear-layer development; and (2) the geometrical transfer, due to cross-sectional changes. The EDM is successfully validated for discharge prediction, but also for water-profile computation, through comparison with existing laboratory and field measurements. The momentum transfer due to turbulent exchanges is then studied experimentally, theoretically and numerically. At first, new experimental data, obtained by using Particle Tracking Velocimetry techniques, are presented : the periodical vortex structures that develop in the shear layer are clearly identified and characterised. Secondly, a hydrodynamic linear stability analysis enables to predict quite successfully the wave length of some observed vortices. Lastly, an Unsteady-RANS numerical method is used to simulate the perturbation development. The estimated vortex wave lengths agree again with the measurements and the theoretical predictions, although vortices merging occurs in the simulation results, which was actually not observed experimentally. The velocity-profile prediction is found improved when the effect of vortices is considered, thanks to the corresponding additional shearing. The geometrical transfer is also investigated experimentally and numerically. Novel experiments are designed, with the measurements of the flow in a compound channel with symmetrically narrowing floodplains. The mass transfer and the evolution of the flow distribution along the channel length are clearly observed. A significant additional head loss due to this transfer is measured, in accordance with the EDM hypothesis. Measured water profiles are finally compared successfully with the EDM predictions. In addition to the EDM development and validation, the so-called Lateral Distribution Method (LDM) is also investigated and the significance of the secondary-currents models proposed by previous authors for this method is discussed. When considering the velocity-profile prediction, the effect of these helical secondary currents is again clearly highlighted, by using dispersion terms in the Saint-Venant equations. However, the actual physical meaning of the related dispersion coefficients remains uncertain. In addition, an extended LDM is also proposed and discussed for non-prismatic flow modelling, using the new narrowing-channel data set./La modélisation des écoulements dans les rivières à plaines inondables est particulièrement complexe. En effet, la vitesse de l'eau étant plus faible sur la plaine d'inondation que dans le lit mineur, une couche de cisaillement se développe à l'interface entre ces sous-sections. La débitance totale de la rivière est dés lors réduite, à cause du transfert de quantité de mouvement qu'occasionne la présence de la couche de cisaillement, mais aussi de part les changements de géométrie qui peuvent se produire dans un lit non-prismatique. La présente thèse propose, pour la représentation de tels écoulements, une nouvelle approche uni-dimensionnelle dénommée Modèle des Débits d'Echange ("Exchange Discharge Model" – EDM). Le transfert de quantité de mouvement entre les soussections de la rivière est pris en compte par l'EDM comme étant proportionnel au gradient de vitesse entre celles-ci et aux débits échangés à travers leur interface. A cette interface, deux phénomènes sont essentiellement présents : (1) un échange turbulent, dû au développement de la couche de cisaillement; et (2) un transfert géométrique, correspondant aux changements de section. L'EDM est validé avec succès pour la prédiction du débit et pour le calcul de lignes d'eau, par comparaison avec des données existantes de laboratoire et de terrain. Le transfert de quantité de mouvement dû à l'échange turbulent est ensuite étudié expérimentalement, théoriquement et numériquement. De nouvelles mesures sont obtenues, au moyen d'une technique de vélocimétrie par suivi de particules. Les structures périodiques qui se développent dans la couche de cisaillement sont clairement identifiées et caractérisées. Deuxièmement, une analyse linéaire de stabilité hydrodynamique permet de prédire théoriquement les longueurs d'onde de quelques tourbillons qui ont été observés expérimentalement, et ce avec succès. Enfin, un modèle numérique, de type "Unsteady-RANS", est utilisé pour simuler la croissance des tourbillons dans la couche de cisaillement. Encore une fois, les longueurs d'onde obtenues correspondent relativement bien avec les valeurs mesurées et prédites théoriquement; bien que les coalescences de tourbillons qui se produisent numériquement n'aient pas été observées expérimentalement. La prédiction des profils de vitesse est améliorée, lorsque l'effet des tourbillons est considéré, grâce à la contrainte de cisaillement additionnelle que ceux-ci génèrent. Les transferts géométriques sont également explorés expérimentalement et numériquement. Une nouvelle campagne expérimentale a été réalisée, en considérant l'écoulement dans un lit composé symétrique, dont les plaines d'inondation se rétrécissent progressivement. Le transfert de masse entre sous-sections et la redistribution des débits qui lui est associée sont clairement observés au long du canal. Une importante perte de charge additionnelle due à ce transfert est mesurée, en concordance avec les hypothèses de l'EDM. Finalement, les lignes d'eau mesurées sont reproduites avec succès par un calcul utilisant l'EDM. En complément au développement et à la validation de l'EDM, la "Lateral Distribution Method" (LDM) est également utilisée, avec pour objectif la clarification du rôle des termes de courants secondaires proposés par différents auteurs. Par rapport à la prédiction du profil de vitesse, l'effet de ces courants secondaires est très marqué. Il est ici reproduit en utilisant des termes de dispersion dans les équations de Saint-Venant. Cependant, le sens physique des valeurs des coefficients de dispersion qui doivent être utilisés est discutable. Par ailleurs, une LDM étendue, pour les écoulement en lits nonprismatiques, est proposée et commentée, en utilisant le nouveau jeu de données pour le canal convergent.

