Search

International audience; Predicting mixing processes, especially transverse mixing, downstream of river confluences, is necessary for assessing and modeling the fate of pollutants transported in river networks, but it is still challenging. Typically, there is a lack of transverse mixing solutions implemented in 1-D hydrodynamical models widely used in river engineering applications. To investigate the mixing processes developing downstream of a medium-sized river confluence, three high-resolution in situ surveys are conducted at the Rhône-Saône confluence in France, based on geolocated specific conductivity and hydroacoustic measurements. Contrasting mixing situations are observed depending on hydrological conditions. In some cases, the two flows mix slowly due to turbulent shear at their vertical interface. This can be modeled by an analytical solution of the advection-diffusion equation. In other cases, the waters from one of the two tributaries move under the waters of the other tributary. The induced local circulation enhances transverse mixing but not vertical mixing and the flow remains stratified vertically, which may be missed when surface or satellite images are analyzed qualitatively. Stratification may be predicted by comparing the time scales for shear and density-driven adjustment. Shear-dominated transverse mixing of depth-averaged concentrations can be predicted analytically and implemented in 1-D hydrodynamical models. However, the initiation of apparently rapid transverse mixing due to density-driven circulation remains to be better understood and quantified. Plain Language Summary Predicting how waters mix downstream of river confluences is necessary for assessing and modeling the fate of pollutants transported in river networks, but it is still challenging. Typically, there is a lack of transverse mixing solutions implemented in models widely used in river engineering applications. To investigate the mixing processes developing downstream of a medium-sized river confluence, three high-resolution in situ surveys are conducted at the Rhône-Saône confluence in France. Contrasting slow or rapid mixing situations are observed depending on hydrological conditions. The transverse mixing of depth-averaged concentrations can be predicted analytically and implemented in 1-D hydrodynamical models. However, the initiation of rapid transverse mixing due to difference in fluid density remains to be better understood and quantified.

International audience; no abstract

High-return-period floods involve complex flows over large floodplains covered with different types of resistive elements, from bed roughness to large emerging obstacles. In order to charac...

This paper provides, through laboratory experiments, the flow discharge passing from a flooded street to a building area through different types of damaged openings (a door, a window and a gate) along with a sensitivity analysis of this discharge to the presence of obstacles or facade details. Four flow regimes in the street are tested, representing a high or low street slope with a high and low water depth. These flows resemble that over rectangular side weirs, with the major difference that obstacles are usually located near the opening in the street or on the facade. 26 configurations of obstacles are then included, one after the other, and their impact on the flow intrusion is measured. For the flow cases without obstacle, the agreement of the semi-analytical equations available in the literature strongly varies from one equation to another. On the other hand, the 220 tested flow cases reveal that the location of the obstacle with regards to the opening strongly modifies the impact of the obstacles, that increasing the Froude number tends to increase the impact of the obstacles and that obstacles have a higher impact on the intrusion discharge through a window than to a door or a gate. Finally we conclude that the largest obstacles (typically parked cars) located in the vicinity of openings should be somehow included in operational numerical models that calculate urban floods for a fair prediction of the intrusion discharge.

Turbulent mixing layers associated with streamwise uniform and non-uniform flows in compound channels (main channel with adjacent floodplains) are experimentally investigated. The experiments start with uniform flow conditions. The streamwise non-uniformity is then generated by imposing an imbalance in the upstream discharge distribution between main channel (MC) and floodplains (FPs), keeping the total discharge constant, which results in a transverse depth-averaged mean flow. This study firstly aims at assessing the effect of a transverse flow on the mixing layer and coherent structures that form at the MC/FP interfaces. A wide range of initial velocity ratio or dimensionless shear between MC and FP is tested. The study secondly aims at assessing the effect of this velocity ratio on the mixing layer, for a fixed vertical confinement of flow. The total discharge was then varied to quantify the confinement effect. The results show that, far from the inlet section, Reynolds-stresses increase with local velocity ratio for a fixed confinement, and decrease with confinement for a fixed velocity ratio. It is also shown that, irrespective of confinement, the existence of quasi-two-dimensional coherent structures is driven by velocity ratio and the direction and magnitude of transverse flow. These structures cannot develop if velocity ratio is lower than 0.3 and if a strong transverse flow towards the MC occurs. In the latter case, the transverse flow is the predominant contribution to momentum exchange (compared with turbulent mixing and secondary currents), convex mean velocity profiles are observed, preventing the formation of quasi-two-dimensional structures.

The authors measured and computed the hydrodynamics and passive scalar dispersion in 90-degree open channel confluences over flat and degraded beds with a dominant upstream or tributary inflow. The present discussion essentially deals with the direction of rotation of the secondary currents, reported for the flat bed configuration with dominant tributary inflow. This rotation direction is indeed surprisingly opposite to the ones reported in the literature, both from calculations and measurements, even if present geometry slightly differs from literature geometries.