No description available.

Two computationally fast fluid mechanical models for vertical axis turbines are the streamtube and the vortex model. The streamtube model is the fastest, allowing three-dimensional modeling of the turbine, but lacks a proper time-dependent description of the flow through the turbine. The vortex model used is two-dimensional, but gives a more complete time-dependent description of the flow. Effects of a velocity profile and the inclusion of struts have been investigated with the streamtube model. Simulations with an inhomogeneous velocity profile predict that the power coefficient of a vertical axis turbine is relatively insensitive to the velocity profile. For the struts, structural mechanic loads have been computed and the calculations show that if turbines are designed for high flow velocities, additional struts are required, reducing the efficiency for lower flow velocities.Turbines in channels and turbine arrays have been studied with the vortex model. The channel study shows that smaller channels give higher power coefficients and convergence is obtained in fewer time steps. Simulations on a turbine array were performed on five turbines in a row and in a zigzag configuration, where better performance is predicted for the row configuration. The row configuration was extended to ten turbines and it has been shown that the turbine spacing needs to be increased if the misalignment in flow direction is large.A control system for the turbine with only the rotational velocity as input has been studied using the vortex model coupled with an electrical model. According to simulations, this system can obtain power coefficients close to the theoretical peak values. This control system study has been extended to a turbine farm. Individual control of each turbine has been compared to a less costly control system where all turbines are connected to a mutual DC bus through passive rectifiers. The individual control performs best for aerodynamically independent turbines, but for aerodynamically coupled turbines, the results show that a mutual DC bus can be a viable option.Finally, an implementation of the fast multipole method has been made on a graphics processing unit (GPU) and the performance gain from this platform is demonstrated.

Thickened tailings production and disposal continue to grow in importance in the mining industry. In particular, the transport of oil sands tailings is of interest in this study. These tailings must be in a homogeneous state (non-segregating) during pipeline flow and subsequent discharge. Tailings are often transported in an open channel or flume. Slurries containing both clay and coarse sand particles typically exhibit non-Newtonian rheological behaviour. The prediction of the flow behaviour of these slurries is complicated by the limited research activity in this area. As a result, the underlying mechanisms of solids transport in these slurries are not well understood. To address this deficiency, experimental studies were conducted with kaolin clay slurries containing coarse sand in an open circular channel.<p> A numerical model has been developed to predict the behaviour of coarse solid particles in laminar, open channel, non-Newtonian flows. The model involves the simultaneous solution of the Navier-Stokes equations and a scalar concentration equation describing the behaviour of coarse particles within the flow. The model uses the theory of shear-induced particle diffusion (Phillips et al., 1992) to provide a number of relationships to describe the diffusive flux of coarse particles within laminar flows. A sedimentation flux has been developed and incorporated into the Phillips et al. (1992) model to account for gravitational flux of particles within the flow. Previous researchers (Gillies et al., 1999) have shown that this is a significant mechanism of particle migration.<p> The momentum and concentration partial differential equations have been solved numerically by applying the finite volume method. The differential equations are non-linear, stiff and tightly coupled which requires a novel means of analysis. Specific no-flux, no-slip and no-shear boundary conditions have been applied to the channel walls and free surface to produce simulated velocity and concentration distributions. The results show that the model is capable of predicting coarse particle settling in laminar, non-Newtonian, open channel flows. The results of the numerical simulations have been compared to the experimental results obtained in this study, as well as the experimental results of previous studies in the literature.