International audience; The drag reduction method using polymer additives is a common strategy to minimize friction losses when carrying fluids (water, oil, or slurries) in pipes over long distances. Previous studies showed that the interactions between the polymer and turbulent structures of the flow tend to modify the streamwise velocity profile close to the walls by adding a so-called elastic sublayer between the classical viscous and log layers. The gain in linear head losses can reach up to 80% depending on the roughness of the walls and the concentration of polymers. The application of this technique to sewers and the subsequent gain in discharge capacity motivated this work to quantitatively measure the drag reduction in classical open-channel flows. Three measurement campaigns were performed in a dedicated long flume for several water discharges and several polymer concentrations: backwater curves over smooth and rough channel walls (including velocity and turbulent shear-stress profiles) and flows around emerging obstacles. The addition of polymers, even in limited concentrations, allowed a high friction decrease with the typical Darcy-Weisbach coefficient reduced by factors of 2 and 1.5, respectively, in smooth and rough walls configurations without obstacles, but without strong modifications of the nondimensional velocity profiles. In contrast, when adding emerging obstacles, the flow was unaffected by the inclusion of polymers, in agreement with the prediction of the literature. The drag reduction method by addition of small concentrations of polymers thus appears to be a promising technique to increase flow conveyance in open-channel flows.

International audience; The flow pattern within a storage zone governs the efficiency of passive scalar accumulation , spreading and homogenization. The present paper aims at characterizing experimentally the evolution of this flow pattern in a simplified open-channel lateral cavity with twenty increasing horizontal aspect ratio, focusing on the transition between the different flow patterns. Four main flow types are observed and detailed herein with, as the aspect ratio gradually increases: two cells aligned along the main stream axis, one unique cell, two cells aligned transversally and finally a more complex 3D pattern in the second cell. Quantitative characteristics of these cells are extracted from the velocity fields measured in the cavity. A 2D angular momentum balance is then applied that explains the transitions between these flow types and proves to be able to predict the flow patterns reported in the literature.

International audience; Lateral cavities are major storage zones in riverine environments for which the mass exchanges with the main stream strongly impact The characteristics of the habitat in these dead zones. An experimental work is presented here with a controlled main stream and aconnected open-channel lateral cavity to assess the processe s responsible for these exchanges and to quantify the exchange capacities. In a first step, the measurements of passive scalar transport allow us to identify the physical processes involved in the exchange of mass from the main stream and its spreading within the cavity. In a second step, the quantitative mass exchange coefficient, representative of the exchange capacity, is measured for 28 flow and cavity configurations. The sensibility analysis to the governing parameters proposed by the dimensional analysis then reveals that changing the geometric aspect ratio of the cavity does not affect the exchange coefficient while increasing the normalized water depth or decreasing the Reynolds number of the main stream tend to increase this coefficient. Indeed, The se parameters modify both the growth rate of the mixing layer width at the interface and the amplitude of the alternating transverse Velocity across the interface, thus affect ing the exchange capacities from the main stream to the cavity

International audience; This paper aims at describing two different recirculation structures observed both numerically and experimentally in the lateral branch of a dividingopen channel flow with identical downstream weirs. The flow conditions leading to each identified recirculation structure are also investigated. Thefirst recirculation pattern is a classic two-dimensional closed recirculation, typically described in the literature. The second one is a three-dimensional helix-shaped recirculation that differs from the first one notably regarding the streamline behaviour.We show that the occurrence of each structure of recirculation can be predicted by means of the Froude number and the aspect ratio in the upstream channel of the bifurcation.

International audience; 12 The present paper aims to identify ways to reduce pollution injected by residents in the urban 13 wastewaters network system. Two approaches are considered. The first one uses flow and 14 pollutant calculation to test whether a polluter can easily be identified in a neighborhood. The 15 second approach uses a survey to examine what incentive would be most effective to 16 influence residents' behavior. Hydrodynamic simulation results show that concentration 17 profiles at the network outlet corresponding to all possible polluters are similar and thus do 18 not point out specific resident source of pollution. Household level survey results show that 19 most socioeconomic and public good-related characteristics do not play a significant role in 20 explaining choices to discard in the home wastewaters network. Apart from the nature of the 21 waste itself, by far, the belief that the respondent has about her neighbors' and relatives' 22 discarding behavior is the main driver of the choice. 23 24

International audience; Seed and fruit dispersal along watercourses favours the long-distance migration of invasive species, not only for aquatic or wetland species, but also for terrestrial wind- dispersed plants, like the Japanese knotweed. The present paper aims at investigating the role of watercourses in the dispersal of the knotweed due to its frequent occurrence on riverbanks and production of fertile achenes (type of fruit of the Japanese knotweed). This dispersal occurs along two steps after the fruits deposit on the water surface: floatation first and then sinking towards the bottom of the watercourse. Regarding the first step, the effects of agitation of the water, temperature, surface tension and luminosity on the achenes floatability are experimentally studied. While no influence of luminosity is observed, an increase of temperature greatly decreases the floating time. Floating time also decreases as the contact between water and the fruit is enhanced (through submersion of achenes, agitation of the water or lower surface tension). Regarding the second step, the fall velocity of the fruits in water at rest is measured and appears to be independent of the seed history (floating time). 3D helical motions are systematically observed with constant tangential velocity with respect to the falling velocity. The trajectory of the fruits in a shear flow is then measured and the evolution of their velocity components along the sinking process is discussed. Finally, the contribution of both steps to the long-distance migration of the seeds is estimated.

International audience; This study aims at investigating the mixing layer located at the interface between the free stream and recirculation zone downstream from a sudden lateral open-channel expansion. Specific attention is paid to the interaction of the shallowness of the flow, characterized by the bed-friction number, with the lateral confinement due to the side wall. The velocity field for four flows with the same geometry but very different bed-friction numbers is measured in detail in order to characterize the mean velocity fields and Reynolds stresses across the mixing layers and evaluate the width of the mixing layers and their growth rates along with the typical oscillation frequencies. In the upstream region of the recirculation zone, the mixing-layer characteristics for the examined configurations are analogous to the ones of classical—laterally unbounded—mixing layers. In this region, the shallowness modifies the shape of the streamwise velocity profiles, extends the mean velocity gradient magnitudes and lowers the Reynolds stress terms, but hardly affects the mixing layer’s expansion rate. On the other hand, in the region near the flow reattachment, the mixing layer adopts very different behavior, with an abrupt drop of the its expansion. This change in behavior is linked to the dynamics of the two-dimensional (2D) vortices within the mixing layer, not a damping effect of the bed friction on these vortices, as the local bed-friction numbers remain much lower than the critical values reported in the literature. It is rather due to the interaction of the coherent structures with the side wall; the characteristics of this interaction itself are influenced by the flow shallowness. It is also shown that the damping effect due to bed friction is not responsible for the huge variations reported on the recirculation zones length. This imposes a distinction between local mixing-layer shallowness—derived using the mixing-layer width as length scale and governing the mixing layer’s characteristics—and a global flow shallowness—derived using the expansion width and governing the recirculation length.

When managing a sewer network, the water discharge has to be measured in real time at several locations in the network. In locations where a simple stage discharge relationship cannot be applied due to an influence of the backwater curve, this measurement is to be performed using sensors that measure the bulk velocity locally in order to estimate the discharge. The aim of this work is to quantify the errors made when such measurement is performed downstream from two types of singularities: a bend and a confluence, which are commonly encountered in sewer networks. Two types of sensors are tested: a time-of-flight flow meters and a acoustic Doppler flowmeter. Both technologies are first detailed along with their possible sources of errors when used in the vicinity of singularities. The second section then introduces the numerical calculation of all flow configurations tested downstream both kinds of singularities. The third section shows the results in terms of error of discharge estimation as a function of the location with regards to the singularities, the type of singularity and the flow conditions. This permits to finally make recommendations for the use of such sensors in a real sewer network.; Les gestionnaires de réseau d'assainissement urbain sont amenés à équiper leurs réseaux de points de mesures débimétriques. Lorsqu'une relation permanente permettant de relier localement la hauteur d'eau et le débit n'existe pas, il est nécessaire de mesurer la vitesse moyenne de l'écoulement afin d'en déduire le débit localement. L'objectif de cet article est de quantifier les erreurs faites lorsqu'une mesure de vitesse est effectuée à l'aval d'une singularité pour des situations réalistes de collecteurs d'assainissement urbain. La première section de cet article présente les deux types de capteurs (cordes et Doppler) en détaillant leur mode d'échantillonnage spatial ainsi que les sources d'erreur associées. La deuxième section présente la démarche de modélisation numérique ainsi que les domaines de calcul associés à l'aval respectivement d'une confluence et d'un coude. Enfin la troisième partie présente les résultats en termes de réponse des capteurs à ces champs de vitesse et discute les erreurs pour les différentes configurations considérées.