Thickened tailings production and disposal continue to grow in importance in the mining industry. In particular, the transport of oil sands tailings is of interest in this study. These tailings must be in a homogeneous state (non-segregating) during pipeline flow and subsequent discharge. Tailings are often transported in an open channel or flume. Slurries containing both clay and coarse sand particles typically exhibit non-Newtonian rheological behaviour. The prediction of the flow behaviour of these slurries is complicated by the limited research activity in this area. As a result, the underlying mechanisms of solids transport in these slurries are not well understood. To address this deficiency, experimental studies were conducted with kaolin clay slurries containing coarse sand in an open circular channel.<p> A numerical model has been developed to predict the behaviour of coarse solid particles in laminar, open channel, non-Newtonian flows. The model involves the simultaneous solution of the Navier-Stokes equations and a scalar concentration equation describing the behaviour of coarse particles within the flow. The model uses the theory of shear-induced particle diffusion (Phillips et al., 1992) to provide a number of relationships to describe the diffusive flux of coarse particles within laminar flows. A sedimentation flux has been developed and incorporated into the Phillips et al. (1992) model to account for gravitational flux of particles within the flow. Previous researchers (Gillies et al., 1999) have shown that this is a significant mechanism of particle migration.<p> The momentum and concentration partial differential equations have been solved numerically by applying the finite volume method. The differential equations are non-linear, stiff and tightly coupled which requires a novel means of analysis. Specific no-flux, no-slip and no-shear boundary conditions have been applied to the channel walls and free surface to produce simulated velocity and concentration distributions. The results show that the model is capable of predicting coarse particle settling in laminar, non-Newtonian, open channel flows. The results of the numerical simulations have been compared to the experimental results obtained in this study, as well as the experimental results of previous studies in the literature.

No description available.

Most flows in nature and engineering are turbulent, and many are wall-bounded. Further, in turbulent flows, the turbulence generally has a large impact on the behavior of the flow. It is therefore important to be able to predict the effects of turbulence in such flows. The Navier-Stokes equations are known to be an excellent model of the turbulence phenomenon. In simple geometries and low Reynolds numbers, very accurate numerical solutions of the Navier-Stokes equations (direct numerical simulation, or DNS) have been used to study the details of turbulent flows. However, DNS of high Reynolds number turbulent flows in complex geometries is impractical because of the escalation of computational cost with Reynolds number, due to the increasing range of spatial and temporal scales. In Large Eddy Simulation (LES), only the large-scale turbulence is simulated, while the effects of the small scales are modeled (subgrid models). LES therefore reduces computational expense, allowing flows of higher Reynolds number and more complexity to be simulated. However, this is at the cost of the subgrid modeling problem. The goal of the current research is then to develop new subgrid models consistent with the statistical properties of turbulence. The modeling approach pursued here is that of "Optimal LES". Optimal LES is a framework for constructing models with minimum error relative to an ideal LES model. The multi-point statistics used as input to the optimal LES procedure can be gathered from DNS of the same flow. However, for an optimal LES to be truly predictive, we must free ourselves from dependence on existing DNS data. We have done this by obtaining the required statistics from theoretical models which we have developed. We derived a theoretical model for the three-point third-order velocity correlation for homogeneous, isotropic turbulence in the inertial range. This model is shown be a good representation of DNS data, and it is used to construct optimal quadratic subgrid models for LES of forced isotropic turbulence with results which agree well with theory and DNS. The model can also be filtered to determine the filtered two-point third-order correlation, which describes energy transfer among filtered (large) scales in LES. LES of wall-bounded flows with unresolved wall layers commonly exhibit good prediction of mean velocities and significant over-prediction of streamwise component energies in the near-wall region. We developed improved models for the nonlinear term in the filtered Navier-Stokes equation which result in better predicted streamwise component energies. These models involve (1) Reynolds decomposition of the nonlinear term and (2) evaluation of the pressure term, which removes the divergent part of the nonlinear models. These considerations significantly improved the performance of our optimal models, and we expect them to apply to other subgrid models as well. / text