Congrès SHF, Hydrométrie 2017, LYON, FRANCE, 14-/03/2017 - 15/03/2017; Sewer network managers are more and more implementing discharge sensors within sewer networks. When no permanent relation exists between the local water depth and discharge, velocity sensors are usually installed. In such case, the velocity values measured by these sensorS must be extrapolated to reach the averaged velocity across the flow section. This extrapolation implies uncertainties which are difficult to assess. 3D modelling of the velocity field allows to simulate the response of sensors for assessing these uncertainties, and reduce them by numerical calibration. In the following text, the general methodology is first introduced and then applied to two case studies, including a bend flow and a complex section configuration; Les gestionnaires de réseau d'assainissement urbain sont amenés à équiper leurs réseaux à l'aide de débitmètres. Lorsqu'il n'existe pas une relation permanente permettant de relier localement la hauteur d'eau et le débit, on recourt fréquemment à la mise en place de capteurs de vitesse. Le passage depuis la vitesse mesurée par ces capteurs vers la vitesse moyenne dans la section implique des incertitudes difficiles à évaluer. La modélisation 3D des écoulements permet de simuler la réponse des capteurs pour évaluer ces incertitudes, et, le cas échéant, de les réduire par étalonnage numérique. Après une présentation de la démarche générale, nous l'illustrerons par un cas générique d'un site à l'aval d'un coude et par le cas spécifique d'une section complexe

The reported experimental study assesses the effects of flow non-uniformity on the momentum flux in straight compound channels. Two flumes were used, featuring vertical and sloping banks. Starting with uniform flow condition, various imbalances in the upstream discharge distribution were introduced. This resulted in a time-averaged lateral flow and advective transport of momentum, which interacted with the shear-layer turbulence generated by the compound geometry. To investigate this interaction, the three contributions to transverse momentum flux (depth-averaged flow, shear-layer turbulence and dispersive term of spanwise velocity) are assessed. The first two contributions were strengthened by the sloping banks, while the third becomes important for the case of the vertical bank. With a lateral flow towards the main channel, the first contribution rises at the expense of the second. With a lateral flow towards the floodplain, the first two contributions have the same order of magnitude, and the Boussines...

Avec l’augmentation de la puissance de calcul des ordinateurs, la simulation 3D devient de plus en plus envisagee en tant qu’outil de conception ou d’instrumentation de sites en reseau d’assainissement. De plus, elle permet la comprehension complete de certains ecoulements particulierement complexes et souvent rencontres en reseau d’assainissement, du fait des nombreux ouvrages speciaux presents. Nous proposons d’utiliser l’approche RANS (Reynolds-Averaged Navier-Stokes) implementee dans plusieurs codes de calcul commerciaux pour simuler ces ecoulements. L’un des defauts majeurs de ces codes de calcul est le manque de fiabilite de certains resultats numeriques, due a la non-connaissance des incertitudes liees au choix d’une strategie de modelisation plutot qu’une autre. Le but de cette etude est de pouvoir quantifier ces incertitudes et definir la plage d’erreur associee aux resultats de simulation, pour le plus de conditions hydrauliques et de strategie de modelisation possible. A terme, l’objectif est de pouvoir etendre l’usage de ces codes aux bureaux d’etude en hydraulique dans le but de mieux concevoir ou mieux instrumenter des sites en vue d’ameliorer la qualite des observations. Les cas d’etudes selectionnes comprennent une canalisation circulaire relativement simple a modeliser mais dont l’ecoulement reste complexe et une jonction a 90°. La canalisation a permis l’etude de la sensibilite des resultats de modelisation aux parametres geometriques. Les premiers resultats montrent que la solution est sensible au changement de rugosite meme en regime torrentiel. La jonction sert a evaluer les performances des nombreuses strategies de modelisation qu’il est possible d’adopter, et leurs effets sur les resultats de simulation. Le modele de proche-paroi « Scalable » semble pertinent pour ce type d’ecoulements.

International audience; An experimental study of the characteristics of dividing critical flows in a 90° open-channel junction formed by three horizontal equal-width branches is presented, conducted for various inflow discharges and downstream boundary conditions. Outflow discharges and flow depths were also measured. Four main flow patterns are identified considering the location and length of the hydraulic jumps that develop across the main and lateral channels. A relationship between the discharge division ratio and the tailwater Froude number is found, reproducing well the experimental data. Practical applications of this relationship include the design of open channel and pipe networks and one-dimensional numerical modelling of flood propagation in urban areas. The study shows also that a two-dimensional depth-averaged numerical model hardly gives better predictions of the discharge ratios because of the presence of three-dimensional features at the junction.

During severe urban flooding, the surface flow can become supercritical near some crossroads if the slope of the streets reaches a critical steepness. The present study aims to validate the capacity of a code solving the two-dimensional shallow water equations to simulate the flows in a four-branch supercritical cross junction. An experimental study showed that five main flow patterns can be observed and are determined by the upstream flow characteristics as well as the slope of the channels. Computed and measured flows were compared on five detailed water depth fields and on more than 200 measured flow rate distributions. It appears that the flow rate distribution to the downstream branches and the flow characteristics, computed by the code using a set of reference parameters, compare well with the experimental data. Nevertheless, some discrepancies appear concerning the prediction of the location and of the thickness of the oblique jumps mainly because jumps are set on one cell in the numerical model. F...

National audience; Dividing flows have been widely studied. These special structures found in sewers are responsible for complex flows. The understanding of these complex flows is the basis to better manage pollutants transport and mixing in sewer systems. This paper describes two flow structures observed experimentally and numerically in the downstream lateral branch of a 90° open channel dividing flow. These structures are recirculation flows: a classical one (2D) and a helix-shaped one (3D). First, the numerical model has been validated, being confronted with experimental measurements of velocities and water depths for two reference cases. Then, a numerical campaign has been led in order to describe in detail these structures. For the two types of recirculation, the flow visualization, notably of the streamlines in 3D, shows quite different characteristics of exchange with the main flow. It should be the same for their influence on the operation in sewers.; Les bifurcations sont des ouvrages spéciaux présents en réseau d'assainissement et qui sont le siège d'écoulements complexes qu'il est important de comprendre et de maîtriser afin de mieux appréhender le transport de polluants particulaires ou dissous. L'étude présentée dans cet article porte sur la description de deux structures différentes observées expérimentalement et numériquement dans la branche aval latérale d'une bifurcation à 90° à surface libre. Ces deux structures sont des recirculations, l'une classique (2D), l'autre en hélicoïde (3D). Après avoir validé le modèle numérique grâce aux données expérimentales de vitesse et hauteur, une campagne de simulations a été menée dans le but de bien décrire ces structures. La visualisation des lignes de courant en 3D montre que leurs caractéristiques d'échange avec l'écoulement principal dans chaque branche sont très différentes, comme devrait l'être ainsi leurs influences sur le fonctionnement du réseau.

This study focuses on the discharge distribution in an intersection of four channels, similar to a city crossroad. The channels and the intersection are all horizontal. Flow enters through two of the channels, and leaves through the other two. The flow is subcritical everywhere, and flow depths are controlled by vertical weirs at the exits of the outlet channels. The main variables that are measured are the flow rates in the four channels. When the weir heights in the outlet channels are the same, the ratio of flow rates in the outlet channels depends only on the ratio of flow rates in the inlet channels; if the outlet conditions are different, other parameters, such as the total flow rate also become important. The flow has also been simulated numerically using a solution of the 1-D Saint Venant equations, with a simple model to predict flow distribution in the intersection. A comparison with the experimental data shows that this model works well for the limited range of experimental conditions studied here. However, further work is needed on a wider range of conditions, closer to real conditions, before the model can be considered valid for practical applications.

Transcritical flows in three and four branch channel intersections have been studied experimentally, and an empirical relationship is derived from the data to express the flow distribution as a function of the Froude number of the flow in the main channel downstream of the intersection, and in the case of the four-branch intersection, the ratio of the incoming flows in the two inlet channels. The empirical relationship agrees closely with the data from this study, and with data from other studies in comparable conditions. The results illustrate an important point which has been overlooked in some previous studies—the existence of a critical section in the lateral outflow channel is not sufficient to isolate flow in the intersection from the downstream control in the lateral channel. It is shown that if drowned flow occurs in the lateral channel, then the downstream control can affect the flow in the intersection.

The motivation for the present work comes from the fact that a few recent publications describing separation flows report a counter-gradient diffusion phenomenon (CDP) along the separating streamline. This CDP is related to (1) a change of sign of the Reynolds shear stress near the separating point and (2) an opposite sign between the Reynolds shear stress and the mean shear rate, leading to a negative eddy-viscosity coefficient. Such CDP is only reported for configurations where the angle between the separating streamline and the main flow is large (referred to as group 1), whereas no CDP was reported for low angle configurations (referred to as group 2). All these flows were analyzed using Cartesian or curvilinear frame axes. The aim of the present paper is to propose a more intuitive frame for analysis, namely the Serret-Frenet frame axis based on the local flow direction, for which the CDP disappears for group 1 flow configurations and that highly improves the characterization of the mixing la...

When two open-channel flows merge in a three-branch subcritical junction, a mixing layer appears at the interface between the two inflows. If the width of the downstream channel is equal to the width of each inlet channel, this mixing layer is accelerated and is curved due to the junction geometry. The present work is dedicated to simplified geometries, considering a flat bed and a $$90^{\circ }$$ angle where two configurations with different momentum ratios are tested. Due to the complex flow pattern in the junction, the so-called Serret–Frenet frame-axis based on the local direction of the velocity must be employed to characterize the flow pattern and the mixing layer as Cartesian and cylindrical frame-axes are not adapted. The analysis reveals that the centerline of the mixing layer, defined as the location of maximum Reynolds stress and velocity gradient, fairly fits the streamline separating at the upstream corner, even though a slight shift of the mixing layer towards the center of curvature is observed. The shape of the mixing layer appears to be strongly affected by the streamwise acceleration and the complex lateral confinement due to the side walls and the corners of the junction, leading to a streamwise increase of the mean velocity along the centerline and a decrease of the velocity difference. This results in a specific streamwise evolution of the mixing layer width, which reaches a plateau in the downstream region of the junction. Finally, the evaluation of the terms in the Reynolds-Averaged-Navier–Stokes equations reveals that the streamwise and normal acceleration and the pressure gradient remain dominant, which is typical of accelerated and rotational flows.