Η παρούσα Διατριβή Μεταπτυχιακού Διπλώματος Ειδίκευσης, πραγματεύεται την ανάλυση της τυρβώδους ροής σε ανοικτούς αγωγούς στον πυθμένα των οποίων ενυπάρχουν σχηματισμοί μορφής θινών (dunes). Μελετήθηκε η περίπτωση 5 θινών οι οποίες τοποθετήθηκαν στον πυθμένα ενός καναλιού βάθους d θεωρώντας μόνιμη ροή. Για την επίλυση χρησιμοποιήθηκαν οι εξισώσεις RANS, ενώ για το κλείσιμο της τύρβης χρησιμοποιήθηκαν τα μοντέλα μιας εξίσωσης Spalart-Allmaras και δύο εξισώσεων k-ε. Η διαχείριση της ελεύθερης επιφάνειας έγινε με την μέθοδο VOF, ενώ η αριθμητική επίλυση βασίστηκε στην μέθοδο των πεπερασμένων όγκων και πραγματοποιήθηκε με τον εμπορικό κώδικα FLUENT 6.1.2. Για την ροή στον ανοικτό αγωγό, στον πυθμένα του οποίου ενυπήρχαν οι σχηματισμοί, θεωρήθηκε αριθμός Reynolds , κλίση πυθμένα και συντελεστή Manning , ο οποίος αντιστοιχεί σε ισοδύναμο ύψος τραχύτητας τοιχωμάτων . Προκειμένου να επαληθευθεί η ακρίβεια της αριθμητικής μεθόδου, επιλύθηκε η περίπτωση του επίπεδου πυθμένα και τα αποτελέσματα συγκρίθηκαν με γνωστά πειραματικά αποτελέσματα καθώς και αποτελέσματα τα οποία προέκυψαν από την μονοδιάστατη ανάλυση της ροής πάνω από επίπεδο πυθμένα. Τα αποτελέσματα βρέθηκαν σε καλή συμφωνία, κυρίως για την κατανομή της ταχύτητας, ενώ για την τύρβη υπήρχε πολύ καλή συμφωνία κυρίως πλησίον του πυθμένα. Για το πρόβλημα των θινών εξετάσθηκαν: (α) τρεις περιπτώσεις με σταθερό άνοιγμα θίνης προς βάθος ροής και διαφορετικά ύψη θινών , 0.25 και (β) τρεις περιπτώσεις με σταθερή αναλογία ανοίγματος προς ύψος και ύψη θινών όπως στην περίπτωση (α). Η ανάλυση έδειξε ότι το μέσο προφίλ της ελεύθερης επιφάνειας μειώνεται στην διεύθυνση της ροής, ενώ πάνω από κάθε θίνη το πλάτος της ανύψωσης της ελεύθερης επιφάνειας αυξάνει με την αύξηση του ύψους και του ανοίγματος των θινών. Η κατανομή των διατμητικών τάσεων παρουσιάζει κυματοειδή μορφή υπεράνω των θινών και αυξάνει αυξανομένου του ύψους τους και με την μείωση του ανοίγματός τους. Πίσω από κάθε θίνη δημιουργείται θύλακας ανακυκλοφορίας της ροής και ο λόγος της απόστασης του σημείου επανακόλλησης προς το ύψος της θίνης είναι . / The spatial development of the turbulent open channel flow over bottom with five dunes is studied. The steady-state flow is described by the RANS equations utilizing either the or the Spalart-Allmaras turbulence models. The free-surface treatment is based on the VOF formulation, while the numerical solution is based on a finite-volume, unstructured-grid discretization. Lengths are rendered dimensionless by the inflow channel depth, while velocities by the mean inflow velocity. The inflow Reynolds number is , the channel slope is and the Manning coefficient is , which results to a roughness height . In order to verify the numerical methods, the flat bottom case is considered and the numerical predictions are compared to known experimental data. We get very good agreement for the velocity distributions, while for turbulence the results are very good close to the bottom and poor close to the free surface. Then, we consider: (a) three cases with constant dune length and differing dune heights 0.15, 0.25, 0.35, and (b) three cases with constant ratio and dune heights as in (a). The spatial development of the free-surface elevation over the dunes presents a negative mean slope for all cases. Locally over each dune, the amplitude of the free-surface elevation increases with increasing dune height and increasing dune length. The spatial development of the wall shear stress presents a wave-like behavior and its amplitude increases with increasing dune height and decreasing dune length. On every dune crest the streamwise velocity profile is steeper than the universal logarithmic profile similar to the behavior in a favorable pressure gradient boundary layer. The detachment at each dune crest is followed by a recirculation region and reattachment at a distance from the dune trough.