The present study deals with turbulent flows in an asymmetrical compound channel with an embankment set on the floodplain, perpendicularly to the longitudinal direction. The main purpose of this study was to assess how a rapidly varied flow affects interaction between the floodplain flow and the main channel flow. In addition to rapid changes in the water level and velocity across the compound channel that have a great influence on the boundary shear stress distribution, the embankment, through two recirculation zones developing upstream and downstream, is also responsible for strong lateral mass exchange between the main channel and the floodplains (channel sub-sections). The lateral velocity can indeed reach 50% of the longitudinal velocity, which modifies the characteristics of the mixing layer developing between the channel sub-sections. Depth-averaged Reynolds shear stresses five times greater than those measured for reference flows are recorded within the mixing layer, which indicates that the turbu...

Summary When simulating urban floods, most approaches have to simplify the topography of the city and cannot afford to include the obstacles located in the streets such as bus stops, trees and parked cars. The aim of the present paper is to investigate the error made when neglecting such singularities in a simple flooded 3-branch crossroad configuration with a specific concern regarding the error in discharge distribution to the downstream streets. Experimentally, the discharge distribution for 14 flows in which nine obstacles occupying 1/6 of the flow section are introduced one after the other is measured using electromagnetic flow-meters. The velocity field for one given flow is obtained using horizontal-PIV. Additionally, all these flows are computed using a CFD methodology. It appears that the modification in discharge distribution is mostly related to the location of the obstacles with regards to the intersection, the location of the separating interface and is strongly impacted by the Froude number of the inflow while the influence of the normalized water depth remains very limited. Overall, the change in discharge distribution induced by the obstacles remains lower than 15% of the inflow discharge even for high Froude number flows.

International audience

International audience; Open-channel junctions are common occurrences in sewer networks and flow rate measurement often occurs near these singularities. Local flow structures are 3-dimensional, impact on the representativeness of the local flow measurements and thus lead to deviations in the flow rate estimation. The present study aims i) to measure and simulate the flow pattern in a junction flow, ii) to analyze the impact of the junction on the velocity distribution according to the distance from the junction and thus iii) to evaluate the typical error derived from the computation of the flow rate close to the junction.

International audience; Intense rainfall on urban areas can generate severe flooding in the city, and if the conditions are right, the flow in the streets can be supercritical. The redistribution of the flow in street intersections determines the flow rates and water levels in the street network. We have investigated the flow that occurs when two supercritical flows collide in a 90° junction formed by streets of identical cross section. Several flow configurations within the intersection are possible, depending on the position of the hydraulic jumps that form in and upstream of the intersection. Previous work has identified three flow types, with Type II flows being further classified into three subregimes. Hydraulic models have been developed, based on the principles of the conservation of flow and momentum flux in the intersection, which predict the angles at which the jumps will form. These models can be used to determine the flow type that will occur. Moreover, additional models have been developed for computing the outflow discharge distribution. For Type I flows, it has not been possible to develop such a hydraulic model for the discharge distribution, but some data are provided for one configuration to indicate the influence of different parameters. For Type II and Type III flows, such models are developed, and their predictions agree with data obtained from the channel intersection facility at the Laboratory of Fluid Mechanics and Acoustics in Lyon.; Les pluies intenses en zone urbaine peut générer des fortes inondations et l'écoulement peut être torrentiel dans les rues. La répartition de l'écoulement aux intersections détermine les niveaux d'eau dans les rues. Nous avons examiné l'écoulement qui se produit lorsque deux écoulements torrentiels se rencontrent dans un croisement à 90° formé de deux canaux de largeur identique. Plusieurs configurations sont possibles dans le croisement en fonction de la position des ressauts hydrauliques. Des travaux précédents avaient identifié trois types d'écoulements, le type II étant subdivisé en trois. Des modèles hydrauliques ont été développés en se fondant sur la conservation de la quantité de mouvement. Ces modèles déterminent le type d'écoulement et l'angle des ressauts. D'autres modèles permettent d'obtenir la répartition des débits pour les types II et III. Ces modèles ont été validés sur les expériences effectuées au LMFA (Lyon, France).

Many investigations have been carried out in order to develop models which allow the understanding of complex physical processes involved in urban flooding. The modelling of the interactions between overland flows on streets and flooding flows from rivers and sewer networks is one of the main objectives of recent and current research programs in hydraulics and urban hydrology. However, the modelling of the discharge distribution in the street network with crossroad needs further research due to the complexity of the flow through junctions. This paper outlines the ability of the improved one-dimensional CANOE software to simulate the street flows through the virtual network (developed under the Hy2Ville French National project framework) with several cross-roads. The improvements are done by adding in CANOE the energy losses coefficients deriving from the calibration phase based on the experimental study of the flow through small scale physical model of cross-road channels. Comparisons between 1D and 2D simulated distribution of discharges through the virtual network show a good agreement for the global distribution. However, large differences are observed focusing on the individual cross-road intersections in the virtual network.

International audience; An unsteady mathematical model for predicting flow divisions at a right-angled open-channel junction is presented. Existing dividing models depend on a prior knowledge of a constant flow regime. In addition, their strong nonlinearity does not guarantee compatibility with the St. Venant solutions in the context of an internal boundary condition treatment. Assuming zero crest height at the junction region, a side weir model explicitly introduced within the one-dimensional St. Venant equations is used to cope with the two-dimensional pattern of the flow. An upwind implicit numerical solver is employed to compute the new governing equations. The performance of the proposed technique in predicting super-, trans-, and subcritical flow bifurcations is illustrated by comparing with experimental data and/or theoretical predictions. In all the tests, lateral-to-upstream discharge ratios (R-q) are successfully reproduced by the present technique with a maximum error magnitude of less than 9%.

This paper investigates energy losses in compound channel under non-uniform flow conditions. Using the first law of thermodynamics, the concepts of energy loss and head loss are first distinguished. They are found to be different within one sub-section (main channel or floodplain). Experimental measurements of the head within the main channel and the floodplain are then analyzed for geometries with constant or variable channel width. Results show that head loss differs from one sub-section to another: the classical 1D hypothesis of unique head loss gradient appears to be erroneous. Using a model that couple 1D momentum equations, called "Independent Sub-sections Method (ISM)", head losses are resolved. The relative weights of head losses related to bed friction, turbulent exchanges and mass transfers between sub-sections are estimated. It is shown that water level and the discharge distribution across the channel are influenced by turbulent exchanges for (a) developing flows in straight channels, but only when the flow tends to uniformity; (b) flows in skewed floodplains and symmetrical converging floodplains for small relative flow depth; (c) flows in symmetrical diverging floodplains for small and medium relative depth. Flow parameters are influenced by the momentum flux due to mass exchanges in all non-prismatic geometries for small and medium relative depth, while this flux is negligible for developing flows in straight geometry. The role of an explicit modeling of mass conservation between sub-sections is eventually investigated. (C) 2009 Elsevier Ltd. All rights reserved.

This paper investigates 1-D modeling of nonuniform flows in compound channels. The issue is how to accurately predict both flow depth and mean velocity in the floodplain. A new model, called "Independent Subsections Method" (ISM), is presented here. Unlike classical 1-D models that solve a dynamic equation on the total cross section, the ISM estimates the water surface profile within each subsection. This enables the water level and the subsection mean velocities to be simultaneously calculated, without priority to any variable. In opposition to the Divided Channel Method (DCM), corrected DCM or the Exchange Discharge Model, the ISM assumes independent evolution of the discharge in each subsection of the compound channel. Indeed, this method does not assume equal head loss gradients in all subsections, and it does not impose the downstream discharge distribution. The ISM consists in a set of three coupled 1-D momentum equations (written within main channel, left-hand, and right-hand floodplains) and a mass conservation equation on the total cross section. Mass and momentum exchanges at the interfaces between subsections are explicitly accounted for. This method is validated against experimental data for developing flows in straight compound channel, flows in skewed compound channel, flows in a symmetric converging or diverging compound channel, and flows in an asymmetrical compound channel with an abrupt floodplain contraction. For the 46 runs, the ISM predicts flow depth and mean velocity in the floodplain with a maximum relative error of 8% and 19%, respectively. The ISM also appears to be a useful theoretical tool to improve our understanding of physical processes governing compound channel flows.

Many investigations have been carried out in order to develop models which allow the linking of complex physical processes involved in urban flooding. The modelling of the interactions between overland flows on streets and flooding flows from rivers and sewer networks is one of the main objectives of recent and current research programs in hydraulics and urban hydrology. This paper outlines the original one-dimensional linking of heavy rainfall-runoff in urban areas and flooding flows from rivers and sewer networks under the RIVES project framework (Estimation of Scenario and Risks of Urban Floods). The first part of the paper highlights the capacity of Canoe software to simulate the street flows. In the second part, we show the original method of connection which enables the modelling of interactions between processes in urban flooding. Comparisons between simulated results and the results of Despotovic et al. or Gomez & Mur show a good agreement for the calibrated one-dimensional connection model. The connection operates likes a manhole with the orifice/weir coefficients used as calibration parameters. The influence of flooding flows from river was taken into account as a variable water depth boundary condition.

International audience; This paper describes the flow structures that occur in a 90° junction of four open channels with supercritical flow in two orthogonal inlet channels. An experimental facility was constructed to permit the measurement of flow rates, water depths and the positions of hydraulic jumps in the channels. The various flow patterns which appear depend on the characteristics of the incoming flows, and can be classified into three main types, depending on the location and shape of the hydraulic jumps that develop. These jumps can either be normal to the flow and located in the upstream channels, or can be oblique and confined within the junction. The explanation for the existence of various flow patterns is derived from previous studies of the rapid deflection of supercritical flow. A detailed description of each flow regime is provided, with information on the surface elevations, the behavior of the hydraulic jumps, the deflections of the incoming flows and the formation and characteristics of recirculation zones.; Cet article décrit les structures des écoulements lorsque quatre canaux se rejoignent avec un angle de 90°. L'installation expérimentale a permis la mesure des débits , hauteurs d'eau et position des ressauts hydrauliques. Les structures qui apparaissent dépendent des caractéristiques des débits entrants et trois types peuvent être distingués selon la localisation et la forme des ressauts hydrauliques. Ces ressauts sont soit normaux et situés dans les canaux amont soit obliques et situés dans le croisement. L'explication de ces structures est liée à la déviation d'un écoulement torrentiel. Une description détaillée de chaque régime est fournie incluant ressauts et recirculations.

International audience; Flooding rivers usually present transition reaches where the floodplain width can significantly vary. The present study focuses on an abrupt floodplain contraction (mean angle 22°) in order to determine whether one-dimensional (1D) models, developed for straight and slightly converging geometry, are equally valid for such a geometry. Experiments on a contraction model were carried out in an asymmetric compound channel flume. Severe mass and momentum transfers from the floodplain towards the main channel were observed, giving rise to a noteworthy transverse slope of the water surface and different head loss gradients in the two subsections. Three 1D models and one 2D simulation were compared to experimental measurements. Each 1D model incorporates a specific approach for the modeling of the momentum exchange at the interface boundary between the main channel and the floodplain. The increase of the lateral mass transfer generates moderate errors on the water level values but significant errors on the discharge distribution. Erroneous results arise because of incorrect estimations of both momentum exchange due to lateral mass transfers and boundary conditions which are imposed by the tested 1D models.; Sur le terrain, les écoulements débordants dits en lit composé sont amenés à traverser des biefs de transition, dont la variation de largeur peut être significative. Une des configurations rencontrées est celle de la contraction brusque des plaines d'inondation. De nouvelles expériences de laboratoires ont été menées dans un canal composé à section assymétrique. Dus à la présence de transferts de masse et de quantité de mouvement relativement marqués entre la plaine d'inondation et le lit mineur, plusieurs phénomènes spécifiques ont été observés parmi lesquels, une inclinaison transversale de la surface libre lorsqu'on s'approche du col du convergent, ainsi qu'une juxtaposition d'un régime supercritique dans la plaine d'inondation et d'un régime subcritique dans le lit mineur. Une évolution différente des charges 1D par lit a aussi été identifiée. Dans ce contexte et d'un point de vue pratique, il nous est apparu fondamental d'évaluer dans quelle mesure les modèles 1D classiques s'éloignaient, le cas échéant, des phénomènes observés et, si les lois établies en canaux composés légèrement non-prismatiques étaient encore valables. Trois modèles 1D, traitant de différentes façons l'échange de quantité de mouvement à l'interface lit mineur/lit majeur, sont testés au regard des données expérimentales. Une modélisation 2D permet de mettre en évidence les phénomènes négligés par les modéles 1D et leurs hypothèses classiques.

Les alimentations classiques des canaux expérimentaux à section composée n'assurent pas une répartition de débit favorisant l'établissement rapide d'un régime uniforme. Puisque la charge totale dans le réservoir amont est la même aux entrées du lit mineur et du lit majeur, on a une distribution homogène de la vitesse sur la section totale amont. Le débit dans la plaine d'inondation est donc trop fort, provoquant des transferts de masse en direction du lit mineur tout au long de l'écoulement. Ce papier sintéresse aux distances d'établissements du régime uniforme, régime d'équilibre entre les sous-sections. Il démontre que ces dernières sont significativement supérieures aux longueurs des canaux existants présentés dans la littérature. / Common inlet design for compound-channel flumes does not ensure a proper upstream discharge distribution. As the total head in the upstream tank is the same for both main-channel and floodplain subsections, the velocity in the upstream section is also the same in both subsections. The floodplain discharge is therefore too large and a mass transfer towards the main channel occurs along the flume. This note investigates how long a compound-channel flume must be to ensure that equilibrium between subsection discharges is achieved. The required length is found to be significant compared to the actual length of experimental flumes reported in the literature.